A novel time-span input neural network was developed to accurately predict the trend of the main steam temperature of a 750-t/d waste incineration boiler.Its historical operating data were used to retrieve sensitive p...A novel time-span input neural network was developed to accurately predict the trend of the main steam temperature of a 750-t/d waste incineration boiler.Its historical operating data were used to retrieve sensitive parameters for the boiler output steam temperature by correlation analysis.Then,the 15 most sensitive parameters with specified time spans were selected as neural network inputs.An external testing set was introduced to objectively evaluate the neural network prediction capability.The results show that,compared with the traditional prediction method,the time-span input framework model can achieve better prediction performance and has a greater capability for generalization.The maximum average prediction error can be controlled below 0.2°C and 1.5°C in the next 60 s and 5 min,respectively.In addition,setting a reasonable terminal training threshold can effectively avoid overfitting.An importance analysis of the parameters indicates that the main steam temperature and the average temperature around the high-temperature superheater are the two most important variables of the input parameters;the former affects the overall prediction and the latter affects the long-term prediction performance.展开更多
In this paper,we developed a hybrid model for the steam turbines of a utility system,which combines an improved neural network model with the thermodynamic model.Then,a nonlinear programming(NLP) model of the steam tu...In this paper,we developed a hybrid model for the steam turbines of a utility system,which combines an improved neural network model with the thermodynamic model.Then,a nonlinear programming(NLP) model of the steam turbine network is formulated by utilizing the developed steam turbine models to minimize the total steam cost for the whole steam turbine network.Finally,this model is applied to optimize the steam turbine network of an ethylene plant.The obtained results demonstrate that this hybrid model can accurately estimate and evaluate the performance of steam turbines,and the significant cost savings can be made by optimizing the steam turbine network operation at no capital cost.展开更多
By combining the Back-Propagation （BP） neural network with conventional proportional Integral Derivative （PID） controller, a new temperature control strategy of the export steam in supercritical electric power pla...By combining the Back-Propagation （BP） neural network with conventional proportional Integral Derivative （PID） controller, a new temperature control strategy of the export steam in supercritical electric power plant is put forward. This scheme can effectively overcome the large time delay, inertia of the export steam and the influencee of object in varying operational parameters. Thus excellent control quality is obtaitud. The present paper describes the development and application of neural network based controller to control the temperature of the boiler＇s export steam. Through simulation in various situations, it validates that the control quality of this control system is apparently superior to the conventional PID control system.展开更多
The real-time fault diagnosis system is very important for steam turbine generator set due serious fault results in a reduced amount of electricity supply in power plant. A novel real-time fault diagnosis system is pr...The real-time fault diagnosis system is very important for steam turbine generator set due serious fault results in a reduced amount of electricity supply in power plant. A novel real-time fault diagnosis system is proposed by using Levenberg-Marquardt algorithm related to tuning parameters of Artificial Neural Network (ANN). The model of novel fault diagnosis system by using ANN are built and analyzed. Cases of the diagnosis are simulated. The results show that the real-time fault diagnosis system is of high accuracy and quick convergence. It is also found that this model is feasible in real-time fault diagnosis. The steam turbine is used as a power generator by SONELGAZ, an Algerian company located at Cap Djinet town in Boumerdes district. We used this turbine as our main target for the purpose of this analysis. After deep investigation, while keeping our focus on the most sensitive parts within the turbine, the weakest and the strongest points of the system were identified. Those are the points mostly adequate for failure simulations and at which the designed system will be better positioned for irregularities detection during the production process.展开更多
This paper introduces a kind of diagnosis principle and learning algorithm of steam turbine fault diagnosis which based on Elman neural network. Comparing the results of the Elman neural network and the traditional BP...This paper introduces a kind of diagnosis principle and learning algorithm of steam turbine fault diagnosis which based on Elman neural network. Comparing the results of the Elman neural network and the traditional BP neural network diagnosis, the results shows that Elman neural network is an effective way to improve the learning speed , effectively suppress the minimum defects that the traditional neural network easily trapped in, and shorten the autonomous learning time. All these proves that the Elman neural network is an effective way to diagnose the steam turbine.展开更多
