Modular Design of the Integrated Meat Plant

By Eben van Tonder and Christa van Tonder-Berger

29 July 2026

Table of Contents

Executive Summary

This paper is the product of Eben van Tonder’s experience in meat factory design over the last sixteen years, drawn from discussions with process engineers, refrigeration engineers, electricians, maintenance specialists, construction companies, and company owners who demanded a healthy return on investment, and from mistakes made and their consequences experienced first hand. Several philosophies exist for how a new meat plant should be designed. The modular framework is one of them. Eben van Tonder and Christa van Tonder-Berger develop this particular model further here.

The paper sets out five design criteria a modular meat plant should meet: expansion without disrupting the existing operation, the shortest practical time to first production, staged growth with no practical ceiling within the site’s own limits, independent shutdown of any part of the plant without affecting the rest, and capacity that can be adjusted down as well as up. These criteria are grounded in established engineering literature, Baldwin and Clark’s theory of modularity, Koren’s Reconfigurable Manufacturing Systems, group technology and cellular manufacturing, and are then worked through every function of an integrated abattoir, deboning, fresh meat, value added processing, and canning operation, from livestock reception to dispatch.

The paper works out what this means in practice at four levels. The flow architecture sets out how people, product, reusable equipment, raw materials, and waste move through the site without incompatible crossings. Refrigeration is treated as an evaluated system architecture rather than a single rule of thumb, addressing the genuine physical limits of a compressor and the buffer and load matching techniques that manage them. Construction staging draws on the architectural idea of shearing layers to explain why a steel frame and drainage sized for the plant’s full future footprint allow later phases to be built beside a live operation without disturbing it. Containerisation is set out as a further, deployable form of modularity, covering not only cold storage but offices, laboratories, new product development, and even core slaughter and processing functions, with the containers able to be redeployed as the operation grows rather than discarded once first production is reached.

Because a modular plant is, for years at a time, simultaneously a working factory and a construction site, the paper also sets out the organisational structure needed to manage that condition, separate production, rollout, and integration teams, a data driven management layer, a staged rollout sequence, and the named companies worth visiting directly as this approach is planned and built. Country case material is drawn from Germany, Austria, Mexico, Brazil, Australia and South Africa, distinguishing regulation and independently audited evidence from supplier documentation throughout. A closing section names the challenges this paper has not yet worked through in the same depth, power supply, drainage engineering, and internal material handling among them, without changing the argument made in the rest of the paper. The approach the paper argues for remains, in the end, simple: decide the full future architecture once, on paper, and fund and build only the current stage.

Part 1. Principles of modular design

A plant that can expand without disruption, start production quickly, grow indefinitely, shut down in pieces, and scale down as well as up needs more than good intentions. It needs a precise definition of what a module is, and criteria specific enough to test a design against.

1.1 What modularity means

Herbert Simon observed that complex systems evolve faster and prove more stable when they are built from nearly decomposable subsystems (Simon, 1962). Carliss Baldwin and Kim Clark developed this observation into a full theory of modularity, defining a module as a unit that is as independent as possible while still functioning as part of a whole, connected to other modules only through a small number of fixed, published interfaces (Baldwin and Clark, 2000). Baldwin and Clark separate the visible design rules of a system, which must remain stable so that modules continue to fit together, from the hidden design choices inside each module, which can change freely so long as the module continues to meet its interface. This separation gives a modular system the ability to evolve one part at a time.

The same principle was reached independently from a manufacturing direction through group technology, the practice of grouping similar parts or products into families so that a small, dedicated group of machines can serve each family with minimal dependence on the rest of the factory. Group technology’s formal treatment is due to Sergei Mitrofanov (Mitrofanov, 1966), and its application to production control is set out by John Burbidge (Burbidge, 1978). Early group technology ideas appeared in both the United States and the Soviet Union during the first half of the twentieth century.

1.2 The purpose of modularity

The purpose of modular design is not only to make future expansion physically possible. Its deeper purpose is to manage uncertainty. A meat plant is built under uncertainty about livestock supply, market demand, product mix, regulation, labour, utilities, finance and technology. A conventional integrated design commits capital to a single forecast made once, at the outset. A modular design instead creates a sequence of decisions, allowing the owner to commission a complete first capability, observe actual operating data, and then expand, substitute, isolate, pause or retire an individual capability without reopening the architecture of the whole plant.

This flexibility has a value that can be stated in the same terms used for a financial option. Baldwin and Clark, and later Gamba and Fusari, describe modularity as creating options because a hidden design decision inside one module can be changed while the stable interface rules connecting it to the rest of the system are preserved (Baldwin and Clark, 2002; Gamba and Fusari, 2009). In a plant, the equivalent options are the option to defer a later line, the option to expand a successful line, the option to replace an unsuitable process, the option to shut down a failed or uneconomic unit, and the option to replicate a proven module at another site. The added cost of creating clear interfaces and reserved capacity should be weighed against the value of these options, not against first phase capital cost alone.

1.3 Five design criteria

Eben van Tonder set the following five criteria for this design method.

Criterion one. Expansion should minimise planned interference with existing operation, and should confine any unavoidable interference to short, defined, controlled connection windows. Structural tie ins, testing, site traffic, dust, fire system impairment, shared utilities, control system changes and competent authority inspection can all touch an operating plant even under the best planned project. What can be planned and verified is that any such effect was bounded in time and controlled.
Criterion two. The design should achieve the shortest practical time to first production through a minimum viable module, the smallest complete unit that can be commissioned and validated on its own ahead of the rest of the planned facility.
Criterion three. The plant should be capable of repeatable, staged expansion within the declared limits of its site, regulatory approvals, utility supply, logistics, environmental permits and commercial demand. Land, access, utilities, effluent capacity, fire water, labour, cold chain capacity, livestock supply, veterinary service capacity and market demand all impose a genuine ceiling. What modular design contributes is a repeated, known, previously costed unit of expansion, a modular platform built from standardised, repeatable units.
Criterion four. Parts of the operation should be capable of independent isolation during downtime or reduced raw material availability, without affecting the rest of the plant, through complete electrical and mechanical lockout, rail or line separation, a defined sanitation boundary and a defined safe access procedure, not by switching off a drive motor alone. Closing a circuit alone does not address stored mechanical energy, product left in the line, or the sanitation status of an idled unit.
Criterion five. Every system in the plant should be designed so that its capacity can be evaluated and, where practical, adjusted both up and down, using an engineered combination of equipment selection, control strategy and, where appropriate, buffer capacity, rather than by simply installing a smaller version of the same equipment. The refrigeration compressor case, treated in full in Part 5, illustrates a genuine physical limit rather than a design preference, and the answer there is an evaluated system architecture, not a single universal technique applied without qualification.

1.4 Defining a module

For this design method, a module is a bounded production capability, not merely a room, a line, a container or a duplicated machine. It has a defined product and process scope, a declared hygienic classification, a nominal capacity, a minimum stable operating capacity, a maximum validated capacity, defined labour and utility requirements, known inputs, outputs, by products, waste streams and data interfaces, and it can be electrically and mechanically isolated. It has a defined sanitation boundary, maintenance boundary, commissioning boundary and regulatory validation boundary. A unit is not modular if its expansion or shutdown requires uncontrolled interruption of another unit, if its utilities cannot be isolated, if its product and waste routes cannot be separated, or if its commissioning requires revalidation of the whole plant. These tests are applied to every function listed in Part 2 below.

1.5 Reconfigurable Manufacturing Systems

Yoram Koren and colleagues at the University of Michigan, working from the mid 1990s onward, defined a category of manufacturing system called Reconfigurable Manufacturing Systems, RMS, built to solve the same problem this project addresses for a meat plant, how to build a production system that can change its capacity and its function quickly and at reasonable cost as demand and product requirements shift, without either the rigidity of a fixed, dedicated line or the excess, pre-built flexibility of a general purpose flexible manufacturing system designed for variation that may never arise (Koren, 2010; Mehrabi, Ulsoy and Koren, 2000).

Koren’s RMS framework rests on six characteristics, three describing the system’s structure and three describing what that structure allows the system to do in operation.

Modularity, meaning the system is built from standardised, functionally complete units.
Integrability, meaning modules can be connected and later reconnected through standardised mechanical, informational and control interfaces.
Customisation, meaning the system’s flexibility is limited to the specific part or product family it actually needs to handle, rather than built for unlimited generality.
Scalability, meaning production capacity can be changed quickly and economically, characteristically by adding further machines into an existing stage of the process rather than by replacing the stage with a larger machine.
Convertibility, meaning the system can be changed over to a new product or a new process quickly.
Diagnosability, meaning the system can quickly identify the source of a quality or reliability problem.

