Energy Prediction Model Research Articles

Development of a Transient Energy Prediction Model for Machine Tools

Machine tools are one of the vital equipment in the manufacturing sector and consume a significant amount of energy during the steady state and transient state. An extensive amount of research work had conducted to model energy consumption in a steady state of machine tools. The researchers generally neglect the transient state because the profile of transient state energy consumption is stochastic, and time duration is concise, which results in a significant error during the prediction of total energy consumption of machine tools. This work presents an experimental study to acquire the transient state energy consumption of a machine tool. The Matlab software is used to process the captured data, and an empirical model is developed for the prediction of energy consumed by a machine tool in a transient state. The coefficient of determination is determined to evaluate the fitness of the empirical model. The results indicate that the model can predict the transient state energy consumption accurately. This approach can be conveniently applied to predict the total energy consumption of machine tools accurately. The accurate energy consumption model of machine tools is the foundation for selection of optimal process parameters of a machining process that leads to sustainable manufacturing

Construction and verification of an environment and energy prediction model for Controlled Environment Housing

In order to satisfy the healthy growth and performance of livestock and poultry, a large amount of energy is often consumed in the production process of modern livestock and poultry breeding, which is used for automatic feeding management and regulation of breeding environment. In this study, based on the ISO 13790 5R1C equivalent resistance capacitance network computing model, using the henhouse heat balance, water balance, balance of gas (ammonia, carbon dioxide concentration) principle, we developed a simple closed loop control henhouse environment and energy consumption prediction model, Through on-site validation tests, the results show that the model is reliable, and it can assist the farmer for breeding planning, house-design process and increase environmental control and energy management efficiency.

Hybrid virtual metering points – a low-cost, near real-time energy and resource flow monitoring approach for production machines without PLC data connection

Transparent energy flows within a factory are the prerequisite for energetic improvements of the involved production machines. With the ongoing digitalization of industrial production, innovative ways of creating energy transparency on the shop floor are emerging. Virtual energy metering points predict the power consumption of a regarded entity and can therefore enable a cost-effective increase in energy transparency on machine level. However, many machines, especially in small and medium-sized enterprises (SMEs), have no external data connection, which prevents the use of data-based energy prediction models. In this paper, a near real-time deployable approach to predict the current energy consumption of production machines without a programmable logic controller (PLC) data connection is presented. By using a Raspberry Pi as low-cost edge analytics device, its integrated camera films the optical signals from light-emitting diodes (LEDs) of different PLC modules, which display the switching state signals of various machine sub-units. In a next step, the filmed PLC information is translated into state signals, which are correlated with temporarily measured electric energy data of the production machine as well as its principal sub-units. After an automated model training and hyperparameter optimization process, the empirical black box model is deployed in a near real-time environment on the Raspberry Pi. Thus, a hybrid virtual energy and resource flow metering point of the production machine as well as its sub-units is generated. In addition, challenges like model training for predicting different production processes as well as the necessary data set size for VMP model generation are addressed. The approach is tested and validated for a metal cutting machine tool and a cleaning machine of the ETA Research Factory at the Technical University of Darmstadt.

Short-term Photovoltaic Power Prediction based on Sparse Representation Method

In the research of solar power prediction, providing accurate prediction data in real time is one of the most effective means to enhance the capacity of wind power acceptance and improve the power reliability and economy. The existing prediction models based on statistical methods are often unavoidable in data preprocessing and model training stage, and their adaptive ability needs to be improved. Considering that the sparse coding method does not require model training, and has the characteristics of high solving efficiency and strong self-adaptability, an online solar energy prediction model using sparse coding is proposed.Firstly, the historical time series data is composed of input-output pairs with delay, and the dictionary is respectively constructed in atomic form. Then, the sparse weight is calculated for the delay input data vector to be predicted, and the corresponding predicted output is obtained by borrowing the dictionary. Taking the actual solar power data of Alberta, Canada as sample, the simulation was carried out in MATLAB. The simulation results show that the model can accurately predict the solar power and improve the effectiveness and practicability of the prediction

