MRPL's Hydrotreater Rupture and the Hidden Risk of Running Refineries Above 100%: What Your Process-Plant Policy Assumes About Nameplate Capacity

A high-pressure cold separator ruptured around noon in the Coker Hydrotreater Unit of the Phase 3 complex at Mangalore Refinery and Petrochemicals Ltd (MRPL) in Mangaluru, according to Oilprice.com's report of 30 September 2026. One worker died and at least eight were injured. The fire that followed took about 2.5 hours to bring under control.

MRPL's own statement, as reported by Hydrocarbon Processing, described "an incident involving rupture of the Cold Separator operating at high pressure." The rest of the roughly 300,000 barrels per day refinery kept running. After the fire, MRPL withdrew three spot export tenders for diesel, jet fuel and reformate.

The context matters for anyone who buys or underwrites process-plant cover. MRPL, an ONGC subsidiary with a 15 million tonnes a year refinery, had planned to run Mangaluru above 100% of capacity through March 2027, citing tight diesel supply and high margins. It had already declared force majeure on gasoline exports earlier in 2026.

This post is written for refiners, petrochemical complexes, fertiliser and chemical plants, and other continuous-process operators who are pushing throughput while fuel is short. The question is simple: what does your property, machinery breakdown and business interruption programme assume about how hard you run the plant, and does that assumption still hold?

Most process-plant programmes never mention nameplate capacity in a single clause, yet the number sits under almost every rating and sum-insured decision. Underwriters price a refinery or cracker using the proposal form, the risk engineering survey and the loss-estimate report. Each of those documents describes the plant at a stated throughput, a stated operating envelope and a stated maintenance regime.

Three things follow from that description:

1. The risk survey sets the baseline hazard. Estimated maximum loss and probable maximum loss figures assume the units run within their design temperatures, pressures and feed rates.
2. The BI sum insured is built from a budget. Gross profit is projected from expected throughput and margins. If the budget assumed rated capacity, the sum insured reflects rated capacity.
3. Maintenance and inspection warranties are tied to a schedule. Turnaround intervals, statutory inspection of pressure vessels and the insurer's own recommendations all assume the plant follows its declared cycle.

When a plant moves to sustained operation above 100%, none of these documents change on their own. The policy keeps describing the plant that existed at inception. That gap is where claims disputes start.

Material change in the risk

Indian fire wordings carry a condition that cover ceases for the affected property if circumstances change in a way that increases the risk of loss by insured perils, unless the insurer agrees by endorsement. Large refineries are usually insured on industrial all risks or mega risk wordings rather than the standard fire policy, and the exact language varies, but most carry an equivalent alteration or increase-in-hazard condition. A planned, months-long move to run above design throughput is the kind of change that should be disclosed, not left for a loss adjuster to discover in the operating logs.

Running above nameplate is rarely a single decision. It is a set of operating choices, each of which has a policy counterpart. A plant manager covering a diesel shortage typically does some combination of the following:

• Extends inspection intervals so units stay online longer between shutdowns.
• Defers a planned turnaround from the current fiscal year into the next.
• Runs hotter or at higher pressure to push conversion and yield.
• Processes different crude slates than the design basis, as availability changes.
• Accepts known minor defects with monitoring rather than an immediate repair.

None of these is unusual, and none is automatically a breach. The issue is that each one moves the plant away from the description the underwriter priced, and each one is visible after a loss in maintenance records, DCS trends and inspection reports.

Where the paper trail ends up

After a major process loss, the insurer's surveyor and loss adjuster will ask for inspection certificates, turnaround history, operating parameter logs for the failed unit, and any deviation approvals. If a turnaround was deferred, they will ask who approved it and whether the insurer was told. If the unit was running outside its design envelope, they will ask whether that contributed to the failure. A cause investigation that clears the operating regime still costs time, and time is the one thing a BI claim cannot spare.

For a fuller view of how these exposures are rated in practice, see our note on [machinery breakdown underwriting for process industries](/underwriting-risk/machinery-breakdown-underwriting-process-industries-india-2026).

Machinery breakdown cover pays for sudden and unforeseen physical damage to plant from internal causes. Indian MB wordings exclude wear and tear, gradual deterioration, corrosion and erosion in some form. A process plant that runs hard, with longer intervals between inspections, accumulates exactly the kind of damage those exclusions describe.

Exclusions that grow with operating intensity

The practical effect is that the boundary between a covered breakdown and an excluded deterioration loss shifts. A vessel or exchanger that fails after an extended run invites the question of whether the failure was sudden or whether it was the end point of a gradual process that inspection would have caught. The exclusion for gradual deterioration does not care why the inspection was deferred.

