选矿过程工艺指标确定方法及应用

Approach for Determining Technical Indices in Mineral Process and its Applications

当前选矿生产实现了对个别工序工艺指标的优化控制,但各工序的工艺指标目标值仍然由选矿工程师根据入选原矿情况来确定,具有较大的模糊性与随意性,各工序之间不能有效协调,无法满足对更好综合生产指标的追求。因此,以某选矿厂为实际背景,设计了一套基于案例推理技术确定各工序工艺指标目标值的系统,为避免人工确定初始案例的缺点,采用动态自组织映射神经网络从历史数据中提取了若干初始案例,该系统与各工序原有的过程优化控制系统相配合在实际生产中取得了较好效果。

The technical index optimization control is realized currently in individual section of mineral processing, but the targets in each section of mineral processing are still determined by the mineral engineers according to the feature of ores. With large extent of fuzziness and randomness, it is hard to cooperate effectively between the sections, and the purpose for better global production indices can not be reached easily. Considering the actual background in mineral processing plant, a system is designed to determine the technical indices of each section based on the technology of case-based reasoning. The system adopts the dynamic self-organizing mapping （SOM） network to pick-up the cases from the historical data. Preferable effect is obtained in practice with the cooperation between this system and the process optimization control system in each section.