Development of Energy Prediction Model for Residential Buildings through Statistical Analysis

Development of Energy Prediction Model for Residential Buildings through Statistical Analysis

Predicting electricity demand profiles of new supermarkets using machine learning

Predicting the electricity consumption of proposed new supermarkets is helpful to design and plan future energy management. Instead of creating complex site-specific thermal engineering models, data-driven energy prediction models can be useful to energy managers. We have designed and implemented a data-driven method to predict the future ’electricity daily load profile’ (EDLP) of new supermarkets using historical EDLPs of existing supermarkets of the same type. The supermarket features used for the prediction are 10 types of floor areas divided by usage (m2) and its location. Four data-driven regression models are used and compared to predict EDLPs: Artificial Neural Networks, Support Vector Machines, k-Nearest Neighbours and OLS. Prediction computational experiments were performed over 1-h electricity readings of 213 UK supermarkets gathered during six years. Prediction error mainly varies between 12 and 20% depending on method, year, supermarket type, and division of the data (season or temperature intervals). EDLPs computed over warm periods are better predicted than over cold periods and supermarkets only with electricity are better predicted than supermarkets with electricity and gas. The three features with more weight in the prediction are Food, Chilled produce and Cafeteria areas. The limitations of machine learning methods to solve this problem are discussed.

Minimum Ignition Energy (MIE) prediction models for ignition sensitive fuels using machine learning methods

Quantitative Structure-Property Relationship (QSPR) regression models for Minimum Ignition Energy (MIE) prediction of 60 flammable compounds have been developed using machine learning algorithms Random Forests (RF), Artificial Neural Networks (ANN) and a non-machine learning algorithm - Genetic Function Approximation (GFA).RF algorithm was implemented for feature selection to identify the 13 molecular descriptors having maximum effect on MIE prediction accuracy (i.e. descriptors affecting MIE prediction accuracy >1%). Thereafter, these descriptors were used as input parameters to develop the RF, ANN and GFA models. The optimized RF algorithm resulted in test set R2 of 0.85 and displayed high internal robustness and external predictability. The ANN and GFA algorithms displayed improved performance only on addition of an additional parameter - Structure Parameter to the existing 13 descriptor set. The optimized ANN and the GFA models displayed a test set R2 of 0.79 and 0.71, respectively. The ANN model resulted in lower machine learning bias as compared to the RF model. The ANN model was observed to be more reliable than RF model for MIE prediction of small datasets. Based on this work, the RF algorithm for feature selection in QSPR modeling and the ANN, RF algorithms were observed to be promising options for MIE prediction applications.

Influence of driving style, infrastructure, weather and traffic on electric vehicle performance

Internal and external circumstances affect a vehicles’ fuel consumption. Various studies reviewed internal combustion engine vehicles (ICEVs) energy consumption. While Battery Electric Vehicles (BEVs) are expected to replace ICEVs, more research is necessary to understand their energy performance. To fill this research gap, a microscopic traffic model and an energy prediction model were combined to estimate the effects of driving styles, weather, traffic and infrastructure on the energy consumption of BEVs. By using a VISSIM model, various scenarios were tested. This resulted in a qualitative insight into the energy consumption and travel time, which was validated by performing 30 driving tests and by using dynamometer data. The results indicate that built environment variables have a large effect on the energy consumption of BEVs. An increasing traffic intensity has been found to not always increase the energy consumption. Moreover, the energy consumption can be decreased by choosing a more energy-efficient route.

Predicting the Energetics of Hydrogen Intercalation in Metal Oxides Using Acid-Base Properties.

The ability to predict intercalation energetics from first principles is attractive for identifying candidate materials for energy storage, chemical sensing, and catalysis. In this work, we introduce a computational framework that can be used to predict the thermodynamics of hydrogen intercalation in tungsten trioxide (WO3). Specifically, using density functional theory (DFT), we investigated intercalation energetics as a function of adsorption site and hydrogen stoichiometry. Site-specific acid-base properties determined using DFT were used to develop linear structure screening models that informed a kernel ridge energy prediction model. These regressions provided a series of hydrogen binding energy predictions across stoichiometries ranging from WO3 to H0.625WO3, which were then converted to equilibrium potentials for hydrogen intercalation. Experimental validation using cyclic voltammetry measurements yielded good agreement with the predicted intercalation potentials. This methodology enables fast exploration of a large geometric configuration space and reveals an intuitive physical relationship between acidity, basicity, and the thermodynamics of hydrogen intercalation. Furthermore, the combination of theoretical and experimental results suggests H0.500WO3 as a maximum stable stoichiometry for the bronzes that arises from competition with hydrogen evolution rather than the inability of WO3 to accommodate additional hydrogen. Our experimental results further indicate hydrogen insertion in WO3 is highly irreversible for low H-stoichiometries, which we propose to be a consequence of the semiconductor-to-metal transition that occurs upon initial H-intercalation. Overall, the agreement between theory and experiment suggests that local acid-base characteristics govern hydrogen intercalation in tungsten trioxide, and this insight can aid the accelerated discovery of redox-active metal oxides for catalytic hydrogenations.

Development of an ANN-based building energy model for information-poor buildings using transfer learning

Accurate building energy prediction is vital to develop optimal control strategies to enhance building energy efficiency and energy flexibility. In recent years, the data-driven approach based on machine learning algorithms has been widely adopted for building energy prediction due to the availability of massive data in building automation systems (BASs), which automatically collect and store real-time building operational data. For new buildings and most existing buildings without installing advanced BASs, there is a lack of sufficient data to train data-driven predictive models. Transfer learning is a promising method to develop accurate and reliable data-driven building energy prediction models with limited training data by taking advantage of the rich data/knowledge obtained from other buildings. Few studies focused on the influences of source building datasets, pre-training data volume, and training data volume on the performance of the transfer learning method. The present study aims to develop a transfer learning-based ANN model for one-hour ahead building energy prediction to fill this research gap. Around 400 non-residential buildings’ data from the open-source Building Genome Project are used to test the proposed method. Extensive analysis demonstrates that transfer learning can effectively improve the accuracy of BPNN-based building energy models for information-poor buildings with very limited training data. The most influential building features which influence the effectiveness of transfer learning are found to be building usage and industry. The research outcomes can provide guidance for implementation of transfer learning, especially in selecting appropriate source buildings and datasets for developing accurate building energy prediction models.

Domestic sector energy demand and prediction models for Punjab Pakistan

The domestic sector consumes ~48% of Pakistan's total energy demand, including biofuels. Pakistan is an emerging economy with 210 million people and growing domestic energy demand, facing economic, geographic, geopolitical, and climate change challenges. This paper presents novel insights into the Punjab, Pakistan domestic sector energy demand, which accounts for over 52% of the Pakistan population, along with energy prediction models derived from a statistically significant 4597 responses obtained from a physical questionnaire survey conducted in 2017-18, which aimed at ascertaining the main domestic energy demand drivers. These models will support future government and energy industry policy in this area, especially the transition to a low carbon economy. Currently, 67% of Pakistan's energy demand is met with non-renewable resources. Analysis of the survey data reveals the key drivers of electrical energy demand per household are the number of appliances, number of lights, and the number of conditioned rooms. In the per capita models, the key drivers are the overall power rating of the appliances, particularly the power rating of the air conditioners for cooling. For annual gas use, weak correlations per household and capita were found only for the floor area. The average annual electricity and gas usages per household are 2401 kWh/a and 5245 kWh/a respectively, and per capita are 391 kWh/a and 770 kWh/a. For electricity, the occupancy, floor area, conditioned rooms, appliances, lights and power rating have predictive power. For gas, only floor area is predictive.

An Approach for Data Treatment of Solar Photovoltaic Generation

A good solar power photovoltaic generation forecast depends on good quality time series data from measurements of Global Horizontal Irradiance and Solar Power Generation. However, measurement system failures and errors in data handling can corrupt data records with gaps and outliers that undermine forecasting accuracy. Therefore, it is important that the fitting of solar energy prediction models must be preceded by a data analysis in order to detect and correct measurement errors. Given that Global Horizontal Irradiance and Solar Power Generation are correlated variables, this paper aims to present the main characteristics of an offline approach developed for the joint treatment of hourly data values of both variables in a photovoltaic plant. In the proposed methodology, the measurements of Global Horizontal Irradiance and Solar Power Generation are analyzed by using reanalysis data and statistical and data mining techniques for the correction of outliers and the filling of data gaps. The application of the approach is illustrated by the analysis of measurements from a real solar PV system.

Solar Energy Prediction Model Based on Artificial Neural Networks and Open Data

With climate change driving an increasingly stronger influence over governments and municipalities, sustainable development, and renewable energy are gaining traction across the globe. This is reflected within the EU 2030 agenda, that envisions a future where there is universal access to affordable, reliable and sustainable energy. One of the challenges to achieve this vision lies on the low reliability of certain renewable sources. While both particulars and public entities try to reach self-sufficiency through sustainable energy generation, it is unclear how much investment is needed to mitigate the unreliability introduced by natural factors such as varying wind speed and daylight across the year. In this sense, a tool that aids predicting the energy output of sustainable sources across the year for a particular location can aid greatly in making sustainable energy investments more efficient. In this paper, we make use of Open Data sources, Internet of Things (IoT) sensors and installations distributed across Europe to create such tool through the application of Artificial Neural Networks. We analyze how the different factors affect the prediction of energy production and how Open Data can be used to predict the expected output of sustainable sources. As a result, we facilitate users the necessary information to decide how much they wish to invest according to the desired energy output for their particular location. Compared to state-of-the-art proposals, our solution provides an abstraction layer focused on energy production, rather that radiation data, and can be trained and tailored for different locations using Open Data. Finally, our tests show that our proposal improves the accuracy of the forecasting, obtaining a lower mean squared error (MSE) of 0.040 compared to an MSE 0.055 from other proposals in the literature.

Prediction of the Energy Consumption of School Buildings

The energy consumption of a constructed facility is a primary concern as a result of its impact on the total energy expenditure. It has been found that up to 70% of the power consumption in Saudi Arabia are caused by building structures and air conditioning (AC). Energy consumption in government-constructed buildings constitutes a considerable ≈13% of the consumption of the total energy in Saudi Arabia. Therefore, the government of Saudi Arabia initiated the Saudi Energy Efficiency Program (SEEP) that goals to lower the domestic energy severity by roughly 30% by 2030. This paper introduces a study carried out in Eastern Province in Saudi Arabia to identify factors influencing the consumption of energy in school facilities (which are built of concrete in hot and humid climate zones), investigate the correlation between those factors and their impacts on the consumption of energy in school facilities, and finally, develop a prediction model for the energy consumption of school facilities. The study was based on the utilization of 352 real-world datasets of energy consumption of operating schools across Eastern Province in Saudi Arabia. The developed energy prediction model considers eleven identified factors that influence the consumption of energy of constructed schools. The identified factors were utilized as input variables to build the model. A systematic search among different neural network (NN) design architectures was conducted to identify the optimal network model. Validation of the developed model on eight real-world cases demonstrated that the accuracy of the developed model was about 87.5%. Moreover, the findings of this study indicate that the weakest correlation between the input variables was recorded as −0.015 between “type of school” and “AC capacity,” while the strongest correlation was recorded as 0.95 between the variables of “number of classrooms” and “total air-conditioned area (sqm),” followed by “total air-conditioned area (sqm)” and “number of students,” which was recorded as 0.90. It is worth noting that “AC capacity” was the most significant predictor, which increased exponentially for high values of energy consumption, followed by “total school roof area.” The study also found that the age of the schools had a very small impact on energy consumption, although the age of the schools varied from 11 to 51 years. This was probably due to a good maintenance system applied by the Ministry of Education. The implication of the developed prediction model was that the model can be used by the Ministry of Education to predict the energy consumption and its associated cost for public school buildings for the purpose of budget allocation. The model may be utilized as a stand-alone application, or it can be integrated with an existing building information module (BIM)-based system.

Big data driven predictive production planning for energy-intensive manufacturing industries

Improving energy and resource efficiency in manufacturing is an important goal for enterprises to sustain their competitive advantages. Predictive production planning is a new solution to achieve such goal, following the direct improvement of energy efficiency and indirect energy savings through better scheduling. With the emergence of new information and communication technologies under the background of Industry 4.0, the amount of various energy and resource data obtained through Internet of Things is reaching the magnitude of big data. It poses a challenge to traditional data processing and mining methods for predictive production. To solve this challenge, in this paper, the big data driven predictive production planning is proposed to improve energy and resource efficiency for energy-intensive manufacturing industries. Additionally, the cube-based energy consumption models and long short-term memory based energy prediction model are established for data preprocessing and mining correspondingly. An industrial case study is presented to process energy big data and predict energy consumption parameters and production status. The performance evaluation results indicate that the proposed long short-term memory models outperform back propagation neural network, autoregressive moving average and support vector regression. Based on data preprocessing and forecasting results, the energy and resource efficiency could be improved during the whole manufacturing process for energy-intensive manufacturing industries.

A novel grey Riccati–Bernoulli model and its application for the clean energy consumption prediction

Accurately forecasting energy demand can better predict future changes in energy demand. This study is aimed to develop more accurate clean energy prediction model from two perspectives of the modeling mechanisms and improvements on the model structure. For this purpose, a novel Riccati–Bernoulli differential equation is established through social practice theory, supply–demand relationship, and the preference of consumption in the energy economy. Considering the limited information in observation data, this differential equation is then transformed into a grey Riccati–Bernoulli model (GRBM(1,1)) according to the differential information principle. With the response function solved on the basis of polynomial equation theory and the power exponent optimized combining the modified flower pollination algorithm, the main process of GRBM(1,1) can be summarized. The four validation examples are provided for confirming the effectiveness and reliability of the new model by comparing with other existing models. Finally, the proposed model is employed to estimate and forecast the clean energy consumption in China and India. The results show that the proposed model demonstrates better estimation in all cases and efficiency in short-term clean energy consumption forecasting. Therefore, by using this optimum model, China’s preference coefficient in total energy consumption is 0.7103, lower than that of India (0.8799), and China’s preference coefficient in clean energy consumption is 0.9665, higher than that of Indian (0.7155), which means China has less interest to increase total energy consumption but more interest in popularizing the clean energy.

Production capacity prediction of hydropower industries for energy optimization: Evidence based on novel extreme learning machine integrating Monte Carlo

As the basic industry of national economic development, the electric power industry is closely associated with the overall economic development and social progress. Hydropower, as a power generation form, has fewer data samples than thermal power generation, which leads to a difficult problem to establish an accurate production and energy prediction model. Therefore, a novel production capacity prediction model using extreme learning machine based on Monte Carlo algorithm is presented for energy optimization and saving. Through using the Monte Carlo algorithm, the small sample data can be expanded. Then, the expansion of the small sample data is utilized as the training set and the testing set for the extreme learning machine to predict the production capacity and optimize the energy configuration. Finally, the proposed method is used to predict the production of a hydropower plant for improving the energy efficiency. Compared with the traditional extreme learning machine, the correctness and the applicability of the proposed method are proved. Moreover, the energy optimal configuration of the hydropower industry production can improve the energy efficiency and save the energy of hydropower industrial processes.

Deep learning‐based household electric energy consumption forecasting

With the advent of various electronic products, the household electric energy consumption is continuously increasing, and therefore it becomes very important to predict the household electric energy consumption accurately. Energy prediction models also have been developed for decades with advanced machine learning technologies. Meanwhile, the deep learning models are still actively under study, and many newer models show the state-of-the-art performance. Therefore, it would be meaningful to conduct the same experiment with these new models. Here, the authors predict the household electric energy consumption using deep learning models, known to be suitable for dealing with time-series data. Specifically, vanilla long short-term memory (LSTM), sequence to sequence, and sequence to sequence with attention mechanism are used to predict the electric energy consumption in the household. As a result, the vanilla LSTM shows the best performance on the root-mean-square error metric. However, from a graphical point of view, it seems that the sequence-to-sequence model predicts the energy consumption patterns best and the vanilla LSTM does not follow the pattern well. Also, to achieve the best performance of each deep learning model, vanilla LSTM, sequence to sequence, and sequence to sequence with attention mechanism should observe past 72, 72, and 24 h, respectively.

Identification of Automotive Cabin Design Parameters to Increase Electric Vehicles Range, Coupling CFD-Thermal Analyses with Design for Six Sigma Approach

<div class="section abstract"><div class="htmlview paragraph">The ongoing global demand for greater energy efficiency plays an essential role in vehicle development, especially in the case of electric vehicles (EVs). The thermal management of the full vehicle is becoming increasingly important, since the Heating, Ventilation, and Air Conditioning (HVAC) system has a significant impact on the EV range. Therefore the EV design requires new guidelines for thermal management optimization.</div><div class="htmlview paragraph">In this paper, an advanced method is proposed to identify the most influential cabin design factors which affect the cabin thermal behavior during a cool down drive cycle in hot environmental conditions. These parameters could be optimized to reduce the energy consumption and to increase the robustness of the vehicle thermal response.</div><div class="htmlview paragraph">The structured Taguchi’s <i>Design for Six Sigma</i> (DFSS) approach was coupled with CFD-Thermal FE simulations, thanks to increased availability of HPC. The first control factors selected were related to the thermal capacity of the panel duct, dashboard, interior door panels and seats. Surface IR emissivity and solar radiation absorptivity of these components were then added to the study. Car glass with absorptive and reflective glazing were finally included in the study. The design space of 18 vehicle configurations was simulated in spring and hot summer conditions, with steady state thermal simulations. A 2-step optimization was then conducted, trying first to increase the robustness of the cabin response and, secondly, to reduce the equivalent temperature actually felt by passengers. The <i>Verify</i> phase was then conducted on the <i>Best Engineering</i> design emerged by the 2-step optimization, through quasi-transient CFD-Thermal FE analyses. The thermal results were then sent to a CFD 1D energy prediction model, confirming the HV battery energy saving and the extended range reached during the cool down drive cycle.</div></div>

A generic energy prediction model of machine tools using deep learning algorithms

Energy prediction of machine tools plays an irreplaceable role in energy planning, management, and conservation in the manufacturing industry. In the era of big machinery data, data-driven energy prediction models of machine tools have achieved remarkable results in the identification of energy consumption patterns and prediction of energy consumption conditions. However, existing data-driven studies towards the energy consumption of machine tools focus on the utilization of handcrafted-feature learning methods, which are inefficient and exhibit poor generalization. Moreover, considering variations in energy consumption characteristics among different machine tools, it is impractical to identify energy consumption features manually for energy model development. Therefore, this paper proposes a novel data-driven energy prediction approach using deep learning algorithms. Here, deep learning is used in an unsupervised manner to extract sensitive energy consumption features from raw machinery data, and in a supervised manner to develop the prediction model between the extracted features and the energy consumption of machine tools. The experiments conducted on a milling machine and a grinding machine are exploited and compared with those conducted in conventional studies. The results show that the proposed method can improve the energy prediction performance from 19.14% to 74.13% for the grinding machine and from 64.89% to 85.61% for the milling machine, and it achieves a better performance than the conventional methods in terms of effectiveness and generalization.