多时段不等长间歇过程运行状态评价及非优状态深层次原因追溯研究

With the urging market requirement for various product types, the manufacturing of higher-value-added products that are mainly produced through batch processes have become increasingly important in many industries. In order to obtain high-quality products and improve the economic benefits, the online assessment of the operating performance and nonoptimal deep-seated cause identification for batch processes have important significance. The project is to propose a systematic operating performance assessment and nonoptimal cause identification strategy for batch processes. According to the sequential of process running and the characteristics of the data distribution changing with phases, the successive phase division method was proposed based on Gaussian mixture model(GMM); through fully consideration about the impact of the correlation and synergy between phases on the process operating performance, the within-phases assessment models and between-phases assessment models were established respectively for evaluating the process operating performance from multiple perspectives; based on the idea of the same performance grade containing similar process variation information, the online assessment strategy are developed for batch processes; for the nonoptimal operating performance, the nonoptimal deep-rooted cause identification method was proposed based on the real time measurements and expert knowledge. The proposed method was validated and improved on the existing experimental platform and gradually applied to one or two typical batch industrial processes.

随着现代社会对多规格产品的迫切市场需求，工业生产更加倚重于生产小批量、高附加值产品的间歇过程。为了获得优质产品以及更高的经济效益，实现间歇过程运行状态评价及非优状态的深层次原因追溯具有重要意义。本项目针对多时段不等长间歇过程，提出一套系统的运行状态评价及非优原因追溯总体解决方案：根据间歇过程生产运行的时序性及过程数据分布随着时段改变的特点，提出基于高斯混合模型的时段逐次划分方法；充分考虑时段间的相关关系与协同作用对间歇过程运行状态的影响，分别建立状态等级时段内和时段间评价模型，为从多角度评价过程运行状态提供参考依据；基于相同状态等级包含相似过程变异信息的思想，提出间歇过程运行状态在线评价方法；将过程实时可测信息与专家知识有机融合，提出非优状态的深层次原因追溯方法；对所提出的方法在已有的实验平台上进行验证与完善，并逐渐将其推广应用到1-2个典型的间歇工业生产过程。

随着人们生活质量的大幅提升，对于小批量、高附加值的间歇过程产品的需求日益增大，间歇过程在工业制造中地位越来越重要，使得面向间歇过程的运行状态评价更具实际意义。间歇过程运行状态评价是指在生产过程运行状态正常，且在综合考虑产品质量、原材料消耗以及能源消耗的基础上，对当前间歇操作的生产运行状态的优劣情况做进一步的区分与识别，同时在运行状态非优的情况下，及时查找出其主导原因，并将评价结果和原因反馈给生产管理者和实际操作人员，以便对生产操作进行调整，最终确保生产企业的综合经济收益。本项目深入研究多时段不等长间歇过程运行状态评价中的关键问题，在离线时段划分、时段内和时段间状态相关特征提取、运行状态在线评价、非优状态深层次原因追溯等关键技术上有所突破，并把提出的方法和技术推广应用到1-2个典型的复杂工业生产过程。