All kinds of reasons are analysed in theory and a fault repository combined with local expert experiences is establishedaccording to the structure and the operation characteristic of steam generator in this paper. At ...All kinds of reasons are analysed in theory and a fault repository combined with local expert experiences is establishedaccording to the structure and the operation characteristic of steam generator in this paper. At the same time, Kohonen algo-rithm is used for fault diagnoses system based on fuzzy neural networks. Fuzzy arithmetic is inducted into neural networks tosolve uncertain diagnosis induced by uncertain knowledge. According to its self-association in the course of default diagnosis. thesystem is provided with non-supervise, self-organizing, self-learning, and has strong cluster ability and fast cluster velocity.展开更多
The slagging/fouling due to the accession of fireside deposits on the steam boilers decreases boiler efficiency and availability which leads to unexpected shut-downs. Since it is inevitably associated with the three m...The slagging/fouling due to the accession of fireside deposits on the steam boilers decreases boiler efficiency and availability which leads to unexpected shut-downs. Since it is inevitably associated with the three major factors namely the fuel characteristics, boiler operating conditions and ash behavior, this serious slagging/fouling may be reduced by varying the above three factors. The research develops a generic slagging/fouling prediction tool based on hybrid fuzzy clustering and Artificial Neural Networks (FCANN). The FCANN model presents a good accuracy of 99.85% which makes this model fast in response and easy to be updated with lesser time when compared to single ANN. The comparison between predictions and observations is found to be satisfactory with less input parameters. This should be capable of giving relatively quick responses while being easily implemented for various furnace types.展开更多
The determination of source-side extracted heating parameters is of great significance to the economic operation of cogeneration systems.This paper investigated the coupling performance of a cogeneration heating and p...The determination of source-side extracted heating parameters is of great significance to the economic operation of cogeneration systems.This paper investigated the coupling performance of a cogeneration heating and power system multidimensionally based on the operating characteristics of the cogeneration units,the hydraulic and thermodynamic characteristics of the heating network,and the energy loads.Taking a steam network supported by a gas-steam combined cycle cogeneration system as the research case,the interaction effect among the source-side prime movers,the heating networks,and the terminal demand thermal parameters were investigated based on the designed values,the plant testing data,and the validated simulation.The operating maps of the gas-steam combined cycle cogeneration units were obtained using THERMOFLEX,and the minimum source-side steam parameters of the steam network were solved using an inverse solution procedure based on the hydro-thermodynamic coupling model.The cogeneration operating maps indicate that the available operating domain considerably narrows with the rise of the extraction steam pressure and flow rate.The heating network inverse solution demonstrates that the source-side steam pressure and temperature can be optimized from the originally designed 1.11 MPa and 238.8°C to 1.074 MPa and 191.15°C,respectively.Under the operating strategy with the minimum source-side heating parameters,the power peak regulation depth remarkably increases to 18.30%whereas the comprehensive thermal efficiency decreases.The operation under the minimum source-side heating steam parameters can be superior to the originally designed one in the economy at a higher price of the heating steam.At a fuel price of$0.38/kg and the power to fuel price of 0.18 kg/(kW·h),the critical price ratio of heating steam to fuel is 119.1 kg/t.The influence of the power-fuel price ratio on the economic deviation appears relatively weak.展开更多
Numerical simulation of complex systems and components by computers is a fundamental phase of any modern engineering activity. The traditional methods of simulation typically entail long, iterative processes which lea...Numerical simulation of complex systems and components by computers is a fundamental phase of any modern engineering activity. The traditional methods of simulation typically entail long, iterative processes which lead to large simulation times, often exceeding transient real time. Artificial neural networks （ANNs） may be advantageous in this context, the main advantage being the speed of computation, the capability of generalizing from the few examples, robustness to noisy and partially incomplete data and the capability of performing empirical input-output mapping without complete knowledge of underlying physics. In this paper, the simulation of steam generator is considered as an example to show the potentialities of this tool. The data required for training and testing the ANN is taken from the steam generator at Abott Power Plant, Champaign （USA）. The total number of samples is 9600 which are taken at a sampling time of three seconds. The performance of boiler （drum pressure, steam flow rate） has been verified and tested using ANN, under the changes in fuel flow rate, air flow rate and load disturbance. Using ANN, input-output mapping is done and it is observed that ANN allows a good reproduction of non-linear behaviors of inputs and outputs.展开更多
