A new generator may save fuel every running hour. Yet buying it may be poor value if it runs only a few hours each year. The same machine can look attractive at a busy factory. At a site that only needs occasional backup, it may seem expensive. The hours, load and required service belong in the financial case from the start.
The decision is also more specific than “old versus new.” Keeping the existing set may require a repair, a later overhaul and continued maintenance. Replacing it may release its sale value but require installation work and a temporary supply. Compare these complete alternatives over the same period.
This guide develops an original replacement example in currency units, abbreviated CU. All prices, fuel rates, operating hours and asset values are invented. They illustrate a method. They are not a supplier quote, a forecast saving or advice to buy equipment for a particular factory.
Establish two technically acceptable alternatives
First decide whether retaining the existing generator is a viable option. Assess its condition and the evidence that it can meet the required duty. A repair may look cheap. It is still not a viable option if it cannot deliver the required service.
Define what the replacement must do as well. Specify the essential and ordinary loads and any permitted interruption. Include expected annual operation and the relevant starting events. A larger nameplate number does not establish that it suits the installation.
Caterpillar's guide to generator ratings lists the questions that affect the rating. These include duty, load factor, maximum load and annual hours. Operation in parallel with the utility is another factor. Use the applicable maker's rating and site conditions. A unit chosen for occasional backup may not suit frequent production supply. Check the rating for that duty.
Keep the options on an equal service basis. The new set may serve extra machines. Separate the cost and benefit of that service from the saving on the existing load. Otherwise, a change in output can be mistaken for better fuel performance.
There may be more than two viable options. A control repair or partial refurbishment may deserve its own case. So may a new dispatch plan or a later replacement date. Start with a clear choice between keeping and replacing the set. Add other viable options while keeping the same service requirement.
Preserve the fuel baseline before removing the old set
Obtain fuel and electrical-output records for the loads that matter. Compare electricity delivered at the same boundary. Treat the generator's auxiliaries consistently in both records. A litres-per-hour figure needs its load figure alongside it. Fuel rate alone cannot price the benefit.
A manufacturer's C13 performance sheet illustrates the point. It reports different fuel values at stated load points, distinguishes duty ratings and states test conditions. Those values belong to that configuration. They are not the rates used in the fictional case below.
For variable factory demand, estimate fuel across a representative load profile. Do not compare the old set at light load with the new set at its most favourable test point. Include starts, test runs and other relevant fuel use in the annual total. Or show them separately in both options.
The U.S. Department of Energy's FEMP measurement and verification guidance stresses the need for a baseline record. It also calls for agreement on how later changes will be handled. Its rules concern federal contracts, but the evidence principle is useful here. Once the old machine is removed, its operating baseline may be impossible to recreate.
Record the source and date of each input. A measured fuel rate has a different evidence base from a supplier estimate or planning allowance. Make those distinctions clear, especially if a small difference determines which option wins.
Include future cash changes once
The old purchase price is part of the asset's history. It does not become a new cash outflow merely because replacement is being considered. An accounting depreciation charge is not another equipment payment. Do not add it to the pretax cash model as a new outflow.
ACCA's relevant-cash-flow guidance focuses on future cash differences, including opportunity costs. Apply that logic to the actual decision. A repair payment avoided by replacement matters. A salary paid at the same level under either option creates no saving.
The old generator's sale value does matter. Selling it releases cash that would not be received if it were retained. Include that receipt once. Do not then add a second identical “opportunity-cost benefit” to the replacement case.
Build the installation allowance from the work needed to make the new unit usable. The project may need transport, foundations, electrical work and changes to its fuel system. It may also need testing, training and a supply during changeover. Obtain the site requirements and relevant quotes. The equipment price may not cover all this work.
Identify future costs by both amount and date. A routine service may already be in the annual contract. Do not count it again as a separate maintenance saving. A later major overhaul can appear separately if the annual allowance clearly excludes it.
Set a common five-year comparison
For the worked case, assume the factory plans to use the selected unit for five years and then sell it. Both options can provide the required service if their stated work plans are followed. No generation benefits after that planned exit are counted.