The correspondence to the five criteria in section 1.3 above runs close. Criterion three, staged expansion, is Koren’s scalability, capacity added by placing further machines into an existing stage without changing the system’s basic architecture. Criterion five, capacity scaling down as well as up, is the same scalability requirement, not a refrigeration specific concern, which is why the compressor case in Part 5 reads as one instance of a general RMS scalability problem rather than a self contained refrigeration topic. Criteria one and four, expansion and shutdown without disturbing the rest of the system, correspond to Koren’s modularity and integrability taken together, since a module can only be added or removed without disturbing its neighbours if its interfaces were standardised in advance.

Case in point. Koren’s own account of the RMS research programme records that the methods and reconfigurable machine technologies developed under the National Science Foundation funded Engineering Research Center for Reconfigurable Manufacturing Systems, which he directed from 1996 to 2012, were implemented in operating factories, principally in the automotive industry, where RMS architecture allows an engine or body assembly line to be scaled by adding reconfigurable machines into an existing stage, and converted to a new product variant by reconfiguring rather than replacing that stage. This case has not been applied to a meat plant in the literature reviewed for this paper; the claim made here is limited to the correspondence between Koren’s six characteristics and the five criteria set out for this project.

1.6 Containerisation as a module boundary

Baldwin and Clark’s interfaces, Koren’s modularity and integrability, and the shearing layers argument developed in Part 6 point toward the same practical conclusion for any function that is not itself a fixed, in-line production step. Where a function can be fully specified, built, tested and even relocated as a single self-contained unit, a shipping container or an equivalent steel framed cabin is the most literal, commercially available form a module boundary can take, since the container’s own structure supplies the isolation, the finish, and the transportability that would otherwise have to be designed from scratch as part of the building.

Eben van Tonder first considered integrating containerised units with physical structures to expedite launch dates in discussions with a Russian company a few years ago. That company had one small processing and chiller room, no slaughtering facility, and lots of available land, and the approach of bringing containerised capability onto that land ahead of any permanent construction was what first suggested containerisation as a way of shortening the time to first production rather than only a way of adding cold storage.

This applies most obviously to cold storage. A container’s floor is a manufactured, replaceable component rather than a poured and coated slab, which keeps it cleanable in a way a damaged in-situ floor is not, and a solar-electric hybrid power supply lets a single unit’s uptime be decoupled from the reliability of the site’s own grid connection without any consequence for the rest of the plant’s power architecture. The same logic extends to every non-core or auxiliary function elsewhere in this document, administration and management offices, staff amenities, security posts, retained and suspect product holding, laboratories and environmental monitoring, and new product development or test kitchen functions, all of which sit close to the operation without needing to be built into its fabric. A containerised unit can be positioned adjacent to the production modules it serves, connected to power, water and data through the same interface discipline argued for in Part 8, and relocated, added to, or removed entirely as the plant’s administrative and support needs change, without structural, drainage or refrigeration consequence for the production modules themselves.

A new product development function is a particularly clear case for this treatment, combining a small footprint, a requirement for its own controlled access separate from production staff traffic, equipment and formulation needs that change faster than a production line’s own equipment does, and a genuine business reason to keep it physically close to the plant it serves without being part of the certified production flow itself. Housing it in a container satisfies the module test given in section 1.4, a defined scope, a known interface, and the ability to be commissioned, altered or removed without disturbing its neighbours. The NPD department is also an example of a broader point, that a container is a deployable asset the operation can redeploy as it grows, not only a way of reaching first production faster. A container bought for NPD today can become a second QC office, a records store, or a canteen extension tomorrow, without the capital being written off against a single fixed use.

Containers can be used for slaughter, processing, and refrigerated storage, and none of this needs to be a temporary or transitional arrangement. A container can hold a long term function within an already established operation, years after the rest of the plant was built in permanent construction. Once a container is accepted as a genuine module rather than a stopgap, its natural range of uses inside an operating plant is wide. A canteen container gives staff a break area without consuming space inside the certified production building. Office space in containers houses supervisors, planners, and administrative staff close to the floor they manage, without the cost or delay of a permanent extension. Filing and records storage, parts storage attached to a workshop, and a QC office with its own bench and sample fridge are all functions that sit naturally in a container, since none of them touch product directly and none of them need the drainage, refrigeration, or structural specification a production room requires. A container is equally useful for separating categories that must not be confused with each other but do not need a full room each, returned or rejected stock kept apart from saleable stock, or different product classes kept in clearly separated units so a physical stocktake can be done by counting containers rather than by sorting mixed stock on the day.

Case in point. A mid size meat processor in Limpopo province has used containers as offices and boardrooms for many years after the plant itself was built, a working example of a container module outliving its original purpose and continuing to serve the operation long after the launch phase that might first justify installing it. Container use for office, canteen, and administrative space at this scale is not unique to meat processing. The Obetz Stadium Plaza development in Ohio assembled 122 shipping containers into offices, conference rooms, a cafeteria, a ticketing office, and public restrooms serving an entire stadium precinct, a scale of container use for exactly these auxiliary functions, canteen, offices, records and support space, that goes well beyond a single unit. The United States military is, by volume, the largest user of shipping containers for offices, sleeping quarters, kitchens, refrigeration, and food and plant processing on its bases worldwide, which is the clearest demonstration available that containers are trusted for sustained, permanent-grade use rather than only short term deployment. On the QC side specifically, container based laboratories are an established commercial offering for the food and beverage industry, and one documented conversion split a 40 foot container into a sealed product testing chamber, a secure instrument room, and a small communal space for staff, the same three way division, testing, instruments, and people, that a plant’s own on-site QC function would need regardless of whether it sits in a container or a permanent room.

Case in point. DC-Supply A/S of Denmark, working with AVS Agric ApS, has delivered a fully functional two-container pork abattoir, built at its Danish workshop and shipped to a client in Malaysia, and has separately delivered a custom built container based salmon processing plant to the Faroe Islands. The manufacturer reports a build time of approximately eight weeks and a site commissioning time of a single day once power and water are connected. AES Food Equipment markets containerised units covering not only modular slaughter lines but also processing and packaging laboratories for meat, fish and dairy products, the direct equipment level precedent for a test or development function. In South Africa, the company trading as Mobile Slaughter Technologies, founded as Mobile Abattoirs in 2015 by the veterinarian Dr Tertius Bergh, has developed solar powered mobile slaughter units for cattle, sheep and goats, together with containerised abattoir designs for pigs and poultry, aimed at remote and Foot and Mouth Disease control zone slaughter in Africa. Eben van Tonder has been in direct contact with Dr Bergh. This data is drawn from supplier and company documentation rather than independent audit, and points to containerisation of both core slaughter functions and non-core support functions as an established, commercially available practice.

Part 2. Full plant breakdown

The same test applies at every stage from live animal reception through to canning: what the module is, how it grows, and how it stops without dragging the rest of the plant down with it.

2.1 Livestock flow and welfare

The one-way flow principle argued for throughout this paper has to work for a live animal moving under its own will before it has to work for anything else, since an animal that senses a dead end, a shadow, or an unfamiliar noise will stop, turn, or reverse regardless of what the layout drawing intended. Temple Grandin’s curved chute and race system (Grandin, 1984) is the clearest documented case of the flow, buffering, and zoning principles used elsewhere in this paper applied to a live animal rather than to product or equipment.

A curved race keeps the animal moving forward voluntarily and removes its sightline to what waits ahead, the same personnel zoning principle argued for in Part 3, not seeing what lies past a boundary reduces the tendency to balk at it.
A round crowd pen filled to no more than half capacity is the same buffer sizing argument made for refrigeration in Part 5. An undersized or wrongly shaped buffer produces the same unstable, cycling behaviour whether it is holding cattle or holding refrigerant load.
Solid sided chutes are the animal handling version of the air and light control treated elsewhere in this paper as its own flow category. An uncontrolled visual disturbance defeats a correctly designed layout the same way an uncontrolled air pressure imbalance defeats a correctly designed hygiene zone.
Lighting, flooring, and the removal of small distractions carry a measurable throughput consequence, not only a welfare one, the same point made throughout this paper, that a design fault shows up as a flow and capacity problem before it shows up as anything else.

