Intelligent dispatching system based on lidar remote sensing atmospheric information and its method thereof

Abstract

The present invention provides an intelligent dispatching system based on LiDAR remote sensing atmospheric information, comprising: a historical database, storing a correlation information related to a historical atmospheric information and a historical solar panel energy generation efficiency; at least one LiDAR device, capturing at least one real-time atmospheric information; a processing unit, optimizing the correlation information to generate an optimized correlation information when a deviation value between the real-time atmospheric information and the historical atmospheric information is greater than the standard range. Besides, according to the deviation value, the processing unit performs a dispatching algorithm for energy dispatch to generate an energy dispatch order; and according to optimized correlation information and the energy dispatch order, an energy storage and energy release device performs the storage of solar energy and / or release of solar energy. The present invention also provides an intelligent dispatching method, thereby improving the energy dispatch efficiency of solar micro grid.

Description

[Technical Field]

[0001] This invention relates to a scheduling system, and more particularly to an intelligent scheduling system and method based on atmospheric information from light telemetry. [Previous Technology]

[0002] To mitigate global climate change caused by greenhouse gases, levying carbon taxes on trading nations has become one of their strategies under international conventions that explicitly stipulate carbon reduction agreements. Amidst the global wave of energy conservation, carbon reduction, and carbon taxation, Taiwanese companies, as an island trading nation, are prioritizing the transformation of local businesses and the development of renewable energy systems such as solar power to reduce carbon emissions. This has become a key development direction for alleviating the carbon tax burden.

[0003] In the manufacturing industry, developing smart microgrids equipped with energy storage systems allows renewable energy to be incorporated into normal factory power supply and operation with maximum utilization efficiency, which not only saves electricity costs but also reduces carbon emissions to a considerable extent. However, when the factory's power supply comes not only from Taiwan Power Company but also from renewable energy, energy storage systems, and other low-carbon energy sources, traditional energy management systems (EMS) face difficulties in dispatching and need to be upgraded to microgrid-level EMS.

[0004] Solar power generation is characterized by its intermittent and random nature. With the impact of global warming, extreme weather events and rapid changes in daily weather, such as sudden cloud cover, intense heat convection caused by the urban heat island effect, and thunderstorms, cause unpredictable fluctuations in the power generation of solar power systems, thus affecting the stability of the power grid. Therefore, how to develop a smart energy dispatching system to assist and improve regional grid resilience and reduce the hidden costs of energy storage systems is a problem that various sectors are striving to solve. [Summary of the Invention]

[0005] The main objective of this invention is to provide an intelligent scheduling system based on atmospheric information obtained through LiDAR telemetry, comprising: a historical database storing correlation data, wherein the correlation data is related to historical atmospheric information and historical solar panel power generation efficiency information; at least one LiDAR unit telemetry of at least one real-time atmospheric information; a processing unit signal-connected to the historical database and the at least one LiDAR unit, wherein when a deviation between the real-time atmospheric information and the historical atmospheric information exceeds a standard range, the processing unit optimizes the correlation data to generate optimized correlation data; furthermore, the processing unit performs an energy scheduling algorithm based on the deviation to generate an energy scheduling command; and an energy storage and release unit signal-connected to the processing unit, wherein the energy storage and release unit stores and / or releases solar power based on the optimized correlation data and the energy scheduling command.

[0006] In some embodiments, the correlation data is obtained through machine learning.

[0007] In some embodiments, the atmospheric information is atmospheric information remotely measured by the light-receiving unit within a radius of 1 to 10 km.

[0008] In some embodiments, the historical atmospheric information includes: cloud coverage, cloud drift path, cloud optical depth (OD), cloud thickness, wind direction, wind speed, humidity or temperature; or the real-time atmospheric information includes: cloud coverage, cloud drift path, cloud optical depth, cloud thickness, wind direction, wind speed, humidity or temperature.

[0009] In some embodiments, the types of the deviation values ​​include: cloud coverage deviation, cloud drift path deviation, cloud optical thickness deviation, cloud height deviation, wind direction deviation, wind speed deviation, humidity deviation, and temperature deviation.

[0010] In some embodiments, the standard range refers to the mean ± standard deviation of the real-time atmospheric information or the historical atmospheric information.

[0011] In some embodiments, the energy storage and release unit includes: a water electrolyzer, a solar battery or a pumped storage device.

[0012] Another object of the present invention is to provide an intelligent scheduling method, comprising: Step 1: using a historical database to store correlation data, wherein the correlation data is related to historical atmospheric information and historical solar panel power generation efficiency information; Step 2: using at least one light-sensing unit to remotely measure at least one real-time atmospheric information; Step 3: using a processing unit to analyze the difference between the real-time atmospheric information and the historical atmospheric information, wherein if a deviation value between the real-time atmospheric information and the historical atmospheric information is greater than a standard range, the processing unit optimizes the correlation data to generate optimized correlation data, and furthermore, the processing unit performs an energy scheduling algorithm based on the deviation value to generate an energy scheduling command; and Step 4: using an energy storage and release unit, wherein the energy storage and release unit stores and / or releases solar power based on the optimized correlation data and the energy scheduling command.

