Hui Ma
Hui Ma has a total of 86 co-authors. Here are the top co-authors by number of shared publications:
- Fei Wang
Department of Power Engineering, North China Electric Power University, Baoding 071003, China
4 shared publications - Shenming Fu
National Key Laboratory of Earth System Numerical Modeling and Application Institute of Atmospheric Physics, Chinese Academy of Sciences Beijing China
4 shared publications Xinyu Meng
4 shared publications- Alexis Kai Hon Lau
Hong Kong University of Science and Technology, Hong Kong
3 shared publications Bangru Xiong
3 shared publicationsDehua Zheng
Microgrid Platform R&D Department, New Energy Power Prediction Products Department, Goldwind Science and Etechwin Electric Co., Ltd. Beijing, China
3 shared publications
Latest Publications
- Open Access
- 14 May 2024
- Published by KEAI Publishing in Atmospheric and Oceanic Science Letters
Long-term variations of the solar energy in different subregions of Northwest China and associated mechanisms
As a type of clean and pollution-free energy source, solar energy plays an important role in achieving the goals of carbon neutrality and global sustainable development. Northwest China occupies an important position in the national energy strategy due to its rich solar energy. Clarifying the long-term variations of Northwest China's solar energy and understanding the associated mechanisms are crucial to improving the layout of new energy sources and the usage efficiency of solar energy within China. In this study, we first divide Northwest China into northwestern and southeastern sections by conducting a rotated empirical orthogonal function analysis on the surface solar radiation (SSR) from 1993 to 2022, and then explore the SSR's variation trends and associated mechanisms within these subregions. It is found that the two subregions, both of which show a significant feature of decadal change, differ notably in their long-term trends: the northwestern section shows a significant increasing trend of ∼8.1 kJ m −2 yr −1 in the annual mean SSR, and in each season the SSR increases significantly, with a maximum/minimum increasing rate of ∼11.2/∼4.6 kJ m −2 yr −1 appearing in summer/autumn. A possible mechanism for the SSR's increasing trend is that global warming results in a lower relative humidity within the northwestern section, which decreases the total cloud cover, as it is harder for the atmosphere to reach saturation state. A decreasing total cloud cover results in an increasing SSR within the northwestern section. In contrast, the southeastern section shows no significant trend in annual mean SSR, as the SSRs in summer and autumn show significant decreasing trends, whereas the trends in spring and winter are not significant.
- 14 May 2024
- Published by KEAI Publishing in Atmospheric and Oceanic Science Letters
Open AccessLong-term variations of the solar energy in different subregions of Northwest China and associated mechanisms
As a type of clean and pollution-free energy source, solar energy plays an important role in achieving the goals of carbon neutrality and global sustainable development. Northwest China occupies an important position in the national energy strategy due to its rich solar energy. Clarifying the long-term variations of Northwest China's solar energy and understanding the associated mechanisms are crucial to improving the layout of new energy sources and the usage efficiency of solar energy within China. In this study, we first divide Northwest China into northwestern and southeastern sections by conducting a rotated empirical orthogonal function analysis on the surface solar radiation (SSR) from 1993 to 2022, and then explore the SSR's variation trends and associated mechanisms within these subregions. It is found that the two subregions, both of which show a significant feature of decadal change, differ notably in their long-term trends: the northwestern section shows a significant increasing trend of ∼8.1 kJ m −2 yr −1 in the annual mean SSR, and in each season the SSR increases significantly, with a maximum/minimum increasing rate of ∼11.2/∼4.6 kJ m −2 yr −1 appearing in summer/autumn. A possible mechanism for the SSR's increasing trend is that global warming results in a lower relative humidity within the northwestern section, which decreases the total cloud cover, as it is harder for the atmosphere to reach saturation state. A decreasing total cloud cover results in an increasing SSR within the northwestern section. In contrast, the southeastern section shows no significant trend in annual mean SSR, as the SSRs in summer and autumn show significant decreasing trends, whereas the trends in spring and winter are not significant.
