Online sequential condition prediction method of natural circulation systems based on EOS-ELM and phase space reconstruction

Chen, Hanying; Gao, Puzhen; Tan, Sichao; Tang, Jiguo; Yuan, Hongsheng

doi:10.1016/j.anucene.2017.08.035

Cited by 14 publications

(5 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It would be interesting to estimate how this architecture would impact on the improvement of performance of some works in the bibliography. As an example, Chen et al [27] used OS-ELM to recognize different types of flow oscillations, and forecast them accurately, as a support for the operation of nuclear plants. They used an ensemble of 15 SLFNs with 40 hidden neurons, 15 input neurons (rolling motion condition), and a re-training frequency of 10 Hz.…”

Section: Discussionmentioning

confidence: 99%

“…These applications use the OS-ELM in the strict real-time sense. As an example, Chen et al [27] used an ensemble of OS-ELMs and phase space reconstruction to recognize different types of flow oscillations and accurately forecast the trend of monitored plant variables. It was intended as a support for the operation of nuclear plants, and provided that the prediction time may not be long for operators to take action, they used a sample interval of 0.1 s, time in which the prediction model can be adjusted according to last newly acquired data.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Moving Learning Machine towards Fast Real-Time Applications: A High-Speed FPGA-Based Implementation of the OS-ELM Training Algorithm

et al. 2018

View full text Add to dashboard Cite

Alfredo.Rosado@uv.es (A.R.-M.); manuel.bataller@uv.es (M.B.-M.); Juan.Barrios@uv.es (J.B.-A.); juan.guerrero@uv.es (J.F.G.-M.)Abstract: Currently, there are some emerging online learning applications handling data streams in real-time. The On-line Sequential Extreme Learning Machine (OS-ELM) has been successfully used in real-time condition prediction applications because of its good generalization performance at an extreme learning speed, but the number of trainings by a second (training frequency) achieved in these continuous learning applications has to be further reduced. This paper proposes a performance-optimized implementation of the OS-ELM training algorithm when it is applied to real-time applications. In this case, the natural way of feeding the training of the neural network is one-by-one, i.e., training the neural network for each new incoming training input vector. Applying this restriction, the computational needs are drastically reduced. An FPGA-based implementation of the tailored OS-ELM algorithm is used to analyze, in a parameterized way, the level of optimization achieved. We observed that the tailored algorithm drastically reduces the number of clock cycles consumed for the training execution up to approximately the 1%. This performance enables high-speed sequential training ratios, such as 14 KHz of sequential training frequency for a 40 hidden neurons SLFN, or 180 Hz of sequential training frequency for a 500 hidden neurons SLFN. In practice, the proposed implementation computes the training almost 100 times faster, or more, than other applications in the bibliography. Besides, clock cycles follows a quadratic complexity O(Ñ 2 ), withÑ the number of hidden neurons, and are poorly influenced by the number of input neurons. However, it shows a pronounced sensitivity to data type precision even facing small-size problems, which force to use double floating-point precision data types to avoid finite precision arithmetic effects. In addition, it has been found that distributed memory is the limiting resource and, thus, it can be stated that current FPGA devices can support OS-ELM-based on-chip learning of up to 500 hidden neurons. Concluding, the proposed hardware implementation of the OS-ELM offers great possibilities for on-chip learning in portable systems and real-time applications where frequent and fast training is required.Electronics 2018, 7, 308 2 of 23 learning of neural networks for the prediction of future opponent robot coordinates; ref.[3] designed an ASIC on-chip learning to learn and extract features existing in input datasets, intended to embedded vision applications; or [4], that implemented a real-time classifier for neurological signals.The Extreme Learning Machine (ELM) algorithm possesses many aspects that makes it suitable for any real-time or custom hardware implementation. It has a reduced and fixed training time along with an extremely fast learning speed that allows determinism in the computation time and, thus, a great advantage compared to previous well-known training...

