论文部分内容阅读
针对采用关联向量机进行软测量建模所存在的多输出建模问题,提出了一种鱼群优化算法(AFSA)—多输出关联向量机(MVRVM)软测量建模方法。通过加权组合全局性Poly核函数和局部性Gauss核函数,形成混合核函数多输出关联向量机模型,有效融合多特征数据信息;然后采用鱼群优化算法对多输出关联向量机模型的相关核参数进行优化,以进一步改善模型的输出精度和稳定性。将该建模方法应用于甲醇制烯烃生产过程(MTO)反应器出口乙烯和丙烯(简称双烯)收率软测量研究中,结果表明:采用该建模方法所建立的软测量模型能有效预测双烯收率变化,具有较高的预测精度和稳定性,这可为复杂化工过程多参数监测与控制提供有力的技术方法支持。
In order to solve the problem of multi-output modeling based on SVM, this paper proposed a fish-swarm optimization algorithm (AFSA) -multi-output output associated vector machine (MVRVM) soft-sensing modeling method. By combining weighted global polynomial function and local Gauss kernel function, a hybrid kernel-based multi-output correlation vector machine model is formed to effectively fuse multi-feature data information. Then, the fish-population optimization algorithm is used to estimate the correlation kernel parameters Optimized to further improve the model’s output accuracy and stability. The modeling method was applied to the soft sensor research on the yield of ethylene and propylene (MLC) at the exit of methanol-to-olefins production (MTO) reactor. The results show that the soft sensor model established by this modeling method can effectively predict The yield of diolefins with high prediction accuracy and stability can provide powerful technical support for the multi-parameter monitoring and control of complex chemical processes.