These are inputs to the reception and lairage module in the table below and to the animal welfare requirement in Part 8, not a separate welfare topic.

Case in point. Curved chute and centre track restrainer systems designed by Grandin have been adopted at meat plants across the United States, Canada, Europe, Mexico, Australia and New Zealand, and roughly half of all cattle in North America are handled through a centre track restrainer system of her design. This is a documented instance of a single, well specified module, the curved race and crowd pen, repeated and adapted across many plants and countries over several decades, a working example of the modular platform argument in section 1.3, a proven, standardised unit reproduced rather than redesigned at every new site.

2.2 Full function breakdown

The unit of analysis is stated for every row, since a module is sometimes a room, sometimes a line, sometimes a department, and sometimes a shared utility. Every growth path in the table below is criterion three from section 1.3 applied to a specific function, and every isolation path is criterion four and criterion five applied to that same function, so the table is the five criteria worked through the plant rather than a separate design exercise.

Function	Unit of analysis	What is modularised	How it scales up	How it scales down or isolates
Livestock reception and lairage	Room and yard group	Pen groups on independent hardstand	Add a further pen group and hardstand extension	Close a pen group for cleaning while livestock is held elsewhere
Stunning and sticking	Line	Stunning box and rail drive per line	Add a parallel rail with its own stunning box	Full electrical and mechanical lockout, not motor circuit isolation alone
Veterinary inspection point	Regulatory control point	One inspection point per rail, subject to competent authority acceptance	A new rail brings its own inspection point and retained carcase rail	Idle rail’s inspection point unstaffed, no consequence for the active rail
Bleeding and dressing	Line	Overhead rail sections per line	Extend rail length or add a parallel rail	Full lockout per criterion four
Evisceration	Line	Synchronised to line speed, per rail	New rail, new synchronised drive	Idle rail stops, no shared drive to disturb
Hide, skin and raw offal separation	Room	Discrete room adjoining the slaughter line	Add a second room	Close one room, route through the remaining room at reduced throughput
Edible offal cleaning and packaging	Department, linked to slaughter by inspection and identity, not merely by space	Cleaning and packing line within the offal department	Add a further line	Idle line closed for cleaning
Carcase chillers	Room	Refrigeration architecture evaluated per Part 5, not assumed to be one compressor per room	Add a further chiller room	Take one room offline for defrost or repair, per the evaluated architecture’s isolation provision
Rendering, blood and hide handling	Department	Standalone, low integration module	Add capacity as a further standalone unit	Operate intermittently
Deboning, per species line	Line	Table, saw and floor scale per line	Add a further species or cut specification line	Idle line closed, others continue
Deboning work in progress chillers	Room	One per line	Add per new line	Idle chiller defrosted or serviced independently
Fresh meat retail cutting and packing	Room	Dedicated room fed from deboning	Add a second packing machine or shift	Close for a shift without affecting other lines
Box frozen line	Bay	Separate bay from retail	Add a further carton line	Close independently of retail
Value added processing lines	Line	Each its own bay and equipment set	Add a further line of the same specification	Idle line cleaned and locked out
Smoking and cooking	Unit	Discrete units, each with its own drive through opening	Add a further unit with its own drive through	Idle unit powered down, drive through sealed
Raw area and post lethality high hygiene packing	Two departments	Already separated by design	Add a further packing bay on either side	Idle bay closed without crossing the boundary
Canning, filling and seaming	Line	Discrete line, physically separate from the retort room	Add a second filling line	Idle line closed, retort room continues on its own schedule
Canning, retorts	Vessel, under a competent thermal process authority	Independently controlled vessels within a room sized for future capacity	Add a further retort vessel into an already provisioned bay	Take one retort out of service for maintenance without stopping others
Dispatch chillers and freezers	Room	One room per product category	Add a further room	Idle room closed independently
Picking and staging	Floor area	Racking and floor space	Add floor area and racking bays	Reduce active staging area without structural change
Central and post lethality wash points	Two clusters minimum, possibly more	See Part 4	Add a further wash point as a cluster is added	Idle wash point closed, others continue
Effluent and wastewater treatment	Shared utility	Package plant or settlement and irrigation system, added in stages	Add a further treatment module	Reduce load during low throughput, provided collection pipework was sized for full future capacity
Steam and hot water	Shared or distributed utility, subject to a thermal utility study	Point of use units or central thermal plant, whichever the study supports	Add capacity per the study’s conclusion	Isolate a department’s unit independently where the architecture allows it
Refrigeration	Evaluated system architecture, see Part 5	Central, distributed or hybrid, selected by evaluation	Add circuits, compressors or branches per the evaluated architecture	Isolate a circuit or branch without affecting gas charge elsewhere, subject to the chosen architecture
Electrical distribution	Shared utility	Reserved physical routes and board space, sized to credible future demand	Install selected trunk capacity as demand becomes credible	Isolate a circuit at its own breaker
Retained, suspect and condemned product handling	Room, under veterinary control	A dedicated, identifiable holding area	Add capacity as throughput grows	Independent of other modules by regulatory necessity
Laboratories and environmental monitoring	Room	A dedicated testing and sampling area	Add bench or instrument capacity	No production consequence if properly separated
Chemical and cleaning agent stores	Room	A dedicated, ventilated, bunded store	Add storage capacity	No production consequence if properly separated
Packaging and ingredient quarantine	Room	A hold area for material pending release	Add capacity	No production consequence if properly separated
Staff amenities	Container or isopanel unit	Per department or shift group	Add a further container unit	Close one unit for cleaning while others remain open
Administration and management offices	Container unit	Not in the product flow	Add a further office container	No operational consequence
Workshop, maintenance and roof void access	Dedicated area and roof void	Bench space and roof void access points	Add bench space or roof void access points	No operational consequence if separated from the production floor per Part 9
Construction isolation for future phases	Temporary zone, not a permanent module	Sealed hoarding, temporary drainage and utility isolation	Applied fresh for each future phase	Isolated from the operating plant until release

Part 3. The flow architecture

A plant built from independent modules still has to move people, product, equipment and waste between them without the categories colliding.

3.1 The objective

A complete plant has external vehicle movement, live animals outside the building, inedible by products, manure, digestive contents, waste, chemicals, contractors, maintenance parts, samples, returned goods, and emergency access, together with air, water, drainage and data, which behave as invisible flows capable of carrying contamination or creating dependence across a boundary even where the visible, physical flow appears correctly separated. Most systems of hygiene regulation, including European Union food hygiene law, permit separation in time as well as in space, provided contamination is prevented. The governing rule is that uncontrolled, incompatible crossings are prohibited, not that no crossing of any kind may ever occur.

3.2 The movement matrix

The following categories are each mapped through the site, from entry to exit, with their compatible and incompatible neighbours identified at every point they might meet another category.

Internal product flow, meaning meat and carcase movement from reception through to finished goods, which should only ever move forward, with any exception treated as a separately authorised route.
Personnel movement, including staff, visitors and any category of person who may legitimately need to cross a hygiene boundary, veterinary and quality staff, maintenance staff, sanitation staff, supervisors, emergency responders and contractors, each requiring their own defined access protocol covering clothing change, hand and footwear hygiene, tool control and a written record of the crossing.
Clean and dirty reusable equipment flow, meaning crates, trolleys and meat buggies moving to and from the wash point, treated fully in Part 4.
Non-meat raw material flow, meaning packaging, ingredients, spices, casings and cans, entering through a goods receiving point physically separate from livestock reception.
Finished goods flow, exiting packing directly to its own dispatch chiller or freezer. Where a validated hazard analysis and traceability system supports it, a controlled, separately defined, segregated and authorised rework or return route may exist, distinct from the ordinary forward product route, and included in the plant’s food safety change control system.
External vehicle and site logistics flow, meaning livestock vehicles, personnel vehicles and finished goods vehicles, routed on site so that they do not cross each other’s yard areas without a controlled reason.
Inedible by product and waste flow, meaning hides, blood, manure, digestive contents, condemned material and general waste, each requiring its own route to its own handling or removal point.
Maintenance, contractor, sample and emergency flow, meaning the movement of parts, tools, test samples and emergency responders through the plant under a defined access protocol.
Air, as its own flow, meaning supply air, extract air, pressure relationships, filtration, humidity and the movement of steam, smoke and odour, which must not be allowed to defeat the physical hygiene zoning built into the walls and doors.
Water and drainage, as its own flow, meaning potable water, process water, recovered water, hot water, blood and fat bearing wastewater, high strength wastewater, manure and digestive content drainage, chemical drains, clean condensate, roof water and storm water, each requiring separation, defined backflow protection, and defined future capacity at every interface.
Data, as its own flow, meaning equipment tags, control signals, alarms, historian data, recipe management, traceability and lot genealogy, treated in Part 8 under digital and control architecture.

