Search

• An improved deep learning model based on tree structure is proposed in VRF system. • Pearson's correlation coefficient combines with isolation forest for data preprocessing • The physical significance of outliers in VRF system is explained. • The IF-CF model does not require complex hyper-parameter optimization strategy. This paper presents a hybrid deep forest approach for outlier detection and fault diagnosis. Isolation forest algorithm is combined with Pearson's correlation coefficient for outlier detection. The physical significance of outliers detected by the proposed algorithm is explained by origin analysis, which is rarely mentioned in existing studies. In addition, a novel non-neural network deep learning model-cascade forest model is proposed to fault diagnosis of HVAC system for the first time to achieve high precision accuracy in low-dimensional features. The proposed approach is validated with the refrigerant charge fault of VRF system. The results show that the isolation forest algorithm can improve the performance of fault diagnosis model and the mainly outliers of VRF system are defrosting data. The IF-CF model has short operation time, and high accuracy in low-dimensional features. When the dimension drops to 6, the accuracy of the IF-CF model is 94.16%, which is 5.26%, 10.02%, 5.87% and 3.34% higher than the IF-MLP, IF-BPNN, IF-SVM and IF-LSTM models, respectively. Moreover, IF-CF model does not require complex hyper-parameter optimization strategy because its maximum accuracy difference in different hyper-parameters is 2.04%. This study is enlightening which may inspire the potential of outlier detection technology and deep learning in HVAC field. [ABSTRACT FROM AUTHOR]

Energy consumption of heat pump system accounts for a large part of the total building energy consumption, and the energy-saving operation of heat pump system has always been the focus of researchers. A promising solution to tackling energy wastes during system operations is anomaly detection. In this study, we propose an anomaly detection method for GSHP system in a public building based on mode decomposition based LSTM and statistical modeling method Grubbs' test. The system energy consumption is predicted using mode decomposition based LSTM algorithm, and the difference between predicted value and actual value is used to detect the abnormal system energy consumption by Grubbs' test. Results show that detected anomalies can be summarily divided into three categories (parabola anomaly, abrupt anomaly and time related anomaly) depending on their characteristics, and the rationality of detected anomalies are evaluated through field investigation and expert knowledge. This work is enlightening and indicates that the proposed method would efficiently detect the abnormal performance of GSHP system, and find out unreasonable operating patterns. [ABSTRACT FROM AUTHOR]

Abstract Short-term load prediction at the district-level is essential for feeders, substations, consumers and transformers starts from 1-h to one-week ahead. Though, the critical problem is in the design of complex load characteristics, switching operations and a large number of junctions at electrical distribution feeders. To overcome the design challenges, this research conducted an analysis of short-term energy requirement forecasting for district-level by applying into two deep learning models. Present, the booming increase of deep learning (DL) methods give them encouraging options to data-driven techniques. However, DL gives exceptional ability in handling the non-linear complex relations, model computation performance and complexity are of attention. Further, the two different climate zones data used for modeling analysis. The input feature parameter sets classified into two parts: i) features selection-I (FS-I); and FS-II. The FS-I comprises based on seven environmental parameters, including various limits of energy consumption data sets and the FS-II contains the different sixteen feature variables. The FS-II mean absolute error is <2.05% and <1.577% of OSSB-NN and BFGS-QNB respectively. The deep learning models achieved higher forecasting accuracy at the proposed structure of the model's network at different hidden neurons. Based on the current trend and the comparison results, a structure of district energy prediction and future research direction proposed for district-level design, planning, and services. Highlights • Two deep-learning models used to forecast future load demand of two climate zones. • Models will play essential role in the future district energy forecasting. • Higher accuracy obtained for short-term horizons leads to economic advantages. • The MAPE of BFGS-QNB forecasting model is <3.679%. • A framework is deep learning proposed for design, planning, and service. [ABSTRACT FROM AUTHOR]

