近年来,基于扫描电镜的矿物自动分析仪如QEMSCAN(Quantitative Evaluation of Minerals by Scanning Electronic Microscopy)、MLA(Mineral Liberation Analyser)和AMICS(Advanced Mineral Identification and Characterization System...近年来,基于扫描电镜的矿物自动分析仪如QEMSCAN(Quantitative Evaluation of Minerals by Scanning Electronic Microscopy)、MLA(Mineral Liberation Analyser)和AMICS(Advanced Mineral Identification and Characterization System)已经广泛应用于矿物的自动识别与测量;激光剥蚀等离子质谱仪(LA-ICP-MS)由于其在痕量上的检测优势,使元素赋存状态的研究更加准确与深入;飞行时间二次离子质谱仪(TOF-SIMS)的应用为研究矿物表面特性及其变化提供了便利和可靠的方法。部分大型矿山企业在地质勘查阶段就对不同区段的样品进行工艺矿物学研究,关注矿石性质的空间变化给选矿工艺及指标带来的影响,为及时优化改进工艺流程提供依据和指导。就目前研究成果而言,矿物自动测量与分析技术的不断完善,建立数学模型预测选矿指标以及矿物三维数据表征技术在工艺矿物学研究中的应用是其持续发展的新趋势和方向。展开更多
State of charge(SOC) is a key parameter of lithium-ion battery. In this paper, a finite difference extended Kalman filter(FDEKF)with Hybrid Pulse Power Characterization(HPPC) parameters identification is proposed to e...State of charge(SOC) is a key parameter of lithium-ion battery. In this paper, a finite difference extended Kalman filter(FDEKF)with Hybrid Pulse Power Characterization(HPPC) parameters identification is proposed to estimate the SOC. The finite difference(FD) algorithm is benefit to compute the partial derivative of nonlinear function, which can reduce the linearization error generated by the extended Kalman filter(EKF). The FDEKF algorithm can reduce the computational load of controller in engineering practice without solving the Jacobian matrix. It is simple of dynamic model of lithium-ion battery to adopt a secondorder resistor-capacitor(2 RC) network, the parameters of which are identified by the HPPC. Two conditions, both constant current discharge(CCD) and urban dynamometer driving schedule(UDDS), are utilized to validate the FDEKF algorithm.Comparing convergence rate and accuracy between the FDEKF and the EKF algorithm, it can be seen that the former is a better candidate to estimate the SOC.展开更多
针对现有MOESP(multiple-input multiple-output output-error state space model identification)和N4SID(numerical algorithm for subspace state space systemidentification)算法在计算状态空间模型系统矩阵(A、B、C、D)时的不足,...针对现有MOESP(multiple-input multiple-output output-error state space model identification)和N4SID(numerical algorithm for subspace state space systemidentification)算法在计算状态空间模型系统矩阵(A、B、C、D)时的不足,提出1种改进的子空间辨识方法。该方法利用MOESP算法可以根据系统观测矩阵直接计算出系统矩阵A和输出矩阵C的优点,先计算矩阵A和C,然后采用N4SID算法计算输入矩阵B和前馈矩阵D。该方法既能够避免MOESP算法在计算矩阵B和D时需要构建大矩阵的缺点,又能避免N4SID算法在计算矩阵A和C时需要求解线性最小二乘的问题,降低了算法的复杂性。将该算法应用于某天然气电站和Alstom气化炉模型的辨识中,通过考核算法的CPU运算时间、CPU浮点数运算次数(floating-pointoperations,FLOPS)和相对误差等指标,将该算法与原有MOESP和N4SID算法进行了比较。计算结果表明,改进的子空间辨识算法能够在保证较好辨识精度的前提下,提高原有算法的计算效率,特别是在大容量数据样本条件下,能够有效降低CPU运算时间和FLOPS。展开更多