3.3 The engineering solution

Routing many distinct categories of movement through one shared building without an incompatible crossing is, in network flow and factory layout theory, a multi-commodity flow problem. The standard engineering answer, used in factory layout design and in complex logistics environments such as hospitals and airports, is to assign each high volume or high risk category its own dedicated physical channel, corridor, lift or rail wherever the volume justifies the capital cost, and to reserve strict time separation, one category moves, the area is then cleared and cleaned, and only then does the next category move, for the lowest volume, lowest risk crossings only. Time separation is a procedural control and depends on continued staff discipline rather than on the architecture itself, and should be treated as the fallback wherever a dedicated channel can instead be justified. A dedicated channel is also what makes criterion one from section 1.3 achievable at the flow level, since a category with its own channel can be extended, rerouted, or temporarily closed for construction without forcing a change on any other category’s route.

Case in point. The Iowa State University Extension guide to designing a small red meat plant sets out published floor plan diagrams built around this logic, a single forward moving product corridor from livestock intake through slaughter, chilling, cutting and cooler storage to a separate dispatch dock, with personnel entry, dry storage and office space held to the opposite side of the building from the product corridor so the two categories of movement never share a doorway (Iowa State University Extension, 2009). The guide addresses a small plant rather than a 300 or 450 head design, but it is a complete, published, dimensioned floor plan built explicitly around a single dominant forward flow with personnel structurally separated from it.

Part 4. The wash architecture

Crates, trolleys and meat buggies are the one category of movement that must legitimately travel in both directions, dirty out to the wash point and clean back to where it is needed.

4.1 The trade off

There are two engineering models. A centralised model routes all crates, trolleys and buggies to a single wash point, achieving higher utilisation of the washing equipment itself. A clusterised model gives each department, or each closely related group of departments, its own smaller wash point, achieving full independence at the cost of lower utilisation of each individual unit. Group technology and cellular manufacturing theory addresses this choice under the general heading of shared versus dedicated resources, and does not resolve it in favour of one model universally, since the correct choice depends on the specific hygiene risk, volume, and consequence of failure in the specific plant. Either model still has to satisfy criterion three, since a wash point of either kind should be a known, repeatable unit that can be added again as the plant grows, not a bespoke design decided afresh at every expansion.

4.2 The recommendation

The one mandatory boundary is between the raw area and the post lethality high hygiene area, since equipment that has touched raw product must never be washed alongside equipment used only on the post lethality side. Where raw poultry, raw red meat, edible offal, post lethality ready to eat production and allergen controlled product are all present in the same facility, two clusters may still be too coarse, since each of these carries a different hygiene risk and a different consequence if its equipment is washed alongside another category’s equipment. The number of wash clusters should be set by a documented risk assessment covering hygienic risk, crate ownership, volume, journey length, wash validation, drying, storage, water and chemical balance, resilience and recontamination risk.

Wherever the raw area and post lethality high hygiene area are separated for washing, this should proceed only with competent authority acceptance and a validated physical or time separation, sanitation regime, drainage design and air control, since food hygiene law generally requires prevention of contamination and permits some separation in time or space, and the exact permitted arrangement differs between jurisdictions and products.

4.3 Overhead rail cleaning equipment

United States Patent 3,935,610 concerns an apparatus for cleaning overhead conveyor rails in a meat packing plant. It illustrates historical overhead rail cleaning equipment, not the design of a central crate or trolley return network, which is a different engineering problem.

Case in point. Kometos’ delivery to the Finnish operator Meatgard is documented by the supplier as a complete nine module solution covering slaughter, meat cutting, packaging and the associated hygiene and cleaning facilities within a single, approximately 510 square metre building. This is supplier documentation of a plant built from the outset as a defined set of modules rather than as one undivided building; no independent audit of the installation’s wash architecture, energy use or water performance was located in this research.

Part 5. Refrigeration

Refrigeration is where criterion five from section 1.3, that capacity scale down as well as up, is hardest to satisfy, since a compressor has a genuine physical minimum below which it cannot run stably.

5.1 The architecture question

Refrigeration modularity should be achieved through a system architecture selected from the plant’s actual load profile and a documented risk assessment, not assumed in advance. A central ammonia plant with multiple compressors, isolated branches, staged condensers and appropriate control architecture can achieve isolation and staged capacity while also achieving better energy performance than a fully distributed system, because a centralised plant allows load to be shared and sequenced across several compressors of matched or graded size. A fully distributed system, one compressor and condenser per room, reduces total refrigerant inventory in any one location and can simplify staged construction, since each room’s refrigeration is commissioned with the room itself. Both architectures should be compared against energy performance, refrigerant charge and its associated safety consequence, redundancy, maintenance access, turndown capability, future expansion, and life cycle cost, and the choice should be documented as an evaluated decision.

5.2 Compressor operating envelope

Every compressor has a manufacturer defined operating envelope and a minimum allowable capacity under the selected operating conditions, and this minimum depends on the compressor type, its control method, oil management, motor cooling and pressure ratio, not on a single figure that applies across all installations. The correct practice is to obtain certified compressor performance data and the control supplier’s stated limits for the specific machine and application selected.

5.3 Liquid slugging

Low suction pressure resulting from an underloaded compressor does not by itself cause liquid slugging. Liquid carryover into a compressor is a distinct failure, arising from inadequate liquid separation, a control failure, poor superheat control, floodback, a defrost event, or another liquid management problem, and is addressed through system design and operational control of liquid handling. Unstable control and short cycling on one hand, and liquid slugging on the other, are separate engineering problems.

5.4 Load matching

Where a compressor’s stable operating range does not match the plant’s actual load profile, the available engineering responses include compressor staging across multiple machines of matched or graded size, sequencing control that brings machines on and off as load changes, variable speed control where the application and machine type supports it, cylinder unloading or slide valve control within the manufacturer’s stated envelope, evaporator zoning so that a subset of the cooling duty can be isolated from the rest, secondary fluid thermal storage such as a chilled water or brine buffer where a secondary fluid circuit already exists and has been properly sized for that duty, and product load scheduling, the timing of chilling or freezing operations planned to keep the refrigeration system within its efficient operating band. A secondary fluid buffer tank is relevant specifically to a secondary fluid system and is not a universal substitute available to every direct ammonia plant, and ice storage is a specific technique for certain freezing duties, not a general substitute for hard freezing capacity elsewhere in the plant. The correct combination of these measures depends on the refrigeration architecture selected in section 5.1, established through engineering evaluation.

5.5 Hot gas bypass

Hot gas bypass can maintain a compressor’s operation at a load below its normal stable minimum, but it does so by wasting the compression energy already invested in the diverted gas, and is a limited control or protection measure for specific situations, not a preferred optimisation method. Right sized parallel compressors, effective sequencing, variable speed control where suitable, unloading within the manufacturer’s envelope, and demand management should be exhausted first, with hot gas bypass reserved for circumstances those measures cannot address.

5.6 The buffer principle

Production variability should not force a piece of equipment to operate outside its validated range, applied through the specific technique appropriate to the specific equipment, not a single buffer concept borrowed from refrigeration and applied loosely elsewhere. A bowl cutter has a validated minimum and maximum batch size, a maximum waiting time for held raw material, and temperature limits that bear on microbial growth and oxidation, together with allergen changeover and traceability requirements, and a work in progress chiller alone does not resolve these, since holding time itself carries a food safety consequence that must be independently validated. A retort’s load configuration is established and accepted by a competent thermal process authority, covering heat distribution, heat penetration, venting, initial product temperature, container orientation and basket configuration, and cannot be set by energy efficiency reasoning alone. In both cases the correct design response is a validated batch size, a validated holding time, and a validated process schedule, of which a buffer chiller is only one supporting element.

5.7 Case in point

The Australian Meat Processor Corporation, funded jointly with Meat and Livestock Australia, produced a staged set of guidebooks for its Refrigeration Plant Energy Improvement Project, titled Industrial Ammonia Systems Part 1 and Part 2 for medium to large abattoirs using centralised ammonia systems, together with a separate handbook for smaller plants using different refrigerant systems and a further new technology handbook (AMPC, Refrigeration Plant Energy Improvement Project). This staged structure reflects the industry’s own finding that plants sit at different starting points in refrigeration control maturity, and that foundational load matching and control measures should be addressed before more advanced optimisation is attempted. Any specific technical claim drawn from these guidebooks should be checked against the specific guidebook and page before it is relied upon.

