Subekti, M.*; Kudo, Kazuhiko*; Nabeshima, Kunihiko; Takamatsu, Kuniyoshi
Atom Indonesia, 43(2), p.93 - 102, 2017/08
Reactor kinetics based on point kinetic model have been generally applied as the standard method for neutronics codes. As the central control rod (C-CR) withdrawal test has demonstrated in a prismatic core of HTTR, the transient calculation of kinetic parameter, such as reactivity and neutron fluxes, requires a new method to shorten calculation-process time. Development of neural network method was applied to point kinetic model as the necessity of real-time calculation that could work in parallel with the digital reactivity meter. The combination of TDNN and Jordan RNN, such as TD-Jordan RNN, was the result of the modeling approach. The application of TD-Jordan RNN with adequate learning, tested offline, determined results accurately even when signal inputs were noisy. Furthermore, the preprocessing for neural network input utilized noise reduction as one of the equations to transform two of twelve time-delayed inputs into power corrected inputs.
Nabeshima, Kunihiko; Subekti, M.*; Matsuishi, Tomomi*; Ono, Tomio*; Kudo, Kazuhiko*; Nakagawa, Shigeaki
Journal of Power and Energy Systems (Internet), 2(1), p.92 - 103, 2008/00
The neural networks have been utilized in on-line monitoring-system of High Temperature Engineering Tested Reactor (HTTR) with thermal power of 30 MW. In this system, several neural networks can independently model the plant dynamics with different architecture, input and output signals and learning algorithm. One of main task is real-time plant monitoring by Multi-Layer Perceptron (MLP) in auto-associative mode, which can model and estimate the whole plant dynamics by training normal operational data only. Other tasks are on-line reactivity prediction, reactivity and helium leak monitoring, respectively. From the on-line monitoring results at the safety demonstration tests, each neural network shows good prediction and reliable detection performances.
Subekti, M.*; Kudo, Kazuhiko*; Nabeshima, Kunihiko
Proceedings of International Conference on Advances in Nuclear Science and Engineering (ICANSE 2007) (CD-ROM), p.53 - 63, 2007/11
The monitoring system for a huge and complex Pressurized Water Reactor (PWR) has some difficulties of monitoring task due to the dynamic system with large number of plant signals. The anomaly detection using neural networks integrated in the monitoring system improves the reactor safety to detect the anomaly faster than conventional methods. The advanced research considered the online anomaly diagnosis using expert system to complete the monitoring system tasks. The combination of neural network and expert system (neuro-expert) has been developed and tested in some anomaly conditions using PWR simulator. In simulation, the neuro-expert system could detect and diagnose the anomalies faster than the conventional alarm system.
Nabeshima, Kunihiko; Matsuishi, Tomomi*; Makino, Jun*; Subekti, M.*; Ono, Tomio*; Kudo, Kazuhiko*; Nakagawa, Shigeaki
Proceedings of 15th International Conference on Nuclear Engineering (ICONE-15) (CD-ROM), 6 Pages, 2007/04
The neural networks have been utilized in on-line monitoring system of High Temperature Engineering Tested Reactor (HTTR) with thermal power of 30MW. In this system, several neural networks can independently model the plant dynamics with different architecture, input and output signals and learning algorithm. One of main task is real-time plant monitoring by Multi-Layer Perceptron (MLP) in auto-associative mode, which can model and estimate the whole plant dynamics by training normal operational data only. Other tasks are on-line reactivity prediction, reactivity and helium leak monitoring, respectively. From the on-line test results, each neural network shows good prediction and reliable detection performances.
Subekti, M.*; Ono, Tomio*; Kudo, Kazuhiko*; Nabeshima, Kunihiko; Takamatsu, Kuniyoshi
Proceedings of 5th American Nuclear Society International Topical Meeting on Nuclear Plant Instrumentation, Controls, and Human Machine Interface Technology (NPIC & HMIT 2006) (CD-ROM), p.75 - 82, 2006/11
The neuro-expert has been utilized in previous monitoring-system research of Pressure Water Reactor (PWR). The research improved the monitoring system by utilizing neuro-expert, conventional noise analysis and modified neural networks for capability extension. The parallel method applications required distributed architecture of computernetwork for performing real-time tasks. The research aimed to improve the previous monitoring system, which could detect sensor degradation, and to perform the monitoring demonstration in High Temperature Engineering Tested Reactor (HTTR). The developing monitoring system based on some methods that have been tested using the data from online PWR simulator, as well as RSG-GAS (30 MW research reactor in Indonesia), will be applied in HTTR for more complex monitoring.
