The use of Artificial Neural Network models for CO2 capture plants

Artificial Neural Networks (ANN) are multifaceted tools that can be used to model and predict various complex and highly non-linear processes. This paper presents the development and validation of an ANN model of a CO₂ capture plant. An evaluation of the concept is made of the usefulness of the ANN model as well as a discussion of its feasibility for further integration into a conventional heat and mass balance programme. It is shown that the trained ANN model can reproduce the results of a rigorous process simulator in fraction of the simulation time. A multilayer feed-forward form of Artificial Neural Network was used to capture and model the non-linear relationship between inputs and outputs of the CO₂ capture process. The data used for training and validation of the ANN were obtained using the process simulator CO2SIM. The ANN model was trained by performing fully automatic batch simulations using CO2SIM over the entire range of actual operation for an amine based absorption plant. The trained model was then used for finding the optimum operation for the example plant with respect to lowest possible specific steam duty and maximum CO₂ capture rate. Two different algorithms have been used and compared for the training of the ANN and a sensitivity analysis was carried out to find the minimum number of input parameters needed while maintaining sufficient accuracy of the model. The reproducibility shows error less than 0.2% for the closed loop absorber/desorber plant. The results of this study show that trained ANN models are very useful for fast simulation of complex steady state process with high reproducibility of the rigorous model.     

A MINLP model including the pressure levels and multiperiods for CHP process optimisation

In this paper a mixed integer nonlinear programming (MINLP) model for small-scale combined heat and power (CHP) plant synthesis and operation is presented. The model includes also the modelling of pressures levels and part load operation unlike the model presented in earlier work. Also, a detailed model of a back-pressure steam turbine is added to the model, taking into account the efficiency changes in turbine stages and the steam extraction pressure dependence on the steam mass flow rate. The sensitivity analysis shows that the developed model can be solved fairly reliably with commercial local MINLP solvers. The model corresponds well with a simulation model based on an existing plant. The model is able to find improved small-scale CHP designs that have higher efficiencies and that are profitable for wide ranges of electricity prices and fossil CO₂ emission permit prices. The developed model is a suitable decision making tool when evaluating the trade-offs between investments, profits, and fossil CO₂ emissions in small-scale CHP plant synthesis and operation. However, the formulations of the model are not limited to small-scale CHP processes and can be used to model also larger CHP processes and energy systems.     

Development of artificial neural network model for a coal-fired boiler using real plant data

Development of artificial neural network (ANN) models using real plant data for the prediction of fresh steam properties from a brown coal-fired boiler of a Slovenian power plant is reported. Input parameters for this prediction were selected from a large number of available parameters. Initial selection was made on a basis of expert knowledge and previous experience. However, the final set of input parameters was optimized with a compromise between smaller number of parameters and higher level of accuracy through sensitivity analysis. Data for training were selected carefully from the available real plant data. Two models were developed, one including mass flow rate of coal and the other including belt conveyor speed as one of the input parameters. The rest of the input parameters are identical for both models. Both models show good accuracy in prediction of real data not used for their training. Thus both of them are proved suitable for use in real life, either on-line or off-line. Better model out of these two may be decided on a case-to-case basis depending on the objective of their use. The objective of these studies was to examine the feasibility of ANN modeling for coal-based power or combined heat and power (CHP) plants.     

Enhancing reliability and lifespan of PEM fuel cells through neural network-based fault detection and classification

In order to maximise fuel cell reliability of operation and useful life span, an accurate online health assessment of the fuel cell system is essential. Existing algorithms for fault detection in fuel cell systems are based on sensing elements, control methods, and statistical/probabilistic models. In this paper, an artificial neural network (ANN) will be developed to detect and classify faults in proton-exchange membrane (PEM) fuel cell systems. As the ANN model developed within the PEM system relies on the input and output current and voltage, additional sensing devices are not required within the system. Based on an experimental setup using a 3-kW fuel cell system, it was found that the proposed model was able to detect faults associated with the reduction/increase of fuel pressure, H₂ consumption rate, and voltage regulation changes in the dc-dc converter with >90% accuracy. In the proposed model, historical data is required to train and validate the ANN algorithm, but after this is complete, no human intervention is required afterward.     

