A Novel Data-Driven Method for Behind-the-Meter Solar Generation Disaggregation With Cross-Iteration Refinement

Pan, Keda; Chen, Zhaohua; Lai, Chun Sing; Xie, Changhong; Wang, Dongxiao; Zhao, Zhuoli; Wang, Xiaomei; Tong, Ning; Lai, Loi Lei; Hatziargyriou, Nikos D.

doi:10.1109/tsg.2022.3171656

Cited by 8 publications

(4 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A data-driven method was proposed to estimate the power generation of invisible solar power stations by using the measured values of a few representative stations [8]. Net load and weather data were used to decompose behind-the-meter solar generation based on the LSTM model, without relying on other photovoltaic agents and physical models of photovoltaic panels [9]. Although the above data-driven method does not depend on the assumption of the physical model of the photovoltaic system, the model needs to input data other than the net load data, such as the measured values of representative photovoltaic stations and accurate numerical weather forecast, etc.…”

Section: Introductionmentioning

confidence: 99%

Behind-the-Meter Solar Generation Disaggregation Based on Attention Mechanism and Bidirectional Gated Recurrent Unit

Xu,

Tan,

Zhang

et al. 2023

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

With the rising installed capacity of rooftop PV, there is an urgent need to improve the accuracy of the behind-the-meter solar generation decomposition to realize the local consumption of distributed PV in order to alleviate the grid stability problem caused by the large-scale rooftop PV access. This paper proposes a bidirectional Gated Recurrent Unit neural network based on the Attention mechanism for a behind-the-meter solar generation decomposition model. First, the temporal characteristics of customer net load data are extracted using bidirectional gated recurrent units. Then, the Attention mechanism is introduced to improve the attention to key net load information. Finally, a nonlinear mapping relationship is constructed from net load data to behind-the-meter solar generation. 184 household PV customers in the SGSC dataset are used for the example analysis. The simulation results show that the proposed method does not rely on accurate physical modeling as well as accurate numerical weather forecast data, and has good generalization in scenarios of different climate zones and good adaptability in scenarios of different seasons.

show abstract

Section: Introductionmentioning

confidence: 99%

Behind-the-Meter Solar Generation Disaggregation Based on Attention Mechanism and Bidirectional Gated Recurrent Unit

Xu,

Tan,

Zhang

et al. 2023

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

show abstract

“…These physical models are heavily dependent on the key design parameters of the PV system, so the model itself is the major source of error. Pan et al [9] proposed a BTM PV generation forecasting method using net load and weather data. Khodayar et al [10] used spatiotemporal graph dictionary learning to extract features to estimate BTM PV generation.…”

Section: Introductionmentioning

confidence: 99%

“…Pan et al. [9] proposed a BTM PV generation forecasting method using net load and weather data. Khodayar et al.…”

Section: Introductionmentioning

confidence: 99%

Short‐term prediction of behind‐the‐meter PV power based on attention‐LSTM and transfer learning

Zhang,

Hong,

Ibrahim

et al. 2023

IET Renewable Power Gen

View full text Add to dashboard Cite

Distributed photovoltaic (PV) systems often lack adequate measurements due to cost considerations, which makes it very difficult to predict them accurately. Here, an approach is proposed for behind‐the‐meter (BTM) PV power prediction using attention‐LSTM neural network and transfer learning. First, the weather is classified into four types based on the deviation ratio β. Second, the correlation analysis algorithm identifies the weather factors that contribute the most to PV power generation as GHI, DNI, humidity, and temperature. Then, attention‐LSTM pre‐trained model is constructed, in which the LSTM network fully extracts the temporal characteristics of PV power generation, while the attention mechanism enhances the attention to the important information in the input. Finally, a novel BTM PV short‐term power prediction method using transfer learning that the upper layer parameters of the pre‐trained model were frozen, and a small amount of DP‐NB data was used to fine‐tune the model. The proposed methodology has higher prediction accuracy compared with other benchmark methods under four weather types, and the time cost is saved by freezing the parameters of the first layer of the network by 37.2%.

show abstract

“…After long-term research at home and abroad, many short-term forecasting methods and algorithms have been used to predict PV power in a short time indirectly. Short-term PV power forecasting methods are divided into three types: (1) a conventional statistical model [8]; (2) an artificial intelligence model [9,10], and (3) a combination forecasting model. Much research has noted that no individual forecasting model is the best for all types of load forecasting.…”

Section: Introductionmentioning

confidence: 99%

A Short-Term Photovoltaic Power Forecasting Method Combining a Deep Learning Model with Trend Feature Extraction and Feature Selection

Peng

et al. 2022

Energies

Self Cite

View full text Add to dashboard Cite

High precision short-term photovoltaic (PV) power prediction can reduce the damage associated with large-scale photovoltaic grid-connection to the power system. In this paper, a combination deep learning forecasting method based on variational mode decomposition (VMD), a fast correlation-based filter (FCBF) and bidirectional long short-term memory (BiLSTM) network is developed to minimize PV power forecasting error. In this model, VMD is used to extract the trend feature of PV power, then FCBF is adopted to select the optimal input-set to reduce the forecasting error caused by the redundant feature. Finally, the input-set is put into the BiLSTM network for training and testing. The performance of this model is tested by a case study using the public data-set provided by a PV station in Australia. Comparisons with common short-term PV power forecasting models are also presented. The results show that under the processing of trend feature extraction and feature selection, the proposed methodology provides a more stable and accurate forecasting effect than other forecasting models.

show abstract

A Novel Data-Driven Method for Behind-the-Meter Solar Generation Disaggregation With Cross-Iteration Refinement

Cited by 8 publications

References 34 publications

Behind-the-Meter Solar Generation Disaggregation Based on Attention Mechanism and Bidirectional Gated Recurrent Unit

Behind-the-Meter Solar Generation Disaggregation Based on Attention Mechanism and Bidirectional Gated Recurrent Unit

Short‐term prediction of behind‐the‐meter PV power based on attention‐LSTM and transfer learning

A Short-Term Photovoltaic Power Forecasting Method Combining a Deep Learning Model with Trend Feature Extraction and Feature Selection

Contact Info

Product

Resources

About