选矿厂设计中采用的磨矿分级数据，一般由实验室模拟生产获取，但由于设备差异较大，在平均粒级分布率相同的情况下，矿物粒度分布与实际生产契合度差，因此可依据实际生产情况进行不断优化。对某含金铜多金属矿石进行了实验室磨矿和生产磨矿方式对比试验。结果表明：采用生产磨矿方式，获得的含金铜精矿铜品位为16.82 %、铜回收率为84.10 %，铜回收率提高了1.30百分点;金品位提高到64.00 g/t、金回收率为67.25 %，金回收率提高了6.80百分点。含金磁铁精矿和含金硫精矿不再磨，直接进行氰化浸出，获得的金浸出率分别为75.89 %和60.67 %，可取消根据实验室磨矿数据设计的含金磁铁精矿和含金硫精矿再磨工艺。研究结果为类似选矿厂优化流程结构及参数，提高流程稳定性，降低能耗，实施技术改造提供了理论依据。

The grinding grading data used in the design of ore-dressing plants are generally obtained from lab simu-lation.However,due to equipment difference,the mineral grain size distribution goes poorly with actual production when average grain size distribution rates are the same.Therefore it can be constantly optimized according to actual production.Lab grinding and actual grinding are compared and tested for a polymetallic ore containing gold and copper.The results show that better indexes can be obtained with actual grinding:copper grade is 16.82 %,copper reco-very rate is 84.10 % for copper concentrates containing gold,copper recovery rate is increased by 1.30 percentage points,gold grade in copper concentrates is increased to 64.00 g/t,gold recovery rate is 67.25 % gold recovery rate is increased by 6.80 percentage points;gold-bearing magnetite concentrate and gold-bearing sulfur concentrate will not be further ground and directly go to cyanidation leaching,and the gold leaching rates are 75.89 % and 60.67 % respectively,which means it is possible to cancel the regrinding process designed according to lab grinding data for magnetite and sulfur concentrates.The research results can be used as reference by similar ore-dressing plants in optimizing flowsheet structure and parameters,improving flowsheet stability,lowering power consumption and implementing technical renovation.