Search

[Display omitted] • The value of transfer learning has been assessed for fault diagnosis in building energy systems. • Four popular model-based transfer learning implementation strategies have been investigated. • Transfer learning can improve the fault diagnosis performance when the modeling data are limited. • The maximum fault diagnosis accuracy improvements are 12.63% and 8.18% in two learning tasks respectively. • The data amount of source domain affects the stability of the transfer learning significantly. Data-driven fault detection and diagnosis (FDD) in building energy systems is typically limited by the quantity and quality of training data. These methods can be only used for individual systems due to the insufficient extrapolation capabilities of most machine learning algorithms. A desirable solution is to utilize transfer learning, which can transfer the knowledge learned from data-rich building energy systems to FDD tasks in data-sparse systems. However, the potential of applying transfer learning to such FDD has not been systematically investigated. Accordingly, this paper proposes a transfer-learning-based methodology for fault diagnosis in building chillers. Experiments were conducted on two water-cooled screw chillers to collect both fault and fault-free data. Transfer-learning-based fault diagnosis experiments were implemented with consideration of different transfer learning tasks, training cases, learning scenarios, and transfer learning implementation strategies. The experimental results validate the value of transfer learning for FDD in building energy systems, especially when the experimental data available for model development are limited. The maximum accuracy improvements were 12.63% and 8.18% in the two learning tasks. The research outcomes provide practical guidelines for developing transfer-learning-based solutions for FDD in building energy systems. [ABSTRACT FROM AUTHOR]

Building energy prediction is a potential tool for benchmarking the future energy uses of individual buildings. However, inevitable gaps between predicted and actual energy uses and discrepancies between buildings make it challenging to quantify energy consumption. Therefore, this paper proposes a systematic methodology of quantitative building energy evaluation based on short-term energy prediction. First, the 24-h ahead building energy prediction model is developed based on a recurrent neural network via the Multi-Input Multi-Output strategy. Second, the quantitative energy evaluation strategy is proposed to quantify prediction gaps based on the 1-D k-means clustering. Third, case studies are conducted on five real buildings to verify the reliability of the proposed methodology. Results show that the energy prediction models achieved outstanding accuracies. Besides, it is necessary to analyze the absolute percentage error (APE) variation of each time step to deeply understand the building energy performance rather than the overall prediction performance evaluation index, such as CV-RMSE and MAPE. Further, customized energy quantification systems are established for buildings per their specific, individual energy performance. The building energy is quantified by labeling APE s into multiple levels. Moreover, building operation characteristics can be further understood by quantifying energy uses. • A systematic methodology is proposed to quantitatively evaluate the individual building energy. • Short-term energy prediction models are established for energy benchmarking based on LSTM algorithm. • Customized energy quantification systems are established for various buildings by clustering analysis. • The building operation characteristics can be understood by quantifying the energy uses. [ABSTRACT FROM AUTHOR]

Timely and accurate fault diagnosis (FD) in building energy systems (BESs) can promote energy efficiency and sustainable development. Especially the heating, ventilating, and air-conditioning (HVAC) systems are diverse and operate under complex and variable operation conditions. System and operation differences lead to great differences in operational data which causes poor adaptability of data-driven FD models that are developed using data from a single HVAC system or limited operation condition. To improve diagnostic performance across different HVAC systems and operation conditions, this study proposes high-adaptability FD models using three deep transfer learning (DTL) strategies including network-based fine-tuning (FT), mapping-based domain-adaptive neural network (DaNN) and adversarial-based domain adversarial neural network (DANN). The effectiveness of DTL-based FD is validated by fault datasets of two typical BESs: one is a 703-kW screw chiller while the other is the 316-kW centrifugal chiller from ASHRAE RP-1043. Two types of TL scenarios (cross-system and cross-operation-condition fault diagnosis) are set up consisting of eight TL tasks. For DTL strategies, both FD performance and transferability are evaluated using metrics like accuracy and accuracy improvement degree (AID). Results indicate that FT obtains 93% FD accuracy averagely for all tasks of the two TL scenarios considered, which is an average 55% AID compared with the non-transfer benchmark model convolutional neural network (CNN). Further, the impacts of source and target data volumes, and TL tasks are analyzed. For cross-operation-condition scenario, DTL-based FD accuracy grows with the increase of target data volume. For cross-system scenario, FT still show high FD performance with less training data. The reason why FT outperforms DANN and DaNN is explained by visualizing classification scatterplots of the last NN layers. Practical application issues of the DTL-based FD strategy for building energy systems are discussed at last. [Display omitted] • Proposed 3 different DTL strategies to validate BES cross-system and cross-operation-condition FD. • The 3 DTL chiller FD strategies: FT outperforms DANN and DaNN with 93% FD accuracy. • FT gets 55% average FD accuracy improvement for 8 cross-system and cross-operation-condition tasks. • DTL gets higher FD accuracy with growing target data for cross-operation-conditions scenario. • FT learns general features in bottom NN layers while specific ones related to the FD task in top layers. [ABSTRACT FROM AUTHOR]

[Display omitted] • 3 DTL strategies obtain 75% PIRs than LSTM for data shortage BEP tasks in OFs and UCs. • Recommend Fine-tune for next few weeks BEP by balance between time cost and performance. • Recommend DANN for nearly a year BEP by outperformed accuracy than Fine-tune, DaNN. • Evaluate DTL prediction accuracy by varying available training data in both source and target buildings. Performing accurate building energy prediction (BEP) is one of the most important foundations for achieving energy resource allocation and developing energy efficiency measures. Buildings are diverse and operate under complex conditions leading to distribution differences of energy-related data among different buildings. Owing to such differences, the already reported data-driven BEP models exhibit poor cross-building prediction performance since they only use insufficient operation data of a single building. Although several deep transfer learning (DTL) strategies have been applied with improved cross-building prediction performance, there is still a lack of comparison of various DTL strategies, which would help to determine the optimal DTL strategy for different scenarios. Hence, three DTL strategies were compared: network-based Fine-tune, adversarial-based domain adversarial neural network (DANN), and mapping-based domain adaptive neural network (DaNN). The usefulness of the DTL strategy was validated using the open source dataset Building Data Genome Project 2 in the scenario of insufficient available data for real buildings. The influences of several factors were analysed, such as the available data volumes of the source and target domains within the training set. The applicability of different DTL strategies was discussed considering both accuracy and computational cost. Results show that the three DTL strategies outperform the traditional long short-term memory (LSTM) with an average BEP performance improvement ratio (PIR) of 0.75. For the extremely limited amount of available training data scenario, Fine-tune was recommended for the next-few-weeks prediction owing to its good balance between time cost and prediction performance. For the data shortage scenario of BEP tasks for nearly a year, DANN was recommended owing to its outperforming prediction accuracy. This provides insights for practical applications of DTL during the development of BEP models for cross-building BEP tasks without sufficient operation data. [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]