针对锅炉过热汽温系统存在的大时延、强非线性和变量耦合等建模难题,建立了一种基于贝叶斯优化时间序列预测神经基扩展分析(neural basis expansion analysis for interpretable time series forecasting,N-BEATS)网络的过热汽温预测模...针对锅炉过热汽温系统存在的大时延、强非线性和变量耦合等建模难题,建立了一种基于贝叶斯优化时间序列预测神经基扩展分析(neural basis expansion analysis for interpretable time series forecasting,N-BEATS)网络的过热汽温预测模型。针对某600 MW超临界火电机组,结合机理分析确定模型的输入和输出变量,通过性能对比实验优化模型的输入/输出时延阶次、Block类型和激活函数,进一步利用贝叶斯优化算法对模型的超参数进行寻优,并与网格搜索、遗传算法的优化效果进行对比。采用该机组仿真运行数据进行建模实验,结果表明所提模型在预测精度方面优于传统优化方法及主流模型。展开更多
Power plants are nonlinear and uncertain complex systems. Reliable control of superheated steam temperature is necessary to ensure high efficiency and high load-following capability in the operation of modem power pla...Power plants are nonlinear and uncertain complex systems. Reliable control of superheated steam temperature is necessary to ensure high efficiency and high load-following capability in the operation of modem power plant. A nonlinear generalized predictive controller based on neuro-fuzzy network （NFGPC） is proposed in this paper. The proposed nonlinear controller is applied to control the superheated steam temperature of a 200MW power plant. From the experiments on the plant and the simulation of the plant, much better performance than the traditional controller is obtained,展开更多
Chemical processes can bene t tremendously from fast and accurate ef uent composition prediction for plant design, control, and optimization. The Industry 4.0 revolution claims that by introducing machine learning int...Chemical processes can bene t tremendously from fast and accurate ef uent composition prediction for plant design, control, and optimization. The Industry 4.0 revolution claims that by introducing machine learning into these elds, substantial economic and environmental gains can be achieved. The bottleneck for high-frequency optimization and process control is often the time necessary to perform the required detailed analyses of, for example, feed and product. To resolve these issues, a framework of four deep learning arti cial neural networks (DL ANNs) has been developed for the largest chemicals production process steam cracking. The proposed methodology allows both a detailed characterization of a naphtha feedstock and a detailed composition of the steam cracker ef uent to be determined, based on a limited number of commercial naphtha indices and rapidly accessible process characteristics. The detailed char- acterization of a naphtha is predicted from three points on the boiling curve and paraf ns, iso-paraf ns, ole ns, naphthenes, and aronatics (PIONA) characterization. If unavailable, the boiling points are also estimated. Even with estimated boiling points, the developed DL ANN outperforms several established methods such as maximization of Shannon entropy and traditional ANNs. For feedstock reconstruction, a mean absolute error (MAE) of 0.3 wt% is achieved on the test set, while the MAE of the ef uent predic- tion is 0.1 wt%. When combining all networks using the output of the previous as input to the next the ef uent MAE increases to 0.19 wt%. In addition to the high accuracy of the networks, a major bene t is the negligible computational cost required to obtain the predictions. On a standard Intel i7 processor, predictions are made in the order of milliseconds. Commercial software such as COILSIM1D performs slightly better in terms of accuracy, but the required central processing unit time per reaction is in the order of seconds. This tremendous speed-up and minimal accuracy loss make the presented framework highly suitable for the continuous monitoring of dif cult-to-access process parameters and for the envi- sioned, high-frequency real-time optimization (RTO) strategy or process control. Nevertheless, the lack of a fundamental basis implies that fundamental understanding is almost completely lost, which is not always well-accepted by the engineering community. In addition, the performance of the developed net- works drops signi cantly for naphthas that are highly dissimilar to those in the training set.展开更多
Early fault warning for nuclear power machinery is conducive to timely troubleshooting and reductions in safety risks and unnecessary costs. This paper presents a novel intelligent fault prediction method, integrated ...Early fault warning for nuclear power machinery is conducive to timely troubleshooting and reductions in safety risks and unnecessary costs. This paper presents a novel intelligent fault prediction method, integrated probabilistic principal component analysis(PPCA), multi-resolution wavelet analysis, Bayesian inference, and RNN model for nuclear power machinery that consider data uncertainty and chaotic time series. After denoising the source data, the Bayesian PPCA method is employed for dimensional reduction to obtain a refined data group. A recurrent neural network(RNN) prediction model is constructed, and a Bayesian statistical inference approach is developed to quantitatively assess the prediction reliability of the model. By modeling and analyzing the data collected on the steam turbine and components of a nuclear power plant, the results of the goodness of fit, mean square error distribution, and Bayesian confidence indicate that the proposed RNN model can implement early warning in the fault creep period. The accuracy and reliability of the proposed model are quantitatively verified.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