Abstract This study presents a case study of public buildings using a novel deep learning method to forecast indoor air temperature. The aim is to explore the potential of long short-term memory (LSTM) model in forecasting indoor temperature, and a novel LSTM model modified by error correction model is established. The performance of the two models is compared with popular prediction methods in the building field. Results show that the proposed novel LSTM model has slight advantages in level indoor temperature prediction performance comparing with other common machine learning methods. However, it outperforms other models including original LSTM in terms of directional prediction accuracy, and accurately predicts the indoor temperature variation trend. This work is enlightening and may have a further reference to the feasibility study of indoor air temperature prediction model. Highlights • We proposed a novel deep learning method for indoor temperature prediction. • Error correction model was used to revise the LSTM model. • The method was compared with three common used machine learning models. • The method showed outstanding effect on improving the performance of indoor temperature prediction. [ABSTRACT FROM AUTHOR]

With the development in information technologies, today's building energy consumption can be well monitored by the building energy management systems. However, in most real applications there is no clear definition of abnormal building energy consumption. To overcome this limitation, this work proposes a novel deep learning based unsupervised anomaly detection framework that includes recurrent neural networks and quantile regression. Moreover, this framework is able to produce a prediction interval to detect and evaluate abnormal building energy consumption. The framework has been applied to analyze the energy data collected from three different residential houses, and anomaly detection results are evaluated by the quantile regression range. The research results can provide promising solutions for building managers to detect abnormal energy performance, and is also valuable to assess the level of anomalies and spot opportunities in energy conservation. [ABSTRACT FROM AUTHOR]

Model predictive control(MPC) achieves great performance in energy management of the building. However, identifying a suitable control-oriented model for MPC is a challenging task. To overcome this problem, we attempt to apply the data-driven models which have universal approximation ability to the MPC task. In this paper, we propose a hybrid optimization algorithm, namely BSAS-LM algorithm, to solve the optimization problem with non-linear or non-convex data-driven models involved in data-driven predictive control(DDPC). To demonstrate the feasibility and scalability of the proposed hybrid optimization method, three case studies are implemented in three buildings with different geometries. The DDPC controllers are developed for each case study in three scenarios, namely constant temperature setpoint, lower temperature setpoint and pre-heating. EnergyPlus is employed to develop the building models and is then exported to Functional Mock-up Units(FMUs) for co-simulation. In the case study #1, the data-driven algorithms such as auto-regressive with external disturbance (ARX) and support vector regression(SVR) are used to develop models for a single-zone building. Those models are then applied in DDPC for climate control of the building. In the case study #2 and #3, the multilayer perceptron(MLP)-based DDPC is applied to two three-zones buildings. Results show that DDPC achieves comparable performance to the grey-box model based MPC. Besides, results also demonstrate the feasibility and scalability of the proposed method in DDPC integrated with various data-driven models. • A hybrid optimization algorithm was proposed to solve the optimization problems in data-driven predictive control(DDPC). • Four widely used machine learning algorithms were employed to develop the prediction models for building thermal dynamics. • Three case studies were implemented to verify the feasibility and scalability of the proposed method in DDPC for buildings. • Results show that the DDPC can achieve comparable control performance to grey-box based MPC. • The proposed method can greatly reduce the cost of building modeling. [ABSTRACT FROM AUTHOR]

The fault diagnosis of air-conditioning systems is of great significance to the energy saving of buildings. This study proposes a novel fault diagnosis approach for building energy saving based on the deep learning method which is deep belief network, and its application potential in the air conditioning fault diagnosis field is investigated. Then, a parameter optimization selection strategy is developed for model optimization. Four kinds of faults of the variable flow refrigerant system under heating mode are used to evaluate the performance of the models. The fault diagnosis results show that the deep belief network model with initial parameters can be used to diagnose the faults of the variable flow refrigerant system. Through the parameter optimization selection strategy, the fault diagnosis correct rate of the optimized model is 97.7%, which is improved by 5.05% compared with the model with initial parameters. The number of hidden layers of the deep belief network model is selected to be 2 layers. This result indicates that the fault diagnosis for variable flow refrigerant systems may not require a very deep model. Additionally, the performance of the optimized deep belief network model is compared with that of the traditional back propagation neural network, and the former is better. This finding also shows that the unsupervised restricted Boltzmann machine layer for data feature reconstruction can improve the fault diagnosis performance. [ABSTRACT FROM AUTHOR]

