从数据中提取生料浆配料过程运行控制的模糊规则

将减法聚类、神经网络、最相邻原则、对提取后的规则进行调整等方法相结合,从过程数据中自动提取出模糊规则,从而实现在氧化铝生料浆配料过程中将生料浆的各项质量指标控制在目标值范围内.某氧化铝厂的应用结果表明,所提取的模糊规则不仅具有良好的完备性和可解释性,同时可根据工况的变化自动调整各个控制回路的设定值,实现了该过程的优化运行.

Fuzzy rules extraction from process data for operation control of the raw slurry blending process

Subtract clustering, neural network, most neighbor principle and regulation for extracted rules are combined
to extract fuzzy rules from process data. During the operation of raw slurry blending process, the quality indices of raw
slurry is kept in their targeted ranges. The proposed method is applied to the operation control of the raw slurry blending
process. Application results show that the extracted fuzzy rules not only have good completeness and interpretation, but also
can realize the operation control objective of the raw slurry blending process.