Nabeshima, Kunihiko; Kurnianto, K.*; Surbakti, T.*; Pinem, Surian*; Subekti, M.*; Minakuchi, Yusuke*; Kudo, Kazuhiko*
Proceedings of ICSC Congress on Computational Intelligence Methods and Applications (CIMA'2005) (CD-ROM), 4 Pages, 2005/12
The ANNOMA (Artificial Neural Network of Monitoring Aids) system is applied to the condition monitoring and signal validation of Multi Purpose Reactor in Indonesia. The feedforward neural network in auto-associative mode learns reactor's normal operational data, and models the reactor dynamics during the initial learning. The basic principle of the anomaly detection is to monitor the deviation between the process signals measured from the actual reactor and the corresponding values predicted by the reactor model, i.e., the neural networks. The pattern of the deviation at each signal is utilized for the identification of anomaly, e.g. sensor failure or system fault. The on-line test results showed that the neural network successfully monitored the reactor status during power increasing and steady state operation in real-time.
Subekti, M.*; Ono, Tomio*; Kudo, Kazuhiko*; Takamatsu, Kuniyoshi; Nabeshima, Kunihiko
Proceedings of International Conference on Nuclear Energy System for Future Generation and Global Sustainability (GLOBAL 2005) (CD-ROM), 6 Pages, 2005/10
In this study, a new full integrated monitoring system scheme based on distributed architecture is proposed. This monitoring system has a distributed architecture; monitoring tasks are assigned to client PCs by the central server. As a result of the distributed architecture, it is expected that the processing capabilities is maximized and the real time consistency is not impaired even if heavy monitoring tasks cause a shortage of bandwidth. And this system integrates signal processing modules in the main system and the main system distributes the monitoring tasks on its client PCs with TCP-IP technology. Signal processing between the main system and the client PCs is optimized so that monitoring tasks are distributed very efficiently. And, each client PC is completely separated, processing condition of one PC never effects on the other PC's processing.
Ono, Tomio*; Subekti, M.*; Kudo, Kazuhiko*; Takamatsu, Kuniyoshi; Nakagawa, Shigeaki; Nabeshima, Kunihiko
Nihon Genshiryoku Gakkai Wabun Rombunshi, 4(2), p.115 - 126, 2005/06
Control-rod withdrawal tests simulating reactivity insertion are carried out in the HTTR to verify the inherent safety features of HTGRs. This paper describes pre-test analysis method using artificial neural networks to predict the changes of reactor power and reactivity. The network model applied in this study is based on recurrent neural networks. The inputs of the network are the changes of the central control rods position and other significant core parameters, and the outputs are the changes of reactor power and reactivity. Furthermore, Time Synchronizing Signal(TSS) is added to input to improve the modeling of time series data. The actual tests data, which were previously carried out in the HTTR, were used for learning the model of the plant dynamics. After the learning, the network can predict the changes of reactor power and reactivity in the following tests.
Ono, Tomio*; Subekti, M.*; Maruyama, Yuta*; Nabeshima, Kunihiko; Kudo, Kazuhiko*
Dai-13-Kai Interijento, Shisutemu, Shimpojiumu Koen Rombunshu, p.212 - 217, 2003/12
In this research, we present nuclear power plant simulation method using Multilayer Perceptron, which is one of the models of Artificial Neural Networks(ANNs). The major characteristics of ANNs are to obtain the model through learning, analogy and very high speed processing. Furthermore, 'time synchronizing signal' and 'progress synchronizing signal' are added as the inputs to adapt the abnormal events with various scales or progress rates. This ANN, learned some sample data, can be flexibly adapted to simulate the abnormal events with different scales including explicit progress rates. In the verification using PWR simulator, we confirmed that this method could model NPP abnormal events by learning data and simulate the data which have different progress rates from learning data.
Subekti, M.*; Ono, Tomio*; Kudo, Kazuhiko*; Nabeshima, Kunihiko; Takamatsu, Kuniyoshi
no journal, ,
The determination of reactivity using alternative method, neural networks has been verified for C-CR withdrawal tests at power level of 9MW, 15MW, and 18MW. The neural network application has demonstrated for pre-test analysis and online reactivity determination as reference data for reactivity anomaly detection. The verification shows the best architecture of neural network that proposed for advanced online application.
Nabeshima, Kunihiko; Nakagawa, Shigeaki; Makino, Jun*; Matsuishi, Tomomi*; Subekti, M.*; Ono, Tomio*; Kudo, Kazuhiko*
no journal, ,
The neural networks have been utilized in on-line monitoring-system of High Temperature Engineering Tested Reactor (HTTR) with thermal power of 30MW. From the real-time test results during "reactivity insertion test; control rod withdrawal test" and "coolant flow reduction test", the monitoring system with neural networks showed good prediction and reliable detection performances.