Heat Load Prediction Research Articles

Comparison of algorithms for heat load prediction of buildings

Achieving precision in the prediction of buildings' dynamic heat load is crucial for the advancement of smart heating systems. This research highlights the urgent need to enhance the accuracy of models predicting dynamic heat load. Through literature review, distinguished machine learning and regression algorithms were chosen to formulate prediction models. These models employ a data time-step adaptive strategy, a physics-guided loss function, and fundamental principles of heat transfer. Optimization algorithms of a mathematical nature were utilized to fine-tune the parameters and the framework of long short-term memory (LSTM) and multi-layer perceptron (MLP) models. An analytical comparison was undertaken between physics-guided models and those not guided by physics. Principal conclusions are: 1) Pelican optimization algorithm (POA)-LSTM model emerges as superior in heat load prediction accuracy of an office building, with percentage errors for actual and simulated datasets ranging from −6.7 % to 5.8 % and −5.2 %–4.5 %, respectively, and the mean absolute percentage error (MAPE) standing at 2.3 % and 1.3 %. 2) The linear regression model exhibits the lowest precision, with a MAPE of 17.5 % and 4.0 % for the 7-day prediction results in the actual and simulated datasets, respectively. These findings provide support for improving heat load prediction in heating systems.

Heat Load Prediction of District Heating Systems Based on SCSO-TCN

Heat Load Prediction of District Heating Systems Based on SCSO-TCN

Optimising building heat load prediction using advanced control strategies and Artificial Intelligence for HVAC system

Amid the dynamic challenges posed by the COVID-19 era, this study offers a nuanced exploration, delving into the complexities of optimising building heat load prediction. This study addresses the imperative challenge of optimising building heat load prediction by implementing advanced control strategies and integrating Artificial Intelligence (AI) into air conditioning systems. The study emphasises the need to design HVAC devices for minimal energy consumption without compromising comfort conditions. The study introduces an intelligent predictive adaptive neuro-fuzzy inference system (ANFIS) model that proves highly capable of accurately predicting HVAC performance outcomes while contributing to the development of a comprehensive dataset. The uncertainty percentage is evaluated across various membership functions, with the trapezoidal membership exhibiting the lowest error rate, followed by the Gaussian membership function. Weather parameters crucial to ventilation efficiency and heat load are examined, leading to the identification of an optimal combination involving 45 % relative humidity, 13 °C dry bulb temperature, and 5 km/h wind speed. The current system demonstrates significant efficiency improvements, with ventilation and heat load rates reaching 93 % and 97 %, respectively, compared to pre-COVID-19 conditions. The findings underscore the importance of considering these parameters in future HVAC designs, particularly in the context of COVID-19 guidelines (75 % and 79 %, respectively).

Limited data-oriented building heating load prediction method: A novel meta learning-based framework

Data-driven models have been widely used in building heating load prediction, but often fail when facing limited data. Previous studies have shown transfer learning can assist model learning of target building under limited data by means of other source building data, however, which is subject to the similarity between source and target building. Selecting similar source building data is not easy, especially when the target building is with limited data. This paper, therefore, proposes a novel meta learning-based framework for building heating load prediction. Using meta learning method, a set of promising model parameters is trained by local and global learning on multiple source buildings data. The obtained model parameters has the ability to get quickly trained with few data in each source building, which is further used as model initialization parameters of target building to assist model learning. Framework validity is confirmed by 550 groups of practical buildings data (50 are as target buildings for testing and 500 are as source buildings). The results showed the proposed framework could reduce the prediction errors by 2.04 %∼61.59 % compared with six common transfer learning methods. The novel meta learning-based framework provides an effective solution for building heating load prediction with limited data.

Short term load prediction of regional heating and heat storage system based on neural network

Accurate heat load prediction is the key to achieve fine control, energy conservation, and carbon reduction of regional hydronics. Taking the regional hydronics of a city in the north of China as the research object, the author, respectively uses back propagation neural network (BPNN), genetic algorithm (GA) optimized BPNN (GA-BPNN), and autoregressive integrated moving average model (ARI?MA) combined BPNN (ARIMA BPNN) to predict its heat load, and compares the accuracy and applicability of each prediction method. The results indicate that GA-BPNN has the smallest prediction error, followed by ARIMA-BPNN, but the latter requires less data for prediction. In practical engineering, if there is a sufficient amount of data related to heat load, it is recommended to use GA-BPNN. If there is a small amount of data, ARIMA-BP prediction method can be used.