基金Project supported by the National Key Research and Development Program of China(No.2018YFC1901300)the Research Project of Multi-data Fusion and Strategy of Intelligent Control and Optimization for Large Scale Industrial Combustion System,China。
文摘A novel time-span input neural network was developed to accurately predict the trend of the main steam temperature of a 750-t/d waste incineration boiler.Its historical operating data were used to retrieve sensitive parameters for the boiler output steam temperature by correlation analysis.Then,the 15 most sensitive parameters with specified time spans were selected as neural network inputs.An external testing set was introduced to objectively evaluate the neural network prediction capability.The results show that,compared with the traditional prediction method,the time-span input framework model can achieve better prediction performance and has a greater capability for generalization.The maximum average prediction error can be controlled below 0.2°C and 1.5°C in the next 60 s and 5 min,respectively.In addition,setting a reasonable terminal training threshold can effectively avoid overfitting.An importance analysis of the parameters indicates that the main steam temperature and the average temperature around the high-temperature superheater are the two most important variables of the input parameters;the former affects the overall prediction and the latter affects the long-term prediction performance.
基金Supported by the Major State Basic Research Development Program of China（2012CB720500）the National Natural Science Foundation of China（U1162202）,the National Natural Science Foundation of China（21276078,61174118,21206037）the National Science Fund for Outstanding Young Scholars（61222303）
文摘In this paper,we developed a hybrid model for the steam turbines of a utility system,which combines an improved neural network model with the thermodynamic model.Then,a nonlinear programming(NLP) model of the steam turbine network is formulated by utilizing the developed steam turbine models to minimize the total steam cost for the whole steam turbine network.Finally,this model is applied to optimize the steam turbine network of an ethylene plant.The obtained results demonstrate that this hybrid model can accurately estimate and evaluate the performance of steam turbines,and the significant cost savings can be made by optimizing the steam turbine network operation at no capital cost.
基金supported by the project of "SDUST Qunxing Program"(No.qx0902075)
文摘By combining the Back-Propagation （BP） neural network with conventional proportional Integral Derivative （PID） controller, a new temperature control strategy of the export steam in supercritical electric power plant is put forward. This scheme can effectively overcome the large time delay, inertia of the export steam and the influencee of object in varying operational parameters. Thus excellent control quality is obtaitud. The present paper describes the development and application of neural network based controller to control the temperature of the boiler＇s export steam. Through simulation in various situations, it validates that the control quality of this control system is apparently superior to the conventional PID control system.
文摘The real-time fault diagnosis system is very important for steam turbine generator set due serious fault results in a reduced amount of electricity supply in power plant. A novel real-time fault diagnosis system is proposed by using Levenberg-Marquardt algorithm related to tuning parameters of Artificial Neural Network (ANN). The model of novel fault diagnosis system by using ANN are built and analyzed. Cases of the diagnosis are simulated. The results show that the real-time fault diagnosis system is of high accuracy and quick convergence. It is also found that this model is feasible in real-time fault diagnosis. The steam turbine is used as a power generator by SONELGAZ, an Algerian company located at Cap Djinet town in Boumerdes district. We used this turbine as our main target for the purpose of this analysis. After deep investigation, while keeping our focus on the most sensitive parts within the turbine, the weakest and the strongest points of the system were identified. Those are the points mostly adequate for failure simulations and at which the designed system will be better positioned for irregularities detection during the production process.
文摘This paper introduces a kind of diagnosis principle and learning algorithm of steam turbine fault diagnosis which based on Elman neural network. Comparing the results of the Elman neural network and the traditional BP neural network diagnosis, the results shows that Elman neural network is an effective way to improve the learning speed , effectively suppress the minimum defects that the traditional neural network easily trapped in, and shorten the autonomous learning time. All these proves that the Elman neural network is an effective way to diagnose the steam turbine.