Part 6. Construction beside a live operation

A modular plant is only as good as its ability to grow while the rest of it keeps running, which is criterion one and criterion two from section 1.3 tested against the building itself rather than against a single piece of equipment.

6.1 Shearing layers

Future construction is itself a food safety and operational hazard, and every later phase should have its own construction segregation plan prepared before the first phase is built. A building is not one object changing at one speed, but several layers changing at different speeds, a distinction set out by the architect Frank Duffy and developed by Stewart Brand into the framework known as shearing layers (Duffy, 1990; Brand, 1994). Brand’s six layers, from the slowest to the fastest, are:

Site, effectively permanent.
Structure, the foundation and load bearing frame, which changes over decades.
Skin, the external cladding and roofing, which changes over a somewhat shorter period.
Services, wiring, plumbing, refrigeration and ventilation, replaced far more often than the structure that contains them.
Space plan, internal walls and room layout.
Stuff, the equipment and furnishings that move on the shortest cycle of all.

Brand’s central design argument is that a building fails to adapt well not when a fast layer changes, which is expected and manageable, but when a fast layer is too tightly embedded in a slow one, so a service or space plan change cannot occur without disturbing the structure itself. His prescription, that an adaptable building must allow slippage between these differently paced layers rather than binding them rigidly together, is the same design principle argued for throughout this document, that the structural steel frame and drainage, the slowest moving layers of a meat plant, should be sized and placed for the full future footprint from the first pour, while the isopanel walls, the refrigeration circuits, and the equipment lines, the faster moving layers, are added, changed or removed in stages against that fixed frame without requiring it to be altered.

6.2 The construction segregation plan

The plan should address sealed hoarding, dust and pest control, contractor entry and exit routes, construction waste removal, pressure relationships between the construction zone and the operating plant, negative pressure control where required, temporary drainage, utility isolation during tie ins, fire protection impairment during the works, testing water, tool and parts control, and the boundary between commissioned and uncommissioned space. No new module should connect directly to an operating hygiene zone immediately on completion of construction. It should first pass a construction clean down, technical commissioning, hygienic inspection, cleaning validation, pest inspection, environmental verification where applicable, and a formal, recorded release, with utility and data connections made during planned windows that have a defined fallback arrangement.

Case in point. MIT’s Building 20, a temporary wartime structure erected in 1943 and in continuous, heavily reconfigured use until its demolition in 1998, is the clearest documented case of a building deliberately designed around fast moving, loosely coupled internal layers inside a slow, cheap, plain structure. Its plywood and concrete block construction was cheap and its structure loose enough that researchers were permitted to cut holes in floors and walls, rewire spaces, and rebuild internal partitions to suit each new project without triggering the formal, slow moving change process a more finished building would have required, and the building is widely credited in the architectural and innovation literature, including Brand’s own account, with having supported an unusually high concentration of significant research over more than fifty years, because its space plan and services layers were free to change quickly while its structure remained cheap, plain and undemanding. A meat plant’s finishes are fixed by food safety regulation in a way a physics laboratory’s are not, but the same deliberate looseness between structure and the faster moving layers above it, achieved through the steel frame and drainage capacity argued for throughout this document, is what allows staged construction to proceed beside a live operation without repeatedly reopening the building’s slowest, most expensive layer.

Part 7. Country case material

Regulation, official project records and independently audited evidence are distinguished throughout this section from equipment supplier marketing material. References are ordered by source hierarchy, applicable law and official standards first, primary academic and recognised engineering handbooks second, official industry funded research third, official project records fourth, independent case studies fifth, supplier documentation sixth, and trade press last, as context only.

7.1 Germany

BANSS Schlacht- und Fördertechnik GmbH, Biedenkopf, is a supplier of configurable slaughter technology, with a company history extending to 1868. Its DDM 120 and DDM 240 model designations refer to discontinuous pig dehairing machines, with configurable dehairing capacity of up to 120 and 240 pigs per hour respectively. The modular, prefabricated on farm slaughter unit developed by Friedrich Sailer of Neu-Ulm, reported in regional trade press and documented by the supplier, is a minimum viable module built specifically to reach first slaughter as quickly as possible.

7.2 Austria

No published engineering case study of Austrian abattoir modular construction at the scale relevant to this project was located in this research. Schlachthof Artmayr is a verified, operating, regional European Union licensed slaughter and cutting establishment; no evidence was found that its building is of modular construction. The University of Veterinary Medicine Vienna publishes Austrian research relevant to animal welfare, meat inspection and plant hygiene, though not modular building design specifically. Direct engagement with the Austrian Landwirtschaftskammer, BOKU, Vetmeduni Vienna, and Austrian abattoir equipment integrators would be the route to genuine Austrian engineering case material.

7.3 Mexico

Mexico’s Tipo Inspección Federal certification system, administered by SENASICA, and the underlying norms, NOM 008 ZOO 1994 and NOM 009 ZOO 1994, together with any amendments published in the Diario Oficial de la Federación, are the primary legal source for Mexican one directional flow and infrastructure requirements; contractor or supplier commentary on these requirements is commercial commentary, not the regulation itself, and the underlying norms were reported to be under active revision in 2026. The municipal project record for the Rastro TIF Municipal in Querétaro, covering 2015 to 2018, documents a staged expansion, including a separately procured and separately certified refrigeration chamber expansion (Municipality of Querétaro, 2015 to 2018). The exact certification notice number and the sequence of individual procurement items should be confirmed against the underlying official project record before further reliance on the specific dates or figures.

7.4 Brazil

Masterboi’s Canhotinho plant, in Pernambuco, was inaugurated on 15 August 2022 with a stated slaughter capacity of 700 head of cattle, sheep, goats and pigs per day, an investment of over R120 million, and a stated daily meat processing figure reported inconsistently across sources as either 250 tonnes, in the state economic development agency’s own account, or 300 tonnes, in several press accounts, a discrepancy that should be resolved against the company’s own published figures (Government of Pernambuco, 2022). Contemporary trade press coverage from the time of the launch, Mercado&Consumo, reported the company’s stated intention to reach 1,400 animals per day, reported as a stated intention at the time of inauguration, not as a confirmed, funded, or engineered second construction phase; no official expansion permit, environmental licence or capital plan confirming that second phase was located in this research. The plant is a single stage, already operating facility whose owner expressed a future growth intention at the time of opening.

7.5 Australia

The Australian Meat Processor Corporation, funded jointly with Meat and Livestock Australia, is the strongest documented case material for the refrigeration and energy dimension of modular design, through its Refrigeration Plant Energy Improvement Project and its staged Industrial Ammonia Systems guidebooks. AMPC has also investigated cogeneration systems for processing facilities; the exact title and project number of that publication should be confirmed directly against AMPC’s own published project record.

7.6 South Africa

The International Finance Corporation, in partnership with the Red Meat Abattoir Association and the Swiss State Secretariat for Economic Affairs, published its benchmarking study of resource efficiency in South African red meat abattoirs in 2020, the strongest, most independently audited source across all six countries (International Finance Corporation, 2020). The study assessed twenty one abattoirs of different sizes, species, locations and operational configurations, and the three descriptions used, Fully Integrated, Slaughter Chill and Bone, and Slaughter and Chill only, are configurations found among those twenty one plants, not twenty one separate categories in their own right. The IFC has also published a companion Practical Guide for Improving Resource Efficiency in Red Meat Abattoirs in South Africa. Both documents should be read alongside applicable South African legal, veterinary, environmental, water, energy and municipal infrastructure requirements before this design method is applied to a specific South African site.

Part 8. Further design requirements

Each item below is a requirement to be addressed in the next stage of design.

