Rooftop Solar Investment Assumptions: A Finance-Team Challenge List for Vietnam Factories

Executive answer: a rooftop-solar investment model is only as credible as the assumptions that connect the roof, the factory load, the electrical interface and the commercial contract. A finance team should not accept one headline payback period. It should require an auditable chain from measured load and solar resource through hourly energy flows, losses, availability, degradation, operating cost, contract terms and downside cases.

This guide is for an overseas investor reviewing a Vietnam factory proposal. It does not quote a market price, promise a yield, interpret tax law or declare a project compliant. The worked figures are deliberately hypothetical and must be replaced by project evidence.

1. Start with the decision, not the spreadsheet

Ask what the model is meant to decide: proceed to detailed feasibility, compare EPC bids, choose a self-funded or third-party structure, approve a capital budget, or verify a performance guarantee. Each decision needs different confidence. A preliminary map-based estimate can screen a site; it should not be treated as a bankable energy forecast.

The finance pack should identify the model owner, base date, currency, tax basis, financing boundary, calculation period, escalation basis and version. It should also state who approves technical inputs. Otherwise an apparently precise internal rate of return can hide unresolved engineering choices.

2. Separate evidence, assumptions and outputs

Do not let an output become an assumption in the next revision without its lineage. A quoted “specific yield” may already include temperature, wiring, inverter and availability losses. Deducting those losses again understates production; adding a separate availability uplift overstates it.

3. Challenge the solar resource and time basis

The World Bank’s Global Solar Atlas is useful for preliminary resource screening and consistent geographic comparison. It is not a substitute for a project-specific assessment of the roof plane, shading, horizon, weather variability and system configuration. Record the dataset version, coordinates, plane-of-array method and period represented.

An annual irradiation number alone cannot calculate factory self-consumption. The model needs a time series at an interval suitable for the decision—hourly at minimum for a first commercial assessment, and often 15 or 30 minutes where demand peaks, export limits or shift changes matter. Solar and load profiles must use the same time zone and daylight-saving convention. Vietnam does not use daylight saving, but imported corporate datasets may have been normalized to another zone.

Check whether leap days, missing intervals and meter resets were handled. Show annual totals before and after cleaning. If a new production line is planned, keep the measured historical load and the future adjustment as separate scenarios.

4. Rebuild the year-one energy bridge

A transparent energy bridge starts with nameplate capacity and specific yield:

E_PV,1 = P_DC × Y_specific

For a hypothetical 2.4 MWp array and a simulated year-one specific yield of 1,420 kWh/kWp, the result is:

2,400 kWp × 1,420 kWh/kWp = 3,408,000 kWh = 3,408 MWh

This is not a Vietnam benchmark or guarantee. It is a unit check. The proposal must state which losses the 1,420 kWh/kWp already contains.

Then split energy at the agreed meter boundary:

E_PV = E_self-consumed + E_exported + E_curtailed

If the same hypothetical case has 92% self-consumption and no separate curtailment because curtailment is already embedded in the simulation, then:

• Self-consumed energy = 3,408 MWh × 92% = 3,135.36 MWh.
• Exported energy = 3,408 MWh × 8% = 272.64 MWh.
• Reconciliation = 3,135.36 + 272.64 = 3,408 MWh.

The model must not assign the same unit value to both streams unless contracts actually support it. If export is prohibited or uncompensated, exported energy may have no modeled revenue and should trigger a sizing or controls review.

5. Test self-consumption instead of assuming it

Self-consumption depends on coincidence, not annual consumption. A factory using 12,000 MWh per year can still export at noon if weekend and holiday demand is low. Conversely, a smaller factory with continuous process loads may absorb more solar.

Request these plots:

• Monthly factory load, PV generation, self-consumption, export and curtailment.
• Representative weekday, weekend and shutdown-day profiles.
• Duration curves for import, export and PV output.
• Metered demand before and after PV at the tariff demand interval.
• Energy balance at the utility meter, PV meter and major internal meters.

Confirm whether auxiliary loads, transformer losses and inverter station consumption sit inside or outside the benefit boundary. The factory electrical metering-boundary guide explains why meter ownership and reconciliation must be fixed before performance reporting.

6. Do not double-count demand-charge savings

Rooftop solar can reduce daytime demand, but it does not automatically reduce the billing maximum. The peak may occur after sunset, during a cloud transient, at production start-up or when the PV system is unavailable. A model claiming demand savings should reproduce the tariff demand interval and apply PV output to the same interval load series.

Keep three outputs separate:

1. Imported energy avoided, in kWh.
2. Maximum or contracted demand change, in kW or kVA under the actual tariff method.
3. Production-resilience value, which rooftop PV without islanding and an appropriately designed energy source may not provide during a grid outage.