- Open Access
- 12 April 2024
- Published by IOP Publishing in Environmental Research Letters
- Vol. 19 (5) , 054009
Increasing risks of the explosive extratropical cyclones over the North Atlantic storm track: a perspective from their surface wind maxima
Extratropical cyclones play a crucial role in balancing the global momentum, energy, and moisture, and also, they shape the extreme weather events over the extra-tropics. As the strongest category of the extratropical cyclones, the explosive extratropical cyclones (EECs) frequently induce severe disasters through their strong surface winds. During the entire lifespan of an EEC, there is a maximum value of its surface wind (i.e. the maximum surface wind; MSW), which processes the greatest destructive power. After nearly a century of research on the EECs, key features about their MSWs still remain vague. In this study, we systematically investigate the EECs’ MSWs over the North Atlantic storm track (NAST) based on the ERA-Interim reanalysis. It is found that, the average intensity of EECs’ MSWs shows a significant increasing trend of ∼0.3 m s−1 per 10a. More importantly, for the last 20 years, even larger increasing rates of 1.5 m s−1 and 3.5 m s−1 per 10a are found in the average and maximum intensities of the EECs’ MSWs, respectively, implying the EECs’ risks increase notably for the NAST. We further clarify the physical mechanisms governing the production of EECs’ MSWs, and then establish a mechanism-based statistical model, which has the potential to predict the MSWs’ annual average intensity. In summary, our study fills a knowledge gap for the EECs’ MSWs, which would have broad implication of the economics and society.
- 12 April 2024
- Published by IOP Publishing in Environmental Research Letters
- Vol. 19 (5) , 054009
Open AccessIncreasing risks of the explosive extratropical cyclones over the North Atlantic storm track: a perspective from their surface wind maxima
Extratropical cyclones play a crucial role in balancing the global momentum, energy, and moisture, and also, they shape the extreme weather events over the extra-tropics. As the strongest category of the extratropical cyclones, the explosive extratropical cyclones (EECs) frequently induce severe disasters through their strong surface winds. During the entire lifespan of an EEC, there is a maximum value of its surface wind (i.e. the maximum surface wind; MSW), which processes the greatest destructive power. After nearly a century of research on the EECs, key features about their MSWs still remain vague. In this study, we systematically investigate the EECs’ MSWs over the North Atlantic storm track (NAST) based on the ERA-Interim reanalysis. It is found that, the average intensity of EECs’ MSWs shows a significant increasing trend of ∼0.3 m s−1 per 10a. More importantly, for the last 20 years, even larger increasing rates of 1.5 m s−1 and 3.5 m s−1 per 10a are found in the average and maximum intensities of the EECs’ MSWs, respectively, implying the EECs’ risks increase notably for the NAST. We further clarify the physical mechanisms governing the production of EECs’ MSWs, and then establish a mechanism-based statistical model, which has the potential to predict the MSWs’ annual average intensity. In summary, our study fills a knowledge gap for the EECs’ MSWs, which would have broad implication of the economics and society.
- 5 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industry Applications
- Vol. 60 (3) , 5130-5141
CNN-LSTM Networks Based Sand and Dust Storms Monitoring Model Using FY-4A Satellite Data
Improving the accuracy of sand and dust storms monitoring can provide effective support for the safety warning of extreme weather for the power system, which in turn can enhance the power system's emergency supply capacity as well as safeguard the normal production and life of human beings. In this paper, a hybrid sand and dust monitoring model based on one-dimensional convolutional neural network (CNN) and long short-term memory (LSTM) network is proposed. The normalized dust index (NDDI), CNN model and 1DCNN-LSTM hybrid model are utilized in conjunction with Meteosat IV (FY-4A). Sand and dust storms in the Taklamakan Desert in southern Xinjiang were monitored and studied using channel scanning imaging radiometer AGRI (Advanced Geostationary Radiometer) data. The results show that the NDDI sand and dust indices determined from images at different times require the use of different thresholds to identify the sand and dust zones. Recognition errors exist in both covered and desert areas. According to several sand and dust storms event tests, the monitoring model based on the 1DCNN-LSTM network can achieve 91.42% recognition accuracy, which is a stronger monitoring capability compared to the CNN model as well as traditional models. In practical applications, the 1DCNN-LSTM model outperforms the CNN model in dealing with sand and dust and non-sand boundaries. In addition, the 1DCNN-LSTM model can recognize dust storms more accurately under a small amount of cloud occlusion.