An artificial neural network‐based condition monitoring method for wind turbines,with application to the monitoring of the gearbox

Major failures in wind turbines are expensive to repair and cause loss of revenue due to long downtime. Condition‐based maintenance, which provides a possibility to reduce maintenance cost, has been made possible because of the successful application of various condition monitoring systems in wind turbines. New methods to improve the condition monitoring system are continuously being developed. Monitoring based on data stored in the supervisory control and data acquisition (SCADA) system in wind turbines has received attention recently. Artificial neural networks (ANNs) have proved to be a powerful tool for SCADA‐based condition monitoring applications. This paper first gives an overview of the most important publications that discuss the application of ANN for condition monitoring in wind turbines. The knowledge from these publications is utilized and developed further with a focus on two areas: the data preprocessing and the data post‐processing. Methods for filtering of data are presented, which ensure that the ANN models are trained on the data representing the true normal operating conditions of the wind turbine. A method to overcome the errors from the ANN models due to discontinuity in SCADA data is presented. Furthermore, a method utilizing the Mahalanobis distance is presented, which improves the anomaly detection by considering the correlation between ANN model errors and the operating condition. Finally, the proposed method is applied to case studies with failures in wind turbine gearboxes. The results of the application illustrate the advantages and limitations of the proposed method. Copyright © 2017 John Wiley & Sons, Ltd.     

A hybrid neural network model for PEM fuel cells

The goal of this paper is to discuss a neural network modeling approach for developing a quantitatively good model for proton exchange membrane (PEM) fuel cells. Various ANN approaches have been tested; the back-propagation feed-forward networks and radial basis function networks show satisfactory performance with regard to cell voltage prediction. The effects of Pt loading on the performance of the PEM fuel cell have been specifically studied. The results show that the ANN model is capable of simulating these effects for which there are currently no valid fundamental models available from the open literature.

Two novel hybrid neural network models (multiplicative and additive), each consisting of an ANN component and a physical component, have been developed and compared with the full-blown ANN model. The results from the hybrid models demonstrate comparable performance (in terms of cell voltage predictions) compared to the ANN model. Additionally, the hybrid models show performance gains over the physical model alone. The additive hybrid model shows better accuracy than that of the multiplicative hybrid model in our tests.     

Hydrogen purification layered bed optimization based on artificial neural network prediction of breakthrough curves

Artificial neural network has generally been used for a quantity of tasks such as classification, prediction, clustering and association analysis in different application fields. To the best of our knowledge, there are few researches on breakthrough curve used artificial neural network. In this paper, an artificial neural network model is established for breakthrough curves prediction in relation to a ternary components gas with a two-layered adsorbent bed piled up with activated carbon (AC) and zeolite, and an optimization is concluded by the artificial neural network. The performance data which acquired by Aspen model has been utilized for training artificial neural network (ANN) model. The ANN model trained has great competence for making prediction of hydrogen purification performance of PSA cycle with impressive speed and rational accuracy. On the strength of the ANN model, we implemented an optimization for seeking first-rank PSA cycle parameters. The optimization is concentrated on the effect of inlet flow rate, pressure and layer ratio of activated carbon height to zeolite height. Furthermore, this paper shows that the PSA cycle's optimal operation parameters can be obtained by use of ANN model and optimization algorithm, the ANN model has been trained according to the data generated by Aspen adsorption model.     

Application of adaptive neuro-fuzzy inference system techniques and artificial neural networks to predict solid oxide fuel cell performance in residential microgeneration installation

This study applies adaptive neuro-fuzzy inference system (ANFIS) techniques and artificial neural network (ANN) to predict solid oxide fuel cell (SOFC) performance while supplying both heat and power to a residence. A microgeneration 5 kW_el SOFC system was installed at the Canadian Centre for Housing Technology (CCHT), integrated with existing mechanical systems and connected in parallel to the grid. SOFC performance data were collected during the winter heating season and used for training of both ANN and ANFIS models. The ANN model was built on back propagation algorithm as for ANFIS model a combination of least squares method and back propagation gradient decent method were developed and applied. Both models were trained with experimental data and used to predict selective SOFC performance parameters such as fuel cell stack current, stack voltage, etc.     

