A spare part deserves priority when its absence could prevent the plant from restoring an important function within an acceptable time. Price and consumption help determine how to manage a part number, but they do not answer that question on their own. A useful decision connects the part to its equipment, the consequence of failure, replenishment time, and alternatives that are actually available.
Asset criticality and part criticality are related, but they are not the same classification. An asset may have redundant, repairable, or interchangeable components. A low-cost part may still be the only item blocking recovery of an essential function.
The U.S. Department of Energy Office of Inspector General (DOE OIG) describes a critical spare as an asset or part needed to return a critical asset to service after wear or plausible failures, including cases where there was not enough warning to buy it in time (DOE OIG, p. 2). That definition comes from the Western Area Power Administration (WAPA) context.
Use it as a prompt to ask how recovery works, not as a legal definition for plants elsewhere.
The practical task is to document the link between the asset and the spare, then agree on an availability policy with maintenance, reliability, operations, and stores. The result should reflect the plant’s evidence and constraints. There is no single formula that can make that decision for every site.
The DOE OIG audit of WAPA’s critical-spares program examined whether the organization could link parts to critical assets, identify them in inventory, set stock levels and reorder points, and keep them in suitable locations. The report concerns a U.S. public agency. It offers useful control questions, but it does not set requirements for plants in other jurisdictions (DOE OIG, report DOE-OIG-24-30, pp. 2-4).
Asset criticality alone cannot choose the spare
An asset ranking helps a team decide where to focus. An inventory decision requires another step: identify which parts support the function and what happens if each one is missing. Components on the same asset can have very different effects. A process pump may be critical yet have an installed standby unit; a specific seal may be the item that extends an outage when no compatible alternative exists.
Start by describing the function and the loss scenario. What service does the equipment provide? Which failure mode could stop or limit it? Would an outage affect production, product quality, worker safety, the environment, compliance, or shared services? This stage does not set stock quantities. It defines the consequence the team is trying to manage and separates a possible hazard from a consequence the plant has assessed.
Identify the part well enough for another person to find it and check compatibility. That may include the manufacturer, part number, revision, material, connections, certification requirements, and equipment where it can be installed. “Spare motor” is not enough when several ratings or configurations exist. The DOE OIG found that incomplete specifications made it harder to identify the parts needed by critical assets.
Its audit points to the need to record both the equipment relationship and the part’s own identity (DOE OIG, pp. 2-3).
Criticality also depends on the function and operating context. Pacific Northwest National Laboratory’s reliability-centered maintenance framework groups systems by the consequences of failure, with categories that distinguish risks to life, health, and safety from operational continuity and mission support. It is a way to prioritize maintenance analysis, not an inventory scale a company should copy without adaptation (PNNL, “An Advanced Maintenance Approach: Reliability Centered Maintenance”).
A local matrix can preserve several dimensions instead of hiding them behind one label. “Severe consequence if missing,” “long supply lead time,” and “infrequent demand” are different observations. A final category may summarize the decision, but retaining the reasons makes it possible to revisit it when redundancy, the process, or the supplier changes.
A 2012 study proposed a multi-criteria spare-parts classification and tested it at a petrochemical plant; it does not offer a set of weights that fits every facility (spare-parts classification study).
Assess what happens when the part is missing and how the function recovers
Describe the failure mode and affected function first. Then check which barriers, redundancies, or alternative equipment would remain available. To estimate recovery time, account for the steps that actually take time: diagnosis, authorization, access, repair, testing, and return to service. This sequence can organize the decision, but it does not replace the plant’s risk analysis or procedures.
The consequence of a shortage is not limited to lost production. A missing part could affect a safety function, a quality specification, a shared service, or an operating obligation. If the decision touches a safety function or regulated equipment, follow the applicable engineering and safety process. A spare-parts management article cannot declare it acceptable to operate without a part, remove a safeguard, or change an inspection interval.
Check redundancy under the conditions in which the team would need it. A duplicate shown on a drawing may be unavailable if both units share power, controls, cooling, or a discharge line. A “common” part may need adaptation, calibration, or approval before installation. Record what is known and what still needs confirmation. Do not count an item as an alternative when compatibility is only assumed.
The U.S. Department of Energy Office of Enterprise Assessments’ maintenance review of Los Alamos National Laboratory describes a nuclear facility that set minimum and maximum levels for critical spares and checked stock against them. It also records problems with missing quality documentation for stored parts and obsolescence among compatible components. Those are lessons from one facility with nuclear controls.
For other plants, they are prompts about traceability and lifecycle, not recommended quantities (DOE Office of Enterprise Assessments, Los Alamos report, pp. 12-14).
