微型电脑应用2026，Vol.42Issue(5)：10-14,5.
基于Dijkstra算法的煤矿矿井水水泵动态调度方法
Dynamic Scheduling Method of Coal Mine Water Pump Based on Dijkstra Algorithm
摘要
Abstract
Because of the traditional control strategy of"avoiding peaks and taking valleys",the pump start-up data can not be adjusted according to the water inflow,which leads to the pump can not be scheduled according to the real-time water level and water inflow changes,thus increasing the scheduling cost.Therefore,a dynamic dispatching method of mine water pump based on Dijkstra algorithm is proposed.Based on the data layer,platform layer,application layer and network layer,the SSM light-weight framework is established to build the digital twin drainage system,and the digital twin drainage system is controlled by the improved"avoiding peaks and valleys"control strategy,and the number of pumps is adjusted according to the water inflow.At the same time,the grey wolf optimization-differential evolution-support vector machine(GWO-DE-SVM)model is used to predict the water inflow,and the problem of pump scheduling is transformed into the problem of solving the minimum electrici-ty charge.Based on the improved control strategy and the all-day water inflow,the mathematical model of pump scheduling is constructed,and the Dijkstra algorithm is used to solve the minimum power consumption scheme,so that the start-stop opera-tion of pumps can be reversed,and the dynamic scheduling of water pumps in coal mines can be realized.The results show that the absolute error and relative error of GWO-DE-SVM prediction are small.When the actual value is 471.6941 m3/h,the abso-lute error and relative error of GWO-DE-SVM model prediction value are 3.9344 m3/h and 0.0083％.After adopting the im-proved strategy,the total pump time is 44 hours,and the electricity bill is 15 541.86 yuan,which is 5 hours and 445.46 yuan lower than the traditional"avoiding peaks and valleys"control strategy.Therefore,the proposed scheduling method can im-prove the dynamic scheduling effect of mine water pumps and reduce the scheduling cost.关键词
数字孪生/动态调度/涌水量预测/避峰就谷/SVMKey words
digital twin/dynamic scheduling/water inflow prediction/avoiding peaks and valleys/SVM分类
信息技术与安全科学引用本文复制引用
杨晓娟,李进,李华..基于Dijkstra算法的煤矿矿井水水泵动态调度方法[J].微型电脑应用,2026,42(5):10-14,5.基金项目
国家自然科学基金(61673128) （61673128）
