Expert System For Boiler Combustion Optimization Based On The Measurement Of The Furnace Temperature Field

Nowadays, the optimization of combustion in the operation of boiler is still mainly relies on the experience of the operation personnel. However, due to the complexity of conditions in boiler operation, it is difficult for operation personnel to make accurate judgment on the conditions of boiler combustion from a large variety of information and to perform corresponding online adjustment. It will be very likely to cause a delay of combustion adjustment, resulting in deviation from the optimal combustion condition. Since furnace temperature field changes consistently with the coal blending modes and air distribution modes, it can reflect the boiler combustion condition rapidly and accurately. Study of the boiler combustion optimization control method based on the furnace temperature field helps partially eliminate the influence of the delay in combustion systems. The boiler combustion optimization control approach based on the artificial intelligence has received great attention recently. This paper studies a kind of combustion optimization expert system that is integrated with the measurement of the furnace temperature field. The main contents are as follows:1. This paper establishes a novel least square support vector machine (SVM) model of NOx emissions and boiler efficiency with the temperature distribution information along the height of related cross-section obtained by an acoustic temperature measurement system and the operation data of the boiler. To deal with the defect of local optimum and premature convergence existing in PSO algorithm, an improved hybrid PSO/DE algorithm is proposed for the optimization of model parameters. The model with regard to the influences of the furnace temperature has higher precision of prediction for the NOx emission and boiler efficiency than the ones without it. Based on the models and the hybrid PSO/DE algorithm, the optimization variables under different condition can be obtained to reach higher boiler efficiency and lower NOx emission.These works are prepared to further establishing of the numerical optimal knowledge base.2. Based on the numerical optimal knowledge base, an approach of establishing the boiler combustion optimization knowledge base combined with the fuzzy c-means clustering is proposed. In the basis of established fuzzy knowledge base, the inference engine of expert system for boiler combustion is constructed with CLIPS development tools, and then a CLIPS fuzzy inference model based on CLIPS is established. A backward tracing method is used for inference interpretation. The test results indicate that the optimization results of expert system is similar to the results of the direct nonlinear optimization approach, but with smaller amount of calculation.3. A general architecture of the boiler combustion optimization system is designed while the functional modules of the system are developed. By calling ClipsNET component, the CLIPS project is integrated into the.Net framework. Reading the data of DCS and acoustic temperature measurement is realized with the application of OPC communication technology. Finally, the interfaces of boiler combustion optimization expert system are designed for further engineering application.