Recovery difficulty can affect the decision even when failure is rare. A standard bearing available locally presents a different problem from a custom rotor, a discontinued card, or an assembly that needs off-site repair. A 1998 study models a policy that reserves some stock for demand associated with critical equipment.
Its results depend on mathematical assumptions; it presents one possible policy, not an instruction for configuring a warehouse (equipment-criticality stocking policy study).
Historical issues are evidence, not a verdict. A part may go years without a recorded issue and still be needed for a rare failure with serious consequences. A zero in the system can also mean that technicians withdrew the part under another code, took it from a kit, did not record the movement, or swapped a component during repair without updating the inventory record.
Check how transactions are recorded and what period the history covers before treating zero demand as proof that a part is unnecessary.
Lead time also needs a clear definition. Separate the supplier’s quoted time from the time observed between order and receipt, and from the time until the part is available for a work order. Include approval, import, incoming inspection, and document-release waits when they apply. If the figure is an estimate and the plant has little delivery history, record that uncertainty instead of presenting the quote as a guaranteed arrival date.
A decision record should show who confirms each assumption: maintenance or engineering for function and compatibility, purchasing for suppliers and lead time, stores for location and condition, and operations for process consequences. If people do not agree, record the uncertainty and send the decision to the responsible authority. Calling a part “critical” does not turn an unverified recovery assumption into an approved condition.
Also check how many viable sources exist, whether there is a sole original-equipment manufacturer, what a supplier change would involve, and whether another site in the group could lend a unit. “Available on the market” does not mean “ready to install.” Confirm technical equivalence, origin, specification, quality requirements, and approvals.
The Los Alamos report shows how documentation and obsolescence can affect the usefulness of stored parts; the DOE OIG also notes that WAPA could not demonstrate strategic locations for its spares (DOE Los Alamos, pp. 12-14; DOE OIG, pp. 2-4).
Finally, consider how a part ages in storage. Rubber goods, batteries, lubricants, electronic components, and calibrated equipment may have storage conditions, expiry dates, or service lives that need tracking. Parts do not all degrade in the same way, and a manufacturer’s guidance does not automatically set how many units to buy.
The responsible team should check the relevant documentation and storage conditions, then consider the cost of preserving a spare or maintaining a repairable unit.
ABC gives an economic or workload view, depending on the variable used to build the classification. It does not measure technical criticality by itself. A cheap part can have severe consequences when missing; an expensive one may be common, have a substitute, or arrive quickly.
The 2025 study combines ABC with criticality and supply criteria, while the 2012 classification study uses several attributes to distinguish policies. Both support keeping the criteria visible before combining them (2025 plastics-industry study; 2012 spare-parts classification study).
Turn the classification into an availability policy
Once the team has described operational risk and logistics, it can agree how to cover the need. Keeping a unit in stores is one option. Others include a repairable part with a known return cycle, an exchange with a supplier, a unit shared across sites, an availability contract, or a purchase tied to planned maintenance. Each option needs an owner, activation conditions, and a response time the team has checked.
The choice should fit the function and available evidence. For a common part with a short lead time and verified alternatives, a team may decide not to tie up another unit in inventory. For a difficult-to-manufacture part linked to a serious consequence, it may choose to hold one, prioritize repair, or secure an exchange source. These are qualitative examples, not recommendations for a real asset.
The plant authority must consider local obligations, specifications, and consequences before approving a policy.
If the plant stocks the part, define which quantities count: available, reserved for a work order, sent for repair, in transit, blocked by quality, or assigned to another site. In its WAPA audit, the DOE OIG found differences between inventory records and physical counts, as well as parts that were not always linked to assets.
A system balance should therefore be checked against the part’s location and condition. That audit result is not an expected discrepancy rate for other warehouses (DOE OIG, pp. 3-5).
There is no universal safety-stock quantity for every critical spare. Demand may be intermittent, lead time may vary, a part may be repairable, and the consequence of a shortage depends on the equipment. Inventory models depend on their assumptions and the policy selected.
This guide leaves general safety-stock formulas to the English resources index, where readers can find quantitative guidance on demand data and its limits. Decide why the part is critical before applying a numerical method, and state which uncertainty the calculation is meant to cover.
The 2025 study examined 3,842 part numbers at a plastics-processing company. The authors classified 69% as having erratic demand, and 39% had no demand recorded in the catalog. They combined ABC with criticality, supply, and maintenance criteria to set policies. Those percentages describe one company’s catalog; they are not a benchmark for another plant (2025 plastics-industry study).
An availability policy should name the agreed option, the applicable level or rule, who replenishes it, where it is held, its preservation requirements, and the response time. For a repairable unit, record whether the failed part returns, who reviews the repair report, and how long the work usually takes.
If the policy depends on a contract, a sister site, or a shared reserve, include a way to confirm that the source is still available when needed.