show abstract

Section: Discussionmentioning

confidence: 99%

mentioning

confidence: 99%

Moving Learning Machine towards Fast Real-Time Applications: A High-Speed FPGA-Based Implementation of the OS-ELM Training Algorithm

et al. 2018

View full text Add to dashboard Cite

show abstract

“…Liu and Xie [ 89 ] established an NPP operating parameter prediction model based on the online training BPNN, and verified the effectiveness of the model by predicting the fluctuating system operating parameters such as the coolant void fraction and the water level of the steam generator and the pressurizer. Chen and Gao [ 90 ] proposed an online-condition approach for flow oscillation pattern recognition and trend forecasting of a natural circulation system based on the ensemble of online sequential extreme learning machine (EOS–ELM), which achieves a significantly high training speed and good accuracy. There are few existing short-term prediction studies on various complex thermal-hydraulic conditions (e.g., severe accidents) in NPP systems, even though such prediction models have great significance from the viewpoint of practical application.…”

Section: Application Of Ai To Nuclear Reactor Oandmmentioning

confidence: 99%

“…However, since the online BPNN still relies on time-consuming gradient descent iterative training, in order to ensure the computation speed, the number of training steps is greatly limited. [ 89 ] Online sequential condition prediction EOS-ELM Fast learning speeding without obvious overfitting problems [ 90 ] Water level prediction DNN-GA DNN model has better performance than cascaded fuzzy neural network [ 91 ] Operating parameters prediction during LOCA DNN/LSTM Proposed methods are 100,000 times faster than the original simulation tool with satisfying accuracy [ 92 ] Prediction of neutron flux and power distributions ROM-ML Able to predict high-dimensional outputs with physics-informed digital twins framework [ 94 ] Compensation for low-precision model deviation K-means and ANN A digital twin model consisting of offline and online stages is proposed, and its calibration results are shown to have good agreement with the ground truth [ 95 ] System-level FD ANN 8 operating conditions can be accurately diagnosed and classified [ 98 ] System-level FD ANN A dynamic architecture was proposed, in which the first network is used to judge whether the system is in an abnormal state, and the second network is used for classification of abnormal conditions [ 99 ] System-level FD PCA PCA enables fast compression of multiple dimensions for transient identification [ 100 ] System-level FD RBFNN Able to recognize the three accidents, even with a noise level up to 10% …”

Section: Introductionmentioning

confidence: 99%

A review of the application of artificial intelligence to nuclear reactors: Where we are and what's next

Huang

Peng

Deng

et al. 2023

Heliyon

View full text Add to dashboard Cite

“…On the other hand, some other studies on the prognostics of NPPs involve the short-term trend prediction of NPP operating conditions. An online condition forecasting method for a natural circulation system with fow instability was developed by Chen et al [9] based on the ensemble of the online sequential extreme learning machine (EOS-ELM). Marseguerra et al [10,11] presented a predictive model based on neurofuzzy techniques to predict the water level in a steam generator of a pressurized water reactor (PWR).…”

Section: Introductionmentioning

confidence: 99%

Prediction of Automatic Scram during Abnormal Conditions of Nuclear Power Plants Based on Long Short-Term Memory (LSTM) and Dropout

Chen

Gao

Tan

et al. 2023

Science and Technology of Nuclear Installations

Self Cite

View full text Add to dashboard Cite

A deep-learning model was proposed for predicting the remaining time to automatic scram during abnormal conditions of nuclear power plants (NPPs) based on long short-term memory (LSTM) and dropout. The proposed model was trained by simulated condition data of abnormal conditions; the input of the model was the deviation of the monitoring parameters from the normal operating state, and the output was the remaining time from the current moment to the upcoming reactor trip. The predicted remaining time to the reactor trip decreases with the development of abnormal conditions; thus, the output of the proposed model generates a predicted countdown to the reactor trip. The proposed prediction model showed better prediction performance than the Elman neural network model in the experiments but encountered an overfitting problem for testing data containing noise. Therefore, dropout was applied to further improve the generalization ability of the prediction model based on LSTM. The proposed automatic scram prediction model can provide NPP operators with an alert to the automatic scram during abnormal conditions.

show abstract

Online sequential condition prediction method of natural circulation systems based on EOS-ELM and phase space reconstruction

Cited by 14 publications

References 25 publications

Moving Learning Machine towards Fast Real-Time Applications: A High-Speed FPGA-Based Implementation of the OS-ELM Training Algorithm

Moving Learning Machine towards Fast Real-Time Applications: A High-Speed FPGA-Based Implementation of the OS-ELM Training Algorithm

A review of the application of artificial intelligence to nuclear reactors: Where we are and what's next

Prediction of Automatic Scram during Abnormal Conditions of Nuclear Power Plants Based on Long Short-Term Memory (LSTM) and Dropout

Contact Info

Product

Resources

About