CFD and machine learning based hybrid model for passive dilution of helium in a top ventilated compartment
-
Pavan K. Sharma
,
Vishnu Verma
Abstract
For hydrogen management in the containment of Nuclear Power Plants (NPPs), besides the Passive autocatalytic Recombiners (PAR), the passive dilution of lighter gas plays an important role. This could be an attractive option to optimize the containment design and to estimate the extent of dilution. Passive dilution has many other applications in nuclear industry. The experimental studies of air entrainment in the upward rising helium plume and the resulting dilution of helium gas by the Canadians in terms of Volume Flow Magnification Factor (VFMF) have been utilized for Computational Fluid Dynamics (CFD) validation. The CFD based Fire Dynamics Simulator (FDS) predicted values of VFMF found to be in good agreement with the test data. After FDS code validation, parametric study has been carried out to generate a data base of VFMF for range of hydrogen injection, side opening area and opening height. In present study various Machine Learning (ML) models are evaluated based on two-parameter relationship i.e. non dimensional hydrogen injection and VFMF using the CFD code generated database. The trained ML models were used for the predictions of the mass flow rate of gas entrainment (through opening) in the rising buoyant plume in terms of VFMF. The ML predictions were in good agreement with the predictions against test data. Multivariate Adaptive Regression Splines (MARS) based ML model found to performed best and discussed in the paper. The paper highlights details of methodology of numerical simulation, results of the CFD studies and machine learning based predictions.
-
Research ethics: Not applicable.
-
Author contributions: The authors have accepted responsibility for the entire content of this manuscript and approved its submission
-
Competing interests: The authors state no conflict of interest
-
Research funding: None declared
-
Data availability: Not applicable.
Step 0: Problem analysis
Variables considered in the model are (i) non dimensional hydrogen injection (ii) VFMF. In this problem, various models are utilized i.e. Liner regression, Decision Tree regression, Lasso regression, Ridge regression, Random Forests regression and finally MARS regression which will study the correlation between the non dimensional hydrogen injection and non-dimensional VFMF.
Step 1: Importing the libraries
The first step will always consist of importing the libraries that are needed to develop the ML model. The NumPy, matplotlib and the Pandas libraries are imported.
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
Step 2: Preparing and importing the dataset
Before importing the data, the data was prepared using the digitisation of graphs available in open literature data. The objective of the present work is to prepare a safe and reasonably conservative correlation for regulatory application. In this step, we shall use pandas to store the data prepared and store it as a Pandas DataFrame using the function ‘pd.read_csv’. In this, we assign the independent variable (X) to the ‘non dimensional hydrogen injection’ column and the dependent variable (y) to the ‘VFMF’ column.
Step 3: splitting the dataset into the training set and test set
In the next step, dataset is splitted as usual into the training set and the test set. In present study, 256 non dimensional VFMF data (as a function of non dimensional hydrogen injection) for training and 86 VFMF data for ML testing.
from sklearn.model_selection import train_test_split
train, test=train_test_split(gdpdata)
Step: 4, 5, 6, 7, 8, 9: training for various models
Liner regression, Decision Tree regression, Lasso regression, Ridge regression, Random Forests regression and finally MARS regression model were used on the training set. The DecisionTreeRegressor class imported from sklearn.tree and assigned it to the variable ‘regressor’. Then X_train and the y_train have been fit to the model by using the regressor.fit function.
#:Linear Regression
from sklearn.linear_model import LinearRegression
model=LinearRegression().fit(train.iloc[:,:1],train.iloc[:,1:])
Step 10: Predicting the results
In this step, the results predicated of the test set with the model trained on the training set values using the regressor.predict function and assign it to ‘y_pred’.
y_pred = regressor.predict(X_test.reshape(-1,1))
Step 11: Comparing the Real Values with Predicted values
In this step, comparison and display of the values of y_test as ‘Real Values’ and y_pred as ‘Predicted Values’ is to be carried out in a Pandas dataframe.