Maintenance conditions and warranties

Many process-plant programmes carry conditions requiring the insured to maintain machinery in efficient working order, follow manufacturer recommendations and comply with statutory inspection requirements. Some are drafted as warranties or conditions precedent, where breach can defeat a claim regardless of whether it caused the loss. Our piece on warranties and conditions precedent covers how Indian courts and insurers treat that distinction.

The fix is administrative, not technical. Tell the insurer what is changing, provide the revised inspection plan and risk-based inspection rationale, and ask for the maintenance condition to be endorsed to reflect the new schedule. Insurers may load the premium or raise the deductible. That is a price, and it is cheaper than a declined claim.

Business interruption is where above-nameplate operation creates its sharpest mismatch. The sum insured is normally gross profit for the indemnity period, projected at the time of renewal. If the renewal budget assumed rated throughput and normal margins, and the plant is now running above 100% with high diesel cracks, actual gross profit is running ahead of the sum insured.

The underinsurance trap

Indian loss of profits wordings apply an average clause: if the sum insured is less than the gross profit that would have been earned over the relevant period, the claim is reduced proportionately. A plant whose actual gross profit has risen well above the declared figure can find every BI claim scaled down, including a claim for a single unit outage while the rest of the complex keeps producing.

Options worth discussing with the broker at mid-term:

• Increase the gross profit sum insured by endorsement to reflect the revised throughput and margin outlook.
• Move to a declaration-linked basis where available, so the sum insured tracks actual results with a premium adjustment at expiry.
• Check the trend or other circumstances clause so the adjuster can recognise the above-nameplate run rate when measuring the loss, rather than reverting to a historical average.

Is the indemnity period long enough?

A high-pressure vessel in a hydrotreater is not an off-the-shelf item. Replacement depends on fabrication slots, metallurgy, inspection and recommissioning. If the plant has also deferred its turnaround, a forced outage may stretch into the next planned shutdown and complicate the restart sequence. An indemnity period sized for a normal operating year can be too short for a plant that has stacked deferred work on top of a long-lead failure. Our guide to business impact analysis and BI sums insured walks through how to stress-test both numbers.

MRPL's withdrawal of three spot export tenders for diesel, jet fuel and reformate is the part of this story that commercial teams tend to miss. Property BI is triggered by insured physical damage and pays loss of gross profit from the resulting reduction in turnover, plus increased cost of working. It does not pay for every commercial consequence that follows.

What BI can reach

If the damaged unit's lost output means less product to sell, the margin on that lost volume is a normal BI claim, subject to the sum insured, deductible or waiting period, and average. Extra costs incurred to reduce the loss, such as buying intermediate feed or rerouting product, can fall under increased cost of working if they are economically justified.

What it usually cannot

• Penalties, liquidated damages or claims by counterparties under term offtake or supply contracts.
• Lost market position with buyers who turn to other suppliers after a cancelled tender.
• Price movements between the planned sale and any later sale.
• Margins on volume that was never physically lost, for example where the complex kept running but the company chose to hold product back for the domestic market.

Each of those is a contract or market risk. It belongs in force majeure drafting, in commercial negotiation and, where available, in specialist contract frustration or trade credit products, not in a property programme. Refiners with term export commitments should read their offtake contracts against the BI wording and identify the gap explicitly.

Supply-side disruption runs the other way too. Downstream buyers who depend on a single refinery's product face their own contingent BI question. See our analysis of the crude import squeeze and contingent business interruption.

The MRPL incident is a prompt for any process operator running hard through a supply shortage to check its programme before the next loss, not after. A practical sequence for the risk manager and broker:

1. Write down the operating change. Target throughput, duration, units affected, revised crude or feed slate, and any change to temperatures or pressures outside the design envelope.
2. Notify insurers formally. Send the change as a material information disclosure under the duty of utmost good faith and ask for written acknowledgement or endorsement.
3. Reconcile the inspection plan. List every deferred inspection or turnaround item, the risk-based rationale, and who signed off. Ask insurers to endorse maintenance conditions against the revised plan.
4. Rebase the BI sum insured and indemnity period. Recompute gross profit at the current run rate and margin outlook, and consider a declaration-linked basis. Map lead times for critical high-pressure equipment and confirm the indemnity period covers fabrication, installation and restart.
5. Separate contract risk from property risk. Review export and offtake contracts for penalty and force majeure terms, and decide whether the residual exposure is retained or transferred elsewhere.
6. Brief the claims team now. Agree who will assemble operating logs and inspection records if a loss happens, so the adjuster's first request does not stall the claim.

For background on how downstream oil and gas risks are structured in the Indian market, our refinery and downstream insurance overview is a good starting point. The underlying principle is the one any underwriter will apply after a loss: the policy covers the plant you described. If you are running a different plant this quarter, describe that one.