Datacenter
Thermal energy storage chilled water tank field guide for data centers
Bank the cold water, hold the thermocline, and prove the room rides through a chiller restart: the chilled-water thermal energy storage tank for the data center.
Direct answer
Thermal energy storage for a data center is a tank of chilled water that holds stored cooling so the room keeps getting cold water through a chiller restart or utility loss. It is the cooling equivalent of the UPS for power. Project specifications, the load, and the required ride-through minutes set the tank size.
Key takeaways
- A chilled-water TES tank banks cold water so the room keeps getting cold water through a chiller restart or utility loss, the cooling equivalent of a UPS.
- Stratification relies on density: cold denser water sits at the bottom, warm return floats on top, separated by a thin thermocline held by low-velocity diffusers.
- A well-built tank holds the thermocline near one meter and delivers 85 to 95 percent of nominal volume as usable cooling; good tanks run a figure of merit above 0.9.
- Tank volume in gallons equals load in tons times ride-through minutes times 24, divided by delta-T and stratification efficiency (0.85 to 0.95); at 20 F delta-T, roughly 72 gallons per ton-hour.
- Acceptance requires a ride-through test under load simulating a chiller trip; a tank that passes only a charge test has not passed.
Thermal energy storage, and what the tank actually holds
Thermal energy storage, or TES, is a tank that banks cooling. In a data center it is almost always a large tank of chilled water that the plant fills with cold water when the chillers have capacity, then draws from when it needs cooling faster than the chillers can make it. The stored cold water is the asset. The tank is just where it lives.
The reason a data center buys one is timing, not energy savings. The IT load makes heat every second it is powered, and it does not pause when the utility blinks. The chillers do pause. A compressor that trips on a power disturbance does not come back instantly, even after the generator picks up, because the machine has its own start sequence and anti-short-cycle timers. That gap, a few minutes, is the problem the tank solves. The stored water keeps cold water flowing to the room until the plant is making cooling again.
Call it what it is. This is the cooling side of the same resilience the UPS gives the electrical side. The UPS carries the IT power through the transfer to generator. The TES tank carries the IT cooling through the same transfer. Lose either one and the room is in trouble within minutes.
Why does a data center need a chilled water tank?
A data center needs a chilled water tank because cooling has to be continuous and the chillers are not. Continuous cooling is the requirement that drives the design. Server inlet air has to stay inside the ASHRAE TC 9.9 envelope, commonly an 18 to 27 C recommended range, and a modern rack drives that air out of range in a short time once the cooling stops. The hotter and denser the load, the shorter that time.
When utility power drops, the building rides on UPS for the IT and waits for the generators to start and accept load. Power transfers in seconds. Mechanical cooling does not. Chillers, pumps, and cooling towers drop out, and the chillers in particular take minutes to restart and reload. For that window the IT is still drawing power off the UPS, still making full heat, with no chiller behind it. The tank is what carries the cooling across that window.
Uptime Institute's outage tracking puts cooling failures at roughly an eighth, about 13 percent, of significant data center outages, second only to power, and the expensive ones run well into six figures. The tank is cheap insurance against being in that statistic. TIA-942 and the project basis of design are where the continuous-cooling and ride-through requirement actually gets written down for a given facility.
The cooling ride-through problem
The ride-through problem is the cooling version of a power outage, and it is easy to underestimate because air feels forgiving. It is not, in a dense room. When the power blips and the chillers stop, the air handlers may keep moving air, but that air is no longer being cooled, so within a minute or two the units are just circulating the room's own rising heat. The IT keeps making heat the whole time.
How fast it climbs depends on the load density and how much thermal mass is in the loop. A lightly loaded room with a lot of chilled water in the pipe rides longer on its own. A high-density room with a tight, low-volume loop heats up fast. The cooling pillar guide covers how the heat leaves the room and where the air and water paths go wrong; the tank is the part that keeps cold water in those paths when the chillers cannot.
The fix is to have enough cold water already stored, at the right temperature, ready to flow the instant the chillers drop. The stored water carries the load for the minutes it takes the chillers to restart on generator power. That is the entire job. Hold the cooling until the plant is back.
What is a stratified chilled water tank?