References
Andreani, M., Putz, F., Dury, T., Gjerloev, C., and Smith, B. (2001). Application of field codes to the analysis of gas mixing in large volumes. In: IAEA technical committee meeting on the implementation of hydrogen mitigation techniques and filtered containment venting, Cologne, Germany, June 2001.Search in Google Scholar
Baum, H.R. and Rehm, R.G. (1982). Numerical computation of large-scale fire induced flows. In: Krause, E. (Ed.), Eighth international conference on numerical methods in fluid dynamics, lecture notes in physics, Vol. 170, Springer-Verlag, Berlin, pp. 124–130.10.1007/3-540-11948-5_9Search in Google Scholar
Baum, H.R., Rehm, R.G., and Mulholland, G.W. (1982). Computation of fire induced flow and smoke coagulation. In: Nineteenth Symposium (International) on combustion/the combustion institute, Vol. 2–22, pp. 921–931, 2–22.10.1016/S0082-0784(82)80268-3Search in Google Scholar
Baum, H.R., McGrattan, K.B., and Rehm, R.G. (1994). Mathematical modelling and computer simulation of fire Phenomena. In: Fire safety science – proceedings of the fourth international Symposium, Ottawa, Ontario, Canada, 13–17 June, pp. 185–193.10.3801/IAFSS.FSS.4-185Search in Google Scholar
Baum, H.R., McGrattan, K.B., and Rehm, R.G. (1996). Large eddy simulations of smoke movement in three dimensions. In: Conference proceedings of the seventh international interflam conference, interscience communications, London, pp. 189–198.Search in Google Scholar
Baum, H.R., McGrattan, K.B., and Rehm, R.G. (1997). Three dimensional simulations of fire plume dynamics. In: Fire safety science – proceedings of the fifth international Symposium, pp. 511–522.10.3801/IAFSS.FSS.5-511Search in Google Scholar
Baum, H.R. (1999). Large eddy simulations of fires – from concepts to computations. Fire Protect. Eng. 6: 36–42.Search in Google Scholar
Clement, J.M. (2000). Experimental verification of the fire dynamics simulator (FDS) hydrodynamic model, Ph.D. Thesis. Christchurch, New Zealand, University of Canterbury.Search in Google Scholar
Ferziger, J.H. (1993). Subgrid-scale modeling. In: Galperin, B. and Orszag, S.A. (Eds.), Large eddy Simulation of complex Engineering and geophysical flows. Cambridge University Press, Cambridge, pp. 37–54.Search in Google Scholar
Friedman, J.H. (1993). Technical report No. 110: fast MARS. Technical Report, Stanford University Department of Statistics. http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Fast+MARS#0.Search in Google Scholar
Galea, E.R., Knight, B., Patel, M.K., Ewer, J., Pitridis, M., and Taylor, S. (1998). SMARTFIRE v2.0 user Guide and technical manual, fire safety engineering group. University of Greenwich, UK.Search in Google Scholar
Hultgren, A., Gallego-Marcos, I., Villanueva, W., and Kudinov, P. (2014). Large scale erosion of a helium stratified layer by a vertical jet using the Gothic Code. In: The 10th international topical meeting on nuclear thermal-hydraulics, operation and safety (NUTHOS10-1290), Okinawa, Japan.Search in Google Scholar
Kerrison, L., MaWhinney, N., Galea, E.R., Hoffmann, N., and Patel, M.K. (1994). A comparison of a FLOW3D based fire field model with experimental compartment fire data. Fire Saf. J. 23: 387–411, https://doi.org/10.1016/0379-7112(94)90005-1.Search in Google Scholar
Kumar, S., Hoffmann, N., Cox, G. (1985). Some validation of jasmine for fires in hospital wards, numerical solution of fluid flow and heat/mass transfer processes. Springer-Verlag, Berlin, p. 159.10.1007/978-3-642-82781-5_12Search in Google Scholar
Max, K. and Johnson, K. (2013). Applied predictive modelling, 1st ed. Springer, New York.Search in Google Scholar
McGrattan, K.B., Baum, H.R., and Rehm, R.G. (1998). Large eddy simulations of smoke movement. Fire Saf. J. 30: 161–178, https://doi.org/10.1016/S0379-7112(97)00041-6.Search in Google Scholar
McGrattan, K.B., Baum, H.R., Rehm, R.G., Forney, G.P., Floyd, J.E., and Hostikka, S. (2001). Fire dynamics simulator (version 2), technical reference guide, Technical Report NISTIR 6783. National Institute of Standards and Technology, Gaithersburg, Maryland, August.10.6028/NIST.IR.6783e2002Search in Google Scholar
py-earth, https://contrib.scikit-learn.org/py-earth/content.html.Search in Google Scholar
Rehm, R.G., McGrattan, K.B., Baum, H.R., and Cassel, K.W. (1997). Transport by gravity currents in building fires. In: Fire safety science – Proceedings of the Fifth international Symposium, pp. 391–402.10.3801/IAFSS.FSS.5-391Search in Google Scholar
