An expert system for the performance control of rotating machinery.

This research presented in this thesis examines the application of feed forward neuralnetworks to the performance control of a gas transmission compressor. It is estimated that aglobal saving in compressor fuel gas of 1% could save the production of 6 million tonnes ofCO2 per year.Current compressor control philosophy pivots around prevention of surge or anti-surgecontrol. Prevention of damage to high capital cost equipment is a key control driver butother factors such as environmental emissions restrictions require most efficient use of fuel.This requires reliable and accurate performance control.A steady state compressor model was developed. Actual compressor performancecharacteristics were used in the model and correlations were applied to determine theadiabatic head characteristics for changed process conditions.The techniques of neural network function approximation and pattern recognition wereinvestigated. The use of neural networks can avoid the potential difficulties in specifyingregression model coefficients. Neural networks can be readily re-trained, once a database ispopulated, to reflect changing characteristics of a compressor.Research into the use of neural networks to model compressor performancecharacteristics is described. A program of numerical testing was devised to assess theperformance of neural networks. Testing was designed to evaluate training set size, signalnoise, extrapolated data, random data and use of normalised compressor coefficient data oncompressor speed estimates. Data sets were generated using the steady state compressormodel. The results of the numerical testing are discussed.Established control paradigms are reviewed and the use of neural networks in controll'Iystems were identified. These were generally to be found in the areas of adaptive or modelpredictive control. Algorithms required to implement a novel compressor performancecontrol scheme are described. A review of plant control hierarchies has identified how theMdwme might be implemented. The performance control algorithm evaluates current!,!'Ocells load and suggests a new compressor speed or updates the neural network model.{'ornpressor speed can be predicted to approximately ± 2.5% using a neural networkh,lt1l'd model predictive performance controller. Comparisons with previous work suggestl'1l1t 'IlUal global savings of 34 million tonnes of CO2 emissions per year. A generic, rotatingmachinery performance control expert system is propose