The foundation review is booked for forty minutes. It runs for two hours and ends by sending the drawing back to the civil designer, because the load table on it belongs to a different container. That failure is not exotic. It happens because a foundation drawing carries four numbers that come from four different documents, and at least one of them is usually stale by the time the concrete is priced.
What the review is approving, and what is outside its boundary
The scope of a foundation review is narrower than it feels. Inside the boundary: the pad or plinth detail, its levelness and flatness tolerances, the anchor positions and their embedment, the cable sleeves and their alignment with the trench, the finished floor elevation against the drainage plan, and the separation to the next unit. Outside the boundary: the container, its internal racking, the electrical protection study, and the suppression system inside the enclosure.
The boundary matters because every item inside it is defined by a document the container supplier holds and the civil designer does not. Anchor positions come from the supplier's anchor plan. Sleeve positions come from the general arrangement drawing. Levelness tolerance comes from the installation manual, not from a concrete code. When the boundary is not written into the review agenda, the meeting approves a shape rather than a load path, and the shape is approved before anyone has confirmed what it has to carry.
The number that stopped the review
The loading table on the drawing showed 21,000 kg. The unit on the order was a 40GP class container with a published net weight of 38,000 kg. Seventeen tonnes is the entire gap between a pad that works and a pad that cracks, and it appeared because the drawing had been priced from the smaller model in the same product line.
Figures are the published datasheet values for the container units described in the linked product pages. Support-point reactions are not published and must be requested separately.
Read the last row carefully, because it is the trap. The 40GP unit is more energy-efficient in kilograms per kilowatt-hour, so a designer comparing the two on specific mass concludes the bigger container is the lighter proposition. In absolute terms it puts 81 percent more mass on the ground. Energy ratings describe the product; they do not describe the load. The published figures for both classes are on the and the , which is why the discrepancy is cheap to catch and expensive to miss.
Net weight is still not the design input. The foundation is sized from the support-point reactions at each leg, combined with the load factors the local code requires for dead, wind and seismic action. ASCE/SEI 7-22 prescribes design loads for those hazards, and ACI CODE-318-25, the current structural concrete code, references ASCE 7-22 for its load combinations. If the reaction table is not in the drawing set, the foundation has been designed for a footprint and not for a load, and no amount of reinforcement added later repairs the missing input.
Three disciplines, three definitions of level
Levelness is the item most likely to be agreed in the room and disputed on site, because three people are using the same word for three different numbers.
The three definitions do not have to be identical, but they have to be reconciled in one document before the pour. If the tolerance is written into the concrete specification rather than taken from the supplier's installation manual, the slab is signed off against the wrong acceptance criterion, and the error surfaces only when the container is set down. If one corner settles more than the others, the enclosure twists, and the sleeve that lines up at the near end is the one that misses at the far end.
Settlement is where this becomes an economic question rather than a technical one. Differential settlement is the dangerous case, which is why the geotechnical report and the supplier's tolerances have to be read together before the reinforcement is detailed. Weak or variable ground is a geotechnical finding, and it has to be reached before the concrete design is finished, because piles and ground improvement are not decisions that can be introduced after the formwork is built.
Where the meeting disagreed
Three items produced real disagreement, and only two of them were resolved in the room.
The first was the support type. A reinforced slab gives continuous support and the easiest tolerance control, and it is the default when the container has a defined footprint. A plinth or pier arrangement pours less concrete and targets the load points directly, and it allows cables to run under the enclosure, but it is unforgiving of layout changes and needs its own corrosion protection and anchor design. The geotechnical report, not preference, decides between them, and if the report indicates weak or variable soils, the conversation moves to piles or ground improvement with an equipment and cost consequence that belongs in the same meeting.
The second was ownership of the tolerance, resolved by assigning it to the supplier's installation manual with the civil specification referencing it.
The third was left open deliberately. Separation between units is usually treated as a fire requirement, but the foundation plan is what fixes it. NFPA 855 in its 2026 edition requires proposed spacing between outdoor enclosures to be analysed using the anticipated wind conditions at the site, since wind governs how released gas disperses, and validated through large-scale fire testing under Section 9.1, with a registered design professional confirming that complete combustion of one enclosure will not propagate to adjacent enclosures. Once the pad layout is poured, that distance is no longer a design variable, so the meeting could not approve the layout and defer the fire analysis at the same time.
What the pad freezes once the concrete is placed
Five decisions stop being reversible at the pour, and each of them is cheap before it and expensive after it.
If the anchor plan arrives after the pour, the anchors are post-installed, and the seismic load path is no longer the one that was analysed. If the finished floor elevation is set from the surrounding grade rather than from the drainage plan, the container sits in the water it was supposed to shed. Levelness and differential settlement behaviour are fixed by the subgrade and the slab detail. Sleeve positions are fixed with the formwork. The separation to the next unit is fixed by the layout, and the fire analysis has to be run against that number rather than the other way round.
That is also what makes quotations incomparable. A price per container says nothing unless it states the design basis and its edition, the geotechnical report it was designed against, the support-point reactions used, the tolerance applied, and whether sleeves, anchor embedment and the as-built survey are inside the scope.
What to require before the next pour
Ask for the reaction table for the ordered unit, at each support point. A supplier who replies with a single net weight figure has answered a different question, and the pad will be designed from a number that describes a product rather than a load.
Ask what levelness and flatness tolerance the pad must meet, and where it is written. The answer that the pad just needs to be level has no acceptance criterion behind it, and the dispute moves to site.
Ask whether the cable entry positions are fixed in the general arrangement drawing and what bend radius applies. An answer that the installer can cut the opening on site means the sleeve positions were never coordinated with the trench.
Ruibit Energy delivered the container project referenced above, and the reaction table for the ordered class is the document that decides the pad, which is why it belongs with the purchase order rather than with the drawing set.
Two items left the review open. The geotechnical interpretation of the differential settlement limit went to the geotechnical engineer, and the spacing confirmation went to the fire protection engineer, both ahead of the reinforcement detail. Nothing was poured until both came back.
FAQs
1. What does a containerized BESS foundation have to carry?
The support-point reactions at each leg, combined with the code-required dead, wind and seismic loads. Net weight alone is not the design input. A published 20GP container unit weighs 21,000 kg and a 40GP unit 38,000 kg, but the pad is sized from each support-point reaction.
2. How level does a BESS container pad have to be?
The tolerance comes from the container supplier's installation manual, not from a concrete code, and it should be written into the civil specification by reference. If it is agreed verbally, the pad is signed off against no acceptance criterion.
3. Why is differential settlement the dangerous case?
Even settlement moves the container down. Differential settlement twists it, which jams doors, strains cable connections and misaligns the sleeves at the far end. The geotechnical report and the supplier's tolerance have to be read together before the reinforcement is detailed.
4. Can a BESS container sit on compacted gravel?
Not as a designed foundation. The loads are concentrated at discrete support points, so settlement and loss of level are the risk. A slab, plinth arrangement or pile foundation designed from the reaction table is the usual answer.
5. What should be requested from the supplier before the pour?
Five documents: support-point reactions per leg, the anchor plan with embedment, the levelness and flatness tolerance, sleeve positions with the cable bend radius, and the container mass and centre of gravity.
6. Why does the foundation plan affect fire compliance?
The layout fixes the separation between units, and NFPA 855 in its 2026 edition requires the proposed spacing to be analysed with site wind conditions and validated by large-scale fire testing. Once the pad is poured, that distance is no longer adjustable.
