At present, the difficulty in formulating the operation plan of the natural gas booster station is how to quickly and accurately adjust the pit pressure according to the change of the downstream gas consumption without exceeding the maximum power parameters of the compressor, and formulate the multi-unit combined operation plan in advance according to the specific requirements.Taking the large-scale natural gas compressor unit of the Tajik booster station of Sino-Petroise's Daniudi gas field as the research object, according to the compressor operation model, the BP neural network algorithm and the support vector machine (SVM) algorithm are used to predict the exhaust temperature, intake pressure and shaft power of compressor numerically through these two intelligent algorithms, and the most accurate prediction method is determined by comparing the prediction errors.
The prediction results of exhaust temperature can effectively avoid the problems of volume efficiency reduction, power consumption increase and lubricant carbon deposit caused by excessive temperature; compressor operation conditions can be adjusted according to the prediction results of intake pressure; compressor unit operation combination can be adjusted according to the prediction results of shaft power, and the optimal unit operation scheme can be formulated.It can be seen from the comparison between the prediction results of the two algorithms and the field measurement results that the prediction curve obtained by the BP neural network algorithm has higher fitting degree, the correlation coefficient R = 1, and the prediction result is more accurate.The relative error between the prediction result of the model and the measured value of the model is less than 2.80%, which verifies the reliability of BP neural network algorithm prediction.This method can help the staff to formulate a multi-unit online operation plan in advance, improve the operation efficiency of the unit, and reduce energy consumption and operation and maintenance costs.Through 4 months of on-site application, it saves 38% of electricity, equivalent to RMB 7,025,113, and saves an average of 1,756,278 yuan per month.