The 2025 NIST lifecycle-cost manual explains why options need a common study period. Their economic assumptions must also be consistent. It also treats the value remaining at the end of the period. Here the terminal amounts are assumed net sale receipts. Selling and removal costs have already been deducted. They do not include another five years of fuel savings under the label of asset value.
Use constant purchasing-power CU and an assumed 8% real annual discount rate. General inflation is excluded, and no real price escalation is assumed. This is a pretax, unfinanced project comparison. The example excludes taxes, borrowing cash flows, grants and insurance discounts. It does not assume that any benefit is available to a real buyer.
The 8% rate is a teaching assumption, not an official rate or a recommended cost of capital. The factory must approve a rate and tax treatment that suit its own decision.
Cash item | Retain existing set | Replace now |
|---|---|---|
Work or installation paid now | 600,000 CU | 6,500,000 CU |
Old unit sold now | No sale | 1,000,000 CU received |
Fuel at the matched service load | 120 L/h | 100 L/h |
Recurring annual non-fuel cost | 400,000 CU | 250,000 CU |
Separate major work at end of year 3 | 900,000 CU | None in this case |
Net unit sale receipt at end of year 5 | 300,000 CU | 2,000,000 CU |
The 600,000 CU retaining work and the later 900,000 CU overhaul are different assumed jobs. Neither is included in the recurring non-fuel allowances. The replacement installation figure includes the assumed changeover costs. Both options start providing the comparison service at the same base date.
Assume each supplies 400 kW of useful electrical output for 1,000 equivalent running hours per year. That is 400,000 kWh annually under either alternative. These equivalent fuel rates cover the whole simplified annual operating package. They include its stated ancillary fuel use. There is no uncounted extra test-fuel allowance waiting outside the model.
Assume fuel costs 120 CU/L throughout the comparison in constant-price terms. The service plans and terminal sale values are also expressed in those same CU. Working-capital needs are the same in this example. An actual difference would need its own cash entries, with dates.
Reconcile the initial and recurring differences
Replacing the set requires 6,500,000 − 1,000,000 = 5,500,000 CU now after selling the old unit. Retaining it requires 600,000 CU now. The additional initial outlay for replacement is therefore:
5,500,000 − 600,000 = 4,900,000 CU.
This comparison includes the old unit's current sale value. It also includes the retaining work that replacement avoids. Neither should be credited again in a later “benefits” list.
At 1,000 equivalent hours, annual fuel use is 120,000 litres for the existing set and 100,000 litres for the replacement. Their fuel costs are 14,400,000 CU and 12,000,000 CU. The annual fuel-cost difference is 2,400,000 CU.
The recurring non-fuel difference adds 400,000 − 250,000 = 150,000 CU each year. Total recurring annual benefit is 2,550,000 CU. It is the difference between the complete recurring costs: 14,800,000 CU for retaining and 12,250,000 CU for replacing.
At the end of year 3, replacement also avoids the separate 900,000 CU major-work payment. At the end of year 5, its net sale receipt exceeds the retained unit's receipt by 1,700,000 CU. Keep these one-time differences in their own years.
No reliability gain has been monetized in this base case. Both options are compared on the stated delivered service. A separate reliability scenario follows later. Any extra costs and benefits must fit this same boundary.
Discount the complete incremental cash flow
Present value converts a future amount to the chosen base-date equivalent. For a net benefit C received at the end of year t, this example uses:
Present value = C ÷ 1.08ᵗ.
Add the discounted benefits, then subtract the additional initial outlay. The result is replacement's net present value, or NPV, relative to retaining. ACCA's investment-appraisal explanation distinguishes discounted cash methods. It explains why cash timing matters and why inflation must be treated consistently.