磷酸铁锂电池(LiFePO_4)凭借其较高的比功率、较高的比能量和使用寿命长等优点,逐渐成为电动汽车领域应用电池的研究热点。准确的电池荷电状态(State-of-Charge,简称SOC)估计离不开合理的等效模型,模型的精度与复杂程度直接影响电池SOC...磷酸铁锂电池(LiFePO_4)凭借其较高的比功率、较高的比能量和使用寿命长等优点,逐渐成为电动汽车领域应用电池的研究热点。准确的电池荷电状态(State-of-Charge,简称SOC)估计离不开合理的等效模型,模型的精度与复杂程度直接影响电池SOC估计。文章首先分析四种等效电路模型参数,基于电池测试系统对LiFePO_4进行混合脉冲功率性能测试(The Hybrid Pulse Power Characterization,简称HPPC),应用最小二乘法进行模型参数辨识,根据辨识结果对比模型精度与复杂程度;然后将电池在整个SOC周期内对三个带有RC网络结构的模型精度进行对比;最后对LiFePO_4电池进行联邦城市运行工况(The Federal Urban Driving Schedule,简称FUDS)测试,对比双极化等效电路模型在不同温度下模型精度。实验结果显示双极化等效电路模型为具有良好精度的等效电路模型,同时带有低温特性。展开更多
文摘近年来,基于扫描电镜的矿物自动分析仪如QEMSCAN(Quantitative Evaluation of Minerals by Scanning Electronic Microscopy)、MLA(Mineral Liberation Analyser)和AMICS(Advanced Mineral Identification and Characterization System)已经广泛应用于矿物的自动识别与测量;激光剥蚀等离子质谱仪(LA-ICP-MS)由于其在痕量上的检测优势,使元素赋存状态的研究更加准确与深入;飞行时间二次离子质谱仪(TOF-SIMS)的应用为研究矿物表面特性及其变化提供了便利和可靠的方法。部分大型矿山企业在地质勘查阶段就对不同区段的样品进行工艺矿物学研究,关注矿石性质的空间变化给选矿工艺及指标带来的影响,为及时优化改进工艺流程提供依据和指导。就目前研究成果而言,矿物自动测量与分析技术的不断完善,建立数学模型预测选矿指标以及矿物三维数据表征技术在工艺矿物学研究中的应用是其持续发展的新趋势和方向。
基金supported by the National Key Research and Development Program of China(Grant No.2017YFB0103100)the Science and Technology Special Project of Anhui Province(Grant No.18030901063)
文摘State of charge(SOC) is a key parameter of lithium-ion battery. In this paper, a finite difference extended Kalman filter(FDEKF)with Hybrid Pulse Power Characterization(HPPC) parameters identification is proposed to estimate the SOC. The finite difference(FD) algorithm is benefit to compute the partial derivative of nonlinear function, which can reduce the linearization error generated by the extended Kalman filter(EKF). The FDEKF algorithm can reduce the computational load of controller in engineering practice without solving the Jacobian matrix. It is simple of dynamic model of lithium-ion battery to adopt a secondorder resistor-capacitor(2 RC) network, the parameters of which are identified by the HPPC. Two conditions, both constant current discharge(CCD) and urban dynamometer driving schedule(UDDS), are utilized to validate the FDEKF algorithm.Comparing convergence rate and accuracy between the FDEKF and the EKF algorithm, it can be seen that the former is a better candidate to estimate the SOC.
文摘针对现有MOESP(multiple-input multiple-output output-error state space model identification)和N4SID(numerical algorithm for subspace state space systemidentification)算法在计算状态空间模型系统矩阵(A、B、C、D)时的不足,提出1种改进的子空间辨识方法。该方法利用MOESP算法可以根据系统观测矩阵直接计算出系统矩阵A和输出矩阵C的优点,先计算矩阵A和C,然后采用N4SID算法计算输入矩阵B和前馈矩阵D。该方法既能够避免MOESP算法在计算矩阵B和D时需要构建大矩阵的缺点,又能避免N4SID算法在计算矩阵A和C时需要求解线性最小二乘的问题,降低了算法的复杂性。将该算法应用于某天然气电站和Alstom气化炉模型的辨识中,通过考核算法的CPU运算时间、CPU浮点数运算次数(floating-pointoperations,FLOPS)和相对误差等指标,将该算法与原有MOESP和N4SID算法进行了比较。计算结果表明,改进的子空间辨识算法能够在保证较好辨识精度的前提下,提高原有算法的计算效率,特别是在大容量数据样本条件下,能够有效降低CPU运算时间和FLOPS。
文摘磷酸铁锂电池(LiFePO_4)凭借其较高的比功率、较高的比能量和使用寿命长等优点,逐渐成为电动汽车领域应用电池的研究热点。准确的电池荷电状态(State-of-Charge,简称SOC)估计离不开合理的等效模型,模型的精度与复杂程度直接影响电池SOC估计。文章首先分析四种等效电路模型参数,基于电池测试系统对LiFePO_4进行混合脉冲功率性能测试(The Hybrid Pulse Power Characterization,简称HPPC),应用最小二乘法进行模型参数辨识,根据辨识结果对比模型精度与复杂程度;然后将电池在整个SOC周期内对三个带有RC网络结构的模型精度进行对比;最后对LiFePO_4电池进行联邦城市运行工况(The Federal Urban Driving Schedule,简称FUDS)测试,对比双极化等效电路模型在不同温度下模型精度。实验结果显示双极化等效电路模型为具有良好精度的等效电路模型,同时带有低温特性。