Do not assign outage-avoidance value to a normal grid-following PV system merely because panels are present. Backup operation requires an approved electrical architecture, controls, protection, operating philosophy and often storage or another grid-forming source.

7. Make every loss explicit and non-overlapping

Ask for a loss diagram that reconciles simulated irradiance to net AC energy at the financial meter. If two bidders use different meter boundaries, normalize them before comparing price or yield.

8. Treat availability as a contract definition

Availability may be time-based, energy-weighted, inverter-based or plant-level. Exclusions for utility outage, owner shutdown, force majeure, communications failure and planned maintenance can materially change the number. A high availability percentage can coexist with lost energy if outages occur during high irradiance.

IEC 61724-1:2021 provides terminology and methods for PV performance monitoring, including monitoring-system classes and data-quality concepts. IEC TS 61724-2:2025 defines a power performance index test for grid-connected PV systems under unconstrained test conditions. These scopes help the owner define measurable evidence; they do not create a project guarantee by themselves.

Require the proposal to state:

• Availability formula, measurement point and time resolution.
• Excluded events and approval process.
• Data substitution rules for failed sensors or communications.
• Response and restoration obligations.
• How availability, performance ratio and energy guarantee interact.

9. Audit degradation without turning it into a slogan

Performance loss is not one universal module number. IEA PVPS research shows that methods, data quality and missing data affect calculated performance-loss rates. The model should identify whether degradation applies to DC module power, system energy or a contractual guaranteed profile.

For illustration, if year-one net energy is 3,408 MWh and a hypothetical annual performance loss rate is 0.45%, year 10 before other scenario changes is:

E_10 = 3,408 × (1 − 0.0045)^9 ≈ 3,272 MWh

The exponent is nine because year one is the reference. Do not apply ten degradation steps to year 10. More importantly, do not use this illustrative rate as a promise. Require technology-, climate-, warranty- and dataset-specific evidence, plus downside sensitivity.

IEC 61215-1:2021 addresses design qualification of terrestrial PV modules, but its public scope expressly notes that qualification test results are not a quantitative prediction of module lifetime. IEC 61730-1:2023 addresses module construction requirements for electrical and mechanical safety. A certificate reference should therefore be matched to the exact module family and does not replace project reliability analysis.

10. Build a finance model that can be audited

The operating cash-flow structure should be visible:

Net operating benefit_t = avoided import value_t + export value_t + verified incentives_t − O&M_t − insurance_t − lease/fee_t − replacement_t − other project costs_t

Then discount project cash flows consistently:

NPV = Σ CF_t / (1 + r)^t

State whether the discount rate is nominal or real and keep inflation treatment consistent. Separate project economics from financing cash flows unless the decision specifically requires equity returns. List taxes, depreciation, withholding, foreign-exchange and accounting treatment as inputs requiring qualified review; do not infer them from an EPC proposal.

Simple payback ignores the timing of cash flows after recovery and should not be the only decision metric. Show at least NPV, IRR where mathematically meaningful, discounted payback and cumulative cash flow, with model checks that flag multiple sign changes or non-convergent IRR.

11. Define avoided cost from the bill, not a headline tariff

The value of one self-consumed kWh is not automatically the total bill divided by total kWh. Separate energy charges, demand charges, fixed charges, taxes and pass-through items. Determine which components actually change when grid import falls. Use bills covering a representative period and document contract changes expected during the model horizon.

If prices are modeled to escalate, show the base source, start date, escalation formula and cap or downside. A foreign investor should also test currency mismatch: revenue or savings may be in VND while modules, inverters, debt service or group reporting may be linked to another currency.

12. Include all life-cycle costs and residual obligations

• Engineering, permits and owner’s engineer.
• Structural reinforcement and roof remediation.
• MV/LV interface changes, switchgear, protection, metering and communications.
• Shutdowns, temporary works and production coordination.
• O&M labour, cleaning, monitoring, testing and vegetation/drainage where relevant.
• Inverter or component replacement assumptions.
• Insurance, security, spare parts and software/service subscriptions.
• Roof access, dismantling/reinstatement and end-of-term removal.
• Waste management and asset disposal obligations.

Do not hide a necessary roof replacement outside the investment case. Conversely, do not charge the solar project for unrelated roof work without identifying the allocation rule.

13. Challenge structural and electrical constraints before approving yield

A finance model should carry gates for roof capacity, waterproofing responsibility, fire access, drainage, lightning protection, earthing, cable routes, inverter locations, transformer loading, protection coordination, power quality and the connection agreement. The rooftop-solar feasibility data guide lists the technical inputs needed before an owner relies on a yield estimate.