ANN-GA based optimization of a high ash coal-fired supercritical power plant

The efficiency of coal-fired power plant depends on various operating parameters such as main steam/reheat steam pressures and temperatures, turbine extraction pressures, and excess air ratio for a given fuel. However, simultaneous optimization of all these operating parameters to achieve the maximum plant efficiency is a challenging task. This study deals with the coupled ANN and GA based (neuro-genetic) optimization of a high ash coal-fired supercritical power plant in Indian climatic condition to determine the maximum possible plant efficiency. The power plant simulation data obtained from a flow-sheet program, “Cycle-Tempo” is used to train the artificial neural network (ANN) to predict the energy input through fuel (coal). The optimum set of various operating parameters that result in the minimum energy input to the power plant is then determined by coupling the trained ANN model as a fitness function with the genetic algorithm (GA). A unit size of 800 MWe currently under development in India is considered to carry out the thermodynamic analysis based on energy and exergy. Apart from optimizing the design parameters, the developed model can also be used for on-line optimization when quick response is required. Furthermore, the effect of various coals on the thermodynamic performance of the optimized power plant is also determined.     

Estimation of sky luminance in the tropics using artificial neural networks: Modeling and performance comparison with the CIE model

An artificial neural network (ANN) model for estimating sky luminance was developed. A 3-year period (2007–2009) of sky luminance data obtained from measurements at Nakhon Pathom (13.82°N, 100.04°E) and a 1-year period (2008) of the same type of data at Songkhla (7.20°N, 100.60°E), Thailand were used in this study. The ANN model was trained using a back propagation algorithm, based on 2 years data (2007–2008) at Nakhon Pathom for clear, partly cloudy and overcast skies. The trained ANN model was used to predict sky luminance at Nakhon Pathom for the year 2009 for the case of clear, partly cloudy and overcast skies. The results were compared with those of the CIE model. It was found that the ANN model performed better than CIE models for most cases. The ANN model trained with Nakhon Pathom data were also used to predict sky luminance at Songkhla and satisfactory results were obtained.     

Artificial neural networks applications in building energy predictions and a case study for tropical climates

This study presents artificial neural network (ANN) methods in building energy use predictions. Applications of the ANN methods in energy audits and energy savings predictions due to building retrofits are emphasized. A generalized ANN model that can be applied to any building type with minor modifications would be a very useful tool for building engineers. ANN methods offer faster learning time, simplicity in analysis and adaptability to seasonal climate variations and changes in the building's energy use when compared to other statistical and simulation models. The model herein is presented for predicting chiller plant energy use in tropical climates with small seasonal and daily variations. It was successfully created based on both climatic and chiller data. The average absolute training error for the model was 9.7% while the testing error was 10.0%. This indicates that the model can successfully predict the particular chiller energy consumption in a tropical climate. Copyright © 2005 John Wiley & Sons, Ltd.     

Application of artificial neural networks (ANN) for modeling of industrial hydrogen plant

A back propagation feed forward artificial neural network (ANN) with three layers is used for modeling of industrial hydrogen plant. The required operating data for training of ANN is obtained by modeling and simulation of an industrial hydrogen plant. The operating data are calculated by changing effective parameters such as feed temperature, reformer pressure, steam to carbon ratio and carbon dioxide to methane ratio in feed stream. Tangent sigmoid transfer function is used in the hidden and output layer and the proposed neural network is trained with a gradient descent algorithm. The optimum number of neurons in hidden layer is determined as optimum value with minimizing of the mean square error (MSE). With changing of effective parameters, the model predicts temperature, pressure and mole fraction of hydrogen and carbon monoxide in the product of the hydrogen plant. The result can be used to gain better knowledge and optimize of the hydrogen production plants.     

