Here in South Australia, back at the beginning of the 1990's average household occupancy was less than 3 persons per household and vehicle ownership at 2 vehicles per household, and around 500,000 dwellings. So could assume on average 1 person per household not yet coupled and with zero population growth further construction of houses would drop average household occupancy to 2 persons per houshold. In detail however, there were a large number of 3 bedroom houses occupied by one person. Whilst "state of housing" report indicated around 5% of houses had inadequate number of bedrooms when assessed against a Canadian quality of life index. This basically indicates when siblings of certain ages and genders ought to have separate bedrooms: not essential but desirable.
One vehicle assembly facility in the state could build 100,000 vehicles per year when operating at full capacity. Therefore it had potential to provide private space to 250,000 couples in 2.5 years, and provide all with a vehicle in 5 years. The building industry statistics indicated it oscillated between 5000 dwellings per year and 15,000 dwellings, and was down at less than 10,000. So assuming could be boosted to 10,000, then it would take 25 years, to provide housing for the 250,000 potential couples. Thus expectation that average household occupancy would be 2 persons, by 2015: assuming zero population growth.
However, basically no change has occurred, as average household occupancy is still less than 3 persons per household but not yet 2. This is due to population growth, mostly derived from immigration. This migration mostly fuelled by incompetent managers, who do not appear to realise the job is to maximise the benefit from the available but otherwise limited resources. Thus they are always claiming shortages, add to this employment policies built around keeping the building industry active. Except now claiming we have a shortage of skilled trades people in the building industry.
So it becomes questionable has to whether we want migrants to build the houses, or create a pressure for more houses to keep the building industry active.
I would argue we do not need more people in the building industry. We may need new people to replace those that retire, but we don't need more. Also we do not need to be releasing more land and do not need to be building more family houses. We already have water rationing, the severity of the restrictions vary on an as needs basis. There has to come a point when we say "enough is enough, no more expansion of our cities, this is as big as it gets". The current push is to increase population density of the inner city areas, however, it is still around population growth.
If we still have average household occupancy of 3 persons per household then we still have on average one person in each house not coupled, and if their coupling is creating demand for more housing, then we push towards 2 people per household. To do this we need to construct new houses equal to half the existing housing stock. To expand the radial reach of the city by releasing more land, typically agricultural land, for housing is ridiculous. Expand population and reduce food production land, how is that sensible?
Allotment Size and Building Size
Most of the existing housing stock comprises of 3 bedroom family houses, that is housing suitable for 4 people. Large numbers of these houses are already occupied by one person. The smallest housing site permitted in the old development plans was 81 sq.m in a caravan park, that is a block 9m x 9m. One requirement for fire safety in caravan parks is 3m between caravans/dwellings on neighbouring sites. So would need a site boundary offset of 1.5m, leaving room for a building 6m x 6m. An old reference indicates minimum area for a sole occupancy dwelling to be 25 sq.m, this can be provided by a 5m x 5m dwelling. This in turn can be divided into 2 strip modules 2.5m x 5m. Comparing against shipping container of 2.4m x 6m, the strip modules are transportable: preferably with width reduced from 2.5m to 2.4, even though 2.5m is compatible with vehicle design rules (Australian Design Rules (ADR)).
Given older housing blocks are larger than in new developments, it is possible to subdivide these blocks, for example a block 32m x 21m, can be divided into two blocks. Allowing for a 9m wide access road, each block would be 16m x 12m, having an area of 192 sq.m, excluding the access road. Each dwelling could be suitable for a couple, thus retaining a limit of 4 persons on the block of land. Though even if such block of land has a 4 bedroom house on it, it is still feasible to locate a 9m x 9m site on the block for a sole occupancy unit, which is otherwise suitable for a couple. Assuming existing house occupied by 5 people, and allow another 2 in the added dwelling, then total site occupancy increased to 7.
Population Density
Though blocks of land typically have 3 bedroom houses which may get extended to have more bedrooms. A 3 bedroom house typically for two parents and two children, but may extend to four children sharing bedrooms. If the resource demand of two children taken to be equal to one adult, then have an equivalent household of 3 adults, for typical family of 4, but for family of 6, then equivalent to 4 adults. However, if have more than two children in such house then places extra demand on schools and other childhood services. However, the demand for some of these services can be reduced by added residence used by part-time caretaker.
