FTR-Based Expert System for Power Generation Units

The present expert system addresses the well-known problem existing on each large size boiler of a power generation utility - cleaning of fouling deposition created on the furnace walls and surfaces of heat exchangers. Continuous cleaning, which is a must, especially on large size units operated on fossil fuels, is very cumbersome and time-consuming procedure, mostly performed manually by technical staff sometime assisted by intelligence systems based on neural network. The expert system for cleaning suggested in the present study offers a new approach when the inference engine of the system is oriented to optimization of heat transfer inside the furnace. The target of optimization is maximization of the overall cleanliness factor of the furnace. On-going calculation of local heat transfer in different zones of the furnace is performed in real time by FTR devices positioned in different locations on the furnace wall and measuring continuously the Fouling Thickness and Reflectivity (FTR), along with local temperature of the flame. The system algorithm defines at each measurement cycle the optimal position where the next cleaning should be done and the corresponding signals are transferred by the system electronic circuitry to the controllers of the cleaning measures (the group of soot blowers to be activated). During the last two years our expert system has been successfully implemented on four coil-firing boilers of 500–600 MW on two power stations in Israel.