| Solar photothermal power generation has the characteristics of large heat storage capacity,simple heat storage process,direct connection of alternating current generated to the grid,easy to complement with conventional power generation mode,and achieve 24 hours of continuous and stable power generation.Scheduled power generation is realized through heat storage,which can be bundled with wind power,photovoltaic and other renewable energy for output.Effectively adjust the randomness and volatility of photovoltaic and wind power,the intermittent solar energy is transformed into high quality clean electricity that can be continuously output and flexibly adjusted,and the consumption level of renewable energy in regional consumption and bundled delivery is improved.The construction of integrated wind power,photovoltaic and solar thermal energy systems is a major strategy to achieve high-quality development of new energy and is of great significance to the optimization and upgrading of the energy structure.In this dissertation,the 50 MW Linear Fresnel concentrating light and heat collecting system of Dunhuang molten solar is taken as the research object,and combined with the actual operation experience in the project construction and commissioning stage,the purpose is to improve the power generation of the molten salt Linear Fresnel photothermal demonstration power station,improve the photothermal conversion efficiency of the concentrating light and heat collecting system and reduce the power consumption of the power plant.By studying the key control technologies such as the variable duty ratio following preheating control algorithm of the empty tube in the heat collection loop of the molten salt Linear Fresnel concentrating light and heat collecting system,the nonlinear compensation algorithm of the target tracking angle of the primary mirror in the heat collection system,and the prediction control algorithm of the molten salt temperature at the outlet of the heat collection loop,etc.Finally,various key control technologies are integrated into the data acquisition and monitoring system of the collecting heat island of the demonstration power station,and the hardware and software design of the monitoring system is setup.Firstly,the requirements and operation measures of anti-coagulant and consumption reduction of molten salt Linear Fresnel power station are proposed.In view of the nonlinear and time-varying characteristics of the temperature rise process of the collector tube in the preheating process of the heat collecting loop,by modeling and analyzing the main factors affecting the temperature rise speed of the heat collection loop,combining the data information such as real-time irradiation,the control algorithm of the variable duty ratio of the heat collecting loop is proposed.The field experiments show that the control effect of the preheating algorithm can meet the constant rate temperature control,and the temperature rise rate error is about 14%,which is far less than the limit safe temperature rise rate of the collector pipe;On the premise of satisfying the safe temperature rise rate of the collector pipe,the whole preheating time of the variable duty cycle method is reduced by 22%compared with that of the fixed duty cycle method.The method has good control effect and high control precision.The theoretical model can also be applied to some subsystems of the trough and tower solar thermal system,and the method has certain universality and practicability.Secondly,according to the structural characteristics of Linear Fresnel concentrating light and heat collecting system,the factors affecting the concentrating accuracy of Linear Fresnel concentrating light and heat collecting system are explored from the perspective of system structure and engineering installation.Through simulation and experimental analysis,the influence mechanism of the primary mirror type error,CPC installation accuracy error,the north-south layout error of the mirror field,the dynamic displacement error caused by the gap between the primary mirror reflection center and the rotating center,and the temperature drift error of the angle measurement sensor on the tracking and concentrating results are determined.Combined with the long-term test records of the actual tracking target angle on site,the error curve between the actual tracking target angle and the theoretical target tracking angle is obtained.According to the trend of the error curve,some influential factors,such as the north-south deviation of the mirror field,the dynamic displacement error of primary mirror reflection center and the temperature drift error of inclination sensor on the tracking concentration result and the influence degree caused by each error are determined.and the compensation algorithm was applied to the practical application of the Dunhuang demonstration project.Through the practical verification in Dunhuang 50 MW molten salt Linear Fresnel demonstration power station,the compensation algorithm can well realize the error compensation of linear Fresnel tracking angle,and the annual error after compensation is less than 0.1°,which meets the engineering application requirements of Linear Fresnel concentrating light and heat collecting system.Thirdly,in view of the characteristics of large fluctuation,non-linearity and large lag in the heating and rising process of molten salt in Linear Fresnel heat collection circuit,the main influencing factors of molten salt temperature at the outlet of the heat collection circuit are determined by analyzing the mathematical model of heat transfer in Linear Fresnel heat collection circuit.K-means method combined with radial basis function(RBF)was used to build a neural network prediction model to realize the prediction of molten salt temperature at the outlet of the heat collection loop.The neural network is trained dynamically by the measured data,and the training samples are pre-processed by the method of adaptive clustering analysis to reduce the complexity and improve the training speed of the network.The gradient descent method is used to dynamically adjust and determine the base function center and expansion constant of the hidden layer,and the network weight output by the base function is determined by the way of pseudo-inverse matrix.The simulation test shows that when the number of hidden layers is selected as 30,the prediction network can get a relatively ideal output result.The prediction model was applied to Linear Fresnel heat collection circuit of Dunhuang molten salt,and the maximum absolute error of network output was 121?by comparing the predicted output with the measured value in different operating environments for 4 days.The neural network prediction model can achieve a good prediction of the molten salt temperature at the outlet of the linear Fresnel concentrating light and heat collecting loop.Finally,according to the structural characteristics of Linear Fresnel concentrating light and heat collecting system,the mirror field control system is modularized and distributed in terms of software and hardware,and the reliability of the control network is improved through the redundancy design of software,hardware and communication network.The security of communication network is improved by VLAN network division.Different communication modes and protocols are adopted for data exchange between different control systems and terminals of demonstration power stations,which improves the timeliness of information exchange.The data packet transmission status of different devices is monitored through IO monitor,and the error rate of device request and response is 0.In the process of communication between master and slave devices,the balance of data receiving and receiving of each port is tested by monitoring the trigger status of each port within 1 min in real time at random,and the maximum unbalance of trigger of each port is 12.5%.After a long term test,the stability of the equipment meets the operating requirements of the system. |
PDF Full Text Request