文摘All kinds of reasons are analysed in theory and a fault repository combined with local expert experiences is establishedaccording to the structure and the operation characteristic of steam generator in this paper. At the same time, Kohonen algo-rithm is used for fault diagnoses system based on fuzzy neural networks. Fuzzy arithmetic is inducted into neural networks tosolve uncertain diagnosis induced by uncertain knowledge. According to its self-association in the course of default diagnosis. thesystem is provided with non-supervise, self-organizing, self-learning, and has strong cluster ability and fast cluster velocity.
文摘The slagging/fouling due to the accession of fireside deposits on the steam boilers decreases boiler efficiency and availability which leads to unexpected shut-downs. Since it is inevitably associated with the three major factors namely the fuel characteristics, boiler operating conditions and ash behavior, this serious slagging/fouling may be reduced by varying the above three factors. The research develops a generic slagging/fouling prediction tool based on hybrid fuzzy clustering and Artificial Neural Networks (FCANN). The FCANN model presents a good accuracy of 99.85% which makes this model fast in response and easy to be updated with lesser time when compared to single ANN. The comparison between predictions and observations is found to be satisfactory with less input parameters. This should be capable of giving relatively quick responses while being easily implemented for various furnace types.
基金Guangdong Province Key Laboratory of Efficient and Clean Energy Utilization(South China University of Technology)(2013A061401005)Research Fund(JMSWFW-2110-044)from Zhongshan Jiaming Electric Power Co.,Ltd.
文摘The determination of source-side extracted heating parameters is of great significance to the economic operation of cogeneration systems.This paper investigated the coupling performance of a cogeneration heating and power system multidimensionally based on the operating characteristics of the cogeneration units,the hydraulic and thermodynamic characteristics of the heating network,and the energy loads.Taking a steam network supported by a gas-steam combined cycle cogeneration system as the research case,the interaction effect among the source-side prime movers,the heating networks,and the terminal demand thermal parameters were investigated based on the designed values,the plant testing data,and the validated simulation.The operating maps of the gas-steam combined cycle cogeneration units were obtained using THERMOFLEX,and the minimum source-side steam parameters of the steam network were solved using an inverse solution procedure based on the hydro-thermodynamic coupling model.The cogeneration operating maps indicate that the available operating domain considerably narrows with the rise of the extraction steam pressure and flow rate.The heating network inverse solution demonstrates that the source-side steam pressure and temperature can be optimized from the originally designed 1.11 MPa and 238.8°C to 1.074 MPa and 191.15°C,respectively.Under the operating strategy with the minimum source-side heating parameters,the power peak regulation depth remarkably increases to 18.30%whereas the comprehensive thermal efficiency decreases.The operation under the minimum source-side heating steam parameters can be superior to the originally designed one in the economy at a higher price of the heating steam.At a fuel price of$0.38/kg and the power to fuel price of 0.18 kg/(kW·h),the critical price ratio of heating steam to fuel is 119.1 kg/t.The influence of the power-fuel price ratio on the economic deviation appears relatively weak.
文摘Numerical simulation of complex systems and components by computers is a fundamental phase of any modern engineering activity. The traditional methods of simulation typically entail long, iterative processes which lead to large simulation times, often exceeding transient real time. Artificial neural networks （ANNs） may be advantageous in this context, the main advantage being the speed of computation, the capability of generalizing from the few examples, robustness to noisy and partially incomplete data and the capability of performing empirical input-output mapping without complete knowledge of underlying physics. In this paper, the simulation of steam generator is considered as an example to show the potentialities of this tool. The data required for training and testing the ANN is taken from the steam generator at Abott Power Plant, Champaign （USA）. The total number of samples is 9600 which are taken at a sampling time of three seconds. The performance of boiler （drum pressure, steam flow rate） has been verified and tested using ANN, under the changes in fuel flow rate, air flow rate and load disturbance. Using ANN, input-output mapping is done and it is observed that ANN allows a good reproduction of non-linear behaviors of inputs and outputs.
