论文部分内容阅读
溶剂回收系统的优化目标是在保证润滑油质量的前提下尽量降低能量消耗，由于外界干扰大、参数波动大、存在较大滞后等因素，能耗优化数学模型难以满足在线优化的要求，为此提出了一种数学建模与优化、专家系统建模与优化相结合的混合优化策略．为满足能耗优化的需要，采用了一种基于ＢＰ神经网络的润滑油质量指标“闪点”的软测量技术和一种保证蒸发塔温度控制的非线性预测算法．实际应用结果证明该混合优化策略是成功的．
The optimization goal of solvent recovery system is to minimize the energy consumption under the premise of ensuring the quality of lubricating oil. Due to the large external disturbance, large parameter fluctuation and large lag, the mathematical model of energy consumption optimization can not meet the requirements of online optimization. A hybrid optimization strategy combining mathematical modeling and optimization, expert system modeling and optimization is proposed. In order to meet the need of energy consumption optimization, a soft-sensing technique based on BP neural network, a “flash point” of lubricating oil quality, and a nonlinear prediction algorithm to ensure the temperature control of the evaporation tower are adopted. The practical application shows that the hybrid optimization strategy is successful.