Also if housing in the vicinity of schools restricted use, and rent only, then need can be concentrated around available services, and housing density can be further increased in these areas by use of 2 to 4 storey apartment blocks. Such apartments being used by older students and teachers. The teachers using such apartments being at either end of the age spectrum, young single teachers, and older couples with no children at home. People raising families, including teachers being in larger detached houses with gardens. Though with school facilities accessible 24 hours per day, the need for gardens can be reduced.
In short we don't need more family dwellings, we need to better manage the housing stock we already have. It seems most of the industrialised world has average occupancy rate less than 3, with some extremely close to 2.
Here in Australia, most of our houses are detached single storey dwellings, and until recently on relatively large blocks of land. Houses are getting bigger and land sizes smaller. Many of these large houses are not practical, and do not meet the long term needs of the occupants, they are also of increasing distance from available services.
If averaging 3 people per household, then constructing a sole occupancy dwelling in the back garden, maintains the allotment occupancy at 3 people, no increase in population density, or increase in demand on services in the area. Rather the third person in each household moves into own dwelling, either at current address or elsewhere. In effect we double the number of dwellings, and average household occupancy drops to 1.5 people per dwelling.
Also hopefully the people constructing these sole occupancy units own the land, and therefore rents can be less, than provided by property speculators who expect others to pay off their mortgage and provide them with a profit.
Production of Modules
As indicated above a single vehicle assembly plant can produce 100,000 vehicles per year, some produce even more. By comparison buildings are trite, most especially in the form of strip modules, or flat panels.
The largest shipping container manufacturer builds 2 million shipping containers per year in 11 factories. That is around 181,818 units per year per factory. A strip module for a dwelling is more complicated than a shipping container, but not has complicated as a vehicle.
As indicated above a sole occupancy dwelling can be provided by 2 strip modules: 2.5m x 5m, or 2.4m x 6m. One strip module being for wet areas containing bathroom, kitchen and laundry. The other providing lounge, dining and sleeping space. If two people couple up, there is no need to replicate the wet area module. Thus two people as a couple only need 3 modules, compared to 4 modules as individuals. A single strip module would be adequate for 2 bedrooms for children. So 4 strip modules can provide a family home. The width of modules can be increased on site with joiner modules and or plain panels.
Also looking at video's of shipping container manufacturing, some of the operations are cumbersome, especially transfer between work stations. Also looking at videos of house manufacturing facilities again some cumbersome operating procedures. A basic strip module is just empty space, with some electrical fixtures (lights/gpo's). The more complicated module is the wet area module with plumbing and electrical fixtures.
So base assumption would be one facility with two production lines can produce 200,000 basic modules per year, and 100,000 wet area modules per year. Which suggests can supply 100,000 dwellings per year, each comprising of 3 strip modules: that is a dwelling suitable for a couple. Alternatively one production line which produces 200,000 basic modules, half of which are converted into wet area modules. So 100,000 dwellings per year comprising of two strip modules.
Now if one vehicle leaves an assembly line at the rate of one every 2 minutes, then in a 480 minute day, produce 240 vehicles, and for 250 productive days in a year, can produce 60,000 units per year. Operating at full capacity is 3 shifts per day, so 180,000 units per year.
Such production does not require a large army of skilled trades people. It does not require electricians or plumbers. If it does then the electrical systems and plumbing systems are poorly designed, and an hinderance to productivity. The electricians and plumbers are needed on site to connect to the mains, and even this can be eliminated with better design of the system connections.
Similarly the assembly of panels and or frames and fabric is not carpentry. Skills maybe required but not the skill sets of people currently in the building industry: their skill set is required for on-site renovation.
Now for a unit of production to leave an assembly line at the rate of 1 unit every 2 minutes, then no operation can take longer than 2 minutes. A roll-forming machine for cold-formed steel can produce at the rate of 5m/minute. That means the bearers along the length of a strip module can be produced in 1 minute.
Dynamic Steel frame video indicates they can assemble the framing for a tiny house on a trailer in 30 minutes. Whilst a video tour of Bailey caravans Part 1 and Part 2, indicates assembly of caravan takes around 16 hours. No indication of rate at which leave the assembly line, which is a different issue.
Starting with the 200,000 units per year, then need 66,667 units for each of 3 shifts, and 266 per day, and 0.55 units/minute, or 1.8 minutes per unit. So no operation can take more than 1.8 minutes, that is no operation on the assembly line.