Formal module definition and capacity envelope. Every module and every shared service should be described by a module data sheet stating nominal capacity, minimum stable throughput, maximum validated throughput, surge allowance, changeover loss, sanitation time, maintenance allowance, labour demand, utility demand, effluent load, by product output, and the product mix and environmental conditions under which these figures apply. A single nameplate capacity figure is not sufficient.
Interface control specification. Every module interface, structural grid and loads, finished floor level, drainage invert and reserve capacity, wall and ceiling junctions, utility connection points, isolation valves, electrical protection, control signals, communications protocol, fire separation, and maintenance and lifting access, should be recorded in a controlled document before detailed design begins, and that interface should remain stable even where the equipment and process inside the module later change.
Dependency and bottleneck map. Every module should be checked against the shared services it depends on, lairage, slaughter rail, chillers, veterinary service, rendering, water, steam, refrigeration, compressed air, electrical supply, effluent, packaging stores, laboratories, dispatch, fire systems, information systems and labour, since the least flexible shared element, not the newest module, sets the plant’s real capacity limit.
Animal welfare and veterinary architecture. Every expansion phase should address lairage density, weather protection, water provision, suspect animal pens, emergency slaughter provision, non-ambulatory animal handling, stunning backup, restraint, line speed, inspection presentation, retained carcase rails, detained meat handling, condemned material handling, and official veterinary facilities, with the competent authority involved in each phase.
Chilling as a validated process. Carcase spacing, loading sequence, air velocity, evaporator placement, defrost cycles, temperature and time targets, surface drying, weight loss, cold shortening risk, hot loading practice, door opening frequency and room recovery time should be defined and validated for each chiller room. Adding a further chiller room of the same size does not by itself guarantee the same chilling performance as the original room; each new room should be commissioned and validated on its own terms.
Post lethality environmental control. For ready to eat product, the module boundary should include air pressure relationships, condensation prevention, environmental monitoring zones, drain design, and controlled equipment, packaging and personnel entry procedures, together with sanitation validation and a documented approach to the control of Listeria monocytogenes.
Allergen and formulation control. Value added processing requires controlled ingredient receipt, segregated storage, dispensing control, recipe control, label verification, rework control, cleaning validation and planned product sequencing, scaling together with the processing modules they serve.
Sanitation as a production system in its own right. Cleaning windows, labour, hot water supply, chemical storage and dilution, foam and rinse capacity, clean out of place and clean in place systems, hose management, drainage, drying, pre-operational inspection and environmental verification should be planned with the same rigour as a production line, since sanitation capacity that does not keep pace with production capacity becomes the plant’s real bottleneck regardless of how well the production modules themselves are designed.
Maintenance and safe isolation. Every module requires lockout points, safe access, lifting points, removable panels, service clearances, drainage, lubrication access, a spare part strategy, and a defined method for bringing tools and parts into a hygienic area without introducing contamination.
Resilience distinguished from modularity. A module can be independent without being redundant, and a redundant item can exist inside an otherwise tightly coupled system. The plant should separately define its credible failure events, the capacity loss it can accept from each, the maximum acceptable recovery time, its duty and standby philosophy, its critical spare parts holding, its manual fallback procedures, and whether partial production is acceptable following each type of failure.
Fire and life safety. Every future expansion phase must preserve escape distances, fire compartmentation, smoke control, fire water supply, hydrant coverage, hazardous chemical separation, ammonia detection where relevant, emergency ventilation, machinery room access, and emergency response routes, checked afresh for each phase.
Ammonia and refrigerant safety. Where ammonia or another hazardous refrigerant is used, the design should address regulatory thresholds, refrigerant inventory limits, machinery room design, relief discharge routing, detection systems, ventilation, emergency isolation, occupied space risk, pressure systems compliance, and a means of expanding refrigeration capacity without an uncontrolled increase in total refrigerant charge on site.
Digital and control architecture. Equipment tagging, programmable controller interfaces, alarm philosophy, historian data, recipe management, traceability and lot genealogy, scale integration, label control, access rights, cyber security, network capacity and spare input and output points should be standardised across modules so a new module integrates into the plant’s control and traceability system without a bespoke redesign each time.
Commissioning and qualification. Each module requires its own factory acceptance testing where relevant, site acceptance testing, dry and wet commissioning, utility balancing, calibration, hygienic inspection, cleaning validation, process validation, performance testing, staff training and a formal handover record before it is brought into production use.
Regulatory change control. Every module addition should trigger a formal review of existing approvals, the HACCP plan, sanitation procedures, environmental monitoring, animal welfare provisions, occupational safety arrangements, pressure systems compliance, fire approval, environmental permits, product labels, filed process schedules and export approvals.
Retort process authority. Every canning module requires a competent thermal process authority to establish validated heat distribution, heat penetration, venting procedure, initial product temperature, container orientation, basket configuration, the critical factors of the process, and a documented deviation procedure.
Economic evaluation. Each staged expansion plan should be compared on capital cost, operating cost, utilisation, labour, energy, water, maintenance, lost production during the expansion itself, the risk of stranded capacity, residual equipment value, and the financial value of the option to defer or expand further.
Climate and infrastructure resilience. Ambient temperature, humidity, flooding, drought, water quality, grid instability, fuel supply, road access, telecommunications and extreme weather should be assessed for the specific site, since modular design is of particular value where infrastructure is uncertain, provided each module is itself capable of operating reliably under the conditions expected at that site.
Environmental performance. Water intensity, energy intensity, refrigerant climate impact, heat recovery potential, biogas potential, blood recovery, rendering, odour, noise, effluent load, sludge, packaging waste and greenhouse gas performance should be added to the module data sheet alongside its production figures, following the benchmarking approach documented in the IFC and AMPC material cited above.
Decommissioning and replacement. A genuinely modular plant permits a module to be removed, replaced or repurposed, and the design should define the disconnection procedure, decontamination procedure, structural removal method, utility capping method, data retirement method and the recovery of reusable equipment for every module, not only its construction and commissioning.
Organisational modularity. A physical module still requires a trained operating team, a maintenance owner, a sanitation owner, quality responsibility, a spare parts holding and a standard method of work. A plant that is physically modular can still fail as an operating system if all of the knowledge needed to run any one module rests with a single person or a single central team rather than being distributed and documented.
Standardisation with controlled variation. Interfaces and repeatable equipment families should be standardised across the plant, but process parameters and internal layouts should be allowed to vary where species, product, hygiene requirement, climate or local skill genuinely require it. Modularity is not identical duplication in every internal detail, only a stable, shared interface and a proven, repeatable capability at the module boundary.
Containerisation of non-core and auxiliary functions. Administration and management offices, staff amenities, security posts, retained and suspect product holding, laboratories and environmental monitoring, and new product development or test kitchen functions should be evaluated for delivery as standalone containerised units rather than construction built into the plant’s fabric, following the treatment in section 1.6 above. Each such unit should still be defined by the same module data sheet, interface specification and commissioning boundary required of any other module.

Case in point. JBT Marel’s DeboFlex system, a shackle based, modular deboning line originally developed for pork fore-end deboning and reported by the company to have been adopted by the United Kingdom pork processor Cranswick, is documented by the manufacturer as giving the customer confidence that new additions will integrate smoothly into their production as the line is expanded. This is supplier and company reported evidence rather than an independently audited account of Cranswick’s own commissioning experience, and is a named, traceable instance of equipment sold and adopted explicitly on the basis of its modular addition capability, the equipment level version of the module definition in section 1.4, a standardised, interface bounded unit that a processor can add to without redesigning the line it joins.

Part 9. Approving a new module

A new module should be approved only when six conditions are jointly satisfied.

The market or strategic need for the added capacity has been demonstrated.
The upstream and downstream capacity balance across the rest of the plant remains acceptable following the addition.
Shared utilities and environmental systems remain within their approved capacity envelopes after the addition.
The hygienic and movement interfaces described in Parts 3 and 4 remain fully controlled.
The competent authority and the plant’s own food safety team have accepted the specific change through the regulatory change control process described in Part 8.
The economic evaluation demonstrates that the value of the added capacity and the flexibility it creates exceeds its full life cycle cost, including disruption during construction, validation, ongoing maintenance and the risk of stranded capacity.

Taken together, these six conditions are the five criteria from section 1.3 read as a single test applied at the moment of decision, rather than as five separate design aspirations checked once at the outset and never again.

Part 10. Closing synthesis

A meat plant should be conceived as a planned architecture of repeatable process units, controlled interfaces, staged civil works, and utilities that can be isolated, expanded and commissioned in defined phases, rather than as one indivisible building and one indivisible production system.