- 5 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industry Applications
- Vol. 60 (3) , 5130-5141
CNN-LSTM Networks Based Sand and Dust Storms Monitoring Model Using FY-4A Satellite Data
Improving the accuracy of sand and dust storms monitoring can provide effective support for the safety warning of extreme weather for the power system, which in turn can enhance the power system's emergency supply capacity as well as safeguard the normal production and life of human beings. In this paper, a hybrid sand and dust monitoring model based on one-dimensional convolutional neural network (CNN) and long short-term memory (LSTM) network is proposed. The normalized dust index (NDDI), CNN model and 1DCNN-LSTM hybrid model are utilized in conjunction with Meteosat IV (FY-4A). Sand and dust storms in the Taklamakan Desert in southern Xinjiang were monitored and studied using channel scanning imaging radiometer AGRI (Advanced Geostationary Radiometer) data. The results show that the NDDI sand and dust indices determined from images at different times require the use of different thresholds to identify the sand and dust zones. Recognition errors exist in both covered and desert areas. According to several sand and dust storms event tests, the monitoring model based on the 1DCNN-LSTM network can achieve 91.42% recognition accuracy, which is a stronger monitoring capability compared to the CNN model as well as traditional models. In practical applications, the 1DCNN-LSTM model outperforms the CNN model in dealing with sand and dust and non-sand boundaries. In addition, the 1DCNN-LSTM model can recognize dust storms more accurately under a small amount of cloud occlusion.
- 23 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 512-517
Wind Farm Abnormal Data Cleaning Method Based on Quartiles
With the continuous expansion of wind farms, cleaning up abnormal data has become increasingly crucial. This article introduces a quartile based method for clearing abnormal data from wind power plants. This method divides the data into two categories: normal and abnormal values by calculating the quartiles of the data, and processes the abnormal values. The experimental results indicate that this method can effectively clean up abnormal data of wind power plants, thereby improving the reliability and accuracy of the data.
- 23 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 512-517
Wind Farm Abnormal Data Cleaning Method Based on Quartiles
With the continuous expansion of wind farms, cleaning up abnormal data has become increasingly crucial. This article introduces a quartile based method for clearing abnormal data from wind power plants. This method divides the data into two categories: normal and abnormal values by calculating the quartiles of the data, and processes the abnormal values. The experimental results indicate that this method can effectively clean up abnormal data of wind power plants, thereby improving the reliability and accuracy of the data.
- 22 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 574-578
Dynamic Wind Farm Power Prediction Method Based on Cluster Analysis
This article studies a dynamic wind farm total power prediction method based on clustering analysis, which includes the following steps: obtaining on-site actual operating data of each wind turbine; Obtain the operating characteristics of each wind turbine based on actual on-site operation data, cluster and analyze all wind turbines based on the operating characteristics, and simultaneously update the clustering results by cycle to obtain the power prediction results of each subregion within each cycle; Superposition the power prediction results of each sub region to obtain the total power of the wind farm. Compared with existing technologies, this article has the advantages of quickly describing rapidly changing wind farms and high accuracy in predicting results.
- 22 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 574-578
Dynamic Wind Farm Power Prediction Method Based on Cluster Analysis
This article studies a dynamic wind farm total power prediction method based on clustering analysis, which includes the following steps: obtaining on-site actual operating data of each wind turbine; Obtain the operating characteristics of each wind turbine based on actual on-site operation data, cluster and analyze all wind turbines based on the operating characteristics, and simultaneously update the clustering results by cycle to obtain the power prediction results of each subregion within each cycle; Superposition the power prediction results of each sub region to obtain the total power of the wind farm. Compared with existing technologies, this article has the advantages of quickly describing rapidly changing wind farms and high accuracy in predicting results.
- 26 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 753-758
Simulation Model and Data Acquisition Method for Wind Turbine Group
This article presents a novel approach to simulating and collecting data for a direct-drive wind turbine array. The simulation model for the direct-drive wind turbine array encompasses various components, including the direct-drive fan, a multiplier amplifier, a controlled current source, the line's equivalent impedance, and the power grid. The direct-drive fan is linked to the controlled current source and the line's equivalent impedance, serving to transmit the fan's current signal to the equivalent impedance of the line. Meanwhile, the multiplier amplifier connects to a controlled current source, tasked with multiplying and amplifying the external system's fan current signal based on the predefined number of fans. Subsequently, it transmits the multiplied and amplified fan current signal back to the controlled current source. The controlled current source then utilizes this amplified fan current signal as a reference value to deliver its controlled current source signal to the equivalent impedance of the line. Ultimately, both the controlled current source signal and the fan current signal are channeled into the power grid through the line's equivalent impedance. This innovative approach is highly applicable to conducting system simulation analyses for large-scale wind power integration, resulting in a significant reduction in computational workload and a substantial increase in practical engineering value.