• Transform segmentation task from classification to regression by making models learn a level set function. • Design a shape-based loss with a convexity constraint term. • Improve segmentation performance at the boundary with weak contrast. The lens is one of the important refractive media in the eyeball. Abnormality of the nucleus or cortex in the lens can lead to ocular disorders such as cataracts and presbyopia. To achieve an accurate diagnosis, segmentation of these ocular structures from anterior segment optical coherence tomography (AS-OCT) is essential. However, weak-contrast boundaries of the object in the images present a challenge for accurate segmentation. The state-of-the-art (SOTA) methods, such as U-Net, treat segmentation as a binary classification of pixels, which cannot handle pixels on weak-contrast boundaries well. In this paper, we propose to incorporate shape prior into a deep learning framework for accurate nucleus and cortex segmentation. Specifically, we propose to learn a level set function, whose zero-level set represents the object boundary, through a convolutional neural network. Moreover, we design a novel shape-based loss function, where the shape prior knowledge can be naturally embedded into the learning procedure, leading to improvement in performance. We collect a high-quality AS-OCT image dataset with precise annotations to train our model. Abundant experiments are conducted to verify the effectiveness of the proposed framework and the novel shape-based loss. The mean Intersection over Unions (MIoUs) of the proposed method for lens nucleus and cortex segmentation are 0.946 and 0.957, and the mean Euclidean Distance (MED) measure, which can reflect the accuracy of the segmentation boundary, are 6.746 and 2.045 pixels. In addition, the proposed shape-based loss improves the SOTA models on the nucleus and cortex segmentation tasks by an average of 0.0156 and 0.0078 in the MIoU metric and 1.394 and 0.134 pixels in the MED metric. We transform the segmentation from a classification task to a regression task by making the model learn the level set function, and embed shape information in deep learning by designing loss functions. This allows the proposed method to be more efficient in the segmentation of the object with weak-contrast boundaries. We propose to incorporate shape priors into a deep learning framework for accurate nucleus and cortex segmentation from AS-OCT images. Specifically, we propose to learn a level set function, where the zero-level set represents the boundary of the target. Meanwhile, we design a novel shape-based loss function in which additional convex shape prior can be embedded in the learning process, leading to an improvement in performance. The IOUs for nucleus and cortex segmentation are 0.946 and 0.957, while the MED that reflects the accuracy of the boundary are 6.746 and 2.045 pixels. The proposed shape-based loss improves the SOTA model for nucleus and cortex segmentation by an average of 0.0156 and 0.0078 in IOU, and 1.394 and 0.134 pixels in MED. We transform segmentation from classification to regression by making the model learn a level set function, resulting in improved performance at the boundary with weak contrast. [ABSTRACT FROM AUTHOR]

• Proposed an interpretable deep learning model using the attention-BiLSTM. • Four typical faults of VRF are well identified with an overall diagnosis accuracy of 97.5%. • Effective fault diagnosis under incomplete information is realized through the integrated model of different feature combinations. • Realize the quantification of diagnosis uncertainty through the ensemble model. • The model performs well on the unknown data set under different working conditions. Developing effective fault detection and diagnosis (FDD) model is of great significance to the energy efficiency and comfort of variable refrigerant flow (VRF) system. In recent FDD studies, the consideration of incomplete information and uncertainty, as well as the generalization of model and the interpretability of fault action mechanism have become a great concern. This paper accordingly proposes a fault diagnosis strategy based on attention-BiLSTM with ensemble feature sets. Three methods including Correlation Analysis, Information Entropy, and Gini Impurity are used for feature selection and ensemble feature sets combination. The proposed model can achieve interpretability of fault diagnosis and quantification of diagnosis uncertainties through attention mechanism and ensemble feature sets. A total of four typical faults under different working conditions are used to verify the accuracy and generalization of the proposed method. The results show that the ensemble attention-BiLSTM model has a fault diagnosis accuracy rate of 98.3% if the first two most likely outputs are considered, and when some sensors fail, the accuracy rate can still remain above 90%. In addition, a generalization verification idea is proposed, that is, the model is trained under specific operating conditions and tested on datasets with different operating conditions or fault levels. The results show that the average generalization of the optimized model reaches 86%. [ABSTRACT FROM AUTHOR]