Decentralised optimisation of cogeneration in virtual power plants

Within several projects we investigated grid structures and management strategies for active grids with high penetration of renewable energy resources and distributed generation (RES & DG). Those ”smart grids” should be designed and managed by model based methods, which are elaborated within these projects. Cogeneration plants (CHP) can reduce the greenhouse gas emissions by locally producing heat and electricity. The integration of thermal storage devices is suitable to get more flexibility for the cogeneration operation. If several power plants are bound to centrally managed clusters, it is called “virtual power plant”. To operate smart grids optimally, new optimisation and model reduction techniques are necessary to get rid with the complexity.There is a great potential for the optimised management of CHPs, which is not yet used. Due to the fact that electrical and thermal demands do not occur simultaneously, a thermally driven CHP cannot supply electrical peak loads when needed. With the usage of thermal storage systems it is possible to decouple electric and thermal production. We developed an optimisation method based on mixed integer linear programming (MILP) for the management of local heat supply systems with CHPs, heating boilers and thermal storages. The algorithm allows the production of thermal and electric energy with a maximal benefit. In addition to fuel and maintenance costs it is assumed that the produced electricity of the CHP is sold at dynamic prices. This developed optimisation algorithm was used for an existing local heat system with 5 CHP units of the same type. An analysis of the potential showed that about 10% increase in benefit is possible compared to a typical thermally driven CHP system under current German boundary conditions. The quality of the optimisation result depends on an accurate prognosis of the thermal load which is realised with an empiric formula fitted with measured data by a multiple regression method.The key functionality of a virtual power plant is to increase the value of the produced power by clustering different plants. The first step of the optimisation concerns the local operation of the individual power generator, the second step is to calculate the contribution to the virtual power plant. With small extensions the suggested MILP algorithm can be used for an overall EEX (European Energy Exchange) optimised management of clustered CHP systems in form of the virtual power plant. This algorithm has been used to control cogeneration plants within a distribution grid.     

Use of artificial neural networks for mapping of solar potential in Turkey

Turkey has sufficient solar radiation intensities and radiation durations for solar thermal applications since Turkey lies in a sunny belt, between 36° and 42° N latitudes. The yearly average solar-radiation is 3.6 kWh/m²day, and the total yearly radiation period is ∼2610 h. The main focus of this study is to determine the solar-energy potential in Turkey using artificial neural-networks (ANNs). Scaled conjugate gradient (SCG), Pola-Ribiere conjugate gradient (CGP), and Levenberg-Marquardt (LM) learning algorithms and a logistic sigmoid transfer function were used in the network. In order to train the neural network, meteorological data for the last 3 years (2000–2002) from 17 stations (namely cities) spread over Turkey were used as training (11 stations) and testing (6 stations) data. Meteorological and geographical data (latitude, longitude, altitude, month, mean sunshine duration, and mean temperature) are used as inputs to the network. Solar radiation is in the output layer. The maximum mean absolute percentage error was found to be less than 6.7% and R² values to be about 99.8937% for the testing stations. However, the respective values were found to be 2.41 and 99.99658% for the training stations. The trained and tested ANN models show greater accuracies for evaluating solar resource posibilities in regions where a network of monitoring stations has not been established in Turkey. The predicted solar-potential values from the ANN were given in the form of monthly maps. These maps are of prime importance for different working disciplines, like those of scientists, architects, meteorologists, and solar engineers in Turkey. The predictions from ANN models could enable scientists to locate and design solar-energy systems in Turkey and determine the appropriate solar technology.     