A stratified chilled water tank stores cold and warm water in the same tank, separated only by their density. Cold water is denser, so it sits at the bottom. Warm return water is lighter and floats on top. This holds across the normal chilled-water range, but water reaches its maximum density near 4 degrees C, about 39 degrees F, so storage taken below that point loses the density gradient and will not stratify the same way, which is one reason standard chilled-water storage stays above it. Between them is a thin transition layer, the thermocline, that keeps the two from mixing. One tank does the work of a separate cold tank and a warm tank, which is why this design won out for chilled water.
Water enters and leaves through diffusers, horizontal manifolds at the top and bottom that spread the flow out to a low velocity. The diffuser is the part that makes stratification possible. Dump water into the tank through an open pipe and the jet stirs everything into one lukewarm mass with no usable cooling. Spread it gently across the full cross-section and buoyancy keeps the layers intact.
The tank stays full of water at all times. It does not empty and refill. What moves is the thermocline. Charge the tank and the band of cold water at the bottom grows and the thermocline rises. Discharge it and the cold zone shrinks and the thermocline falls. The total volume never changes. Only the split between cold and warm does.
What is a thermocline?
The thermocline is the thin layer where the temperature swings from cold storage at the bottom to warm return at the top. It is the dividing line that lets one tank hold both. The thinner you keep it, the more of the tank is genuinely cold water you can use, and the closer the tank performs to its rated capacity.
A thick or smeared thermocline is wasted tank. The water inside the transition band is neither cold enough to use nor warm enough to count as discharged, so it is dead volume. A well-built tank holds the thermocline to roughly a meter and delivers 85 to 95 percent of its nominal volume as usable cooling. A poorly designed or abused tank can fall to 60 to 70 percent, which means you paid for a tank a third bigger than what you get.
Engineers score this with the half-cycle figure of merit, a number that compares the cooling you actually got out against the ideal for a perfectly stratified tank. Good tanks run a figure of merit above 0.9. The thermocline thickness and that figure of merit are the two numbers a commissioning agent watches, because together they say whether the tank will deliver its rated ton-hours or quietly come up short on the day it matters.
Charging and discharging the tank
Charging and discharging are the same flow path run in opposite directions. Charging puts cold water in. The chillers send cold supply water to the bottom diffuser, and an equal volume of warm water leaves the top. The cold zone grows from the bottom and the thermocline rises toward the top. The plant charges when it has spare capacity, which in a data center usually means whenever the load is below the chillers' output, and the tank is held full and ready.
Discharging takes cold water out. Cold water leaves the bottom diffuser to feed the cooling loop, and warm return water comes back into the top. The cold zone shrinks and the thermocline falls. This is the mode that matters in an emergency. The instant the chillers drop, the tank starts feeding the loop from its cold bottom while the warm return stacks on top.
Flow rate is where stratification gets won or lost. Push water too fast through the diffusers and the jet velocity rises enough to stir the layers and thicken the thermocline. The diffusers are sized for a flow that stays below that mixing threshold, and the control valves are slow-acting on purpose so a fast valve stroke does not slam flow through the tank and churn the layers. Slow and gentle keeps the cold water cold.
How big does a TES tank need to be?
A TES tank is sized to carry the cooling load for the required ride-through time at the system delta-T, plus a margin for the volume that is not usable. Three inputs set it: the load in tons, the ride-through minutes, and the delta-T the tank works across. A fourth, the stratification efficiency, accounts for the dead volume in the thermocline and at the tank ends.
Start from the stored cooling. A gallon of water gives up about 8.33 BTU for every degree F it warms, and a ton-hour is 12,000 BTU, so the usable ton-hours in a tank are the volume in gallons times 8.33 times the delta-T, divided by 12,000. Rearranged, one gallon stores about the delta-T divided by 1441 ton-hours. At a 20 degree F delta-T that is roughly 72 gallons per ton-hour. At a weak 10 degree F delta-T it doubles to about 144.
To go straight to a volume, multiply the load in tons by the ride-through minutes by 24, then divide by the delta-T and the stratification efficiency. A 1,000 ton load for 15 minutes at a 16 degree F delta-T and 0.9 efficiency works out near 25,000 gallons of usable storage, and the tank itself is built larger than that to leave the thermocline and the diffuser zones out of the usable count. These numbers are starting math. The project basis of design and the tank designer set the final size.
