Yixin Yu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yixin Yu has a total of 53 co-authors. Here are the top co-authors by number of shared publications:
- Wenpeng Luan
School of Electrical and Information Engineering, Tianjin University, Tianjin, China
6 shared publications - Bo Liu
School of Electrical and Information Engineering, Tianjin University, Tianjin, China
5 shared publications - Bochao Zhao
School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China
3 shared publications Chao Qin
3 shared publicationsBing Sun
Tianjin University
2 shared publications- Chao Qin
School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China
2 shared publications
Latest Publications
- 25 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73  (00189456) , 1-13
Automatic Appliance Labeling for Unsupervised NILM Based on Hierarchical Decision-Making
The unsupervised non-intrusive load monitoring (NILM) can determine fine-grained energy consumption information at the appliance level in unseen scenarios, by autonomously analyzing aggregate electricity data. However, a challenge for unsupervised NILM is to determine the actual physical names of discovered appliances, which hinders the large-scale implementation of NILM technologies. For addressing this problem, an automatic appliance labeling method is proposed in this paper, which can assign physical names to unknown appliances identified by unsupervised NILM. The general operation characteristics shared by same type of appliances are summarized. On this basis, a hierarchical decision-making process for appliance labeling is proposed. Rough set theory is adopted to select the potential candidate appliance set by evaluating general operation characteristics, followed by the most likely candidate name determination through voting on parametric characteristics. Furthermore, association pattern mining is carried out to discover recurring operation mode combinations and identify those associated modes corresponding to multi-mode appliances. Experimental results on real-measured and public datasets show that the proposed method can achieve accurate labeling of NILM results in various unseen scenarios with limited prior knowledge, which demonstrates the potential for practical application.
- 25 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73  (00189456) , 1-13
Automatic Appliance Labeling for Unsupervised NILM Based on Hierarchical Decision-Making
The unsupervised non-intrusive load monitoring (NILM) can determine fine-grained energy consumption information at the appliance level in unseen scenarios, by autonomously analyzing aggregate electricity data. However, a challenge for unsupervised NILM is to determine the actual physical names of discovered appliances, which hinders the large-scale implementation of NILM technologies. For addressing this problem, an automatic appliance labeling method is proposed in this paper, which can assign physical names to unknown appliances identified by unsupervised NILM. The general operation characteristics shared by same type of appliances are summarized. On this basis, a hierarchical decision-making process for appliance labeling is proposed. Rough set theory is adopted to select the potential candidate appliance set by evaluating general operation characteristics, followed by the most likely candidate name determination through voting on parametric characteristics. Furthermore, association pattern mining is carried out to discover recurring operation mode combinations and identify those associated modes corresponding to multi-mode appliances. Experimental results on real-measured and public datasets show that the proposed method can achieve accurate labeling of NILM results in various unseen scenarios with limited prior knowledge, which demonstrates the potential for practical application.
- 8 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73  (00189456) , 1-14
Adaptive Multitimescale Event Detection in Nonintrusive Load Monitoring Based on Minimum Description Length Principle
Load event detection is the fundamental step for the event-based nonintrusive load monitoring (NILM). However, existing event detection methods with fixed parameters may fail in coping with the inherent multitimescale characteristics of events and their detection accuracy is easily affected by the load fluctuations. In this regard, this article extends our previous two-stage event detection framework, and proposes a novel multitimescale event detection method based on the minimum description length (MDL) principle. Following the completion of step-like event detection, a long-transient event detection scheme with variable length sliding window is designed, which is intended to observe and characterize the same event at different timescales. In that, the context information in the aggregated load data is mined by motif discovery, then based on the MDL principle, the proper observation scales are selected for different events and the corresponding detection results are determined. For the postprocessing, a load fluctuation segment location method based on voice activity detection (VAD) algorithm is proposed to identify and remove the unreasonable events caused by load fluctuations. Based on new evaluation metrics, the comparison tests on public and private datasets demonstrate that our method achieves higher detection accuracy and integrity for events of various appliances across different scenarios.