[0013] In some embodiments, the correlation data is obtained through machine learning.

[0014] In some embodiments, the atmospheric information is atmospheric information remotely measured by the light-receiving unit within a radius of 1 to 10 km.

[0015] In some embodiments, the historical atmospheric information or the real-time atmospheric information includes: cloud coverage, cloud drift path, cloud optical properties, wind direction, wind speed, humidity or temperature.

[0016] The advantages of the present invention are that it optimizes the predicted value of solar power generation by using light telemetry and monitoring atmospheric information, and dynamically allocates energy storage and energy release units through energy dispatching algorithms to enhance the resilience of solar power dispatching and energy utilization.

Implementation Method

[0017] Regarding the technical means of us inventors, several preferred embodiments are described in detail below with reference to the accompanying drawings, so that you may gain a deeper understanding and acceptance of the present invention.

[0018] Given that the large-scale atmospheric information collected by satellite radar often fails to reflect regional weather changes, and that while forecasting modules can provide trends in future solar power generation, they often fail to reflect the actual solar power generation under extreme weather conditions, the inventors propose a solar intelligent dispatching system based on atmospheric information from LiDAR telemetry to enhance the resilience of regional solar photovoltaic microgrid systems in terms of energy dispatching, energy storage, and / or release when facing solar power generation overload or below a threshold.

[0019] A solar power generation system mainly consists of solar photovoltaic cell arrays, batteries, power regulation equipment, and wiring boxes. The principle of solar photovoltaic power generation is to utilize solar panels to absorb sunlight with wavelengths from 0.2μm to 1μm (different materials of solar panels absorb different wavelengths of sunlight), directly converting light energy into electrical energy output. The power generation and efficiency of a solar panel are related to its area size, weather changes (e.g., sunlight intensity), and seasonal changes (e.g., sunshine duration).

[0020] In predicting solar power generation, understanding changes in atmospheric information becomes quite important. Among these, instruments for measuring the optical properties of clouds include, but are not limited to, passive observation instruments such as satellites and photometers, actual sampling equipment such as aircraft and radiosondes, and mobile (active) observation instruments such as radar and light-emitting diodes.

[0021] Solar Panels

[0022] Currently, solar panels can be composed of silicon or compounds as the main materials. In practice, solar panels composed of monocrystalline silicon and polycrystalline silicon are the most common on the market, and their conversion efficiency is also better.

[0023] Among them, monocrystalline silicon solar photovoltaic cells are currently the most efficient crystalline silicon solar photovoltaic cells (approximately 15% to 24%). Both monocrystalline and polycrystalline silicon solar photovoltaic cells are composed of silicon crystalline semiconductors. The difference lies in the following: polycrystalline silicon cells are directly mixed and pressurized after high-purity crystalline silicon is extracted to form a crystalline block, which is then cut into wafers; in contrast, monocrystalline silicon solar photovoltaic cells incorporate a crystal pulling (crystal growth) process during manufacturing, causing the crystallization process to proceed in the same direction, thus resulting in higher photoelectric conversion efficiency, but also relatively higher cost.

[0024] Among them, although the photoelectric conversion efficiency of polycrystalline silicon solar photovoltaic cells (approximately 10 to 17%) is not as high as that of monocrystalline silicon solar photovoltaic cells, the cost is lower and the manufacturing process is relatively simple.

[0025] Guangda

[0026] The atmospheric light scattering signal collected by the LiDAR is basically proportional to the sum of the backscattering coefficients of air molecules and particles in the atmosphere. The light scattering signal received by the LiDAR unit 20 can be described by the "LiDAR Equation", and the optical thickness of gas molecules and particles can be expressed by the relationship, see TW202248579A.

[0027] In addition, the calculation of the total atmospheric backscattering coefficient uses the inversion formula of Klett (1981):

[0028] The background air backscattering coefficient and the total backscattering coefficient at a certain reference height are parameters that must be established before calculation. The parameters are calculated based on daily radiosonde data released by the meteorological bureau, using light intensity measurements under cloudless conditions. Cloud height distribution will be statistically analyzed for the height and optical thickness of clouds below 5 kilometers in the Tainan area, and its correlation with the power generation efficiency of solar panels in the same area will be studied.

[0029] Please refer to Figure 1, which is a schematic diagram of the effect of atmospheric environment on the power output of solar panels (Figure 1 is from the Solar Photovoltaic Information Network of the Energy Bureau of the Ministry of Economic Affairs of Taiwan). It can be seen that the power generation efficiency of solar panels will vary under different cloud cover and solar wavelengths. The cloud height, the optical thickness of clouds in different seasons, and the probability of cloud formation also vary in time, space and region. Therefore, measuring the optical thickness of clouds, their height distribution ratio, cloud characteristics and seasonal distribution will help predict the power generation of solar panels.