TH = (Vgal × 8.33 × ΔT) / 12,000Vgal = (Qtons × tmin × 24) / (ΔT × η)V / TH = 1441 / ΔT- V
- Stored water volume in gallons, the usable cold volume the tank delivers
- Q
- Cooling load in tons of refrigeration the tank has to carry during the ride-through
- t
- Required ride-through time in minutes, from the project basis of design
- delta-T
- Temperature difference in degrees F between the warm return and the cold supply
- eta
- Stratification efficiency, the usable fraction of the tank volume, commonly 0.85 to 0.95
Delta-T, capacity, and low delta-T syndrome
The delta-T is the lever that decides how much cooling a gallon holds, and it is the one most often given away on a real plant. Stored cooling is volume times delta-T, so a tank that runs at half its design delta-T stores half the cooling, full stop. You can build a beautiful tank and starve it by feeding it a weak temperature difference.
Low delta-T syndrome is the chronic version of this. The return water comes back colder than design, usually from coils and air handlers set up for more flow than they need, three-way valves left in mixing positions, or a loop that short-circuits supply into return. A plant designed for a 16 degree F delta-T that actually runs at 10 has lost more than a third of its tank before anyone looks at the tank. The ride-through you sized for 15 minutes now ends at 9.
This is why the TES conversation and the chilled-water delta-T conversation are the same conversation. Chase the delta-T at the coils and you protect the tank's capacity for free. Ignore it and no amount of tank volume buys back the minutes, because the water leaving the tank is doing only part of the work each gallon was supposed to do.
Ride-through resilience vs load-shifting economics
TES does two different jobs, and a data center buys it for one of them. Load-shifting is the commercial and district-cooling use: charge the tank at night on cheap off-peak power, discharge it during the expensive afternoon peak, and shrink the chiller plant because it no longer has to meet the peak head-on. The payback is in the energy bill and the demand charge.
Resilience is the data center use. The tank is there to ride through a chiller restart or a utility event, and it is sized in minutes, not hours. A facility may run its tank inline, with the chilled water passing through the tank on the way to the load so it is always charged and instantly available for an emergency discharge, rather than cycled deeply every day. The value is uptime, not the power bill.
The two are not mutually exclusive. A site on a utility with steep time-of-use rates or demand charges can size the tank to do both, taking the off-peak charging economics on normal days and keeping the ride-through capacity in reserve for the bad one. But when you have to pick, the data center picks resilience, and the sizing follows the ride-through minutes, not the rate schedule. Cross-check the cooling pillar guide for where this sits in the overall plant.
Tying the tank into the chilled-water loop
The tank lives in the chilled-water loop, and how it ties in decides how fast it can take over. The common arrangement puts the tank between the chillers and the load with its own pumps and a set of control valves that decide whether the plant is charging the tank, discharging it, or running straight through. The building management system runs those valves off the plant state and the tank temperature sensors.
The inline scheme is popular in data centers because it keeps the tank continuously charged and ready. Chilled water passes through the tank on its way to the load during normal operation, so the cold zone is always topped up and the tank can discharge the instant the chillers drop without waiting for a mode change. The trade is that the tank sees daily flow and the controls have to protect the thermocline against that everyday traffic.
Sensor placement is not an afterthought. A string of temperature sensors up the height of the tank is how the controls and the operators see the thermocline, know the state of charge, and confirm a discharge is actually happening. Without that string you are flying blind on the one asset that has to work during an outage. Tie those points into the BMS and the alarms, not just a local gauge.
Controls and the failover sequence
The failover sequence is the whole point of the tank, and it has to be automatic and fast. On a chiller trip or a loss of normal power, the controls switch the tank to discharge before the loop loses its cold water. The cold supply at the bottom of the tank feeds the loop, the secondary pumps keep the water moving to the room, and the warm return stacks back into the top. None of that can wait for an operator to notice and throw a valve.
Timing is the design problem. The valves are slow-acting to protect the thermocline, but the switch to discharge has to happen fast enough that inlet temperatures never leave the envelope. Reconciling those two is real engineering. The loop has to stay fed during the seconds the valves are stroking, often by keeping the tank already in the flow path so there is no dead time. The chiller-restart sequence runs in parallel, with the machines coming back on generator power and reloading while the tank covers the gap.
The handoff back matters too. When the chillers are making cooling again, the controls return the tank to charging so it is ready for the next event. A tank that discharged and never recharged is a tank that will not be there the second time. The sequence of operations should spell out the trip response, the hold, and the recharge, and commissioning should prove all three.
