选取150家火电企业的煤质检测数据,通过分析泛在煤质化验数据信息,构建L-M算法下的煤质发热量的预测模型。实验结果表明:(1)煤质化验数据中仅碳(Cd)、灰分(Aad)与发热量(Qgr,ad)的线性关系较为显著,相关系数 R2为0.8768和0.6880;(2)主成分分析法挖掘出影响煤质发热量的主成分特征值、特征矩阵及得分,实现了由六维矩阵降至四维矩阵的降维效果,增强了神经网络在训练过程中收敛的稳定性;(3)基于L-M算法下,改进的BP神经网络预测模型(LMBP)的训练集系数Rt、验证集系数Ra和测试集系数Rm分别为0.9957、0.9942和0.9963,总体系数为0.9931,同时通过待测20组数据进一步验证了LMBP预测模型可靠,预测精度较高,更符合实际预测需求。
The coal quality test data of about 150 thermal power enterprises are selected, and the coal calorific value prediction model based on L-M algorithm is built by analyzing the ubiquitous coal quality test data information. The experimental results show that: (1) the linear relationships between calorific value and carbon, ash are significant respectively, and the correlation coefficients are 0.8768 and 0.6880; (2) principal component analysis(PCA) method excavates the information of the principal component eigenvalue, characteristic matrix and score, which could realize the dimension reduction effect from six-dimensional matrix to four-dimensional matrix, and enhance the convergence stability of neural network in the training process; (3) based on the L-M algorithm, the training coefficient , verification coefficient and test coefficient of the Levenberg-Marquardt back propagation neural network prediction method(LMBP) model are 0.9957, 0.9942 and 0.9963 respectively, and the overall coefficient is 0.9931. Through the further verification of 20 groups of data to be tested, the LMBP prediction model is more reliable, the prediction accuracy is higher, and better to meet the actual forecast demand.
[1] 李大虎, 韦鲁滨, 朱学帅 . 基于 SVR 与特征变量选择方法的煤炭发热量预测[J]. 煤炭学报, 2019, 44(Suppl 1):278-288.
[2] 宋宜猛 . 基于工业分析预测煤质发热量研究[J]. 西安科技大学学报, 2022, 42(1): 71-75.
[3] 谭鹏, 李鑫, 张小培, 等. 基于工业分析的煤质发热量预测[J]. 煤炭学报, 2015, 40(11): 2642-2645.
[4] 姜英, 涂华, 陈亚飞, 等. 我国商品褐煤低位发热量回归式的推导[J]. 煤炭学报, 2004, 29(4): 477-480.
[5] Feng Q, Zhang J, Zhang X, et. al. Proximate analysis based pre-diction of gross calorific value of coals: A comparison of support vector machine, altering conditional ex⁃pectation and artificial neural network[J]. Fuel Processing Technology, 2015, 129: 120-129.
[6] 陈文敏, 姜宁 . 利用工业分析、元素分析数据计算我国煤炭热值的新回归式[J]. 煤炭分析及利用, 1993(1): 16-18.
[7] Majumder A K, Jain R, Banerjee P, et al. Development of a new proximate analysis based correlation to predict calorific value of coal[J]. Fuel, 2008, 87(13/14): 3077-3081.
[8] Sözer M, Haykiri-acma H, Yaman S. Prediction of calorific value of coal by multilinear regression and analysis of variance[J]. Journal of Energy Resources Technology, 2021, 144: 1-11.
[9] 李祥, 杜政烨, 刘翠茹, 等 . 基于主成分-线性回归分析的煤炭热值预测模型研究[J]. 煤炭工程, 2018, 50(7): 142-146.
[10] 周孑民, 朱再兴, 刘艳军, 等 . 基于 Elman 神经网络的动力配煤发热量及着火温度的预测[J]. 中南大学学报(自然科学版): 2011, 42(12): 3871-3875.
[11] 郭凯维, 郭传超, 史耀凡, 等 . 基于主成分分析的 GA-BP神经网络地表下沉系数预测[J]. 北京测绘, 2021, 35(11): 1374-1378.
[12] 张靖苑 . 基于主成分分析和随机配置网络的矿井突水水源判别方法研究[J]. 煤炭工程, 2020, 52(Suppl 1): 101-104.
[13] 王汉元, 贾宝山, 李守国, 等 . 基于 PCA 的煤炭自然发火预测预报研究[J]. 洁净煤技术, 2015, 21(6): 119-122.
[14] Livet C, Rouvier T, Sauret C, et al. A penalty method for constrained multibody kinematics optimisation using a Levenberg-Marquardt algorithm[J]. Computer Methods in Biomechanics and Biomedical Engineering, 2023, 26(7): 864-875.
[15] 肖舜仁, 胡青春, 李妮妮, 等 . 基于 Levenberg-Marquardt 算法的无人机多传感器校正方法研究[J]. 机床与液压, 2022, 50(10): 13-18.
[16] 杨承午, 刘志平, 徐永明. L-M算法优化的灰色模型在GPS卫星钟差预报中的应用[J]. 河南理工大学学报(自然科学版), 2020, 39(2): 48-51.
[17] 黄天炎, 唐莲 . L-M 优化改进的 BP网络模型在水环境承载力评价中的应用研究[J]. 中国农村水利水电, 2019(5): 49-51.
[18] 王华, 王连华, 葛岭梅 . 主成分分析与 BP 神经网络在煤耗氧速度预测中的应用[J]. 煤炭学报, 2008, 33(8): 921-925.
[19] 陈家慧 . 灰色理论与人工神经网络在电力系统中长期负荷预测中的应用[D]. 苏州: 苏州大学机电工程学院, 2019.
[20] 骆仲泱, 殷春根, 姚强, 等 . 人工神经网络在优化动力配煤中应用的理论研究[J]. 煤炭学报, 1997, 22(4): 337-342.