MODELING OF SOLAR DOMESTIC WATER HEATING SYSTEMS USING ARTIFICIAL NEURAL NETWORKS

Artificial Neural Networks (ANN) are widely accepted as a technology offering an alternative way to tackle complex and ill-defined problems. They can be trained to predict results from examples, are fault tolerant, are able to deal with non-linear problems, and once trained can perform prediction at high speed. ANNs have been used in diverse applications and they have shown to be particularly useful in system modeling and for system identification. The objective of this work was to train an ANN to learn to predict the useful energy extracted and the temperature rise in the stored water of solar domestic water heating (SDHW) systems with the minimum of input data. An ANN has been trained based on 30 known cases of systems, varying from collector areas between 1.81 m² and 4.38 m². Open and closed systems have been considered both with horizontal and vertical storage tanks. In addition to the above, an attempt was made to consider a large variety of weather conditions. In this way the network was trained to accept and handle a number of unusual cases. The data presented as input were the collector area, storage tank heat loss coefficient (U-value), tank type, storage volume, type of system, and ten readings from real experiments of total daily solar radiation, mean ambient air temperature, and the water temperature in the storage tank at the beginning of a day. The network output is the useful energy extracted from the system and the temperature rise in the stored water. The statistical R²-value obtained for the training data set was equal to 0.9722 and 0.9751 for the two output parameters respectively. Unknown data were subsequently used to investigate the accuracy of prediction. These include systems considered for the training of the network at different weather conditions and completely unknown systems. Predictions within 7.1% and 9.7% were obtained respectively. These results indicate that the proposed method can successfully be used for the estimation of the useful energy extracted from the system and the temperature rise in the stored water. The advantages of this approach compared to the conventional algorithmic methods are the speed, the simplicity, and the capacity of the network to learn from examples. This is done by embedding experiential knowledge in the network. Additionally, actual weather data have been used for the training of the network, which leads to more realistic results as compared to other modeling programs, which rely on TMY data that are not necessarily similar to the actual environment in which a system operates.     

Artificial neural networks used for the performance prediction of a thermosiphon solar water heater

Artificial Neural Networks (ANN) are widely accepted as a technology offering an alternative way to tackle complex and ill-defined problems. They can learn from examples, are fault tolerant, are able to deal with non-linear problems, and once trained can perform prediction at high speed. ANNs have been used in diverse applications and they have shown to be particularly effective in system modelling as well as for system identification. The objective of this work is to train an artificial neural network (ANN) to learn to predict the performance of a thermosiphon solar domestic water heating system. This performance is measured in terms of the useful energy extracted and of the stored water temperature rise. An ANN has been trained using performance data for four types of systems, all employing the same collector panel under varying weather conditions. In this way the network was trained to accept and handle a number of unusual cases. The data presented as input were, the storage tank heat loss coefficient (U-value), the type of system (open or closed), the storage volume, and a total of fifty-four readings from real experiments of total daily solar radiation, total daily diffuse radiation, ambient air temperature, and the water temperature in storage tank at the beginning of the day. The network output is the useful energy extracted from the system and the water temperature rise. The statistical coefficient of multiple determination (R²-value) obtained for the training data set was equal to 0.9914 and 0.9808 for the two output parameters respectively. Both values are satisfactory because the closer R²-value is to unity the better is the mapping. Unknown data for all four systems were subsequently used to investigate the accuracy of prediction. These include performance data for the systems considered for the training of the network at different weather conditions. Predictions with maximum deviations of 1 MJ and 2.2°C were obtained respectively. Random data were also used both with the performance equations obtained from the experimental measurements and with the artificial neural network to predict the above two parameters. The predicted values thus obtained were very comparable. These results indicate that the proposed method can successfully be used for the estimation of the performance of the particular thermosiphon system at any of the different types of configuration used here. The greatest advantage of the present model is the capacity of the network to learn from examples and thus gradually improve its performance. This is done by embedding experimental knowledge in the network.     

Prediction of hydrogen production by magnetic field effect water electrolysis using artificial neural network predictive models