- 8 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73  (00189456) , 1-14
Adaptive Multitimescale Event Detection in Nonintrusive Load Monitoring Based on Minimum Description Length Principle
Load event detection is the fundamental step for the event-based nonintrusive load monitoring (NILM). However, existing event detection methods with fixed parameters may fail in coping with the inherent multitimescale characteristics of events and their detection accuracy is easily affected by the load fluctuations. In this regard, this article extends our previous two-stage event detection framework, and proposes a novel multitimescale event detection method based on the minimum description length (MDL) principle. Following the completion of step-like event detection, a long-transient event detection scheme with variable length sliding window is designed, which is intended to observe and characterize the same event at different timescales. In that, the context information in the aggregated load data is mined by motif discovery, then based on the MDL principle, the proper observation scales are selected for different events and the corresponding detection results are determined. For the postprocessing, a load fluctuation segment location method based on voice activity detection (VAD) algorithm is proposed to identify and remove the unreasonable events caused by load fluctuations. Based on new evaluation metrics, the comparison tests on public and private datasets demonstrate that our method achieves higher detection accuracy and integrity for events of various appliances across different scenarios.
- 28 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 14  (6) , 4312-4321
Accurate Identification of Critical Boundary Hyperplanes of Practical Steady-State Security Region in Distribution Grids
The practical steady-state security region (PSSR) in distribution grids describes the region where the power flow can be managed securely. The PSSR is irrelevant to operation states and its boundaries can be approximated by hyperplanes (HPs). Benefit from these, the PSSR can dramatically simplify complicated optimal scheduling problems, such as deliverable energy flexibility scheduling. However, calculating the PSSR boundaries corresponding to all the power flow security constraints indiscriminately will theoretically produce lots of HPs to bound the whole PSSR and thus make solving of the PSSRbased optimal scheduling challenging. In fact, the whole PSSR is the intersection of the regions bounded by all the HPs within practical power injection ranges, and only a few HPs, called the critical boundary hyperplanes (CHPs), can bound it. Therefore, it is necessary to identify the CHPs before applying them to the optimal scheduling as power flow security constraints, so as to accelerate the solution. In this paper, a fast and accurate method is first proposed to identify the CHPs. The identified results are irrelevant to the specific objective and can be applied to different optimal scheduling problems. The proposed method is applied to several case-study networks and the results demonstrate its effectiveness and speed.
- 28 March 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 14  (6) , 4312-4321
Accurate Identification of Critical Boundary Hyperplanes of Practical Steady-State Security Region in Distribution Grids
The practical steady-state security region (PSSR) in distribution grids describes the region where the power flow can be managed securely. The PSSR is irrelevant to operation states and its boundaries can be approximated by hyperplanes (HPs). Benefit from these, the PSSR can dramatically simplify complicated optimal scheduling problems, such as deliverable energy flexibility scheduling. However, calculating the PSSR boundaries corresponding to all the power flow security constraints indiscriminately will theoretically produce lots of HPs to bound the whole PSSR and thus make solving of the PSSRbased optimal scheduling challenging. In fact, the whole PSSR is the intersection of the regions bounded by all the HPs within practical power injection ranges, and only a few HPs, called the critical boundary hyperplanes (CHPs), can bound it. Therefore, it is necessary to identify the CHPs before applying them to the optimal scheduling as power flow security constraints, so as to accelerate the solution. In this paper, a fast and accurate method is first proposed to identify the CHPs. The identified results are irrelevant to the specific objective and can be applied to different optimal scheduling problems. The proposed method is applied to several case-study networks and the results demonstrate its effectiveness and speed.
- 22 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72  (00189456) , 1-13
Nonintrusive Load Monitoring Based on Self-Supervised Learning
Deep learning models for non-intrusive load monitoring (NILM) tend to require a large amount of labeled data for training. However, it is difficult to generalize the trained models to unseen sites due to different load characteristics and operating patterns of appliances between data sets. For addressing such problems, self-supervised learning (SSL) is proposed in this paper, where labeled appliance-level data from the target data set or house is not required. Initially, only the aggregate power readings from target data set are required to pre-train a general network via a self-supervised pretext task to map aggregate power sequences to derived representatives. Then, supervised downstream tasks are carried out for each appliance category to fine-tune the pre-trained network, where the features learned in the pretext task are transferred. Utilizing labeled source data sets enables the downstream tasks to learn how each load is disaggregated, by mapping the aggregate to labels. Finally, the fine-tuned network is applied to load disaggregation for the target sites. For validation, multiple experimental cases are designed based on three publicly accessible REDD, UK-DALE, and REFIT data sets. Besides, state-of-the-art neural networks are employed to perform NILM task in the experiments. Based on the NILM results in various cases, SSL generally outperforms zero-shot learning in improving load disaggregation performance without any sub-metering data from the target data sets.
