Optimization On Inital Pressure Of A Steam Turbine Based On Symbiotic Organisms Search Algorithm

In recent years,the social power industry has undergone relatively large changes,and the load on the power grid has changed day and night.According to national policy requirements,large generator sets are required to participate in the peaking task.However,if a large steam turbine of a power plant is operated in a peak-shaving operation for a long period of time,the unit will be operated at a low load,thereby greatly reducing the thermal economic efficiency.In order to ensure that the steam turbine can still operate effectively under variable load,it is necessary to make a deeper optimization of the unit's sliding pressure running curve and reduce the heat consumption rate of the unit.Optimizing the unit's sliding pressure running curve is to find the optimal initial pressure under each working condition,and then guide the normal operation of the unit.Therefore,this paper takes the optimal initial pressure of steam turbine as the research subject,and studies it from two aspects: optimization strategy and heat rate modeling.Firstly,the symbiotic organisms search(SOS)is studied.In view of the shortcomings of slow convergence,an ameliorated symbiotic organisms search(ASOS)is proposed.The algorithm can accurately find the optimal value and is the key to the optimization of steam turbine initial pressure.In order to verify the efficiency of the algorithm,six standard test functions are used for testing and compared with other classical intelligent algorithms.The test results show that the ASOS algorithm has a good optimization effect and can solve the complicated optimization problem well.Secondly,using the global search ability of ASOS algorithm,the activation function of extreme learning machine(ELM)is optimized,and a new modeling method ASOS-ELM is proposed.By comparing the UCI standard dataset with other classic extreme learning machines,the results show that the ASOS-ELM model has a better modeling effect.The ASOS-ELM is used to model the heat rate and compare with other classical modeling methods.The test shows that the heat rate model established by ASOS-ELM has better predictive tracking ability.Finally,based on the heat rate model,the ASOS algorithm is used to optimize the main steam pressure of the input variable of the heat rate model.The optimum main steam pressure obtained by this method can cause a corresponding decrease in the heat rate value.Then,the pressure after optimization is fitted to a sliding pressure curve,which is compared with the fixed sliding pressure running curve designed by the manufacturer.It is found that the heat rate decreases by about 58.51 kJ/(kW·h)on average according to the optimal initial pressure curve.Prove that the curve can better guide the actual operation of the plant.