文摘针对锅炉过热汽温系统存在的大时延、强非线性和变量耦合等建模难题,建立了一种基于贝叶斯优化时间序列预测神经基扩展分析(neural basis expansion analysis for interpretable time series forecasting,N-BEATS)网络的过热汽温预测模型。针对某600 MW超临界火电机组,结合机理分析确定模型的输入和输出变量,通过性能对比实验优化模型的输入/输出时延阶次、Block类型和激活函数,进一步利用贝叶斯优化算法对模型的超参数进行寻优,并与网格搜索、遗传算法的优化效果进行对比。采用该机组仿真运行数据进行建模实验,结果表明所提模型在预测精度方面优于传统优化方法及主流模型。
基金This work was supported by the Natural Science Foundation of Beijing (No. 4062030)National Natural Science Foundation of China (No. 50576022,69804003)Scientific Research Common Program of Beijing Municipal Commission of Education (KM200611232007).
文摘Power plants are nonlinear and uncertain complex systems. Reliable control of superheated steam temperature is necessary to ensure high efficiency and high load-following capability in the operation of modem power plant. A nonlinear generalized predictive controller based on neuro-fuzzy network （NFGPC） is proposed in this paper. The proposed nonlinear controller is applied to control the superheated steam temperature of a 200MW power plant. From the experiments on the plant and the simulation of the plant, much better performance than the traditional controller is obtained,
文摘Chemical processes can bene t tremendously from fast and accurate ef uent composition prediction for plant design, control, and optimization. The Industry 4.0 revolution claims that by introducing machine learning into these elds, substantial economic and environmental gains can be achieved. The bottleneck for high-frequency optimization and process control is often the time necessary to perform the required detailed analyses of, for example, feed and product. To resolve these issues, a framework of four deep learning arti cial neural networks (DL ANNs) has been developed for the largest chemicals production process steam cracking. The proposed methodology allows both a detailed characterization of a naphtha feedstock and a detailed composition of the steam cracker ef uent to be determined, based on a limited number of commercial naphtha indices and rapidly accessible process characteristics. The detailed char- acterization of a naphtha is predicted from three points on the boiling curve and paraf ns, iso-paraf ns, ole ns, naphthenes, and aronatics (PIONA) characterization. If unavailable, the boiling points are also estimated. Even with estimated boiling points, the developed DL ANN outperforms several established methods such as maximization of Shannon entropy and traditional ANNs. For feedstock reconstruction, a mean absolute error (MAE) of 0.3 wt% is achieved on the test set, while the MAE of the ef uent predic- tion is 0.1 wt%. When combining all networks using the output of the previous as input to the next the ef uent MAE increases to 0.19 wt%. In addition to the high accuracy of the networks, a major bene t is the negligible computational cost required to obtain the predictions. On a standard Intel i7 processor, predictions are made in the order of milliseconds. Commercial software such as COILSIM1D performs slightly better in terms of accuracy, but the required central processing unit time per reaction is in the order of seconds. This tremendous speed-up and minimal accuracy loss make the presented framework highly suitable for the continuous monitoring of dif cult-to-access process parameters and for the envi- sioned, high-frequency real-time optimization (RTO) strategy or process control. Nevertheless, the lack of a fundamental basis implies that fundamental understanding is almost completely lost, which is not always well-accepted by the engineering community. In addition, the performance of the developed net- works drops signi cantly for naphthas that are highly dissimilar to those in the training set.
基金the National Natural Science Foundation of China(No.51875209)the Guangdong Basic and Applied Basic Research Foundation(No.2019B1515120060)the Open Funds of State Key Laboratory of Nuclear Power Safety Monitoring Technology and Equipment。
文摘Early fault warning for nuclear power machinery is conducive to timely troubleshooting and reductions in safety risks and unnecessary costs. This paper presents a novel intelligent fault prediction method, integrated probabilistic principal component analysis(PPCA), multi-resolution wavelet analysis, Bayesian inference, and RNN model for nuclear power machinery that consider data uncertainty and chaotic time series. After denoising the source data, the Bayesian PPCA method is employed for dimensional reduction to obtain a refined data group. A recurrent neural network(RNN) prediction model is constructed, and a Bayesian statistical inference approach is developed to quantitatively assess the prediction reliability of the model. By modeling and analyzing the data collected on the steam turbine and components of a nuclear power plant, the results of the goodness of fit, mean square error distribution, and Bayesian confidence indicate that the proposed RNN model can implement early warning in the fault creep period. The accuracy and reliability of the proposed model are quantitatively verified.
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。
