基于IGWO-BPNN的露天矿卡车故障预测方法

An IGWO-BPNN-based method for open-pit mine truck failure prediction

为有效解决露天矿中卡车的故障预测问题,提出了一种基于改进灰狼算法的BP神经网络模型,并成功应用于预测露天矿卡车故障次数和故障持续时间。首先,针对传统灰狼算法的不足,引入了新的非线性更新机制和基于线性插值的种群更新机制,提出了融合多策略的改进灰狼优化算法。其次,将IGWO应用于BP神经网络的权值和阈值搜索中,形成了基于IGWO的BP神经网络模型(IGWO-BPNN)。最后,以宝日希勒露天煤矿卡车故障数据为例,成功将该模型应用于卡车故障预测研究。结果表明,在相同实验条件下,与其他算法相比,IGWO-BPNN具有更高的模型预测性能和分类精度,可帮助露天矿山科学制定卡车预防性检修计划,并为智慧露天矿山建设提供科学有效的基础决策数据。

To effectively address the truck failure prediction problem in open-pit mines, we proposed an improved grey wolf algorithm-based BP neural network model, which was successfully applied to predict the frequency and duration of truck failures in open-pit mines. Firstly, considering the limitations of the traditional grey wolf algorithm, new nonlinear updating mechanisms and population updating mechanisms based on linear interpolation were introduced, leading to the development of a multi-strategy integrated improved gray wolf optimizer(IGWO). Secondly, IGWO was applied to search for the weights and thresholds of the BP neural network, forming the IGWO-based BP neural network model(IGWO-BPNN). Finally, using the failure data of trucks from the Baorixile open-pit mine, the IGWO-BPNN model was successfully employed in truck failure prediction research. Results demonstrate that under the same experimental conditions, compared to other algorithms, IGWO-BPNN exhibited superior predictive performance and classification accuracy, effectively aiding open-pit mining enterprises in developing scientifically sound truck preventive maintenance plans and providing scientifically valid foundational decision-making data for the construction of intelligent open-pit mining.