The cited studies examine different policies. The 2025 paper applies periodic review and rules adapted to the demand stability observed at one plastics company. The 1998 study models reserving some units for critical demand. Neither gives a replenishment level that can be transferred without checking demand, lead time, repair options, and local consequences (2025 plastics-industry study; equipment-criticality stocking policy study).
Keep the record useful and review exceptions
A list only helps if people can act on it. Link each spare to the compatible assets and keep a specification that stores staff can check when receiving or issuing a unit. Record its physical location, available quantity, condition, inspection or expiry dates where relevant, and storage restrictions. Preserve the reason for the decision too: shortage consequence, supply source, approved alternative, and review date.
Keeping the reason is more useful than filling in a category with no explanation. “Critical for safety” should say which function or assessment supports the label and who can confirm it. “No alternative” should include the search or supplier response and its date. “Long lead time” should identify whether the figure is an estimate or based on observed deliveries.
That record helps the next review distinguish a confirmed fact from an open decision.
Data ownership is usually shared. Maintenance may identify the component and failure mode; engineering can validate the specification; purchasing can maintain supplier and lead-time information; stores can track balance and location; operations can describe the consequence of a shortage. Assign these roles to fit the organization. The DOE OIG found that incomplete fields, separate records, and a missing tracking procedure made WAPA’s spares harder to control (DOE OIG, pp. 2-5).
Set a regular review and reopen the decision when an important assumption changes. Possible triggers include an equipment modification, an unexpected failure, removal of redundancy, a new supplier, a lead time that is no longer met, an end-of-life notice, or a delivered part with a different specification. Choose the review interval to fit risk, catalog size, and team capacity.
Do not treat the schedule in the DOE OIG audit as a required calendar for other jurisdictions. The Los Alamos report describes monthly minimum and maximum reviews in a specific regulated setting (DOE OIG, pp. 2-4; DOE Los Alamos, pp. 12-14).
When the list is large, a pilot can show whether the required data exist before the team scales up its method. Choose a bounded equipment family that maintenance, stores, and purchasing can review together. Track how many part numbers lack a clear asset link, how many have an unreliable location, and how many decisions depend on missing information. These measures describe process quality.
Do not present them as savings or improved availability without a valid comparison and enough observation time.
Prioritization should not become a fixed list. If a manufacturer changes a design, the group approves a substitute, or the plant installs new redundancy, revisit the part-to-equipment link. If a critical unit is used, check whether the temporary exposure while it is replenished calls for an approved measure. If an alternative is no longer available, reopen the analysis.
A classification that never changes describes the conditions on the day it was made, not the plant as it operates now.
Frequently asked questions
How do you identify a critical spare part?
Link the part to the equipment and function it supports. Assess what happens if it is unavailable, how long recovery takes, and which approved alternatives exist. Then consider lead time, repairability, and demand records. Record the result and review it when equipment or supply conditions change.
The DOE OIG recommends managing spares so an organization can identify the parts needed for each critical asset, the items in inventory, stock levels and reorder points, and where the parts are located (DOE OIG, pp. 2-4).
Should every spare for critical equipment be kept in stock?
No. Equipment criticality signals that its consequences need attention, but the decision depends on the specific part, redundancy, recovery time, available alternatives, and lead time. Repair, exchange, or a documented supply agreement may cover some needs. The responsible people should confirm the option and make sure it fits the specification and applicable safety process (2012 spare-parts classification study; equipment-criticality stocking policy study).
Can ABC classification identify critical spare parts?
ABC can rank part numbers using the criterion a company chooses, often consumption value. It does not, by itself, describe the operational consequence of a shortage or the difficulty of replenishment. Spare-parts studies use multi-criteria classifications because price, demand, and technical criticality answer different questions.
The 2025 plastics study combines criticality, supply, and maintenance in its classification, but its weights belong to that case (2025 plastics-industry study; 2012 spare-parts classification study).
How should you prioritize a spare with no consumption history?
No recorded issues do not prove that a part is unnecessary. Check whether the data are complete, the period covered, compatible equipment, failure modes, documented manufacturer guidance, and actual procurement lead time. State the uncertainty and agree how to handle the part until there is enough evidence. In the plastics study, the authors found part numbers with no recorded demand and adjusted policies by class.
That finding does not determine what another plant should do with one SKU (2025 plastics-industry study).
How often should a critical-spares list be reviewed?
Set a regular review and also review after an event: an equipment or design change, a change in function, a failure that exposes a gap, a new supplier, obsolescence, or a change in lead time. A DOE assessment describes monthly reviews of minimum and maximum levels in one nuclear setting. That frequency is not a universal rule for other plants.
Base the interval on how stable the assumptions are and on your site’s controls (DOE Los Alamos, pp. 12-14).