So if have two roll forming machines, producing the needed 2 bearers, in one minute, and takes 30 minutes to assemble framing, have a problem. Roll forming all the framing components, takes around 1 hour at rough estimate, so this has to be done aside from the main assembly line. So the 30 minutes assembly time as shown in the video, is manually assembly and not optimised, this video show the process more clearly. The workers are running all over the place, no mechanical handling for the wall panels, whilst mechanical handling for roof panels comes from elsewhere in factory, and working at height provided by temporary means. In a factory optimised for such production, timing is likely to be faster, additionally, wouldn't just install wall framing, would have fully clad wall frame panels, and roof panels. So can do more in the 30 minutes than simply assemble a bare skeleton. Whilst the 16 hours for a caravan includes full interior fitout, cabinets, beds, kitchen, bathroom, electrical and plumbing.
One thing noticed with videos of manufactured housing is that the use of "vacuum lifts" is a clumsy and cumbersome process, with time wasted on alignment. Better mechanical handling systems would reduce production time at each workstation.
It noted that many of South Australia's builders of transportable homes, simply construct them outside in a yard in same manner as construct on site. Whilst this avoids the trade people running all over the state, it is still a low quality production process. The building industry does not fully design anything. Full design involves design of product and process. With the design of process involving the design of tooling, jigs and fixtures, minimising handling, minimising process delays and waiting time between processes.
As indicated above the electrical and plumbing systems are poorly designed. For example there should be no stripping of wires and wrapping around terminals: industrial product design would replace these with moulded plugs and sockets and lockable couplings. When it comes to houses, on site electricians and plumbers are largely responsible for design, and that design is dependent on the components available to them. Full system design changes the components, and if components are changed then production processes can be changed.
The building industry is largely based on "standard industry practice", which equates to no one knowing what any one does, but what ever it is it'll be good enough. To increase productivity this is not acceptable. Just look at video's of builders and the time they waste on site, working things out, which should have been worked out before they got to site, preferably by the designers and then checked.
And the its "one-off", customised to the site. Even if the product is designed for the site, it is most likely a concept adapted to fit the site, rather than starting with the site. Either way though, the production process is the same. Whilst builders may manage the process, they don't design the process, the process is left to individual trades. So electricians and plumbers may have to turn up at the site twice, the so called first fix and second fix processes. For example first electrician puts wiring in the wall framing and ceiling space, then when clad, they return and attach fixtures to the wall and ceiling faces. This is imposed by the product design, modify the product, and can modify the process. For example wall cladding panels can be cut out by CNC machines, and then inserted over installed electrical and plumbing fixtures. So only one fix for electrical, though plumbing may still require two fixes.
So consider a box, one floor panel to which 4 wall panels attached and a single roof panel. There are 4 base joints at bottom of walls, 4 joints at top of walls, and 4 corner joints. A total of 12 joints. Assume 12 workstations and 1 minute for the longest joint, then boxes leave line at rate of one every minute. However, that requires 12 partially completed boxes in the production line, and it takes 12 minutes to produce one box.
Consider 3 work stations, one installs two side walls, with temporary bracing. It takes one minute for each wall. If done sequentially it takes two minutes at the work station, if done in parallel then it only takes one minute. Next station, two end walls installed, each wall has one base connection and two corner connections. These are shorter joints so should take less than 1 minute to complete, but assume 1 minute is the limit. So if each wall installed sequentially it takes 6 minutes, if in parallel then 3 minutes. The last work station install the roof, for which have 4 joints, so time is 4 minutes.
So 3 stations with sequential times have [2,6,4] or parallel times [1,3,4]. So if done sequentially the boxes exit the line at the rate of one every 6 minutes, not 4 minutes. The last workstation after completing its 4 minute task has to wait 2 minutes for preceding workstation to complete its task 6 minute task. If have parallel tasks at each workstation then time can drop to 4 minutes.
But also note that total production time for sequential is 12 minutes, but with parallel times it is 8 minutes. The 3 work stations is preferable has have less inventory tied up in the line, but not achieving the rate of one unit per minute. However it is possible to install the roof with all joints in parallel. The two side wall joints in one operation and the two end wall joints in one operation, and these two operations simultaneously. So total time at end station is reduced from 4 minutes to 1 minute, the second work station now has the longest time at 3 minutes.
