温室高效生产依赖于适宜的温室小气候环境,建立高精度的温室小气候多步预测模型对实现温室环境优化调控具有重要意义。本研究提出一种基于滚动的反向传播神经网络(Rolling Back Propagation Neural Network,R-BP)的温室小气候多步滚动...温室高效生产依赖于适宜的温室小气候环境,建立高精度的温室小气候多步预测模型对实现温室环境优化调控具有重要意义。本研究提出一种基于滚动的反向传播神经网络(Rolling Back Propagation Neural Network,R-BP)的温室小气候多步滚动预测模型。模型主要包括两个阶段:(1)建立初始的BP神经网络。采用自动编码器无监督学习方法获取初始网络参数,并利用改进的粒子群算法优化网络参数。(2)建立滚动的BP神经网络群。将前一个网络输出作为后一个网络的部分输入进行滚动训练和预测,实现温室小气候多步滚动预测。为验证R-BP模型的有效性,在阿拉伯联合酋长国阿布扎比市的自控温室和中国苏州市的非自控温室分别进行试验。验证试验表明,与传统BP神经网络相比,在阿布扎比温室试验中,采用R-BP模型预测未来6h室内温度,其平均误差降低69.9%,预测相对湿度,其平均误差降低47%;在苏州温室试验中,采用R-BP模型预测未来6h室内温度,其平均误差降低43.3%,预测相对湿度,其平均误差降低55.6%。说明R-BP模型能够较准确预测未来6h内温室小气候环境变化,可为制定温室小气候优化调控方案提供依据。展开更多
Automatic width control is a key issue in hot strip rolling process. The edge rolling has been widely used in the roughing stand of hot strip mill to control the width of the slab. However, the edge rolling and conseq...Automatic width control is a key issue in hot strip rolling process. The edge rolling has been widely used in the roughing stand of hot strip mill to control the width of the slab. However, the edge rolling and consequent hori- zontal rolling will cause a significant width change in the head part and tail part of the slab, which have to be trimmed before the finishing stands. Based on the width reduction deformation curve of the head and tail along the longitudinal direction of slab, the short stroke control （SSC） technology has been developed to overcome this problem. The finite element method has been used to simulate the unsteady edge rolling process. Three short stroke control curves have been compared in order to obtain the best width control result. The optimized short stroke control curve has been ap- plied to the automatic width control system of industrial hot rolling mill, and good performance is obtained.展开更多
Highly wind power integrated power system requires continuous active power regulation to tackle the power imbalances resulting from the wind power forecast errors. The active power balance is maintained in real-time w...Highly wind power integrated power system requires continuous active power regulation to tackle the power imbalances resulting from the wind power forecast errors. The active power balance is maintained in real-time with the automatic generation control and also from the control room, where regulating power bids are activated manually. In this article, an algorithm is developed to simulate the activation of regulating power bids, as performed in the control room, during power imbalance between generation and load demand. In addition, the active power balance is also controlled through automatic generation control, where coordinated control strategy between combined heat and power plants and wind power plant enhances the secure power system operation. The developed algorithm emulating the control room response,to deal with real-time power imbalance, is applied and investigated on the future Danish power system model. The power system model takes the hour-ahead regulating power plan from power balancing model and the generation and power exchange capacities for the year 2020 into account.The real-time impact of power balancing in a highly wind power integrated power system is assessed and discussed by means of simulations for different possible scenarios.展开更多
Including servo valve, hydraulic cylinder, mill and sensor and ignoring nonlinear factors, the linear dynamic model of hydraulic automatic gage control(HAGC) system of a temper rolling mill was theoretically derived. ...Including servo valve, hydraulic cylinder, mill and sensor and ignoring nonlinear factors, the linear dynamic model of hydraulic automatic gage control(HAGC) system of a temper rolling mill was theoretically derived. The order of the model is 4/4, and can be reduced to 2/2. Based on modulating functions method, utilizing numerical integration, we constructed the equivalent identification model of HAGC, and the least square estimation algorithm was established. The input and output data were acquired on line at temper rolling mill in Shangshai Baosteel Group Corporation, and the continuous time model of HAGC system was estimated with the proposed method. At different modulating window intervals, the estimated parameters changed remarkably. When the frequency bandwidth of modulating filter matches that of estimated system, the parameters can be estimated accurately. Finally, the dynamic model of the HAGC was obtained and validated based on the spectral analysis result.展开更多
Conventional techniques to control variations within one plate have been based on preset models and constant automatic position control (APC or pressure feedback automatic gauge control(PAGC).However.because of the r...Conventional techniques to control variations within one plate have been based on preset models and constant automatic position control (APC or pressure feedback automatic gauge control(PAGC).However.because of the rolling force prediction error in the preset models and of the inadequate response speed of dynamic system and of the eccentricity, etc.,the conventional method has not given satisfactory results, the statistics'variations within one plate are in range of 0.25-0.60 mm The authors have developed the techniques to control the variations, which are dynamic intelligent control of hydraulic screwdown system. synchronism control of hydraulic screwdown, eccentricity control method by rotary encoder and the curve of modulus of mill measured automatically, etc., The techniques were fully and successfully industrialized in The Plate Mill of Maanshan Iron and Steel Company and good results that variations are in range of 0.08-0.15min hare been obtained in that mill.展开更多