- 26 August 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 753-758
Simulation Model and Data Acquisition Method for Wind Turbine Group
This article presents a novel approach to simulating and collecting data for a direct-drive wind turbine array. The simulation model for the direct-drive wind turbine array encompasses various components, including the direct-drive fan, a multiplier amplifier, a controlled current source, the line's equivalent impedance, and the power grid. The direct-drive fan is linked to the controlled current source and the line's equivalent impedance, serving to transmit the fan's current signal to the equivalent impedance of the line. Meanwhile, the multiplier amplifier connects to a controlled current source, tasked with multiplying and amplifying the external system's fan current signal based on the predefined number of fans. Subsequently, it transmits the multiplied and amplified fan current signal back to the controlled current source. The controlled current source then utilizes this amplified fan current signal as a reference value to deliver its controlled current source signal to the equivalent impedance of the line. Ultimately, both the controlled current source signal and the fan current signal are channeled into the power grid through the line's equivalent impedance. This innovative approach is highly applicable to conducting system simulation analyses for large-scale wind power integration, resulting in a significant reduction in computational workload and a substantial increase in practical engineering value.
- Open Access
- 18 January 2023
- Published by Frontiers Media SA in Frontiers in Earth Science
- Vol. 10
Evolutionary mechanisms of the strong winds associated with an intense cold wave event and their effects on the wind power production
Cold wave events (CWEs) often cause major economic losses and serious casualties in the cold seasons, making CWEs among the most significant types of disastrous weather. Previous studies have mainly focused on disasters due to abrupt drops in surface temperatures, with less discussion of the strong winds associated with CWEs. Based on an intense CWE that occurred in late December 2020, we investigated the evolutionary mechanisms of the associated strong winds in terms of kinetic energy (KE) budget and evaluated the effects of this CWE on wind power production based on quantitative comparisons with the mean state. The results showed that the CWE occurred under favorable background conditions, which were characterized by a southward-moving transversal trough and a southeastward-moving shortwave trough in the middle troposphere. The surface high ridge that formed around Lake Baikal and the cold front around the southern periphery of the ridge were key factors related to the CWE evolution. The positive work carried out on the horizontal wind by the pressure gradient force that linked a lower tropospheric high-pressure ridge inland and a low-pressure trough offshore and the downward momentum transportation due to the descending motions behind the cold front dominated the enhancement and sustainment of the CWE-associated strong winds. The CWE contributed to wind power production by 1) increasing the wind power density (by an average of ∼1.05-fold) and 2) improving the availability of the wind to generate power, as it reduced the percentage of zero wind power generation by ∼6.4%, while maintaining the high-wind-velocity cut-out percentage.
- 18 January 2023
- Published by Frontiers Media SA in Frontiers in Earth Science
- Vol. 10
Open AccessEvolutionary mechanisms of the strong winds associated with an intense cold wave event and their effects on the wind power production
Cold wave events (CWEs) often cause major economic losses and serious casualties in the cold seasons, making CWEs among the most significant types of disastrous weather. Previous studies have mainly focused on disasters due to abrupt drops in surface temperatures, with less discussion of the strong winds associated with CWEs. Based on an intense CWE that occurred in late December 2020, we investigated the evolutionary mechanisms of the associated strong winds in terms of kinetic energy (KE) budget and evaluated the effects of this CWE on wind power production based on quantitative comparisons with the mean state. The results showed that the CWE occurred under favorable background conditions, which were characterized by a southward-moving transversal trough and a southeastward-moving shortwave trough in the middle troposphere. The surface high ridge that formed around Lake Baikal and the cold front around the southern periphery of the ridge were key factors related to the CWE evolution. The positive work carried out on the horizontal wind by the pressure gradient force that linked a lower tropospheric high-pressure ridge inland and a low-pressure trough offshore and the downward momentum transportation due to the descending motions behind the cold front dominated the enhancement and sustainment of the CWE-associated strong winds. The CWE contributed to wind power production by 1) increasing the wind power density (by an average of ∼1.05-fold) and 2) improving the availability of the wind to generate power, as it reduced the percentage of zero wind power generation by ∼6.4%, while maintaining the high-wind-velocity cut-out percentage.