The two corner joints can be completed in parallel. So 1 minute for corner joints and 1 minute for base joint, drops to 2 minutes. If can do the base joint in parallel with wall corners, then dropped to 1 minute. The station times become [1,1,1] and total time for box is 3 minutes. With boxes exiting the production line at the rate of one every minute.
Assuming bolted joints and maximum length is 6000mm and spacing minimum of 200mm, then have 30 spaces or 31 bolts. The objective is thus installing such bolts in less than 1 minute, so would need to install each bolt in less than 1.91 seconds. Assuming not possible, then increase bolt size and install at 600mm to say 1200mm centres, thus reducing number of bolts and allowing more time per bolt. Thus 5 spaces, requires 6 bolts, so 10 seconds per bolt. Introduce multiple spindle drivers, and can install all bolts along a line in one operation.
Also note with modular construction, that 2 modules 6m long placed end to end, create a 12m wide building, made from 2.4m wide segments. Though 12m long modules can be transported, though may have manoeuvrability problems in some suburban streets, and then at driveways. However if feasible then end to end have 24m wide building. In Australia our timber framing code (AS1684) and simplified wind classification system limits building widths to 16m. So would want to replicate this using other materials and systems. So would need 8m long modules. Assume maximum site width 21m and length 32m. So in width can get maximum length of 2400*8 = 19200, or in length allowing 6m offsets from boundaries, then also limited to 8 modules. So total of 16 modules in one building. Assuming it takes 30 minutes to install a module on site, then will take 480 minutes to install whole building: 16*19.2 = 307.2 sq.m. Which is larger than typical single storey house. Earlier versions of timber framing code had 12m limit on width, however few buildings have such widths let alone the now permitted 16m. Maximum length is 5 times the width, so at 16m can make 80m long, and unlikely to find a block of land suitable for such length. Also development plans typically have large offsets from the front boundary and from the rear boundary, offsets from side boundaries typically around 900mm to 1000mm.
So consider a 3m square grid, and 12 x 12m dwelling. Then have 16 potential rooms. But only need around 8 rooms: kitchen, laundry, bathroom, lounge, dining, 3 bedrooms. So such a dwelling potentially has 8 rooms surplus to requirements. Though some rooms may want larger than 3m x 3m, whilst others are smaller.
Consider 3.6m wide broad loom carpet. So two side by strips and 1.2m hallway between, gives 3.6*2+1.2 = 8.4m width, and put 4 rooms on each side, at say 3.6m long, so 14.4m long, which would require 6 modules, in length, whilst width provide by single module, so total of 6 modules at installation of 30 minutes each, is 180 minutes or 3 hours.
Or consider a popular 4 bedroom floor plan, L-shaped dwelling, 15m legs and each leg 7.5m wide. So one leg take to be 15m, the other 7.5m, so 6.25 modules one leg and 3.125 modules the other leg. So we could round the number of modules down, or up, alternatively change the width to something less than 2.4m. For example adopt 8 modules and 4 modules for each leg, and width of 1875mm. So total of 12 modules at 30 minutes installation each, so 360 minutes or 6 hours.
So assuming modules arrive on schedule then expect the typical house can be assembled in one working day. Also expecting need less than 16 modules, and expecting fabrication of module is less than 30 minutes. So expect all modules fabricated in one day, and delivered and assembled on site the next day. Or for some distant site, the modules all arrive in sequence on the same day.
If expect to roll of the production line at the rate of one every minute, then expecting can produce 480 modules per day, per shift, and if using 16 modules per dwelling then providing 30 dwellings per day.
One example of the potential is BOXABL, so 300,000 sq.ft of factory space (so a 547 ft square, or around 167m square), building a home every 4 hours, aiming for home every 30 to 40 minutes, producing around 6000 and 7000 homes per year and currently with order book for 120,000 homes. Aiming for new factory, with production capacity of home every 10 to 15 minutes. Part of problem here is what is their timing referring to, total production time, or time flowing from the line.
They are no where near 200,000 units per year from one factory. But like the caravan company above, they are producing a fully fitted out box: wet areas, kitchens, bathrooms, and cabinets. Also they are not producing two strip modules, but one expandable unit. So separate the empty box from the fit out, and also consider wet area modules from basic empty box. Then it does seem like there is potential for modular production to hit 200,000 empty units per year.
Related Posts
Revisions:
[3/01/2024] : Original
