Prediction of primary quality variables in real time with adaptation capability for varying process conditions is a critical task in process industries.This article focuses on the development of non-linear adaptive so...Prediction of primary quality variables in real time with adaptation capability for varying process conditions is a critical task in process industries.This article focuses on the development of non-linear adaptive soft sensors for prediction of naphtha initial boiling point(IBP)and end boiling point(EBP)in crude distillation unit.In this work,adaptive inferential sensors with linear and non-linear local models are reported based on recursive just in time learning(JITL)approach.The different types of local models designed are locally weighted regression(LWR),multiple linear regression(MLR),partial least squares regression(PLS)and support vector regression(SVR).In addition to model development,the effect of relevant dataset size on model prediction accuracy and model computation time is also investigated.Results show that the JITL model based on support vector regression with iterative single data algorithm optimization(ISDA)local model(JITL-SVR:ISDA)yielded best prediction accuracy in reasonable computation time.展开更多
Ceria(CeO2)supports,synthesized by hydrothermal treatment with different synthesis time(CeO2-X h,where X is the synthesis time in h)in the presence of the surfactant cetyltrimethyl ammonium bromide,were used as suppor...Ceria(CeO2)supports,synthesized by hydrothermal treatment with different synthesis time(CeO2-X h,where X is the synthesis time in h)in the presence of the surfactant cetyltrimethyl ammonium bromide,were used as supports for gold(Au)catalysts.The synthesis time significantly affects the morphological structure and crystallite size of CeO2,where CeO2-2 h has the smallest crystallite size with coexisting nanorods and nanoparticles.Transmission electron microscopy analysis confirms the morphology of CeO2 with distinctive(110),(100)and(111)planes,in agreement with interplanar spacings of 0.19,0,27 and 0.31,respectively.However,the morphology of CeO2-8 h and CeO2-48 h is mainly a truncated octahedral with crystal planes(111)and(100)accompanied by an interplanar spacing of 0.31 and0.27 nm,respectively.The CeO2-X h supports and those with a 3 wt%Au loading(Au/CeO2-X h)were investigated in the oxidative steam reforming of methanol at temperatures between 200 and 400℃.The Au/CeO2-2 h gave the highest methanol conversion level and hydrogen yield at a low temperature of 250℃.This superior catalytic performance results from the good interaction between the metal and support and the well-distributed Au species on the CeO2 support.展开更多
A generalized formulation for short-term scheduling of steam power system in iron and steel industry under the time-of-use（TOU）power price was presented,with minimization of total operational cost including fuel cos...A generalized formulation for short-term scheduling of steam power system in iron and steel industry under the time-of-use（TOU）power price was presented,with minimization of total operational cost including fuel cost,equipment maintenance cost and the charge of exchange power with main grid.The model took into account the varying nature of surplus byproduct gas flows,several practical technical constraints and the impact of TOU power price.All major types of utility equipments,involving boilers,steam turbines,combined heat and power（CHP）units,and waste heat and energy recovery generators（WHERG）,were separately modeled using thermodynamic balance equations and regression method.In order to solve this complex nonlinear optimization model,a new improved particle swarm optimization（IPSO）algorithm was proposed by incorporating time-variant parameters,a selfadaptive mutation scheme and efficient constraint handling strategies.Finally,a case study for a real industrial example was used for illustrating the model and validating the effectiveness of the proposed approach.展开更多
