| At present,China’s external dependence on natural gas has exceeded 45%.As one of the four major channels of natural gas import,liquefied natural gas plays an important role in China’s energy security.Natural gas liquefaction technology is an important link in the LNG industry chain,and its energy consumption and process control have a great impact on the industrial chain.In this paper,based on the operation data of the natural gas liquefaction unit,genetic algorithm and artificial neural network technology are used to optimize the liquefaction process and improve the control.In the aspect of natural gas liquefaction process optimization,in this paper,according to the experimental data of dual nitrogen expansion liquefaction process with R22 pre-cooling and dual mixed refrigerant liquefaction process,the corresponding steady-state models are established respectively.The specific power consumption is taken as the optimization objective,and the liquefaction process is optimized by genetic algorithm,and the optimization results are compared with the experimental results of the liquefaction device.The results show that: due to the differences in design stage,construction stage and operation stage,there are always differences in performance parameters and operation parameters between experimental device and simulation device,so it is necessary to use the method based on historical data to guide the optimal operation of liquefaction device.Based on the historical operation data of two sets of liquefaction units,the BP neural network model of liquefaction unit is established to predict the specific power consumption of liquefaction process under corresponding operation variables.Then the BP neural network model of the liquefaction unit is optimized by genetic algorithm.The data test results show that the prediction accuracy of the optimized BP neural network model is significantly improved.Finally,on the basis of the trained GA-BP neural network model,the specific power consumption is taken as the optimization objective,and the liquefaction process is optimized by genetic algorithm,the optimization results which are more suitable for the actual operation of the device are obtained.In the aspect of control improvement of natural gas liquefaction process,the dynamic model of natural gas liquefaction process is established,and on this basis,the sensitivity experiments of feed gas inlet temperature,pressure and composition are carried out.Through the dynamic simulation of natural gas liquefaction process,the foundation is laid for the optimization of control system of natural gas liquefaction device.The dynamic simulation models of natural gas throttling temperature control,compressor outlet pressure control and mixed refrigerant level control are established respectively.In order to reduce the integral value of control variable error to time,the PID controller parameters of natural gas dynamic model are optimized by genetic algorithm,and the optimized parameters of the controller are compared with the empirical parameters in the control effect.Furthermore,the traditional PID control strategy and neural network technology are combined to construct a BP neural network PID control system for natural gas throttling temperature control.The dynamic simulation is carried out,and the simulation results verify the superiority of neural network PID control in natural gas liquefaction process control.In this paper,the mechanism model and data model of natural gas liquefaction process are established,and the difference between the theoretical simulation and the actual operation is explained.The new model not only reduces the energy consumption of natural gas liquefaction,but also improves the stability of the plant operation.It has practical guidance and application value for the design,construction and operation of natural gas liquefaction device. |
