200-bed hospital · 240 kW BESS

The site and the problem

A 200-bed multi-specialty hospital in Tier-2 NCR operated with a sprawling backup power architecture: a 60 kVA diesel set for general ward load, an ageing flooded-lead-acid UPS (48 x 2V blocks, 100 kWh nominal) for critical care and imaging, and a separate 40 kW inverter/charger for the server room. The lead-acid UPS was 11 years old, with cycle life exhausted and cell stratification causing charge imbalance. Every 3-4 months, the facility ran into low-power-mode alerts during peak afternoon consumption.

The deeper problem: imaging departments (CT, MRI) draw 80-120 kW during scan cycles. When the grid sagged during peak hours, the main UPS dipped below minimum voltage, triggering an automatic DG startup. This 5-10 second transfer time was acceptable for ward loads but risky for imaging cold starts and ICU monitors. Hospital administrators faced a choice: buy a new lead-acid UPS for ₹18 lakh or architect a modern, integrated solution.

Sizing the system

We analysed 6 weeks of 1-minute granularity power data. Critical load breakdowns: ICU/IMCU 45 kW, imaging suite 55 kW, OT 40 kW, server/pharmacy/surgical lights 30 kW. Total critical load: 170 kW. The hospital wanted 2 hours of autonomy to ride through an extended grid outage (common in NCR during summer GRAP phases). A 240 kW BESS (IZ-240K-3P) with 480 kWh usable capacity was sized. This allowed the critical load to run for 2.8 hours on battery alone, with the DG available as a last resort for extended outages.

The business case was clean: ₹52 lakh capex (BESS + installation) versus ₹60 lakh for a new lead-acid UPS plus a larger DG set. The BESS alternative meant no CPCB CEMS compliance (hospital DG is above 62.5 kVA threshold) and lifespan advantage (16 years vs 5 years for lead-acid).

Engineering details

The IZ-240K-3P cabinet houses LiFePO4 cells in 48 V, 200 Ah modules (9.6 kWh per module, 50 modules total). The PCS integrates a native three-phase 240 kW AC source to the hospital's distribution panel. The critical innovation was grid-to-battery transfer: when the hospital's main line voltage drops below 200V (three-phase), the system switches to island mode in under 14 ms. This is faster than the 5-10 second DG ATS changeover, preventing imaging cold-starts and ICU monitor glitches.

The system was integrated with the hospital's SCADA via Modbus TCP. The cloud EMS pulls real-time UPS load data and pre-charges the battery bank during off-peak hours (10 PM to 6 AM) when electricity is cheapest. Critical care loads have priority dispatch: if total demand exceeds 240 kW, non-critical loads (water chiller, HVAC) are managed by a secondary relay.

Core technical specifications for this deployment:

• LiFePO4 chemistry, 6,000-cycle design life with 80% DoD (10+ years in a hospital duty cycle)
• Grid-to-island transfer: 14 ms (vs 5-10 s for DG changeover)
• Three-phase 240 kW AC output, peak 300 kW for 2-second bursts (imaging cold start)
• Modbus TCP integration for SCADA real-time monitoring
• Active thermal management rated to 40°C ambient (hospital HVAC keeps facility 20-25°C)
• Cell-level BMS firmware monitors voltage, current, temperature, and SOH (state-of-health)

What changed after commissioning

DG runtime fell by 72%. Previously, the DG fired up 8-12 times daily for 30-60 minute intervals during peak grid stress. After the BESS, the DG runs only during extended outages (2-3 times per month, lasting 2+ hours). Diesel fuel consumption dropped from ₹4.8 lakh per year to ₹1.4 lakh. CPCB compliance ceased to be a burden: no more emission-testing notices, no annual engine overhaul bills, no fuel pilferage audits.

The 14 ms transfer meant imaging departments removed the risk buffer from CT/MRI scan protocols. Previously, technicians added 10-minute soak-in time before each scan to ensure stable power. Now, transitions are seamless. The hospital also retired its old lead-acid UPS and the 40 kW inverter/charger, recovering floor space in the basement. Uptime climbed to 99.99% (calculated as (8760 hours - 52 minutes of unplanned downtime) / 8760), a marked improvement from the previous 98.4% marred by DG startup delays and lead-acid cell failures.

Lessons we carried into the next deployment

This deployment taught us that hospitals are not generic commercial sites. Critical-care facilities demand speed and redundancy thinking that differs from factories or residential settings.

• Milliseconds matter. A 100 ms transfer time sounds fine until you learn that an ICU monitor can lose lock on cardiac output in 80-150 ms. We now specify sub-15 ms as a hard requirement for healthcare sites, which requires in-house firmware tuning and testing with actual hospital SCADA vendors (not generic test rigs).
• Cold-start surges are larger than nameplate load. Imaging cold starts (CT scanner warm-up) draw 1.5-2x rated power for 2-4 seconds. Our initial load survey missed this. Now we always ask imaging teams to run a scan cycle during site audit.
• Compliance headroom is valuable in negotiations. Removing CPCB burden meant the hospital could defer DG replacement by 5 years, freeing capex for other priorities. This made BESS approval easier than expected with hospital finance teams.