- Open Access
- 4 December 2022
- Published by MDPI AG in Atmosphere
- Vol. 13 (12) , 2034
Improving the Wind Power Density Forecast in the Middle- and High-Latitude Regions of China by Selecting the Relatively Optimal Planetary Boundary Layer Schemes
As the power generation mode with the lowest carbon emissions, wind power generation plays an indispensable role in achieving the goal of carbon neutralization. To optimize the wind power density (WPD), forecasting is crucial to improve wind power utilization and power system stability. However, because near-surface wind is characterized by notable randomness, diversity, intermittence, and uncontrollability, accurately forecasting the WPD on wind farms remains a challenging task. In this study, we attempted to improve the WPD forecast in the middle- and high-latitude regions of China (wind energy resources are abundant there) by selecting the relatively optimal planetary boundary layer (PBL) scheme, as the PBL processes exert notable effects on the near-surface wind field directly. Based on a whole month in the summer (July 2021), seven PBL schemes were compared quantitatively by using the Weather Research and Forecasting (WRF) model for a total of 70 runs (for each run, the forecast period was 3 days). The results show that no PBL schemes could always show the best performance in forecasting all variables, and the forecast accuracy showed a notable dependence on the evolution of the weather systems. Among the seven PBL schemes, the Medium-Range Forecast (MRF) scheme showed the overall best performance in forecasting the 100 m wind speed, sea level pressure, and 2 m temperature, which ensured that it had the highest forecast skill for the WPD in the middle- and high-latitude regions of China. Further analyses indicate that the background conditions were also well forecasted by the MRF scheme (ranked first or second). This was a crucial reason why the WPD forecast was the best for the MRF scheme.
- 4 December 2022
- Published by MDPI AG in Atmosphere
- Vol. 13 (12) , 2034
Open AccessImproving the Wind Power Density Forecast in the Middle- and High-Latitude Regions of China by Selecting the Relatively Optimal Planetary Boundary Layer Schemes
As the power generation mode with the lowest carbon emissions, wind power generation plays an indispensable role in achieving the goal of carbon neutralization. To optimize the wind power density (WPD), forecasting is crucial to improve wind power utilization and power system stability. However, because near-surface wind is characterized by notable randomness, diversity, intermittence, and uncontrollability, accurately forecasting the WPD on wind farms remains a challenging task. In this study, we attempted to improve the WPD forecast in the middle- and high-latitude regions of China (wind energy resources are abundant there) by selecting the relatively optimal planetary boundary layer (PBL) scheme, as the PBL processes exert notable effects on the near-surface wind field directly. Based on a whole month in the summer (July 2021), seven PBL schemes were compared quantitatively by using the Weather Research and Forecasting (WRF) model for a total of 70 runs (for each run, the forecast period was 3 days). The results show that no PBL schemes could always show the best performance in forecasting all variables, and the forecast accuracy showed a notable dependence on the evolution of the weather systems. Among the seven PBL schemes, the Medium-Range Forecast (MRF) scheme showed the overall best performance in forecasting the 100 m wind speed, sea level pressure, and 2 m temperature, which ensured that it had the highest forecast skill for the WPD in the middle- and high-latitude regions of China. Further analyses indicate that the background conditions were also well forecasted by the MRF scheme (ranked first or second). This was a crucial reason why the WPD forecast was the best for the MRF scheme.
- 1 November 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industry Applications
- Vol. 59 (2) , 2210-2224
Prediction Interval Estimation and Deterministic Forecasting Model Using Ground-Based Sky Image
Due to the explosive penetration of photovoltaic (PV) systems in the power grid, accurate and well-informed solar PV power forecasting has become exceptionally crucial, while accurate irradiance prediction is the basis of PV power forecasting. Yet most research on PV power/irradiance forecasting focus on deterministic forecasting, whereas it is difficult to express the credibility of the forecasting results, especially in the scene of instantaneous fluctuation of irradiance and photovoltaic power caused by cloud movement. Therefore, in order to accurately estimate and track the rapid fluctuations of solar irradiance, this paper takes the ground-based sky image information into consideration and obtains the probability-level prediction results along with the deterministic point forecasting, proposes a comprehensive evaluation function constructed with the average coverage and the average width as the fitness function to assist in the optimization of the deterministic forecasting results, and then constructs a three-output model, naming it DenLUBE. In addition, a framework for irradiance variation pattern prediction is proposed to provide pattern information for the forecasting time period to further improve the performance of the ultra-short-term forecasting model. The results show that the proposed model can achieve high quality and strongly robust deterministic forecasting results while producing tight prediction intervals independent of weather conditions. In addition, the performance of our model is greatly enhanced compared to the traditional approach of taking the middle position of the interval as the deterministic forecasting result, as well as other benchmark models.