With the development of power plants towards high power and intelligent operation direction,the vibrations or failures of blades,especially the last stage blades in steam turbines,happen more frequently due to the uns...With the development of power plants towards high power and intelligent operation direction,the vibrations or failures of blades,especially the last stage blades in steam turbines,happen more frequently due to the unstable operating conditions brought by flexible operation.A vibration measuring method for the shrouded blades of a steam turbine based on eddy current sensors with high frequency response is proposed,meeting the requirements of non-contact heath monitoring.The eddy current sensors produce the signals which are related to the area changing of every blade’s shroud resulting from the rotation of stator.Then an improved blade tip timing(BTT)technique is proposed to detect the vibrations of shrouded blades by measuring the arrival time of each area changing signal.A structure of eddy current sensors is developed in steam turbines and an amplitude modulation/demodulation circuit is designed to improve the response bandwidth up to 250 kHz.Vibration tests for the last stage blades of a steam turbine were carried out and the results validate the efficiency of the improved BTT technique and the high frequency response of the eddy current sensors presented.展开更多
This paper presents a novel method to solve old problem of water level control system of pressurized water reactor (PWR) steam generator (SG) of nuclear power plant (NPP) .The level control system of SG plays an impo...This paper presents a novel method to solve old problem of water level control system of pressurized water reactor (PWR) steam generator (SG) of nuclear power plant (NPP) .The level control system of SG plays an important role which effects the reliablity,safty,cost of SG and its mathematical models have been solved.A model of the conventional controller is presented and the existing problems are discussed. A novel rule based realtime control technique is designed with a computerized water level control (CWLC) system for SG of PWR NPP.The performance of this is evaluated for full power reactor operating conditions by applying different transient conditions of SG′s data of Qinshan Nuclear Power Plant (QNPP).展开更多
蒸汽管网流量负荷的精准预测与不确定性量化分析是优化能源调度和保障系统安全运行的关键。针对传统预测模型存在的预测精度不足和不确定性量化不全面等问题,提出了一种基于贝叶斯神经网络与补偿预测的融合模型。通过季节性-趋势分解(S...蒸汽管网流量负荷的精准预测与不确定性量化分析是优化能源调度和保障系统安全运行的关键。针对传统预测模型存在的预测精度不足和不确定性量化不全面等问题,提出了一种基于贝叶斯神经网络与补偿预测的融合模型。通过季节性-趋势分解(STL)将原始负荷数据解耦为周期项、趋势项和噪声项,分别采用贝叶斯神经网络与门控循环单元神经网络(Gated Recurrent Unit Neural Netwer,GRU)补偿预测模型进行多分量建模,并结合贝叶斯信息融合与不确定度合成方法,同步实现预测结果的认知不确定性和任意不确定性的动态量化。实测实验表明:相较于传统BP神经网络模型和LSTM模型,STL-BNN模型预测精度显著提升,均方误差和平均绝对误差分别降低2.29%和1.58%;在不确定性量化方面,通过认知-任意不确定性的分层解析与合成,STL-BNN模型预测值的不确定度估计值的平均绝对误差控制在实际计算数据的7.08%左右,补充并完善了预测结果在线不确定性实时分析和量化功能。展开更多
文摘Prediction of primary quality variables in real time with adaptation capability for varying process conditions is a critical task in process industries.This article focuses on the development of non-linear adaptive soft sensors for prediction of naphtha initial boiling point(IBP)and end boiling point(EBP)in crude distillation unit.In this work,adaptive inferential sensors with linear and non-linear local models are reported based on recursive just in time learning(JITL)approach.The different types of local models designed are locally weighted regression(LWR),multiple linear regression(MLR),partial least squares regression(PLS)and support vector regression(SVR).In addition to model development,the effect of relevant dataset size on model prediction accuracy and model computation time is also investigated.Results show that the JITL model based on support vector regression with iterative single data algorithm optimization(ISDA)local model(JITL-SVR:ISDA)yielded best prediction accuracy in reasonable computation time.
基金Project supported by the Ratchadaphiseksomphot Endowment Fund,Chulalongkorn University(CU-GES-60-04-63-03)the Thammasat University Research Fund under the Research University Network Initiative(8/2560)
文摘Ceria(CeO2)supports,synthesized by hydrothermal treatment with different synthesis time(CeO2-X h,where X is the synthesis time in h)in the presence of the surfactant cetyltrimethyl ammonium bromide,were used as supports for gold(Au)catalysts.The synthesis time significantly affects the morphological structure and crystallite size of CeO2,where CeO2-2 h has the smallest crystallite size with coexisting nanorods and nanoparticles.Transmission electron microscopy analysis confirms the morphology of CeO2 with distinctive(110),(100)and(111)planes,in agreement with interplanar spacings of 0.19,0,27 and 0.31,respectively.However,the morphology of CeO2-8 h and CeO2-48 h is mainly a truncated octahedral with crystal planes(111)and(100)accompanied by an interplanar spacing of 0.31 and0.27 nm,respectively.The CeO2-X h supports and those with a 3 wt%Au loading(Au/CeO2-X h)were investigated in the oxidative steam reforming of methanol at temperatures between 200 and 400℃.The Au/CeO2-2 h gave the highest methanol conversion level and hydrogen yield at a low temperature of 250℃.This superior catalytic performance results from the good interaction between the metal and support and the well-distributed Au species on the CeO2 support.