Ice storage vs chilled water storage
Ice storage and chilled water storage solve the same problem with different physics. Chilled water stores cooling as sensible heat, about one BTU per pound per degree F, so it leans on volume. Ice stores cooling as the latent heat of fusion, roughly 144 BTU per pound to melt, so it packs far more cooling into far less space. Ice systems run around 2 to 4 cubic feet per ton-hour against roughly 15 for chilled water, so an ice tank can be a fraction of the size for the same capacity.
For a data center the usual answer is still chilled water, and the reason is simplicity. The stored medium is the same water already in the loop, the tank has no heat exchangers or refrigerant-side complexity, and discharge is just pumping cold water. Ice buys its density with lower chiller efficiency, because making ice means running the chillers at a colder, less efficient suction temperature, plus the added machinery to freeze and melt it.
Footprint is what flips the decision. When the site cannot fit a chilled water tank, which for big loads runs to 30 ft tall and thousands of square feet, ice or a phase-change material earns its complexity. Most data centers with room for the tank take the simpler chilled water route and live with the size.
Tank construction, insulation, and diffusers
The tank itself is a welded steel or a concrete structure, sized to the volume and built to hold a tall column of water. Field-welded steel tanks are common, and large installations use prestressed concrete. Either way, the build details that decide whether the tank performs are on the inside and the outside, not the shell.
Inside are the diffusers, the top and bottom manifolds that spread flow to a low velocity, and the temperature sensor string. Get the diffuser geometry and the inlet velocity right and the tank stratifies. Get them wrong and it mixes no matter how big it is. The water gets a treatment program before it ever rides through anything, because this is a large volume of water that mostly sits still.
Outside is insulation, and it is not optional on a chilled tank. Skip it or skimp it and the tank gains heat from ambient, the cold water warms toward the room, and the stored cooling bleeds away while you wait for the event that may never come that day. The shell gets a hydrostatic test to prove it holds before it is insulated and filled for service, the same fill, pressurize, hold, and document discipline as the rest of the chilled-water piping. The chilled-water hydro test guide covers how that package is built and witnessed.
How is a chilled water TES tank commissioned?
Commissioning a TES tank proves it stores and delivers the cooling it was sold to deliver, and it ends with a test under load. The sequence is fill and treat the water, confirm the instrumentation reads true up the height of the tank, then run the tank through its modes and measure what comes out.
The stratification test charges the tank and watches the thermocline form and stay thin. The sensor string shows the cold zone building from the bottom with a sharp transition on top. A smeared profile means a diffuser or flow problem to fix before anything else. Then full charge and discharge cycles at design flow and off-design flow confirm the thermocline holds across the operating range and let you compute the half-cycle figure of merit and the usable capacity.
The acceptance test is the ride-through. Simulate a chiller trip with the room under real or artificial load, let the controls switch the tank to discharge, and prove the cooling holds for the rated minutes with inlet temperatures inside the envelope. This is also where capacity gets verified for real. Count the ton-hours actually delivered and the minutes the tank carried the load, and compare them to the rated numbers. A tank that charges fine but cannot prove its ride-through under load has not passed, no matter how good the paperwork looks.
The water inside the tank: treatment and freeze
The water sitting in a TES tank is a large, mostly stagnant volume, and stagnant water grows things. Without treatment a tank becomes a reservoir for biological growth and corrosion that fouls the diffusers, loads the loop with debris, and eats the steel. The treatment program is part of the design, not an operations afterthought, and it has to suit a volume that turns over slowly.
The tank also needs the loop's water chemistry to stay matched, because the tank is a big slug of water that can dilute or shift treatment when it cycles. An owner who lets the treatment lapse finds out through a fouled diffuser or a corrosion problem years later, long after the commissioning agent is gone.
Outdoor tanks add freeze risk. A chilled-water tank holds cold water by design, and in a cold climate an outdoor or partially exposed tank, its piping, and its instrument lines can freeze when they sit without flow or heat trace. The insulation, heat tracing on the vulnerable lines, and the freeze-protection design get checked the same way the rest of the cold-side work does. This is a topic to confirm against the project's climate and the tank designer, not to assume.