The individual disciplines drawn on here, product and process modularity theory following Baldwin and Clark (2000), reconfigurable manufacturing systems following Koren (2010), group technology and cellular manufacturing following Mitrofanov (1966) and Burbidge (1978), livestock flow design following Grandin (1984), shearing layers following Brand (1994) and Duffy (1990), and industrial refrigeration and utility load matching following AMPC’s own published guidebooks, are each well established in their own right, and this paper does not originate any one of them. What this paper attempts, and what no single source located in this research states as a combined method for an abattoir, deboning, fresh meat, processing and canning facility together, is the integration of modular architecture, cellular layout, staged civil works, a multi-category flow architecture, an evaluated refrigeration and utility strategy, containerised delivery of non-core functions, and a defined validation and change control boundary, into one planning method for this specific type of facility.

Part 11. Challenges still to be overcome

Several parts of the plant carry the same weight as the ones treated in full above, and have not yet been given the same treatment in this paper. Each is named here plainly rather than left as a silent gap.

Power supply. The refrigeration architecture in Part 5 assumes a power supply capable of running that architecture continuously, and this paper has not yet set out how that supply itself should be designed. A grid connection, a standby generator set sized and fuelled for the plant’s critical loads, and a solar or hybrid supply for specific containerised or remote modules are three different answers with different capital costs, different fuel logistics, and different failure modes, and the choice interacts directly with the refrigeration architecture decision in section 5.1, since a distributed refrigeration system tolerates a partial power failure differently to a centralised one. Load shedding and grid instability, a live constraint in South Africa and much of the African continent this paper draws case material from, is not a footnote to this question but a design input, and the standby and changeover philosophy set out for refrigeration redundancy in Part 8 should be extended to cover the electrical supply itself, not only the compressors it drives.

Drainage. This paper has treated drainage at the level of a design principle, that underground collection pipework should be sized for the full future footprint from the first pour, and at the level of a movement category, water and drainage as their own flow in Part 3. It has not addressed drainage as its own engineering discipline, pipe sizing and gradient calculations, grease and fat trap placement and sizing, oil and water separation ahead of any effluent treatment stage, backflow prevention devices at every cross connection, manhole and inspection chamber spacing, and the capacity check against local municipal or on-site treatment infrastructure that any actual drainage design requires. This is a genuine gap, not a detail folded into the principle already stated.

Internal material handling. The whole paper assumes product, crates, and reusable equipment move through the plant, but has not set out how. Overhead rail speed and capacity, conveyor selection between belt, roller and chain systems, forklift and pallet truck routes and their interaction with the personnel and product flows in Part 3, automated guided vehicles where volume justifies them, and the vertical handling, hoists, lifts, and dumbwaiters, between a ground floor production area and any first floor office or plant room, are all still open. This interacts directly with the shearing layers argument in Part 6, since a conveyor or rail system sits in the services layer and should be plannable for future change in the same way refrigeration and electrical services are, and this paper has not yet worked that interaction through in the same depth given to refrigeration.

Water supply security. Refrigeration, sanitation, and drainage all assume a secure water supply, and this paper has not addressed where that water comes from, municipal connection, borehole yield and its seasonal variation, or a combination with on-site storage, nor the water treatment step, softening, filtration, or disinfection, that a given source may need before it can be used on product contact surfaces.

Structural and seismic design. The steel frame and staged construction argument in Part 6 assumes a structural design exists to reserve capacity within, but this paper has not addressed wind loading, snow or roof loading where relevant, seismic design category, or the specific structural engineering that would size that frame for a given site and climate.

Security, IT and network infrastructure. Perimeter security, access control at the personnel and vehicle boundaries described in Part 3, and the physical network cabling and wireless infrastructure that the digital and control architecture item in Part 8 assumes already exists, have not been designed here.

Workforce, training, and transport. A modular plant that can be built and expanded in stages still needs a workforce trained to operate each new module to the same standard as the one beside it, and, in many of the rural and remote settings this paper’s case material is drawn from, staff transport to and from the site is itself an operational dependency as real as any utility.

Byproduct and rendering offtake markets. Rendering, blood, and hide handling are treated in Part 2 as a standalone, low integration module, but this paper has not addressed the commercial side of that module, whether a market exists locally for tallow, meat and bone meal, blood meal, or hides, and what happens to that byproduct stream where no such market exists.

Financing and capital structure. Part 1.2 makes the case that modularity has an option value, but this paper has not translated that into an actual financing structure, staged debt drawdown matched to staged construction, equity versus debt mix, or the specific covenants a lender might attach to a phased rather than a single-completion capital project.

None of these gaps changes the argument made in Parts 1 through 10. They are the next layer of detail a specific project would need to add, in roughly the order a feasibility study would normally take them up.

Part 12. Managing the plant as an ongoing construction site

Every part of this paper up to now has treated construction beside a live operation as a physical and civil problem, addressed in Part 6 through shearing layers and a construction segregation plan. A separate and, in practice, larger problem is organisational. A plant that grows in stages over several years is, for that whole period, simultaneously a working factory and a construction site, and the two have different chains of command, different risk tolerances, and different definitions of a good day. Getting the physical segregation right, drainage, structure, and services staged correctly, does not by itself solve the management problem of running both at once.

12.1 Team structure

A minimum of three teams, each with a distinct mandate, is needed from the outset.

A production team, responsible for the plant as it currently stands, with no role in the rollout of new modules beyond the notice and consultation the change control process in Part 8 requires of it.
A modular rollout team, responsible for designing, procuring, and building each new module, with no authority to touch an operating area except through the construction segregation plan and the formal handover process set out in Part 6.
An integration and commissioning team, sitting between the two, whose sole function is to take a completed module from the rollout team, commission and validate it against the module data sheet in Part 8, and hand it to the production team as a going concern. This team is the single controlled gate between construction and operation, and no module should cross from one side to the other except through it.

Beyond these three, a project of this kind needs at least four further, more specialised roles.

A process and equipment design function, held by a specific named process engineer for each module, responsible for line design, machine selection, and the interfaces set out in section 1.4, so that no module’s equipment specification is decided informally on site.
A commercial function, held by the marketing and sales side of the business, responsible for the final product specification each module is built to meet, since a processing line’s design cannot be finalised before the product it is meant to produce is defined.
A procurement and logistics function, since equipment lead times, import and customs processes, and construction material supply are frequently the actual pacing item in a staged build, not the construction work itself.
A regulatory and compliance liaison function, a single point of contact with the competent authority responsible for carrying every module addition through the regulatory change control process in Part 8, so that approval is sought once, by one team, in a consistent form, rather than separately by whichever team happens to be closest to a given module at the time.

12.2 A data driven management layer over independent subunits

Above these teams, a single management layer should hold the overall programme data, current status, cost, schedule, and risk for every module in progress, and should be staffed by people capable of resolving a problem using a shared, documented method rather than by improvisation specific to one team’s habits. This is the same principle behind Koren’s integrability characteristic in section 1.5, applied to people rather than machines. A subunit, whether the rollout team on one module or the production team on one line, should be free to solve its own problems independently, but should reach for the same standard method of diagnosing and resolving a problem that every other subunit uses, so that a lesson learned on one module transfers to the next one without having to be relearned.

12.3 How work in progress is kept from touching operations

This is answered in full by two mechanisms already set out earlier in this paper, brought together here. The construction segregation plan in Part 6 answers the question physically, sealed hoarding, separate access, and a formal release before a new module connects to anything operating. The decision rule in Part 9 answers it organisationally, since a module is not handed to the production team until all six conditions there are satisfied, including that the hygienic and movement interfaces in Parts 3 and 4 remain fully controlled. The integration and commissioning team in section 12.1 is the people who apply both of these checks at every handover, which is why that team, and not the rollout team directly, should be the one with authority to declare a module ready.

12.4 Stages of rollout

A workable sequence, consistent with criterion two in section 1.3, is to commission the smallest complete module first, then add capacity in a fixed order rather than all at once.

Planning and benchmarking, including the visits described in section 12.5, before any construction begins.
The first minimum viable module, commissioned and run long enough to generate real operating data before the next stage starts.
A first parallel or adjoining module, built under the full construction segregation plan while the first module continues operating, used to test and refine the handover process itself while the stakes are still small.
Progressive further modules, each approved individually through the decision rule in Part 9, with the integration and commissioning team retained as a standing function rather than disbanded after the first handover.
A steady state in which rollout capability is kept, at reduced size, as a permanent part of the organisation, since a plant built on this method is never really finished, only at its current stage.

