论文部分内容阅读
丙烯精馏塔关键组分塔顶的丙烷浓度和塔釜的丙烯浓度的准确测量是乙烯生产企业提高丙烯收率的关键。鉴于丙烷浓度和丙烯浓度分析仪经常出现故障,提出以RBF神经网络加协同随机粒子群优化(PSO)算法的软测量建模法,即利用RBF神经网络的局部逼近能力来获得模型的结构,利用协同随机PSO算法的全局搜索能力来优化模型的参数,提高模型的逼近能力和泛化能力。该方法克服了BP网络对初始值和网络结构敏感,容易陷入局部最优的缺陷,以及RBF网络全局逼近能力差的缺点。仿真结果表明,此方法所得软测量模型精度高,泛化能力强。
The accurate measurement of the propane concentration at the top of the main component of the propylene rectification column and the propylene concentration in the bottom of the reactor is the key for ethylene production enterprises to increase the propylene yield. In view of the frequent failures of propane concentration and propylene concentration analyzers, a soft-sensing modeling method based on RBF neural network and PSO algorithm is proposed, which uses the local approximation ability of RBF neural network to obtain the model structure. Cooperate with the global search ability of random PSO algorithm to optimize the parameters of the model and improve the approximation ability and generalization ability of the model. This method overcomes the shortcomings that the BP network is sensitive to the initial value and the network structure, easily falls into the local optimum and the global approximation ability of the RBF network is poor. Simulation results show that the soft sensor model obtained by this method has high precision and strong generalization ability.