Experimental evidence of very short power decay lengths in H-mode discharges in the COMPASS tokamak

Analysis of heat fluxes in scrape-off layer (SOL) plasma is important for the prediction of divertor tile heat loads in future reactor-sized tokamak machines. Typically, the radial heat flux profile can be accurately characterized by the decay length parameter λ q . The predictions are then based on the dependence of λ q on plasma and machine parameters. In recent years, several empirical scaling models were derived from a collective database of decay lengths observed in several large tokamaks as well as two spherical tokamaks. Most recently, a report from the TCV tokamak showed a deviation from several of the proposed models (Maurizio et al (TCV team) 2021 Nucl. Fusion 61 024003). In this work, edge plasma profiles of ELMy H-mode discharges were collected from the COMPASS tokamak database for heat flux analysis. The COMPASS tokamak has a similar machine size and plasma parameters to the TCV tokamak, offering a valuable comparison. The λ q was measured in both downstream and upstream SOL utilizing a divertor probe array validated by an IR camera and Thomson scattering diagnostics. A comparison with the predictions of several scaling models indicated an occurrence of anomalously short (by a factor of 2–5) inter-ELM power decay lengths in both upstream and downstream, as well as a difference in λ q between these two regions. Possible causes related to the accuracy of magnetic reconstruction were eliminated and physics-based sources of apparent compression effects were hypothesized as a motivation for future research.

Digital twin modeling for district heating network based on hydraulic resistance identification and heat load prediction

As a vital infrastructure in modern cities, district heating (DH) systems provide stable high-quality heat sources. To satisfy the demands of energy-saving, emission-reduction policies, and user preferences, a high-precision digital simulation platform is essential for district heating network (DHN) scheduling. In this paper, a method for digital twin (DT) modeling based on hydraulic resistance identification using an online adaptive particle swarm optimization (APSO) algorithm is proposed. Additionally, Singular Spectrum Analysis and back propagation artificial neural network algorithms (SSA-BP) are proposed to achieve heat load prediction in order to provide sufficient conditions for DHN scheduling. The load prediction for the substation is completed by predicting the secondary return water temperature at the subsequent time step based on the historical outdoor temperature time series and secondary supply water temperature of the substation. Accordingly, simulation analysis has been conducted on its application to an existing DHN in Tianjin. With the proposed methods, the simulation errors are reduced for about 80 % substations in the case DHN compared to mechanistic models when simulating hydraulic conditions. Among these, 77.49 % of substations were reduced by 0 %–3 %, and 1.43 % of substations were reduced by 3 %–5 %. The prediction errors of secondary return water temperature for case substations A and B exhibit random variation trends, with maximum absolute errors within the range of 0.02 °C and 0.06 °C, respectively.

Prediction of Heat and Cold Loads of Factory Mushroom Houses Based on EWT Decomposition

Load forecasting has significant implications on optimizing the operation of air conditioning systems for industrial mushroom houses and energy saving. This research paper presents a novel approach for short-term load forecasting in mushroom houses, which face challenges in accurately modeling cold and heat loads due to the complex interplay of various factors, including climatic conditions, mushroom growth, and equipment operation. The proposed method combines empirical wavelet transform (EWT), hybrid autoregressive integrated moving average (ARIMA), convolutional neural network (CNN), and bi-directional long short-term memory (BiLSTM) with an attention mechanism (CNN-BiLSTM-Attention) to address these challenges. The first step of this method was to select input features via the Boruta algorithm. Then, the EWT method was used to decompose the load data of mushroom houses into four modal components. Subsequently, the Lempel–Ziv method was introduced to classify the modal components into high-frequency and low-frequency classes. CNN-BiLSTM-Attention and ARIMA prediction models were constructed for these two classes, respectively. Finally, the predictions from both classes were combined and reconstructed to obtain the final load forecasting value. The experimental results show that the Boruta algorithm selects key influential feature factors effectively. Compared to the Spearman and Pearson correlation coefficient methods, the mean absolute error (MAE) of the prediction results is reduced by 14.72% and 3.75%, respectively. Compared to the ensemble empirical mode decomposition (EEMD) method, the EWT method can reduce the decomposition reconstruction error by an order of magnitude of 103, effectively improving the accuracy of the prediction model. The proposed model in this paper exhibits significant advantages in prediction performance compared to the single neural network model, with the MAE, root mean square error (RMSE), and mean absolute percentage error (MAPE) of the prediction results reduced by 31.06%, 26.52%, and 39.27%, respectively.