12.5 Role model companies

Several of the companies named earlier in this paper as case material are also the companies this project should actually visit, not only cite. BANSS in Germany, Kometos and its Meatgard installation, DC-Supply and AVS Agric in Denmark, and Marel with its documented Cranswick installation, are all named, traceable operators of exactly the modular or containerised approach this paper argues for, and a supplier’s own marketing claim is not the same thing as having stood in the building and asked its staff how the handover between construction and operation actually went. Mobile Slaughter Technologies in South Africa is a further, closer case, and Eben van Tonder’s existing direct contact with Dr Bergh there is a real advantage, since it is a relationship rather than a cold introduction. Direct, sustained interaction with a small number of these operators, concentrated in the planning phase but continued through implementation, should be treated as a project cost in its own right, not an optional extra, since the single largest risk in a staged build is discovering a handover problem for the first time on your own site rather than having already seen it solved, or badly solved, somewhere else.

12.6 Bringing the scope down to what it costs

Everything in sections 12.1 to 12.5 describes a full, mature version of this system, and built all at once it would be well beyond what most operators could afford or manage. The way scope is brought down to an affordable, executable project is the same mechanism argued for throughout this paper, the full future architecture is decided once, on paper, structural frame, drainage, and team roles all sized for where the plant is eventually going, but only the current stage is actually funded and built. A client does not need three full teams and a regulatory liaison function on day one. A client needs one person who owns the current operation, one person who owns the current build, and one clear rule for how a finished piece of work moves from the second person to the first. Every further team, and every further stage in section 12.4, is added only when the previous stage’s own operating data shows it is needed, which is the same evaluated, one module at a time discipline argued for in Part 9, applied now to the organisation itself rather than only to the building.

The end of this is, and should remain, a simple, manageable strategy. One current operating team. One current build team. One handover rule between them, applied the same way every time. Everything else in this Part is detail that fills in around that simple structure as the plant grows, not a replacement for it.

References

References are ordered by source hierarchy, applicable law and official standards first, primary academic publications and recognised engineering handbooks second, official industry funded research third, official project records fourth, independent case studies fifth, and supplier or trade material last.

Simon, H.A., 1962. “The Architecture of Complexity.” Proceedings of the American Philosophical Society, 106(6), pages 467 to 482.
Baldwin, C.Y. and Clark, K.B., 2000. “Design Rules, Volume 1, The Power of Modularity.” MIT Press, Cambridge, Massachusetts.
Baldwin, C.Y. and Clark, K.B., 2002. “The Option Value of Modularity in Design.” Harvard Business School Working Paper 02-078.
Gamba, A. and Fusari, N., 2009. “Valuing Modularity as a Real Option.” Management Science, 55(11), pages 1877 to 1896.
Mitrofanov, S.P., 1966. “Scientific Principles of Group Technology.” English translation, National Lending Library for Science and Technology. Original Russian edition, 1959.
Burbidge, J.L., 1978. “The Principles of Production Control,” 4th edition. Macdonald and Evans, London.
Koren, Y., 2010. “The Global Manufacturing Revolution, Product-Process-Business Integration and Reconfigurable Systems.” Wiley.
Mehrabi, M.G., Ulsoy, A.G. and Koren, Y., 2000. “Reconfigurable Manufacturing Systems and Their Enabling Technologies.” Journal of Manufacturing Science and Engineering.
Grandin, T., 1984. “Race System for Cattle Slaughter Plants with a 1.5-cm Radius Curve.” Applied Animal Behaviour Science, 13, pages 295 to 299.
Grandin, T., “Recommended Cattle Handling Facility Layouts and Design Guidance.” grandin.com, Colorado State University Department of Animal Science.
Duffy, F., 1990. “Measuring Building Performance.” Facilities, 8(5), pages 17 to 21.
Brand, S., 1994. “How Buildings Learn, What Happens After They’re Built.” Viking Press.
European Parliament and Council, Regulation (EC) No 853/2004, laying down specific hygiene rules for food of animal origin, current consolidated text.
Codex Alimentarius Commission, CXC 58-2005, “Code of Hygienic Practice for Meat.”
North American Meat Institute, “Facility Design Principles.”
North American Meat Institute, “Food Safety Equipment Design Principles.”
United States Department of Agriculture, Food Safety and Inspection Service, “Small and Very Small Plant Guidance,” including current HACCP models.
United States Department of Agriculture, Food Safety and Inspection Service, “Cooking and Stabilization Guidelines for Meat and Poultry Products,” current version.
Iowa State University Extension, “Guide to Designing a Small Red Meat Plant,” 2009.
International Finance Corporation, “Environmental, Health, and Safety Guidelines for Meat Processing,” 2007.
International Finance Corporation, in partnership with the Red Meat Abattoir Association and the Swiss State Secretariat for Economic Affairs, 2020. “Benchmarking Study, Resource Efficiency in Red Meat Abattoirs in South Africa.”
International Finance Corporation, 2020. “Practical Guide for Improving Resource Efficiency in Red Meat Abattoirs in South Africa.”
Australian Meat Processor Corporation, “Refrigeration Plant Energy Improvement Project,” including the Industrial Ammonia Systems guidebooks, Parts 1 and 2, and the New Technology Handbook.
Australian Meat Processor Corporation, “Investigation into Cogeneration Systems for Processing Facilities,” project reference to be confirmed against AMPC’s own published project record.
ASHRAE, “ASHRAE Handbook, Refrigeration,” current edition, relevant compressor, ammonia and refrigerated facility design chapters.
International Institute of Ammonia Refrigeration, technical papers and standards relevant to compressor capacity control, machinery rooms, relief systems, detection and safe ammonia system design.
Stoecker, W.F., 1998. “Industrial Refrigeration Handbook.” McGraw-Hill.
United States Food and Drug Administration, guidance and regulations for thermally processed low acid foods in hermetically sealed containers.
SENASICA, official requirements and certification process for Tipo Inspección Federal establishments, Mexico.
Diario Oficial de la Federación, NOM-008-ZOO-1994 and NOM-009-ZOO-1994, together with current amendments or replacement standards.
Municipality of Querétaro, “Modernización y Equipamiento del Rastro TIF Municipal 2015 a 2018,” official municipal project record.
Government of Pernambuco, Agência de Desenvolvimento Econômico de Pernambuco, and Diário de Pernambuco, official and press record of the Masterboi Canhotinho plant inauguration, 15 August 2022.
Max Rubner-Institut, Institute for Safety and Quality of Meat and International Competence Centre for Meat Quality, relevant project and publication records where process quality, hygiene and meat quality claims are added in future work.
University of Veterinary Medicine Vienna, Austrian slaughterhouse research and official control studies, used for animal welfare, meat inspection and plant hygiene context.
BANSS Schlacht- und Fördertechnik GmbH, product catalogue, supplier documentation, with DDM 120 and DDM 240 identified as discontinuous pig dehairing equipment.
Kometos, “Meatgard, Modular Slaughterhouse,” supplier case documentation.
Friedrich Sailer, modular slaughter facility product documentation, supported by regional trade press reporting.
JBT Marel, “DeboFlex Modular Pork Deboning System,” company product and customer documentation, including the reported Cranswick installation.
Jiangsu Sanye Freezer Equipment Co., Ltd., product and reel documentation on solar-hybrid mobile cold storage and containerised cold rooms.
DC-Supply A/S, in partnership with AVS Agric ApS, “Mobile Slaughterhouse Container,” including the reported Malaysia pork abattoir and Faroe Islands salmon processing container installations.
AES Food Equipment, “Modular Abattoir and Mobile Slaughter Units,” including its stated capability for containerised processing and packaging laboratories.
Mobile Slaughter Technologies, formerly Mobile Abattoirs (Pty) Ltd, South Africa, founded 2015 by Dr Tertius Bergh, company documentation on solar powered mobile slaughter units and containerised pig and poultry abattoir designs.
Structure Magazine, “Shipping Container Design,” including the Obetz Stadium Plaza development, Ohio, and United States military use of shipping containers for offices, kitchens, refrigeration, and food and plant processing.
Norlab Furniture LLC, “Shipping Container Laboratory,” including its stated capability for food and beverage quality control laboratories, cited as supplier documentation.
Container Container, “40ft Laboratory and Product Testing Facility, Shipping Container Conversion,” case study, cited as supplier documentation.
Mid size meat processor, Limpopo province, South Africa, direct professional observation by Eben van Tonder of sustained office and boardroom use of containers many years after original construction.

Sources numbered 35 to 45 are supplier or company documentation, cited for named, traceable installations or commercial configurations rather than independently audited performance data. Reference 46 is a direct professional observation rather than a published source. Where an exact project detail, page reference, or specific figure depends on the underlying official document,

References

Share this: EarthwormExpress Article

Like this: