Case Studies

Too much dependent on software
It is an ‘overkill’ these days when senior management at times thinks that outputs from a reputed software are always beyond doubt and correct while they miss to verify the inputs accuracy. There are several instances, the professionals of AshPhil have come across, where they challenged the outputs from industry proven software run by young engineers showing with pride of results obtained. Here are a few examples when software can deceive unless management place a check on.
1. Fluid Catalytic Cracking Unit in a refinery was being revamped for higher intake. The process technology licensor was actively involved too with bright engineers from client organization. Several modifications were suggested to be within safe operating window while almost all equipment in the unit was stretched to entitle limits. Licensor with refinery engineer made presentation to management on path forward and overall cost benefit. The refinery was supplying Clarified slurry oil (CLO) produced at refinery which had a minimum density specification of 1.07gm/cc @ 15 deg C. The licensor and simulation engineer jointly declared that post revamp, the CLO density will fall down to below 1.04 gm/cc implying refinery scheduling team needs to find an alternate home for CLO which on several occasion became a bottleneck to raise the unit intake. Refinery management was not keen to get to depth as to why the CLO density should fall down when feed quality and operational severity were the same barring increased intake. Just a simple query to the licensor on CLO density used in base case as input to software solved the mystery. Refinery pushed FCC thruput and continued to be the favored supplier of CLO to Carbon Black manufacturer.
2. Power Loss in Back Pressure HP Steam Turbine placed Propylene Recovery Unit (PRU) in a refinery for heat pump compressor was reported and operations team along with process technologist wanted a simulation to be done as they found excess steam consumption. In fact, the compressor end process data (flow, suction and discharge pressure, temperature, etc.) were as per plant design operating window. Still insistence was on carrying out simulation runs. When asked about accuracy of the process data on turbine end viz. HP steam quality, back pressure of the LP seam header, etc. both operation team and PRU technologist reported in unison as all correct. Here came the catch. The temperature of HP steam header in upstream unit showed about 20 deg C more than the indication of steam within PRU. Similarly the LP steam header pressure showed variance with downstream users. Post recalibration of the instruments, the difference was resolved and no simulation was required.