Performance of tungsten plasma facing components in the stellarator experiment W7-X: Recent results from the first OP2 campaign

The transition to reactor-relevant materials for the plasma facing components (PFCs) is an important and necessary step to provide a proof of principle that the stellarator concept can meet the requirements of a future fusion reactor by demonstrating high performance steady-state operation. As a first step to gain experience with tungsten as plasma-facing material in the Wendelstein 7-X (W7-X) stellarator, graphite tiles coated with an approximately 10 µm MedC tungsten layer (NILPRP Bucharest) were installed to complete the ECRH beam dump area in two of the five plasma vessel modules over an area of approximately one square meter each. In addition, tungsten baffle tiles are installed (40 tiles in total) with either bulk tungsten as part of NBI shine-through target or with a tungsten heavy alloy (W95-Ni3.5-Cu1.5) to replace the graphite tiles that were previously thermally overloaded. Based on an advanced diffusive field line tracing method and EMC3-Eirene simulations, the overloaded baffle tiles were redesigned by making the tiles thinner (i.e. moving the plasma-facing surface (PFS) away from the hot plasma region) and by reducing the local angle of incidence through toroidal displacement of the watershed. Significant erosion of the tungsten tiles can only be expected if sputtering by impurity ions such as carbon or oxygen ions contributes. However, the resulting central concentration of tungsten and the corresponding radiation losses are expected to be marginal. The expected deposition of carbon on the tungsten surfaces in the baffle regions mitigates further the possible tungsten enrichment in the core plasma. In OP2.1, no adverse effects of the installed tungsten PFCs on the plasma performance were observed during normal plasma operation. With the design changes made in the baffle area, the predicted heat load reductions could be experimentally confirmed.

Hourly Heat Load Prediction for Residential Buildings Based on Multiple Combination Models: A Comparative Study

The accurate prediction of residential heat load is crucial for effective heating system design, energy management, and cost optimization. In order to further improve the prediction accuracy of the model, this study introduced principal component analysis (PCA), the minimum sum of squares of the combined prediction errors (minSSE), genetic algorithm (GA), and firefly algorithm (FA) into back propagation (BP) and ELMAN neural networks, and established three kinds of combined prediction models. The proposed methodologies are evaluated using real-world data collected from residential buildings over a period of one year. The obtained results of the PCA-BP-ELMAN, FA-ELMAN, and GA-BP models are compared with the neural network before optimization. The experimental results show that the combined prediction models have higher prediction accuracy. The Mean Absolute Percentage Error (MAPE) evaluation indices of the three combined models are distributed between 5.95% and 7.05%. The FA-ELMAN model is the combination model with the highest prediction accuracy, and its MAPE is 5.95%, which is 2.25% lower than the MAPE of an individual neural network. This research contributes to the field by providing a comprehensive and effective framework for residential heat load prediction, which can be valuable for building energy management and optimization.

Temporal graph attention network for building thermal load prediction

Machine learning models have seen widespread application in predicting building thermal loads. Yet, these existing models generally predict the thermal load for a single thermal zone or an entire building, overlooking buildings with multiple thermal zones. Moreover, interactions between thermal zones are seldom accounted for these methods. Addressing this gap, our study introduces a graph-based approach that employs a Temporal Graph Attention Network (TGAT) model for building thermal load prediction. This model accommodates both the spatial and temporal dependence across different thermal zones. Specifically, the TGAT model initially utilizes graph attention neural networks (GATs) to identify the spatial relationships between thermal zones. The spatial representation thus obtained is subsequently processed via Gated Recurrent Units (GRUs) to ascertain temporal dependencies. Lastly, a fully connected layer is engaged to decode the representation for the final output. Our model incorporates 11 features as inputs to depict the characteristics of outdoor environment and thermal zones. Testing the proposed TGAT model in a small office virtual environment for heating load prediction revealed that a configuration of four GAT attention heads achieved optimal result for predicting the ensuing hour’s heating load with a six-hour lookback. Additionally, we scrutinized and discussed the efficacy of feature selection and the GAT module. This research thus establishes a novel groundwork for forthcoming studies on zone-based building energy management optimization.

