Flood-Proofing a 33/11 kV Substation That Feeds a District Hospital
A 33/11 kV substation (2 × 10 MVA) on the right bank of a hill river feeds a town of 90,000 people and the district hospital. In the 2019 cloudburst the yard was under 1.6 m of water for 40 hours and supply to the hospital failed for 52 hours, because the 11 kV feeder pillars and the control room were at ground level. Options before the utility: A, raise the yard, plinth, cable trenches and control room by 2.0 m - ₹9.6 crore; B, flood-proof in place with sealed cable entries, a first-floor control room, four submersible pumps and waterproof switchgear - ₹4.2 crore, though 33 kV structures, current transformers and transformer radiators remain exposed; C, relocate 6 km inland - ₹26 crore plus ₹8 crore of feeder reconfiguration, with 14–18 months of land acquisition. Flood levels: 1-in-10 year +0.8 m, 1-in-25 year +1.6 m, 1-in-100 year +2.6 m. Present expected annual damage is ₹2.8 crore. Climate projections add about 20 per cent to extreme flows by 2050.
1. Compute simple payback for A, B and C, stating the residual damage you assume for each.
Read the worked answer
Assume residual expected annual damage falls to the level of the unprotected exposure each option leaves:
- A (raised 2.0 m): protected up to just above the 1-in-25 level, so damage is confined to rarer events - residual ₹0.15 crore. Benefit ₹2.65 crore; payback 9.6/2.65 = 3.6 years.
- B (flood-proofed in place): equipment protected, but the yard still floods above roughly 0.8 m and transformer radiators and 33 kV CTs remain exposed - residual ₹1.1 crore. Benefit ₹1.70 crore; payback 4.2/1.70 = 2.5 years.
- C (relocation): residual ₹0.05 crore. Benefit ₹2.75 crore; payback 34/2.75 = 12.4 years, before counting 14–18 months of land acquisition and higher feeder losses from running 6 km to the load centre.
The ranking on payback alone is B, then A, then C. But payback ignores tail events: B's residual damage is not a smooth annual figure, it is a lumpy transformer replacement of ₹3–4 crore per unit with a 6–9 month lead time after any event above 1.6 m, which is exactly the 2019 event.
2. Choose a course of action and defend it against the cheapest-payback option. Bring in the hospital outage and the 2050 climate projection.
Read the worked answer
Adopt A with selected elements of B - sealed entries, submersible pumps, waterproof switchgear - at about ₹10.8 crore, staged over two dry seasons, and defer C.
Defence against B: B protects equipment but leaves the yard flooding, so the substation still goes out at the 1-in-25 level that has already occurred once in recent memory. Its payback looks best only because the residual damage is treated as an average; in reality it is a single large, uninsurable loss with a long lead time. A's payback of 3.6 years against an asset life of 35–40 years is comfortably dominant, and the ₹5.4 crore premium over B buys protection of the transformers, not just the switchgear.
Hospital: a 52-hour outage to a district hospital is not a damage line item. Value it with a loss-of-load estimate and note that a dedicated dual-source feeder plus a 72-hour generator fuel contract costs a fraction of any of these options and should be done regardless.
2050: a 20 per cent increase in extreme flows will downgrade A's 1-in-25 protection. Design the raised yard to +2.6 m (1-in-100) plus 0.3 m freeboard even if construction is phased; the incremental cost is far below a second rebuild.
3. Name the non-structural measures that must accompany the chosen option, and state why each is not optional.
Read the worked answer
- Flood forecasting and a de-energisation SOP: an upstream gauge with two hours' lead time allows orderly shutdown and protects equipment from energised immersion. Physical works do not remove the need to act before the water arrives.
- Hospital redundancy: a dedicated feeder from a second source and a fuel contract for 72 hours of generation. The hospital's failure is a life-safety issue that no substation upgrade fully covers.
- Back-feed capability: a mobile transformer and ring-main switching so the town can be supplied from an adjacent substation during restoration. Raising the yard does not shorten repair time after a rarer event.
- Pre-positioned spares and an insurance/asset register: bushings, CTs and cables have long lead times; without spares, availability is decided by procurement.
- Joint inspections with the irrigation department, mock drills and training: flood levels are a function of upstream reservoir operation as much as rainfall, and the operating discipline is what failed in 2019.
- Land-use control on the riverbank: without it, encroachment raises flood levels around the substation and quietly erodes the design assumption.
Each of these is cheap relative to the capital works and covers the gap between the design flood and the actual event.