Developing an efficient water electrolysis (WE) configuration is essential for high-efficiency hydrogen evolution reaction (HER) activity. In this regard, it has been proven that adding a magnetic field (MF) to the electrolysis system greatly improves the hydrogen output rate. In this study, we developed a method based on a machine learning approach to further improve the hydrogen production (HP) system with MF effect WE. An artificial neural network (ANN) model was developed to estimate the effect of input parameters such as MF, electrode material (cathode type), electrolyte type, supplied power (onset voltage), surface area, temperature, and time on HP in different electrolyzer systems. The network was built using 104 experimental data sets from various electrolysis studies. In the study, the percentage contributions of the input parameters to the HP rate and the optimum network architecture to minimize computation time and maximize network accuracy are presented. The model architecture of 7–12–1 was obtained using the best-hidden neurons. The Levenberg-Marquardt (LM) algorithm was used to train the multi-layer feed-forward neural network. Moreover, the utilization of a range of categorical variables to improve ANN prediction accuracy is a significant novelty in this work. Results demonstrated that the output of the trained ANN model fitted well with the experimental data. The test's correlation coefficient (R) and mean squared error (MSE) were 0.973 and 0.01125, respectively, confirming its powerful predictive performance. This ANN application is the first novel viable model to perform prediction using a neural network algorithm in the electrolysis process for MF effect HP using both categorical and continuous data inputs.     

Use of artificial neural networks for transport energy demand modeling

The paper illustrates an artificial neural network (ANN) approach based on supervised neural networks for the transport energy demand forecasting using socio-economic and transport related indicators. The ANN transport energy demand model is developed. The actual forecast is obtained using a feed forward neural network, trained with back propagation algorithm. In order to investigate the influence of socio-economic indicators on the transport energy demand, the ANN is analyzed based on gross national product (GNP), population and the total annual average veh-km along with historical energy data available from 1970 to 2001. Comparing model predictions with energy data in testing period performs the model validation. The projections are made with two scenarios. It is obtained that the ANN reflects the fluctuation in historical data for both dependent and independent variables. The results obtained bear out the suitability of the adopted methodology for the transport energy-forecasting problem.     

Non-iterative estimation of heat transfer coefficients using artificial neural network models

The Inverse Heat Conduction Problem (IHCP) dealing with the estimation of the heat transfer coefficient for a solid /fluid assembly from the knowledge of inside temperature was accomplished using an artificial neural network (ANN). Two cases were considered: (a) a cube with constant thermophysical properties and (b) a semi-infinite plate with temperature dependent thermal conductivity resulting in linear and nonlinear problem, respectively. The Direct Heat Conduction Problems (DHCP) of transient heat conduction in a cube and in a semi-infinite plate with a convective boundary condition were solved. The dimensionless temperature-time history at a known location was then correlated with the corresponding dimensionless heat transfer coefficient/Biot number using appropriate ANN models. Two different models were developed for each case i.e. for a cube and a semi-infinite plate. In the first one, the ANN model was trained to predict Biot number from the slope of the dimensionless temperature ratio versus Fourier number. In the second, an ANN model was developed to predict the dimensionless heat transfer coefficient from non-dimensional temperature. In addition, the training data sets were transformed using a trigonometric function to improve the prediction performance of the ANN model. The developed models may offer significant advantages when dealing with repetitive estimation of heat transfer coefficient. The proposed approach was tested for transient experiments. A ‘parameter estimation’ approach was used to obtain Biot number from experimental data.     

Artificial neural network based optimization for hydrogen purification performance of pressure swing adsorption

A pressure swing adsorption (PSA) cycle model is implemented on Aspen Adsorption platform and is applied for simulating the PSA procedures of ternary-component gas mixture with molar fraction of H₂/CO₂/CO = 0.68/0.27/0.05 on Cu-BTC adsorbent bed. The simulation results of breakthrough curves and PSA cycle performance fit well with the experimental data from literature. The effects of adsorption pressure, product flow rate and adsorption time on the PSA system performance are further studied. Increasing adsorption pressure increases hydrogen purity and decreases hydrogen recovery, while prolonging adsorption time and reducing product flow rate raise hydrogen recovery and lower hydrogen purity. Then an artificial neural network (ANN) model is built for predicting PSA system performance and further optimizing the operation parameters of the PSA cycle. The performance data obtained from the Aspen model is used to train ANN model. The trained ANN model has good capability to predict the hydrogen purification performance of PSA cycle with reasonable accuracy and considerable speed. Based on the ANN model, an optimization is realized for finding optimal parameters of PSA cycle. This research shows that it is feasible to find optimal operation parameters of PSA cycle by the optimization algorithm based on the ANN model which was trained on the data produced from Aspen model.