AI and high-density loads shorten the clock
High-density and AI compute change the ride-through math by shrinking the clock. A traditional rack at a few kW gives the room some grace when cooling stops. An AI training rack at tens of kW does not. The same loss of cooling drives a dense rack out of its temperature envelope far faster, because the heat per unit of air and water in the loop is much higher and there is less thermal mass relative to the load.
That cuts both ways on the tank. The minutes of ride-through you need may be shorter because the chillers and controls have to react faster regardless, but the tank has to deliver its cooling faster, at a higher flow, the instant the load is uncovered. A tank sized in volume for an old load but plumbed for a slow discharge will not move cold water fast enough for a dense room.
Liquid-cooled high-density gear pulls the storage closer to the chip and raises the premium on a high delta-T, since the same tank does more work per gallon when the water comes back genuinely warm. The denser the load gets, the less forgiving the cooling becomes, and the more the tank's discharge rate, not just its volume, decides whether the room rides through.
The maintenance the owner takes on
The tank does not run itself after turnover, and the maintenance is the kind that gets skipped because the tank looks fine until the day it does not. Water treatment is first. The program that kept the tank clean during commissioning has to keep running, with the chemistry checked and dosed on a schedule, or the stagnant volume goes bad on its own timeline.
The diffusers and the sensor string need to stay clean and calibrated. A fouled diffuser thickens the thermocline and quietly erodes usable capacity. A drifted sensor string lies to the controls about the state of charge. The insulation has to stay intact, because a breached or wet insulation jacket bleeds cooling and can hide corrosion underneath.
The item owners most often drop is the test. A TES tank that is never re-tested under load after the first year is a tank nobody actually knows still works. The ride-through capacity degrades as the delta-T drifts, the thermocline smears, and the water fouls, and the only way to know the tank still carries its rated minutes is to run the discharge test again on a schedule. Put it on the preventive maintenance calendar with the generators and the UPS, because it is the same class of asset.
What to document
The record on a TES tank is what tells the next engineer whether the cooling will hold, and the numbers that matter are the rated ones and the measured ones side by side. Capture the tank's volume and type, the delta-T it was designed and tested at, the usable ton-hours, the ride-through minutes proven under load, and the thermocline performance, along with the treatment program and the test date so the owner knows when it was last verified.
| Field to record | Why it matters |
|---|---|
| Tank type and volume | Sets the nominal storage and the footprint |
| Design and measured delta-T | Stored cooling is volume times delta-T |
| Usable ton-hours | Nominal volume minus the dead thermocline and end zones |
| Ride-through minutes at load | The number the tank exists to deliver |
| Thermocline thickness and figure of merit | Says how much of the tank is really usable |
| Discharge and failover response time | Proves the cooling switches over fast enough |
| Water treatment program and last test date | Tells the owner what to maintain and when it was verified |
Common mistakes
- Sizing the tank for the load but not for the ride-through minutes the basis of design requires.
- Designing for a high delta-T the plant never achieves, so the tank stores less cooling than its rating.
- Letting low delta-T at the coils quietly shrink the usable ton-hours below the rated ride-through.
- Pushing flow through the diffusers too fast and smearing the thermocline into dead volume.
- Leaving the discharge on a slow or manual sequence so the loop loses cold water before the tank takes over.
- Skipping the ride-through test under load and accepting the tank on a charge test alone.
- Neglecting water treatment on a large stagnant volume until the diffusers foul or the steel corrodes.
- Under-insulating the tank so stored cooling bleeds off to ambient before it is ever needed.
Field checklist
Want this checklist to run itself on every job — with photo proof and a signed record crews can hand the customer? That's FieldOS.
Standards and references
The TES design framework comes from ASHRAE, whose handbook material on thermal storage covers the stratified-tank approach, diffuser design, and the sizing logic. The project's mechanical engineer applies it to the specific plant. There is no single code number that mandates a tank, so treat ASHRAE as the design reference and the project basis of design as the document that actually sets the capacity and the ride-through.
The continuous-cooling and ride-through requirement, where it is formalized, lives in the data center infrastructure standards. TIA-942 sets infrastructure requirements by topic and aligns its thermal envelopes with ASHRAE TC 9.9, and the Uptime Institute Tier framework addresses cooling continuity as part of redundancy and concurrent maintainability. Which level of ride-through a given facility owes is a function of its Tier or rating target and its own basis of design, so confirm the requirement against the project documents rather than a rule of thumb.