基金Sponsored by National Natural Science Foundation of China(51304053)International Science and Technology Cooperation Program of China(2013DFA10810)
文摘A generalized formulation for short-term scheduling of steam power system in iron and steel industry under the time-of-use（TOU）power price was presented,with minimization of total operational cost including fuel cost,equipment maintenance cost and the charge of exchange power with main grid.The model took into account the varying nature of surplus byproduct gas flows,several practical technical constraints and the impact of TOU power price.All major types of utility equipments,involving boilers,steam turbines,combined heat and power（CHP）units,and waste heat and energy recovery generators（WHERG）,were separately modeled using thermodynamic balance equations and regression method.In order to solve this complex nonlinear optimization model,a new improved particle swarm optimization（IPSO）algorithm was proposed by incorporating time-variant parameters,a selfadaptive mutation scheme and efficient constraint handling strategies.Finally,a case study for a real industrial example was used for illustrating the model and validating the effectiveness of the proposed approach.
基金National Natural Science Foundation of China(No.51775377)National Key Research and Development Plan(No.2017YFF0204800)+2 种基金Natural Science Foundation of TianJin City(No.17JCQNJC01100)Young Elite Scientists Sponsorship Program by Cast of China(No.2016QNRC001)Open Project of Key Laboratory of Underwater Information and Control(No.6142218081811)
文摘With the development of power plants towards high power and intelligent operation direction,the vibrations or failures of blades,especially the last stage blades in steam turbines,happen more frequently due to the unstable operating conditions brought by flexible operation.A vibration measuring method for the shrouded blades of a steam turbine based on eddy current sensors with high frequency response is proposed,meeting the requirements of non-contact heath monitoring.The eddy current sensors produce the signals which are related to the area changing of every blade’s shroud resulting from the rotation of stator.Then an improved blade tip timing(BTT)technique is proposed to detect the vibrations of shrouded blades by measuring the arrival time of each area changing signal.A structure of eddy current sensors is developed in steam turbines and an amplitude modulation/demodulation circuit is designed to improve the response bandwidth up to 250 kHz.Vibration tests for the last stage blades of a steam turbine were carried out and the results validate the efficiency of the improved BTT technique and the high frequency response of the eddy current sensors presented.
文摘This paper presents a novel method to solve old problem of water level control system of pressurized water reactor (PWR) steam generator (SG) of nuclear power plant (NPP) .The level control system of SG plays an important role which effects the reliablity,safty,cost of SG and its mathematical models have been solved.A model of the conventional controller is presented and the existing problems are discussed. A novel rule based realtime control technique is designed with a computerized water level control (CWLC) system for SG of PWR NPP.The performance of this is evaluated for full power reactor operating conditions by applying different transient conditions of SG′s data of Qinshan Nuclear Power Plant (QNPP).
文摘蒸汽管网流量负荷的精准预测与不确定性量化分析是优化能源调度和保障系统安全运行的关键。针对传统预测模型存在的预测精度不足和不确定性量化不全面等问题,提出了一种基于贝叶斯神经网络与补偿预测的融合模型。通过季节性-趋势分解(STL)将原始负荷数据解耦为周期项、趋势项和噪声项,分别采用贝叶斯神经网络与门控循环单元神经网络(Gated Recurrent Unit Neural Netwer,GRU)补偿预测模型进行多分量建模,并结合贝叶斯信息融合与不确定度合成方法,同步实现预测结果的认知不确定性和任意不确定性的动态量化。实测实验表明:相较于传统BP神经网络模型和LSTM模型,STL-BNN模型预测精度显著提升,均方误差和平均绝对误差分别降低2.29%和1.58%;在不确定性量化方面,通过认知-任意不确定性的分层解析与合成,STL-BNN模型预测值的不确定度估计值的平均绝对误差控制在实际计算数据的7.08%左右,补充并完善了预测结果在线不确定性实时分析和量化功能。
