High-Density Racks Overheat Despite N+1 CRAHs
A 180 kW IT load is distributed across 40 racks. Most racks draw 3-5 kW, but ten new racks draw 12 kW each. The room has two 200 kW chilled-water CRAHs in N+1, a 12/18°C chilled-water supply, and a 24°C cold-aisle setpoint. Hot-aisle containment is installed, but roof panels are missing above three high-density racks to clear cable trays, and about 20 U of blanking panels are missing near the new racks. CRAH A runs at 100% fan speed; CRAH B runs at 55%. Cold-aisle temperature is 22°C at floor level but reaches 29°C at the top of several high-density racks. Hot-aisle return is 36°C. The building automation system shows no humidity alarm; relative humidity is 35%. The site has had two overtemperature events in six weeks, both during utility peaks. The operations team wants to lower the chilled-water supply temperature and add two portable fans.
1. Diagnose whether capacity, airflow, or containment is the dominant problem. Use approximate calculations.
Read the worked answer
Total IT load is 180 kW, and the N+1 CRAH plant has 200 kW of capacity in the surviving unit, so raw cooling capacity is not the first explanation, though it must be verified at actual entering-water temperature and airflow. The pattern points to airflow and containment. A 12 kW rack with a 12 K air temperature rise needs roughly 0.8 m3/s, about 1,700 cfm, so the high-density racks need concentrated airflow, not just average room airflow. Missing roof panels above three racks create a direct hot-air path back to the cold aisle; missing blanking panels allow bypass air. The 29°C top-of-rack inlet versus 22°C at the floor is a classic recirculation signature. CRAH B at 55% may be unloaded because return-air sensors are seeing mixed temperatures or because the control sequence is not maintaining static pressure. Portable fans would mix air and can make recirculation worse; lowering chilled-water temperature may mask symptoms without curing the distribution problem. Investigate CRAH airflow, sensor placement, and containment before adding capacity.
2. Recommend immediate corrective actions and permanent design changes, with acceptance criteria.
Read the worked answer
Immediate: restore all roof panels with proper cable pass-throughs, install blanking panels, seal floor and wall penetrations, and balance CRAH fans to maintain cold-aisle pressure. Move rack inlet sensors to the top, middle, and bottom of the high-density racks and alarm at 27°C. If any inlet exceeds the allowable range, reduce the high-density load or add temporary spot cooling with condensate management until containment is complete. Permanent: design the high-density rows for the actual airflow, using in-row or rear-door heat exchangers if air containment alone cannot hold the 12 kW racks within the envelope; provide dual-fed CRAHs, isolation valves, VFD and static-pressure control, and cable management that does not breach containment. Acceptance criteria should include a steady-state thermal test at design load with all CRAHs running, a single-CRAH failure test, and documented rack inlet temperatures at the top of every high-density rack. Lowering the chilled-water supply temperature is not a substitute for fixing bypass and recirculation.
3. The business rejects capital work for six months. What operating limits, monitoring, and contingency plan would you set, and how would you communicate the risk?
Read the worked answer
Without capital, set operating limits: cap the ten high-density racks at a tested power level, prohibit new loads until the containment repairs are complete, and require blanking panels to be replaced after any work. Install temporary rack-inlet temperature monitoring with alarms to the NOC, and trend CRAH fan speed, valve position, and differential pressure. Pre-cool or shift batch workloads during utility peaks if the central plant is capacity-limited. Keep portable spot coolers on standby with power and condensate plans, and define the trigger for shedding noncritical load. Communicate in writing to the business: the site is operating outside its design airflow distribution, the risk is localized hotspot-driven shutdown, and the residual risk remains until containment and airflow are corrected. Record the acceptance decision and review weekly; do not treat portable fans or lower water temperature as a permanent fix.