The electricity-hydrogen integrated energy system(EH-IES)enables synergistic operation of electricity,heat,and hydrogen subsystems,supporting renewable energy integration and efficient multi-energy utilization in futu...The electricity-hydrogen integrated energy system(EH-IES)enables synergistic operation of electricity,heat,and hydrogen subsystems,supporting renewable energy integration and efficient multi-energy utilization in future low carbon societies.However,uncertainties from renewable energy and load variability threaten system safety and economy.Conventional chance-constrained programming(CCP)ensures reliable operation by limiting risk.However,increasing source-load uncertainties that can render CCP models infeasible and exacerbate operational risks.To address this,this paper proposes a risk-adjustable chance-constrained goal programming(RACCGP)model,integrating CCP and goal programming to balance risk and cost based on system risk assessment.An intelligent nonlinear goal programming method based on the state transition algorithm(STA)is developed,along with an improved discretized step transformation,to handle model nonlinearity and enhance computational efficiency.Experimental results show that the proposed model reduces costs while controlling risk compared to traditional CCP,and the solution method outperforms average sample sampling in efficiency and solution quality.展开更多
针对风力发电机组变桨系统故障诊断模型参数难以优化问题,提出了基于状态转移算法优化多类最优间隔分布机(multi-class Optimal Margin Distribution Machine optimized by the State Transition Algorithm,mcODM-STA)的风电机组变桨系...针对风力发电机组变桨系统故障诊断模型参数难以优化问题,提出了基于状态转移算法优化多类最优间隔分布机(multi-class Optimal Margin Distribution Machine optimized by the State Transition Algorithm,mcODM-STA)的风电机组变桨系统故障诊断方法.该方法选择风电机组功率输出作为主要状态参数,利用Pearson相关系数对风电数据采集与监视控制系统中风电机组历史运行数据进行相关性分析,剔除与功率输出状态参数相关性较低的特征,对余下特征进行二次分析,减少样本特征.将数据集分为训练集和测试集,训练集用来训练所提故障诊断模型,测试集用来进行测试.利用国内风电场实际运行数据进行实验验证.实验结果表明,与其他多种参数优化方法相比,所提方法故障诊断准确率和Kappa系数更高.展开更多
蚁群算法拥有良好的全局性、自组织性、鲁棒性,但传统蚁群算法存在许多不足之处。为此,针对算法在路径规划问题中的缺陷,在传统蚁群算法的状态转移公式中,引入目标点距离因素和引导素,加快算法收敛性和改善局部最优缺陷。在带时间窗的...蚁群算法拥有良好的全局性、自组织性、鲁棒性,但传统蚁群算法存在许多不足之处。为此,针对算法在路径规划问题中的缺陷,在传统蚁群算法的状态转移公式中,引入目标点距离因素和引导素,加快算法收敛性和改善局部最优缺陷。在带时间窗的车辆路径问题(vehicle routing problem with time windows,VRPTW)上,融合蚁群算法和遗传算法,并将顾客时间窗宽度以及机器人等待时间加入蚁群算法状态转移公式中,以及将蚁群算法的解作为遗传算法的初始种群,提高遗传算法的初始解质量,然后进行编码,设置违反时间窗约束和载重量的惩罚函数和适应度函数,在传统遗传算法的交叉、变异操作后加入了破坏-修复基因的操作来优化每一代新解的质量,在Solomon Benchmark算例上进行仿真,对比算法改进前后的最优解,验证算法可行性。最后在餐厅送餐问题中把带有障碍物的仿真环境路径规划问题和VRPTW问题结合,使用改进后的算法解决餐厅环境下送餐机器人对顾客服务配送问题。展开更多
基金Project(2022YFC2904502)supported by the National Key Research and Development Program of ChinaProject(62273357)supported by the National Natural Science Foundation of China。
文摘The electricity-hydrogen integrated energy system(EH-IES)enables synergistic operation of electricity,heat,and hydrogen subsystems,supporting renewable energy integration and efficient multi-energy utilization in future low carbon societies.However,uncertainties from renewable energy and load variability threaten system safety and economy.Conventional chance-constrained programming(CCP)ensures reliable operation by limiting risk.However,increasing source-load uncertainties that can render CCP models infeasible and exacerbate operational risks.To address this,this paper proposes a risk-adjustable chance-constrained goal programming(RACCGP)model,integrating CCP and goal programming to balance risk and cost based on system risk assessment.An intelligent nonlinear goal programming method based on the state transition algorithm(STA)is developed,along with an improved discretized step transformation,to handle model nonlinearity and enhance computational efficiency.Experimental results show that the proposed model reduces costs while controlling risk compared to traditional CCP,and the solution method outperforms average sample sampling in efficiency and solution quality.
文摘针对风力发电机组变桨系统故障诊断模型参数难以优化问题,提出了基于状态转移算法优化多类最优间隔分布机(multi-class Optimal Margin Distribution Machine optimized by the State Transition Algorithm,mcODM-STA)的风电机组变桨系统故障诊断方法.该方法选择风电机组功率输出作为主要状态参数,利用Pearson相关系数对风电数据采集与监视控制系统中风电机组历史运行数据进行相关性分析,剔除与功率输出状态参数相关性较低的特征,对余下特征进行二次分析,减少样本特征.将数据集分为训练集和测试集,训练集用来训练所提故障诊断模型,测试集用来进行测试.利用国内风电场实际运行数据进行实验验证.实验结果表明,与其他多种参数优化方法相比,所提方法故障诊断准确率和Kappa系数更高.
文摘蚁群算法拥有良好的全局性、自组织性、鲁棒性,但传统蚁群算法存在许多不足之处。为此,针对算法在路径规划问题中的缺陷,在传统蚁群算法的状态转移公式中,引入目标点距离因素和引导素,加快算法收敛性和改善局部最优缺陷。在带时间窗的车辆路径问题(vehicle routing problem with time windows,VRPTW)上,融合蚁群算法和遗传算法,并将顾客时间窗宽度以及机器人等待时间加入蚁群算法状态转移公式中,以及将蚁群算法的解作为遗传算法的初始种群,提高遗传算法的初始解质量,然后进行编码,设置违反时间窗约束和载重量的惩罚函数和适应度函数,在传统遗传算法的交叉、变异操作后加入了破坏-修复基因的操作来优化每一代新解的质量,在Solomon Benchmark算例上进行仿真,对比算法改进前后的最优解,验证算法可行性。最后在餐厅送餐问题中把带有障碍物的仿真环境路径规划问题和VRPTW问题结合,使用改进后的算法解决餐厅环境下送餐机器人对顾客服务配送问题。