TCVN 7447-7-712:2019, equivalent to IEC 60364-7-712:2017, covers low-voltage electrical installations for solar PV supply systems. The project team must confirm the applicable standard set, Vietnamese regulations, utility or industrial-park requirements and contract editions at the design date. A reference in the model is not proof of compliance.

14. Model production disruption and HSE as decision risks

Rooftop work can affect live production through crane movements, roof access, dropped-object risk, water ingress, hot work, electrical tie-ins and planned outages. The investment pack should include a construction sequence, shutdown assumptions and responsibility for production coordination.

Minimum HSE planning should address fall prevention, rescue, lifting, roof fragility, weather, fire, DC hazards, lockout/tagout, test-for-dead, energized-work prohibition or controlled exception, and emergency access. Only competent personnel using approved procedures should perform isolation, testing and commissioning. A low-price model that omits safe access and shutdown cost is not comparable with a complete scope.

15. Tie the assumptions to FAT, SAT and acceptance evidence

The acceptance plan should state what happens when data are unavailable or the test window is constrained. Do not let a short commissioning snapshot become an unqualified 20-year yield promise.

16. Require base, downside and break-even cases

A robust investment committee pack should show more than one deterministic case.

Run variables individually and in a coherent downside case. Solar resource, clipping, export and self-consumption can be correlated; combining independent “best cases” produces an unrealistic result.

17. Compare commercial structures on the same boundary

Self-funded EPC, equipment lease, rooftop lease, energy-service agreement and third-party power structures allocate ownership, operating responsibility and risk differently. Normalize them on:

• Energy and meter boundary.
• Contract term and termination rights.
• Roof access and reinstatement.
• Performance and availability definitions.
• Tariff/indexation and currency.
• Change in factory load or site closure.
• Insurance, damage, step-in and lender rights.
• Asset transfer or removal at term end.

Do not compare an EPC capex number with a service price as if they are equivalent cash flows. For a third-party proposal, legal, tax and electricity-market treatment require project-specific professional review.

18. RFQ return schedule for a finance-grade proposal

Ask bidders to return a structured assumption register with:

• Hourly energy model and readable annual summary.
• Weather/resource source, coordinates, period and uncertainty.
• Factory load dataset, cleaning log and future-load scenarios.
• Loss diagram with meter boundary and non-overlap check.
• Self-consumption, export and curtailment calculation.
• Availability, degradation and performance-test definitions.
• CAPEX/BOQ inclusions, exclusions, taxes and currency.
• O&M, replacement, spare and end-of-term assumptions.
• Programme, permits, shutdowns and owner dependencies.
• Base, downside, upside and break-even outputs.
• Model file with formulas accessible for audit.

For scope benchmarking, the industrial rooftop-solar service page describes QuangAnhcons’ owner-side electrical and construction interfaces. BESS should be evaluated separately; see the industrial BESS owner-requirements brief before assigning storage value to a solar case.

19. FAQ

Can the Global Solar Atlas be used for investment approval?

It is valuable for preliminary screening. Investment approval should document project-specific roof, shading, weather, design, loss and uncertainty inputs at the confidence appropriate to the decision.

Is a higher DC/AC ratio always better?

No. It may improve inverter utilization but can increase clipping and thermal or interface constraints. Evaluate hourly energy and commercial value, not only installed DC capacity.

Does module qualification prove a 25-year output?

No. IEC 61215 design qualification is not a quantitative lifetime prediction. Review exact product evidence, warranty terms, degradation assumptions, environment and operating strategy.

Should export energy be valued?

Only under a verified contractual and regulatory basis. If export is constrained or uncompensated, model it accordingly and consider sizing or controls changes.

Can rooftop solar provide backup during a grid outage?

Not by assumption. The architecture must be designed and approved for islanded or backup operation with appropriate sources, controls, protection and operating procedures.

What is the most important audit file?

The assumption register linking each model input to its source, owner, date, uncertainty and approval. It lets the finance team understand what changed and why.

Owner action checklist

• Freeze the decision purpose, model boundary and base date.
• Reconcile interval load, simulated PV, self-consumption, export and curtailment.
• Eliminate duplicated losses and unsupported demand savings.
• Define availability, degradation and performance tests contractually.
• Include roof, electrical interface, HSE, outages, O&M and end-of-term costs.
• Review base, downside and break-even cases.
• Require formulas and evidence—not only a PDF summary.

QuangAnhcons can help an FDI owner structure the technical assumption register, electrical interface, BOQ boundary, ITP and commissioning evidence for a factory rooftop-solar evaluation in Vietnam. The starting package should include 12 months or more of interval load data, bills, roof drawings, structural information, the proposed single-line diagram and the commercial term sheet. Final legal, tax, financing and acceptance decisions remain with the owner’s authorized advisers and project representatives.

No quick estimate yet. You can still submit a project quotation request.