- 22 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72  (00189456) , 1-13
Nonintrusive Load Monitoring Based on Self-Supervised Learning
Deep learning models for non-intrusive load monitoring (NILM) tend to require a large amount of labeled data for training. However, it is difficult to generalize the trained models to unseen sites due to different load characteristics and operating patterns of appliances between data sets. For addressing such problems, self-supervised learning (SSL) is proposed in this paper, where labeled appliance-level data from the target data set or house is not required. Initially, only the aggregate power readings from target data set are required to pre-train a general network via a self-supervised pretext task to map aggregate power sequences to derived representatives. Then, supervised downstream tasks are carried out for each appliance category to fine-tune the pre-trained network, where the features learned in the pretext task are transferred. Utilizing labeled source data sets enables the downstream tasks to learn how each load is disaggregated, by mapping the aggregate to labels. Finally, the fine-tuned network is applied to load disaggregation for the target sites. For validation, multiple experimental cases are designed based on three publicly accessible REDD, UK-DALE, and REFIT data sets. Besides, state-of-the-art neural networks are employed to perform NILM task in the experiments. Based on the NILM results in various cases, SSL generally outperforms zero-shot learning in improving load disaggregation performance without any sub-metering data from the target data sets.
- Open Access
Multi-timescale Event Detection in Nonintrusive Load Monitoring based on MDL Principle
Load event detection is the fundamental step for the event-based non-intrusive load monitoring (NILM). However, existing event detection methods with fixed parameters may fail in coping with the inherent multi-timescale characteristics of events and their event detection accuracy is easily affected by the load fluctuation. In this regard, this paper extends our previously designed two-stage event detection framework, and proposes a novel multi-timescale event detection method based on the principle of minimum description length (MDL). Following the completion of step-like event detection in the first stage, a long-transient event detection scheme with variable-length sliding window is designed for the second stage, which is intended to provide the observation and characterization of the same event at different time scales. In that, the context information in the aggregated load data is mined by motif discovery, and then based on the MDL principle, the proper observation scales are selected for different events and the corresponding detection results are determined. In the post-processing step, a load fluctuation location method based on voice activity detection (VAD) is proposed to identify and remove the unreasonable events caused by fluctuations. Based on newly proposed evaluation metrics, the comparison tests on public and private datasets demonstrate that our method achieves higher detection accuracy and integrity for events of various appliances across different scenarios.
Open AccessMulti-timescale Event Detection in Nonintrusive Load Monitoring based on MDL Principle
Load event detection is the fundamental step for the event-based non-intrusive load monitoring (NILM). However, existing event detection methods with fixed parameters may fail in coping with the inherent multi-timescale characteristics of events and their event detection accuracy is easily affected by the load fluctuation. In this regard, this paper extends our previously designed two-stage event detection framework, and proposes a novel multi-timescale event detection method based on the principle of minimum description length (MDL). Following the completion of step-like event detection in the first stage, a long-transient event detection scheme with variable-length sliding window is designed for the second stage, which is intended to provide the observation and characterization of the same event at different time scales. In that, the context information in the aggregated load data is mined by motif discovery, and then based on the MDL principle, the proper observation scales are selected for different events and the corresponding detection results are determined. In the post-processing step, a load fluctuation location method based on voice activity detection (VAD) is proposed to identify and remove the unreasonable events caused by fluctuations. Based on newly proposed evaluation metrics, the comparison tests on public and private datasets demonstrate that our method achieves higher detection accuracy and integrity for events of various appliances across different scenarios.
- 31 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 13  (6) , 4690-4703
The Balanced Window-Based Load Event Optimal Matching for NILM
Load event matching is the key for event-based non-intrusive load monitoring (NILM). It aims to find the load event sequence corresponding to the appliance’s operation cycle from all detected load events. Firstly, a novel combinatorial optimization model for load event matching is established. To reduce the dimension of the estimation matrix in the proposed optimization model, the concept of balanced window (BW) is introduced to indicate a period of load operation within which events can be fully matched. Then an iterative load event optimal matching strategy based on nested BWs is proposed, where most inner BWs are processed firstly. For the event matching in one specific BW, a two-stage optimization algorithm is designed. In the first stage, a set of candidate load event sequences is generated by using the depth first search algorithm with pruning, which reduces the solution space of the combinatorial optimization problem. In the second stage, the optimal load event sequences that result in the optimal fit of the aggregated load power is selected. Comparative experiments on the publicly-accessible REDD datasets have shown that the proposed method achieves better performance against the existing unsupervised NILM methods, and is promising for practical implementation.