Heat Load Prediction for District Heating Systems with Temporal Convolutional Network and CatBoost

Heat Load Prediction for District Heating Systems with Temporal Convolutional Network and CatBoost

Bayesian Optimization-Based LSTM for Short-Term Heating Load Forecasting

With the increase in population and the progress of industrialization, the rational use of energy in heating systems has become a research topic for many scholars. The accurate prediction of heat load in heating systems provides us with a scientific solution. Due to the complexity and difficulty of heat load forecasting in heating systems, this paper proposes a short-term heat load forecasting method based on a Bayesian algorithm-optimized long- and short-term memory network (BO-LSTM). The moving average data smoothing method is used to eliminate noise from the data. Pearson’s correlation analysis is used to determine the inputs to the model. Finally, the outdoor temperature and heat load of the previous period are selected as inputs to the model. The root mean square error (RMSE) is used as the main evaluation index, and the mean absolute error (MAE), mean bias error (MBE), and coefficient of determination (R2) are used as auxiliary evaluation indexes. It was found that the RMSE of the asynchronous length model decreased, proving the general practicability of the method. In conclusion, the proposed prediction method is simple and universal.

Daily residential heat load prediction based on a hybrid model of signal processing, econometric model, and support vector regression

Daily residential heat load prediction based on a hybrid model of signal processing, econometric model, and support vector regression

Forecasting heating and cooling loads in residential buildings using machine learning: a comparative study of techniques and influential indicators

Residential buildings are a significant source of energy consumption and greenhouse gas emissions, making it crucial to accurately predict their energy demand for reducing their environmental impact. In this study, machine-learning techniques such as linear regression, decision tree classification, logistic regression, and neural networks were applied to forecast the heating and cooling loads of 12 different building types using their area and height attributes. The correlation coefficient was utilized to assign weights to the predictors in linear regression, and the models’ performance was evaluated using metrics such as equations of R2, MAE, and RMSE. The decision tree technique demonstrated the highest accuracy of 98.96% and 93.24% for predicting heating and cooling loads, respectively, among the classification methods. Notably, the cooling load prediction was more accurate than the heating load prediction. The height and area of the roof and floor, along with the relative compactness of the building, were identified as the most influential factors in the heating and cooling loads. These findings have significant implications for optimizing energy efficiency in residential buildings and mitigating their impact on climate change.

Transfer learning-based adaptive recursive neural network for short-term non-stationary building heating load prediction

Building energy consumption is a non-stationary time series, and its distribution law changes over time. Traditional machine-learning models are prone to model shift, which leads to a reduction of their prediction accuracy when applied to building energy consumption forecasting. Therefore, in this paper, an adaptive neural network framework is proposed for non-stationary building energy consumption prediction based on transfer learning, in which the training datasets are divided into the most dissimilar periods, and based on the transfer learning mechanism, the different periods of data are learned to obtain the minimum overall loss to improve the model generalization. Taking LSTM and GRU models as examples, adaptive long short-term memory (adaLSTM) and adaptive gated recurrent unit (adaGRU) building energy consumption prediction models are established. The models were trained and verified using heating load data from Xi'an, China. The results show that compared with LSTM model, the coefficient of determination, root mean square error, the coefficient of variation of the root mean squared error and mean absolute error of the adaLSTM model were improved by 0.61%, 37.78%, 38.05% and 30.69%, respectively, and the over-fitting degree was reduced by 227.7%. Compared with the traditional GRU model, the corresponding evaluation indexes of adaGRU were improved by 2.50%, 70.58%, 70.64% and 68.83%, respectively, and the over-fitting degree was improved by 505.7% points. The adaptive recurrent neural network framework proposed in this paper is a generalized approach which can be applied to other non-stationary time series prediction models.