The tank itself is built and proven against the pressure-vessel and tank standards for its construction type, the manufacturer's diffuser and stratification design, and the commissioning plan's acceptance criteria. The hydrostatic proof of the shell and the connected piping follows the chilled-water pressure-test discipline. Cite the document that controls the specific point, and let the project specification and the tank designer govern the numbers.
Units, terms, and conversions
TES carries its own vocabulary, and the same quantity shows up in a few units across a drawing set and a manufacturer's submittal. Stored cooling is in ton-hours in the US and in kWh thermal or kJ in metric sources. Tank size is in gallons or cubic feet here and cubic meters elsewhere. Delta-T is in degrees F on US plants and degrees C on metric ones, and the temperature envelope for the room is almost always quoted in degrees C from ASHRAE TC 9.9.
- TES
- Thermal energy storage, here a chilled water tank that banks cooling capacity
- Ton-hour
- 12,000 BTU of cooling, one ton of refrigeration delivered for one hour
- Delta-T
- The temperature difference between the warm return and the cold supply the tank works across
- Thermocline
- The thin transition layer between the cold stored water and the warm return
- Stratification efficiency
- The fraction of the tank volume that is usable cooling rather than dead thermocline and end zones
- Ride-through
- The minutes of cooling the tank carries while the chillers restart or generators load
- Figure of merit
- A score comparing delivered cooling to the ideal for a perfectly stratified tank
- Diffuser
- The top and bottom manifolds that spread flow to a low velocity to preserve stratification
FAQ
What is thermal energy storage in a data center?
Thermal energy storage in a data center is a large insulated tank of chilled water that banks cooling capacity. Chillers charge it when they have spare capacity, and it discharges chilled water to the cooling loop when the chillers trip or restart, holding server inlet temperatures while generators and chillers come back.
Why does a data center need a chilled water tank?
A data center needs a chilled water tank because the IT load keeps making heat the instant power blips, but the chillers stop and take minutes to restart on generator. The stored cold water carries the cooling load through that gap so server inlet temperatures stay inside the ASHRAE envelope.
What is a thermocline?
A thermocline is the thin transition layer between the cold water at the bottom of a stratified storage tank and the warm return at the top. Keeping it thin is the whole game. A sharp thermocline means most of the tank volume is usable cold water, often 85 to 95 percent.
How big does a TES tank need to be?
A TES tank is sized to carry the cooling load for the required ride-through minutes at the system delta-T. As a rough figure, one gallon stores about delta-T divided by 1441 ton-hours, so a 20 degree F delta-T needs roughly 72 gallons per ton-hour, before stratification losses.
Ice storage or chilled water: which is better for a data center?
Chilled water storage is the common data center choice because the tank and controls are simpler and the water is already the cooling medium. Ice stores roughly four to eight times more cooling per cubic foot, so it wins on footprint, but it adds complexity and lowers chiller efficiency.
How many minutes of cooling ride-through should a data center have?
Most data centers target 10 to 15 minutes of cooling ride-through, enough to cover a chiller restart after a utility event or generators picking up full load. Mission-critical sites spec 30 minutes or more, and less redundant edge sites may accept 5 to 7. The project basis of design controls the number.
What happens if a TES tank fails to discharge on a chiller trip?
If the tank does not switch to discharge the instant the chillers trip, the cooling loop loses its flow of cold water while the IT load keeps producing heat, and inlet temperatures climb within minutes. This is why the discharge valve sequence is automatic, fast, and tested under load, not left to an operator.
What is low delta-T syndrome and why does it waste a TES tank?
Low delta-T syndrome is when the return water comes back colder than design, shrinking the temperature difference the tank works across. Since stored cooling is volume times delta-T, a tank sized for 16 degrees F but running at 10 delivers far fewer ton-hours and falls short of its rated ride-through.
How is a chilled water TES tank commissioned?
Commissioning a chilled water TES tank runs the fill and water treatment, a stratification check, full charge and discharge cycles at design and off-design flow, and a ride-through test that simulates a chiller trip under real or artificial load. The acceptance proof is that cooling holds for the rated minutes.
People also ask
Codes cited in this guide
This guide is written and reviewed against the published standards below. Always confirm the current adopted edition with the authority having jurisdiction.