- 31 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 13  (6) , 4690-4703
The Balanced Window-Based Load Event Optimal Matching for NILM
Load event matching is the key for event-based non-intrusive load monitoring (NILM). It aims to find the load event sequence corresponding to the appliance’s operation cycle from all detected load events. Firstly, a novel combinatorial optimization model for load event matching is established. To reduce the dimension of the estimation matrix in the proposed optimization model, the concept of balanced window (BW) is introduced to indicate a period of load operation within which events can be fully matched. Then an iterative load event optimal matching strategy based on nested BWs is proposed, where most inner BWs are processed firstly. For the event matching in one specific BW, a two-stage optimization algorithm is designed. In the first stage, a set of candidate load event sequences is generated by using the depth first search algorithm with pruning, which reduces the solution space of the combinatorial optimization problem. In the second stage, the optimal load event sequences that result in the optimal fit of the aggregated load power is selected. Comparative experiments on the publicly-accessible REDD datasets have shown that the proposed method achieves better performance against the existing unsupervised NILM methods, and is promising for practical implementation.
- 3 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 71  (00189456) , 1-14
An Adaptive Two-Stage Load Event Detection Method for Nonintrusive Load Monitoring
Load event detection is fundamental to event-based nonintrusive load monitoring (NILM) solution. It is directly related to whether the transient and steady-state signatures of appliances can be accurately extracted. In the real world, the load composition is random and complicated, the power consumption patterns of appliances are diverse, and multiple events may occur simultaneously or near each other, all these make it difficult for any one single event detection method to achieve satisfactory robustness. In coping with this, an adaptive two-stage event detection method is proposed in this article. First, an adaptive detection threshold whose value can be adjusted adaptively according to the load fluctuations is adopted, thus improving the ability to detect events of different amplitudes. Then, considering the different event geometric features, an improved edge detection method and a window-based detection method combining moving average with moving t-test are proposed for step-like events and long-transient events, respectively, and they are executed consecutively to achieve the effective capture of the complete transient process of appliances. Furthermore, an event separation step is taken to locate and separate the near-simultaneous events which often appear in low-frequency data. Specific event detection performance evaluation metrics are also designed. Comparison test results on private and public datasets show that the proposed method achieves high detection accuracy for different events of various appliances and maintains strong robustness in different operation scenarios.
- 3 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 71  (00189456) , 1-14
An Adaptive Two-Stage Load Event Detection Method for Nonintrusive Load Monitoring
Load event detection is fundamental to event-based nonintrusive load monitoring (NILM) solution. It is directly related to whether the transient and steady-state signatures of appliances can be accurately extracted. In the real world, the load composition is random and complicated, the power consumption patterns of appliances are diverse, and multiple events may occur simultaneously or near each other, all these make it difficult for any one single event detection method to achieve satisfactory robustness. In coping with this, an adaptive two-stage event detection method is proposed in this article. First, an adaptive detection threshold whose value can be adjusted adaptively according to the load fluctuations is adopted, thus improving the ability to detect events of different amplitudes. Then, considering the different event geometric features, an improved edge detection method and a window-based detection method combining moving average with moving t-test are proposed for step-like events and long-transient events, respectively, and they are executed consecutively to achieve the effective capture of the complete transient process of appliances. Furthermore, an event separation step is taken to locate and separate the near-simultaneous events which often appear in low-frequency data. Specific event detection performance evaluation metrics are also designed. Comparison test results on private and public datasets show that the proposed method achieves high detection accuracy for different events of various appliances and maintains strong robustness in different operation scenarios.