Rehm, R.G., Tang, H.C., Baum, H.R., Sims, J.S., and Corley, D.M. (1991). A Boussinesq algorithm for enclosed buoyant convection in two dimensions, NISTIR 4540. U.S. Department of Commerce, National Institute of Standards and Technology, Computing and Applied Mathematics Laboratory, Gaithersburg, MD 20899, U.S.A.10.6028/NIST.IR.4540Search in Google Scholar
Sharma, P.K., Markandeya, S.G., Ghosh, A.K., Kushwaha, H.S., and Venkat Raj, V. (2001). A computational fluid dynamics study of a buoyant plume in a cross-flow condition. In: 28th national conference on fluid mechanics and fluid power. Punjab University, Chandigarh, Dec. 13–15.Search in Google Scholar
Sharma, P.K., Ghosh, A.K., and Kushwaha, H.S. (2006). A Computational fluid dynamics study of pool Fire and its pulsating behaviour. In: National conference on advances in mechanical engineering AIME-2006, Aaligarh Muslim University.Search in Google Scholar
Smagorinsky, J. (1963). General circulation experiments with the primitive equations. Mon. Weather Rev. 91: 99–165, https://doi.org/10.1175/1520-0493(1963)091<0099:gcewtp>2.3.co;2.10.1175/1520-0493(1963)091<0099:GCEWTP>2.3.CO;2Search in Google Scholar
Spearpoint, M., Mowrer, F.W., and McGrattan, K. (1999). Simulation of a compartment flashover fire using hand calculations, zone models and a field model, International Conference on Fire Research and Engineering (ICFRE3). In: Third (3rd) proceedings. Society of fire protection engineers (SFPE), national institute of standard and technology (NIST) and international association of fire safety science (IAFSS), Oct 4–8, 1999, Chicago, IL, society of fire protection engineers, Bostan, MA, pp. 3–14.Search in Google Scholar
Trevor, H., Tibshirani, R., and Friedman, J. (2009). The elements of statistical learning: data mining, inference, and prediction, 2nd ed. Springer, New York.Search in Google Scholar
Welch, S. and Rubini, P.A. (1996). SOFIE 2.1 user’s manual. Cranfield University, Cranfield, United Kingdom.Search in Google Scholar
Yadav, M.K., Khandekar, S., and Sharma, P.K. (2016). An integrated approach to steam condensation studies inside reactor containments: a review. Nucl. Eng. Des. 300: 181–209, https://doi.org/10.1016/j.nucengdes.2016.01.004.Search in Google Scholar
© 2023 Walter de Gruyter GmbH, Berlin/Boston
Articles in the same Issue
- Frontmatter
- CFD modeling of natural convection in pebble bed geometry with finite volume method
- Experimental investigation of heat transfer characteristics of inclined aluminium two phase closed thermosyphon
- Advances in triple tube heat exchangers regarding heat transfer characteristics of single and two-phase flows in comparison to double tube heat exchangers part 1
- Advances in triple tube heat exchangers regarding heat transfer characteristics of single and two-phase flows in comparison to double tube heat exchangers part 2
- Neutronic and thermal-hydraulic assessment of the TRR with new core designed based on tubular fuels
- A non-dominated discrete differential evolution for fuel loading pattern optimization of a nuclear research reactor
- CFD and machine learning based hybrid model for passive dilution of helium in a top ventilated compartment
- Optimization strategy for SAM in nuclear power plants based on NSGA-II
- Probing 6He induced reactions with nuclear level density
- An application for nonlinear heterogeneity-based isotherm models in characterization of niobium sorption on clay rocks and granite
- Calendar of events
Articles in the same Issue
- Frontmatter
- CFD modeling of natural convection in pebble bed geometry with finite volume method
- Experimental investigation of heat transfer characteristics of inclined aluminium two phase closed thermosyphon
- Advances in triple tube heat exchangers regarding heat transfer characteristics of single and two-phase flows in comparison to double tube heat exchangers part 1
- Advances in triple tube heat exchangers regarding heat transfer characteristics of single and two-phase flows in comparison to double tube heat exchangers part 2
- Neutronic and thermal-hydraulic assessment of the TRR with new core designed based on tubular fuels
- A non-dominated discrete differential evolution for fuel loading pattern optimization of a nuclear research reactor
- CFD and machine learning based hybrid model for passive dilution of helium in a top ventilated compartment
- Optimization strategy for SAM in nuclear power plants based on NSGA-II
- Probing 6He induced reactions with nuclear level density
- An application for nonlinear heterogeneity-based isotherm models in characterization of niobium sorption on clay rocks and granite
- Calendar of events