- 1 November 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industry Applications
- Vol. 59 (2) , 2210-2224
Prediction Interval Estimation and Deterministic Forecasting Model Using Ground-Based Sky Image
Due to the explosive penetration of photovoltaic (PV) systems in the power grid, accurate and well-informed solar PV power forecasting has become exceptionally crucial, while accurate irradiance prediction is the basis of PV power forecasting. Yet most research on PV power/irradiance forecasting focus on deterministic forecasting, whereas it is difficult to express the credibility of the forecasting results, especially in the scene of instantaneous fluctuation of irradiance and photovoltaic power caused by cloud movement. Therefore, in order to accurately estimate and track the rapid fluctuations of solar irradiance, this paper takes the ground-based sky image information into consideration and obtains the probability-level prediction results along with the deterministic point forecasting, proposes a comprehensive evaluation function constructed with the average coverage and the average width as the fitness function to assist in the optimization of the deterministic forecasting results, and then constructs a three-output model, naming it DenLUBE. In addition, a framework for irradiance variation pattern prediction is proposed to provide pattern information for the forecasting time period to further improve the performance of the ultra-short-term forecasting model. The results show that the proposed model can achieve high quality and strongly robust deterministic forecasting results while producing tight prediction intervals independent of weather conditions. In addition, the performance of our model is greatly enhanced compared to the traditional approach of taking the middle position of the interval as the deterministic forecasting result, as well as other benchmark models.
- 9 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 8-14
Sand and Dust Storms Monitoring Using FY-4A Satellite Data based on Convolutional LSTM Networks
Improving the sand and dust storms monitoring capability can provide early warning information for sandstorms, and effectively improve the PV power prediction accuracy under sand and dust storms weather. In this paper, a hybrid model dust monitoring method based on a one-dimensional convolutional neural network (CNN) and a long short-term memory (LSTM) network is proposed. Using the normalized difference dust index (NDDI), CNN model, and 1DCNN-LSTM hybrid model, combined with the number four meteorological satellite (FY-4A). The channel scanning imaging radiometer AGRI (Advanced Geostationary Radiation Imager) data is used to monitor and study the sand and dust storms of the Taklimakan Desert in southern Xinjiang. The results show that the NDDI dust index established by images at different times needs to take different thresholds to identify the dust area. There are misidentifications in the coverage area as well as in the desert area. The sand and dust storms monitoring model established based on CNN network and 1DCNN-LSTM network, the accuracy (Accuracy) and loss function (Loss) of training samples and test samples are 99.9% and 1%, respectively, which has strong sand and dust storms monitoring capabilities. In practical applications, the 1DCNN-LSTM model is better than the CNN model in processing the boundary between sand and non-sand dust. In addition, the 1DCNN-LSTM model can also more accurately identify sand and dust storms in the case of a small amount of cloud occlusion area.
- 9 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 8-14
Sand and Dust Storms Monitoring Using FY-4A Satellite Data based on Convolutional LSTM Networks
Improving the sand and dust storms monitoring capability can provide early warning information for sandstorms, and effectively improve the PV power prediction accuracy under sand and dust storms weather. In this paper, a hybrid model dust monitoring method based on a one-dimensional convolutional neural network (CNN) and a long short-term memory (LSTM) network is proposed. Using the normalized difference dust index (NDDI), CNN model, and 1DCNN-LSTM hybrid model, combined with the number four meteorological satellite (FY-4A). The channel scanning imaging radiometer AGRI (Advanced Geostationary Radiation Imager) data is used to monitor and study the sand and dust storms of the Taklimakan Desert in southern Xinjiang. The results show that the NDDI dust index established by images at different times needs to take different thresholds to identify the dust area. There are misidentifications in the coverage area as well as in the desert area. The sand and dust storms monitoring model established based on CNN network and 1DCNN-LSTM network, the accuracy (Accuracy) and loss function (Loss) of training samples and test samples are 99.9% and 1%, respectively, which has strong sand and dust storms monitoring capabilities. In practical applications, the 1DCNN-LSTM model is better than the CNN model in processing the boundary between sand and non-sand dust. In addition, the 1DCNN-LSTM model can also more accurately identify sand and dust storms in the case of a small amount of cloud occlusion area.