Time | Incremental benefit or outlay | Present value, nearest CU |
|---|---|---|
Now | −4,900,000 CU | −4,900,000 CU |
End of year 1 | 2,550,000 CU | 2,361,111 CU |
End of year 2 | 2,550,000 CU | 2,186,214 CU |
End of year 3 | 3,450,000 CU | 2,738,721 CU |
End of year 4 | 2,550,000 CU | 1,874,326 CU |
End of year 5 | 4,250,000 CU | 2,892,479 CU |
Using unrounded values, the discounted benefits total 12,052,851 CU, rounded to the nearest CU. Subtracting the initial 4,900,000 CU gives an NPV of about 7,152,851 CU. On these assumptions, replacement has the lower lifecycle cost of the two options.
This result does not select the best generator on the market. Nor does it select the best replacement date. It establishes the result of one specified comparison. A cheaper viable refurbishment or a different installation offer could change the ranking.
Read the year-3 and year-5 entries carefully
Year 3 contains the usual 2,550,000 CU recurring difference plus 900,000 CU of avoided work. It is not a permanently higher annual saving. Year 5 contains the recurring difference plus the 1,700,000 CU difference in net sale receipts.
Support the terminal receipt with evidence. Can the unit be sold in its expected condition, and what will disposal cost? The factory may plan to keep the set beyond the study period. In that case, a value-in-use approach may suit the decision better than a planned sale. Do not count a sale receipt and the same remaining service value in one terminal entry.
Use payback as a timing indicator
Dividing the initial difference by recurring annual benefit gives 4,900,000 ÷ 2,550,000 = 1.92 years, rounded. This simple annual-rate calculation assumes the benefit arrives evenly. It ignores discounting, the later overhaul and terminal receipts. It is a screening number rather than the full investment result.
The actual table places annual cash differences at year-end. With that timing, total undiscounted benefits first exceed the initial outlay at the end of year 2. They total 5,100,000 CU by then, leaving a 200,000 CU surplus over the initial difference.
Discounted year-1 and year-2 benefits total about 4,547,325 CU, which remains below 4,900,000 CU. The next year-end receipt closes the gap. On this timing basis, discounted recovery occurs at the end of year 3.
These timing measures differ from annual profit and NPV. If a proposal says an upgrade “pays back in two years,” state the method and cash dates. The label determines what that statement actually proves.
Test the standby-use case separately
Now reduce equivalent annual operation to 100 hours, keeping the same service load and fuel-rate difference. For this sensitivity test only, keep the calendar-based work plans and non-fuel allowances unchanged. Keep the same terminal sale values too. This shows the effect of running hours alone. A real forecast must review the other assumptions too.
The fuel-cost benefit falls to 20 L/h × 100 hours × 120 CU/L = 240,000 CU per year. Adding the unchanged 150,000 CU non-fuel difference gives 390,000 CU of recurring annual benefit.
With the same initial outlay, year-3 work and terminal receipts, NPV is now about −1,471,403 CU. On this financial basis, keeping the existing set costs less. The positive high-use result cannot simply be carried into a standby application.
For this five-year, 8% case, the sum of the annual discount factors is 3.992710. If H is equivalent annual running hours, recurring benefit is:
Annual recurring benefit = 2,400H + 150,000 CU.
The 2,400 CU per hour comes from the 20 L/h fuel difference and the assumed price of 120 CU/L. Leave the other assumptions unchanged. The operating-hours break-even point is then about 253.55 hours per year. This is where the full five-year NPV reaches zero, including major work and terminal receipts.
This is not a service limit. It does not guarantee that replacement wins above that value. A different fuel curve, sale value, installation cost or work schedule moves the threshold. Fewer running hours may change the overhaul date. They may also change the condition of the unit when it is sold.
Value reliability without assuming a perfect replacement
A backup system's value is not exhausted by its fuel saving. But a new engine does not make every interruption disappear. The electrical distribution, transfer equipment, controls and fuel supply remain part of the service system.
NREL's 2023 backup-power reliability report separates three issues: availability when needed, failure to start, and continued operation during an outage. It warns against assuming that generation equipment is completely reliable. Its published data do not predict failures at this fictional factory or in any unit offered for sale.
