一种基于IBKA-GBDT的火控系统故障预测方法  

A fire control system fault prediction method based on IBKA-GBDT

作　　者：于昂 李英顺 郭占男 曹胜冲 赵恒 YU Ang;LI Yingshun;GUO Zhannan;CAO Shengchong;ZHAO Heng(School of Information Engineering,Beijing Institute of Petrochemical Technology,Beijing 102617,China;School of Control Science and Engineering,Dalian University of Technology,Dalian 116024,China)

机构地区：[1]北京石油化工学院信息工程学院,北京102617 [2]大连理工大学控制科学与工程学院,辽宁大连116024

出　　处：《兵器装备工程学报》2024年第12期169-177,共9页Journal of Ordnance Equipment Engineering

基　　金：辽宁省“兴辽英才计划”项目(XLYC1903015)。

摘　　要：火控系统是坦克作战的核心组件,通过提供高精度目标打击、快速反应以及全天候作战支持,显著提升坦克的战场生存能力和作战效能,因此对其进行故障预测极为重要。为提高故障预测的准确性并减少成本,提出了一种基于混合策略改进的黑翅鸢算法优化梯度提升决策树的模型预测方法。采用灰色关联度方法处理原始数据,以减少数据冗余和降低维度,并选择关联度高的属性来构建数据集。引入Logistic混沌映射、螺旋搜索策略以及三角形游走策略对黑翅鸢算法进行改进,进一步优化梯度提升决策树关键参数,构建故障预测模型实现对预测数据的故障预测。同时,选取火控系统电气部件试验台采集的信号数据作为实验对象,设置相同参数与传统梯度提升决策树、鲸鱼优化算法和黑翅鸢优化算法优化的梯度提升决策树模型进行实验对比。实验结果表明,该方法能够快速准确地对处理后的数据集进行故障预测,平均准确率达到了96.74%,为火控系统的后续维护和维修提供了重要依据。The fire control system is an essential element of tank operations,greatly enhancing battlefield survivability and tactical efficiency through precise strikes,swift responses,and all-conditions support.As such,accurately predicting faults within this system is crucial.To enhance fault prediction accuracy and reduce operational costs,a model prediction method based on a hybrid strategy-improved Black-winged Kite Algorithm to optimize the Gradient Boosting Decision Tree(GBDT)is proposed.The grey relational analysis method is used to process the raw data to reduce data redundancy and dimensionality,and highly correlated attributes are selected to construct the dataset.Logistic chaotic mapping,spiral search strategy,and triangular walk strategy are introduced to improve the Black-winged Kite Algorithm,further optimizing the key parameters of the Gradient Boosting Decision Tree and constructing a fault prediction model to achieve fault prediction for the predicted data.Additionally,signal data collected from the fire control system’s electrical component test bench was used as the experimental subject,setting the same parameters to conduct comparative experiments with traditional gradient boosting decision trees,whale optimization algorithms,and Black-winged Kite optimized gradient boosting decision tree models.Experimental results demonstrate that this method can quickly and accurately predict faults in the processed dataset,achieving an average accuracy rate of 96.74%,providing a crucial basis for subsequent maintenance and repair of the fire control system.

关 键 词：火控系统 故障预测 黑翅鸢优化算法 梯度提升决策树 灰色关联度分析 

分 类 号：TJ811.2[兵器科学与技术—武器系统与运用工程]