• Deep reinforcement learning (DRL) based models are proposed for HVAC system energy consumption prediction. • Autoencoder (AE) algorithm is used to extract the high-level features of state space of DDPG. • The DDPG based models are applied for energy consumption forecasting over short-term time horizon. • Improvement in prediction performance is observed when combine AE with DDPG. • The DDPG based prediction models, outperform common supervised models like BP neural network and support vector machine. Short-term energy consumption prediction has fundamental importance in many HVAC system management tasks, such as demand-side management, short-term maintenance, etc. Currently, the prevailing data-driven techniques, especially supervised machine learning methods, are widely applied for short-term energy consumption prediction. Deep reinforcement learning (DRL), as the state-of-the-art machine learning techniques, have been applied for HVAC system control, but rarely for energy consumption prediction. In this paper, a DRL algorithm, namely Deep Deterministic Policy Gradient (DDPG), is firstly introduced for short-term HVAC system energy consumption prediction. Moreover, Autoencoder (AE), which is powerful in processing data in their raw form, is incorporated into DDPG method to extract the high-level features of state space and optimize the prediction model. The operation data of the ground source heat pump (GSHP) system of an office building in Henan province, China is used to train and assess the proposed models. The results demonstrate that the proposed DDPG based models can achieve better prediction performance than common supervised models like BP Neural Network and Support Vector Machine. This study is an enlightening work which may inspire other researchers to tap the potential of DRL algorithms in this field. [ABSTRACT FROM AUTHOR]

• The potentials of deep reinforcement learning (DRL) techniques are investigated. • Three commonly-used DRL techniques, i.e. A3C, DDPG and RDPG are utilized for building energy consumption forecasting. • The prediction performances of six popular predictive algorithms are studied. • DDPG and RDPG can evidently enhance prediction performance while accounting for more computation time. • A3C doesn't show advantage in forecasting building energy consumption. Reliable and accurate building energy consumption prediction is becoming increasingly pivotal in building energy management. Currently, data-driven approach has shown promising performances and gained lots of research attention due to its efficiency and flexibility. As a combination of reinforcement learning and deep learning, deep reinforcement learning (DRL) techniques are expected to solve nonlinear and complex issues. However, very little is known about DRL techniques in forecasting building energy consumption. Therefore, this paper presents a case study of an office building using three commonly-used DRL techniques to forecast building energy consumption, namely Asynchronous Advantage Actor-Critic (A3C), Deep Deterministic Policy Gradient (DDPG) and Recurrent Deterministic Policy Gradient (RDPG). The objective is to investigate the potential of DRL techniques in building energy consumption prediction field. A comprehensive comparison between DRL models and common supervised models is also provided. The results demonstrate that the proposed DDPG and RDPG models have obvious advantages in forecasting building energy consumption compared to common supervised models, while accounting for more computation time for model training. Their prediction performances measured by mean absolute error (MAE) can be improved by 16%-24% for single-step ahead prediction, and 19%-32% for multi-step ahead prediction. The results also indicate that A3C performs poor prediction accuracy and shows much slower convergence speed than DDPG and RDPG. However, A3C is still the most efficient technique among these three DRL methods. The findings are enlightening and the proposed DRL methodologies can be positively extended to other prediction problems, e.g., wind speed prediction and electricity load prediction. [ABSTRACT FROM AUTHOR]