Base the reliability comparison on events that cause loss at the site. A set may start successfully but still be too slow for a process that cannot pause. A brief stop may cause little cash loss if work resumes without scrap or a missed delivery. State the event, affected load and recovery path before assigning a monetary value.
The NREL description of its Customer Damage Function Calculator distinguishes outage costs by type and duration. The useful principle is to identify the site's damage pattern. One lost-sales-per-hour figure can miss key costs. Restart costs may be fixed, spoilage may depend on duration, and some work can be recovered later.
For lost production, first consider the ways to recover it. Then estimate the cash contribution actually lost. Sales revenue alone ignores costs that would not be incurred. The work may be recovered on a later shift. In that case, extra overtime or other recovery costs may matter more than a claim for permanently lost output.
One event may have several possible responses. Do not add a full lost margin, full recovery cost and full rental avoidance when those responses are alternatives. Trace what each option would actually do. Exclude unrelated failures. Also exclude losses that replacement would not prevent.
Find the reliability benefit the low-use case needs
The 100-hour scenario has a discounted shortfall of about 1,471,403 CU. Assume an added reliability benefit, if supported, is the same net amount at each year-end. Deduct any extra fuel, service, labour or other costs not already counted.
The annual amount needed to close that shortfall is about 1,471,403 ÷ 3.992710 = 368,522 CU, using unrounded values. This threshold tells the analyst what evidence is needed. It does not establish that the benefit exists.
For comparison, an independently supported net reliability benefit of 300,000 CU per year would add about 1,197,813 CU in present value. The combined low-use case would still have an NPV of about −273,590 CU. Reliability may improve and still not save enough to justify the quoted investment.
Minimum service requirements remain a separate gate. If the retaining option cannot meet them, it is not the appropriate economic alternative. Compare replacement with another option that meets the requirements. A speculative cash value cannot make an unacceptable option viable.
Challenge the assumptions that can change the decision
Review operating hours, matched-load fuel rates and fuel price separately. They multiply together, so a small error in each can produce a material combined error. Use scenarios with plausible operating conditions. Choosing only the most favourable value from each range can create an unrealistic forecast.
Challenge the old unit's condition and sale offer as well. The 1,000,000 CU receipt today assumes an actual disposal route at that net value. A dealer's gross indication is not the same as cash after transport, removal and selling costs.
The 2,000,000 CU terminal receipt for the replacement is also uncertain. Obtain evidence for the relevant age, condition and market, and show the result under a lower value. Explain this value. It should not be a plug figure that makes a weak project look worthwhile.
Check the timing of installation and benefits. A delay in commissioning also delays savings. The new unit cannot save fuel before it enters service. Temporary-power costs and any retained running period need their own entries. The worked example assumes service starts at once. A real proposal must use its actual project schedule.
For a financed purchase, also show how it will be funded and when cash is needed. Keep this view consistent with the project appraisal. An unfinanced project cash flow excludes loan interest. Do not charge the same financing effect again through both cash flow and the discount rate. Tax allowances and grants depend on current local rules. Check eligibility and timing; this example does not provide them.
ApparelDiary's capacity investment decision-tree guide helps organize uncertain outcomes. Its critical-spares stocking guide covers another part of the reliability plan. It explains the evidence needed to keep essential support available. Neither replaces the generator's specific service and supply arrangements.
Make the proposal reviewable after commissioning
Keep the service definition and the evidence behind each input. Retain the dated cash-flow schedule too. The approval record should explain each change. Which changes should save money, and which are needed to meet the duty? A single ROI percentage cannot carry all of that information.
After installation, verify the promised operating states and the agreed fuel boundary. Record actual hours and output alongside fuel use. Review changes in demand, price, maintenance and service performance. The new machine may not explain every change.
The replacement decision should remain open to later checks against its own assumptions. It needs a viable baseline, equal service and complete future cash flows. Uncertainty must remain clear. A generator's age or one attractive payback figure cannot supply those missing pieces.
Discussion0
Loading the discussion…