Efficient tool condition monitoring techniques help to realize intelligent management of tool life and reduce tool usage costs.In this paper,the influence of different wear degrees of ball-end milling cutters on the t...Efficient tool condition monitoring techniques help to realize intelligent management of tool life and reduce tool usage costs.In this paper,the influence of different wear degrees of ball-end milling cutters on the texture shape of machining tool marks is investigated,and a method is proposed for predicting the wear state(including the position and degree of tool wear)of ball-end milling cutters based on entropy measurement of tool mark texture images.Firstly,data samples are prepared through wear experiments,and the change law of the tool mark texture shape with the tool wear state is analyzed.Then,a two-dimensional sample entropy algorithm is developed to quantify the texture morphology.Finally,the processing parameters and tool attitude are integrated into the prediction process to predict the wear value and wear position of the ball end milling cutter.After testing,the correlation between the predicted value and the standard value of the proposed tool condition monitoring method reaches 95.32%,and the accuracy reaches 82.73%,indicating that the proposed method meets the requirement of tool condition monitoring.展开更多
交叉熵法可显著加速电网可靠性评估,但往往聚焦于独立随机变量,若将其拓展至相关性变量可进一步提升加速性能。为有效获取相关性变量的重要抽样密度函数以实现其重要抽样,针对相关性建模中广泛使用的核密度估计模型(kernel density esti...交叉熵法可显著加速电网可靠性评估,但往往聚焦于独立随机变量,若将其拓展至相关性变量可进一步提升加速性能。为有效获取相关性变量的重要抽样密度函数以实现其重要抽样,针对相关性建模中广泛使用的核密度估计模型(kernel density estimation,KDE)开展了交叉熵优化研究。因KDE模型不属于指数分布家族,传统交叉熵优化难以实施,故利用复合抽样算法特点提出了新颖的直接交叉熵优化方法,推导出KDE模型最优权重参数的解析表达式。因权重参数数量级较小,直接优化易导致准确性退化,故基于子集模拟思想进一步提出间接交叉熵优化方法,将较小的权重参数优化转换成较大的条件概率优化,提升了优化准确性。通过MRTS79和MRTS96可靠性测试系统的评估分析,验证了所提方法在含相关性变量电网可靠性评估中的高效加速性能。展开更多
超短期电力负荷预测作为电力系统的基本组成,能为生产调度计划的制定提供重要依据。然而,电力负荷具有非线性、时变性和不确定性,充分挖掘其潜在特征并分别预测,是提升预测准确性的关键。提出一种基于自适应局部迭代滤波(adaptive local...超短期电力负荷预测作为电力系统的基本组成,能为生产调度计划的制定提供重要依据。然而,电力负荷具有非线性、时变性和不确定性,充分挖掘其潜在特征并分别预测,是提升预测准确性的关键。提出一种基于自适应局部迭代滤波(adaptive local iterative filtering,ALIF)的BiGRU-Attention-XGBoost电力负荷组合预测模型。该模型基于ALIF-SE实现将历史负荷序列分解重组为周期序列、波动序列和趋势序列;通过Attention机制对BiGRU模型进行改进,并结合XGBoost模型构建基于时变权重组合的电力负荷预测模型。实验分析表明,输入模型数据经过ALIF-SE处理后预测精度有明显提升;所提组合模型在工作日和节假日均具有较好的预测效果,预测误差大部分在5%以下;通过在不同负荷数据集下进行实验对比,验证了所提预测方法的可迁移性。实验结果证明,所提模型具有有效性、准确性和可行性。展开更多
基金Project(51975169)supported by the National Natural Science Foundation of ChinaProject(LH2022E085)supported by the Natural Science Foundation of Heilongjiang Province,China。
文摘Efficient tool condition monitoring techniques help to realize intelligent management of tool life and reduce tool usage costs.In this paper,the influence of different wear degrees of ball-end milling cutters on the texture shape of machining tool marks is investigated,and a method is proposed for predicting the wear state(including the position and degree of tool wear)of ball-end milling cutters based on entropy measurement of tool mark texture images.Firstly,data samples are prepared through wear experiments,and the change law of the tool mark texture shape with the tool wear state is analyzed.Then,a two-dimensional sample entropy algorithm is developed to quantify the texture morphology.Finally,the processing parameters and tool attitude are integrated into the prediction process to predict the wear value and wear position of the ball end milling cutter.After testing,the correlation between the predicted value and the standard value of the proposed tool condition monitoring method reaches 95.32%,and the accuracy reaches 82.73%,indicating that the proposed method meets the requirement of tool condition monitoring.
文摘交叉熵法可显著加速电网可靠性评估,但往往聚焦于独立随机变量,若将其拓展至相关性变量可进一步提升加速性能。为有效获取相关性变量的重要抽样密度函数以实现其重要抽样,针对相关性建模中广泛使用的核密度估计模型(kernel density estimation,KDE)开展了交叉熵优化研究。因KDE模型不属于指数分布家族,传统交叉熵优化难以实施,故利用复合抽样算法特点提出了新颖的直接交叉熵优化方法,推导出KDE模型最优权重参数的解析表达式。因权重参数数量级较小,直接优化易导致准确性退化,故基于子集模拟思想进一步提出间接交叉熵优化方法,将较小的权重参数优化转换成较大的条件概率优化,提升了优化准确性。通过MRTS79和MRTS96可靠性测试系统的评估分析,验证了所提方法在含相关性变量电网可靠性评估中的高效加速性能。
文摘超短期电力负荷预测作为电力系统的基本组成,能为生产调度计划的制定提供重要依据。然而,电力负荷具有非线性、时变性和不确定性,充分挖掘其潜在特征并分别预测,是提升预测准确性的关键。提出一种基于自适应局部迭代滤波(adaptive local iterative filtering,ALIF)的BiGRU-Attention-XGBoost电力负荷组合预测模型。该模型基于ALIF-SE实现将历史负荷序列分解重组为周期序列、波动序列和趋势序列;通过Attention机制对BiGRU模型进行改进,并结合XGBoost模型构建基于时变权重组合的电力负荷预测模型。实验分析表明,输入模型数据经过ALIF-SE处理后预测精度有明显提升;所提组合模型在工作日和节假日均具有较好的预测效果,预测误差大部分在5%以下;通过在不同负荷数据集下进行实验对比,验证了所提预测方法的可迁移性。实验结果证明,所提模型具有有效性、准确性和可行性。