文摘温室高效生产依赖于适宜的温室小气候环境,建立高精度的温室小气候多步预测模型对实现温室环境优化调控具有重要意义。本研究提出一种基于滚动的反向传播神经网络(Rolling Back Propagation Neural Network,R-BP)的温室小气候多步滚动预测模型。模型主要包括两个阶段:(1)建立初始的BP神经网络。采用自动编码器无监督学习方法获取初始网络参数,并利用改进的粒子群算法优化网络参数。(2)建立滚动的BP神经网络群。将前一个网络输出作为后一个网络的部分输入进行滚动训练和预测,实现温室小气候多步滚动预测。为验证R-BP模型的有效性,在阿拉伯联合酋长国阿布扎比市的自控温室和中国苏州市的非自控温室分别进行试验。验证试验表明,与传统BP神经网络相比,在阿布扎比温室试验中,采用R-BP模型预测未来6h室内温度,其平均误差降低69.9%,预测相对湿度,其平均误差降低47%;在苏州温室试验中,采用R-BP模型预测未来6h室内温度,其平均误差降低43.3%,预测相对湿度,其平均误差降低55.6%。说明R-BP模型能够较准确预测未来6h内温室小气候环境变化,可为制定温室小气候优化调控方案提供依据。
基金Item Sponsored by National Science and Technology Support Programfor 11th Five-Year Plan of China (2006BAE03A13)International Cooperation Research Project of Shanxi Province (2009081013)
文摘Automatic width control is a key issue in hot strip rolling process. The edge rolling has been widely used in the roughing stand of hot strip mill to control the width of the slab. However, the edge rolling and consequent hori- zontal rolling will cause a significant width change in the head part and tail part of the slab, which have to be trimmed before the finishing stands. Based on the width reduction deformation curve of the head and tail along the longitudinal direction of slab, the short stroke control （SSC） technology has been developed to overcome this problem. The finite element method has been used to simulate the unsteady edge rolling process. Three short stroke control curves have been compared in order to obtain the best width control result. The optimized short stroke control curve has been ap- plied to the automatic width control system of industrial hot rolling mill, and good performance is obtained.
基金a part of Ph.D.project funded by Sino-Danish centre for education and research(SDC)
文摘Highly wind power integrated power system requires continuous active power regulation to tackle the power imbalances resulting from the wind power forecast errors. The active power balance is maintained in real-time with the automatic generation control and also from the control room, where regulating power bids are activated manually. In this article, an algorithm is developed to simulate the activation of regulating power bids, as performed in the control room, during power imbalance between generation and load demand. In addition, the active power balance is also controlled through automatic generation control, where coordinated control strategy between combined heat and power plants and wind power plant enhances the secure power system operation. The developed algorithm emulating the control room response,to deal with real-time power imbalance, is applied and investigated on the future Danish power system model. The power system model takes the hour-ahead regulating power plan from power balancing model and the generation and power exchange capacities for the year 2020 into account.The real-time impact of power balancing in a highly wind power integrated power system is assessed and discussed by means of simulations for different possible scenarios.
文摘Including servo valve, hydraulic cylinder, mill and sensor and ignoring nonlinear factors, the linear dynamic model of hydraulic automatic gage control(HAGC) system of a temper rolling mill was theoretically derived. The order of the model is 4/4, and can be reduced to 2/2. Based on modulating functions method, utilizing numerical integration, we constructed the equivalent identification model of HAGC, and the least square estimation algorithm was established. The input and output data were acquired on line at temper rolling mill in Shangshai Baosteel Group Corporation, and the continuous time model of HAGC system was estimated with the proposed method. At different modulating window intervals, the estimated parameters changed remarkably. When the frequency bandwidth of modulating filter matches that of estimated system, the parameters can be estimated accurately. Finally, the dynamic model of the HAGC was obtained and validated based on the spectral analysis result.
文摘Conventional techniques to control variations within one plate have been based on preset models and constant automatic position control (APC or pressure feedback automatic gauge control(PAGC).However.because of the rolling force prediction error in the preset models and of the inadequate response speed of dynamic system and of the eccentricity, etc.,the conventional method has not given satisfactory results, the statistics'variations within one plate are in range of 0.25-0.60 mm The authors have developed the techniques to control the variations, which are dynamic intelligent control of hydraulic screwdown system. synchronism control of hydraulic screwdown, eccentricity control method by rotary encoder and the curve of modulus of mill measured automatically, etc., The techniques were fully and successfully industrialized in The Plate Mill of Maanshan Iron and Steel Company and good results that variations are in range of 0.08-0.15min hare been obtained in that mill.
