基于模糊综合与神经网络的制丝工艺多因素评价模型

为解决将制丝工艺质量、设备运行状态、生产消耗作为影响因素联合对制丝工艺进行综合等级判定的问题,建立了基于层次分析法、线性投影法的多因素模糊综合评价模型和神经网络评价模型。通过对甲乙丙3个班组9~12月份质量系数、断料情况等进行建模,表明:神经网络评价模型既可以用于验证多因素模糊综合评价模型的合理性与准确性,也可独立对制丝工艺综合等级进行判定。两种方法相结合,互相验证,为提高制丝工艺综合生产水平提供了科学、简洁的依据,对查找工艺质量、设备运行状态、生产消耗指标下的各项不良因素提供了支持。

Application of fuzzy comprehensive and Neural network evaluation models on tobacco primary processing line quality

To solve the problems on determining comprehensive level of pipe tobacco technology via pipe tobacco quality, equipment operation and production consumption, the Fuzzy comprehensive and neural network evaluation models were built based on analytic hierarchy process, linear projection method. The modeling results of the three groups based on the mass coefficient and cutting material condition from September to December indicated that neural network evaluation model could not only verify the rationality and accuracy of multi-factor fuzzy comprehensive evaluation model, but also could independently determine the comprehensive level of pipe tobacco technology. It not only offers scientific evidence to improve production levels of pipe tobacco, but also provides support to find negetive factors in process quality, equipment operation and production consumption by combing with the two methods.