- 25 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 12  (3) , 2136-2152
A Comprehensive Resilience-Oriented FLISR Method for Distribution Systems
Under extreme events, distribution systems may suffer blackouts and numerous restoration tasks will compete for limited crews and facilities. To enhance the resilience of distribution systems, multiple resources should be fully coordinated, coupling relationships between fault location, fault isolation, and service restoration should be considered, and flexible strategies should be formed for different situations faced by distribution systems. To this end, a comprehensive resilience-oriented fault location, fault isolation, and service restoration (RO-FLISR) method is proposed in this paper, where, the fault location, fault isolation, and service restoration under extreme events are comprehensively considered at the same time for the first time, and multiple resources are coordinated. Three basic modules are formulated in the RO-FLISR method, i.e., the possible faulted zone determination (PFZD) module which could determine all the possible faulted zones when there are multiple faults and DGs, the line patrolling (LP) module which supports the optimal scheme for line paroling based on crews, and the fault isolation and service restoration (FISR) module in which the operation of MSs and RCSs are coordinated for comprehensive fault isolation and service restoration. The modules are formulated as MILP models. Effectiveness of the proposed method is verified with the IEEE 123-bus system.
- 25 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 12  (3) , 2136-2152
A Comprehensive Resilience-Oriented FLISR Method for Distribution Systems
Under extreme events, distribution systems may suffer blackouts and numerous restoration tasks will compete for limited crews and facilities. To enhance the resilience of distribution systems, multiple resources should be fully coordinated, coupling relationships between fault location, fault isolation, and service restoration should be considered, and flexible strategies should be formed for different situations faced by distribution systems. To this end, a comprehensive resilience-oriented fault location, fault isolation, and service restoration (RO-FLISR) method is proposed in this paper, where, the fault location, fault isolation, and service restoration under extreme events are comprehensively considered at the same time for the first time, and multiple resources are coordinated. Three basic modules are formulated in the RO-FLISR method, i.e., the possible faulted zone determination (PFZD) module which could determine all the possible faulted zones when there are multiple faults and DGs, the line patrolling (LP) module which supports the optimal scheme for line paroling based on crews, and the fault isolation and service restoration (FISR) module in which the operation of MSs and RCSs are coordinated for comprehensive fault isolation and service restoration. The modules are formulated as MILP models. Effectiveness of the proposed method is verified with the IEEE 123-bus system.
- 15 November 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 11  (3) , 2381-2395
Enhancing Distribution System Resilience With Proactive Islanding and RCS-Based Fast Fault Isolation and Service Restoration
- 15 November 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 11  (3) , 2381-2395
Enhancing Distribution System Resilience With Proactive Islanding and RCS-Based Fast Fault Isolation and Service Restoration
- 14 January 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 10  (6) , 5920-5929
Estimation of Target Appliance Electricity Consumption Using Background Filtering
Measurement of the electricity consumption of major appliances in different time segments is of crucial significance to demand-side management and energy conservation. Non-intrusive load monitoring (NILM) can infer the target appliances’ power use information by only collecting and analyzing the aggregate power data at the single power entrance point. Inspired by the success of deep neural network in other fields, some researchers have applied it to NILM with promising results. However, existing studies require labeled real aggregate data to train the networks, while time-synchronized measurement of the target appliance for labelling is hard to achieve in practice. This paper proposes to train networks with only synthetic aggregate data. Furthermore, a training data generation method via background filtering is proposed, and the obtained training data is used to train the network for estimating electricity consumption. This generation method only needs unlabeled real aggregate data and the target appliance’s operation curves, which reduces the difficulty of training data acquisition. The proposed estimation method achieves higher accuracy than current methods in tests on a public dataset which also demonstrates the effectiveness of background filtering.
- 14 January 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Smart Grid
- Vol. 10  (6) , 5920-5929
Estimation of Target Appliance Electricity Consumption Using Background Filtering
Measurement of the electricity consumption of major appliances in different time segments is of crucial significance to demand-side management and energy conservation. Non-intrusive load monitoring (NILM) can infer the target appliances’ power use information by only collecting and analyzing the aggregate power data at the single power entrance point. Inspired by the success of deep neural network in other fields, some researchers have applied it to NILM with promising results. However, existing studies require labeled real aggregate data to train the networks, while time-synchronized measurement of the target appliance for labelling is hard to achieve in practice. This paper proposes to train networks with only synthetic aggregate data. Furthermore, a training data generation method via background filtering is proposed, and the obtained training data is used to train the network for estimating electricity consumption. This generation method only needs unlabeled real aggregate data and the target appliance’s operation curves, which reduces the difficulty of training data acquisition. The proposed estimation method achieves higher accuracy than current methods in tests on a public dataset which also demonstrates the effectiveness of background filtering.