Optimal rule based double predictive control for the management of thermal energy in a distributed clean heating system

Solar energy coupled with electric heat storage is a kind of promising energy saving technology for distributed building heating. Precise and quick heat load prediction for the demand side of heating system and renewable energy prediction for the supply side are imperative in realizing the flexibility of building clean energy supply systems. A predictive control based on back propagation (BP) artificial neural network is built to predict the demand heat load of an office building and the heat supply quantity of solar collector (SC) system. The heat dissipation of the SC system is considered to improve the prediction accuracy of its heat supply quantity. A double predictive (DP) control is introduced and combined with dynamic adjustment to optimize the operation of renewable energy, phase change heat storage and valley electric heating system. The results indicate that predictive model is of high precision and the variation coefficient of root mean squared errors corresponding to each prediction parameter are all greater than 0.9. Furthermore, the optimization results indicate that the DP control would save, in March, about 59.3% of charge heat quantity of phase change material (PCM) tank and terms of cost (30.4%). Meanwhile, maintaining indoor temperature between 20 °C and 22 °C.

Industrial steam consumption analysis and prediction based on multi-source sensing data for sustainable energy development

Centralized heating is an energy-saving and environmentally friendly way that is strongly promoted by the state. It can improve energy utilization and reduce carbon emissions. However, Centralized heating depends on accurate heat demand forecasting. On the one hand, it is impossible to save energy if over producing, while on the other hand, it is impossible to meet the heat demand of enterprises if there is not enough capacity. Therefore, it is necessary to forecast the future trend of heat consumption, so as to provide a reliable basis for enterprises to reasonably deploy fuel stocks and boiler power. At the same time, it is also necessary to analyze and monitor the steam consumption of enterprises for abnormalities in order to monitor pipeline leakage and enterprise gas theft. Due to the nonlinear characteristics of heat load, it is difficult for traditional forecasting methods to capture data trend. Therefore, it is necessary to study the characteristics of heat loads and explore suitable heat load prediction models. In this paper, industrial steam consumption of a paper manufacturer is used as an example, and steam consumption data are periodically analyzed to study its time series characteristics; then steam consumption prediction models are established based on ARIMA model and LSTM neural network, respectively. The prediction work was carried out in minutes and hours, respectively. The experimental results show that the LSTM neural network has greater advantages in this steam consumption load prediction and can meet the needs of heat load prediction.

A control method combining load prediction and operation optimization for phase change thermal energy storage system

Due to the high latent heat and load-shifting capacity, phase-change thermal energy storage technology is an effective way to reduce energy costs under time-of-use electricity pricing. A favorable operation strategy is essential to exploit the advantage of the phase change thermal energy storage system. Previous studies on the operation strategy lack consideration of load prediction, which could reduce the matching degree of heat supply and demand. This study proposed a control method combing load prediction and operation optimization based on an electric boiler-phase change thermal energy storage heating system. A deep learning-based heating load prediction model was built; on this basis, an operation optimization method using dynamic programming was formulated subsequently. The proposed method was applied to a case building located in Beijing, China. By applying the proposed control method, the energy costs of the system were saved by 10% compared to the original operation strategy in the whole heating season, and the matching degree of heat supply and demand could be close to 100%, demonstrating that the method can provide users with cheaper and the best thermally comfortable heating. The proposed method is capable of achieving optimal operation of a phase change thermal energy storage system in practical applications.

A multivariate time series graph neural network for district heat load forecasting

Heat load prediction is essential for energy efficiency and carbon reduction in district heating systems. However, heat load is influenced by many factors, such as building characteristics, consumption behavior, and climate, making its prediction challenging. Traditional methods based on physical models are complex and insufficiently accurate, whereas most data-driven statistical methods ignore customer energy consumption behaviors and their correlation, and do not account for the temporal inertia of consumption. This paper proposes a graph ambient intelligence (GAIN) method for heat load prediction, which classifies customers based on their load profiles and uses collaborative attention on temporal graphs to capture their associations and the weather impact on heat loads. GAIN also incorporates recursive and autoregressive methods to model the temporal inertia of consumption. The proposed method is evaluated on four metrics and compared with fifteen baseline methods. The results show that GAIN achieves the lowest daily forecasting errors in terms of RMSE, MAE, and CV-RMSE, with values of 6.972, 4.442, and 0.191, respectively. Besides, the proposed method achieves a maximum reduction of 25%, 29%, and 25% in RMSE, MAE, and CV-RMSE, respectively, compared to other methods when taking meteorological factors into account.