[0030] One embodiment of the smart grid of the present invention includes: a solar power generation system, an energy storage system, a mains power system, an emergency load, a non-emergency load, and a control system. Furthermore, a photovoltaic unit is installed adjacent to the solar power generation system. Preferably, the control system is a power control system (PCS). Further, the control system is electrically connected to the solar power generation system, the energy storage system, the mains power system, the emergency load, and the non-emergency load, and the control system is signal-connected to the photovoltaic unit.

[0031] Please refer to Figure 2, which is a schematic diagram of system S according to one embodiment of the present invention. As shown in the figure, the intelligent scheduling system for atmospheric information based on light-based telemetry includes: a historical database 10, at least one light-based unit, a processing unit 30, and an energy storage and release unit 40.

[0032] Please refer to Figure 3, which is a schematic diagram of a method according to an embodiment of the present invention. As shown in the figure, a method for intelligent scheduling includes: Step 1 (S1): using a historical database 10 to store correlation data, wherein the correlation data is related to historical atmospheric information and historical solar panel power generation efficiency information; Step 2 (S2): using at least one light-sensing unit to remotely measure at least one real-time atmospheric information; Step 3 (S3): using a processing unit 30 to analyze the difference between the real-time atmospheric information and the historical atmospheric information. If a deviation value between the real-time atmospheric information and the historical atmospheric information is greater than a standard range, the processing unit 30 optimizes the correlation data to generate optimized correlation data. In addition, the processing unit 30 performs an energy scheduling algorithm based on the deviation value to generate an energy scheduling command; and Step 4 (S4): using an energy storage and release unit 40, wherein the energy storage and release unit 40 stores and / or releases solar power according to the optimized correlation data and the energy scheduling command.

[0033] The historical database 10 stores correlation data, which is related to historical atmospheric information and historical solar panel power generation efficiency information. In some embodiments, the historical database 10 is a server or the cloud, but is not limited thereto. In some embodiments, the historical database 10 is a cloud database with regional attributes (e.g., but not limited to Tainan, Taiwan) to record atmospheric information remotely measured by LiDAR in the long term or short term, and to identify key data affecting solar power generation (e.g., but not limited to cloud cover, cloud thickness, and cloud movement speed) through statistical analysis, and to analyze the impact of changes in these data on solar power generation and the smart grid.

[0034] In some embodiments, the historical atmospheric information includes, but is not limited to: solar irradiance, aerosol concentration, cloud cover, cloud drift path, cloud optical thickness, cloud height, wind direction, wind speed, humidity, or temperature. The solar panel power generation efficiency information is a power output efficiency curve of the solar panel over time, but is not limited thereto. The correlation data is a correlation curve of the atmospheric environment's effect on the power output of the solar panel, including, but not limited to, the relationship between the dynamic changes of clouds in the atmosphere after sunrise and before sunset and the solar power generation efficiency. In some preferred embodiments, the correlation data is obtained through machine learning, specifically by establishing a correlation curve through machine learning that describes the relationship between atmospheric environmental data (such as cloud optical and physical properties) and the power output of the solar panel, while simultaneously establishing a predictive model to predict the future trend of solar power generation.

[0035] Machine Learning includes, but is not limited to: Regression Models, Time-Series Forecasting, or Time-Series Forecasting.

[0036] In order to facilitate the acquisition of regional atmospheric information, the atmospheric information is the atmospheric information remotely measured by the light-emitting unit within a radius of 1 to 10 km, preferably within a radius of 5 km, so as to provide the solar power generation prediction model as the optimization parameter for the prediction value and to closely match the actual value of regional solar power generation.

[0037] The LiDAR unit remotely measures at least one real-time atmospheric information. In some embodiments, a scanning LiDAR system (installed at Kun Shan University of Science and Technology) is used to perform a 3-D cloud scan of the area. The remotely measured regional, real-time atmospheric information includes cloud coverage, cloud drift path, cloud optical thickness, cloud height, wind direction, wind speed, humidity, or temperature. At the same time, solar panel power generation data is collected, and the correlation between the aforementioned atmospheric information and the power generation of the solar power generation area is further statistically analyzed.

[0038] Referring again to Figures 2 and 3, the processing unit 30 is connected to the historical database 10 and the at least one light-emitting unit signal respectively. When the deviation between the real-time atmospheric information and the historical atmospheric information is greater than the standard range (for example, the deviation is greater than a preset range and / or the duration of the deviation is greater than a preset period), the processing unit 30 optimizes the correlation data to generate optimized correlation data. In addition, the processing unit 30 performs an energy scheduling algorithm based on the deviation value to generate an energy scheduling instruction.

[0039] In some embodiments, the processing unit 30 includes, but is not limited to, a microcontroller unit (MCU), an embedded processor (ARM Cortex-A, FPGA), and an industrial PC (IPC). When the processing unit 30 performs a monitoring procedure, such as using a Supervisory Control and Data Acquisition (SCADA) system, and detects that real-time atmospheric information deviates from historical atmospheric information (e.g., the real-time atmospheric information deviates from the standard range of historical atmospheric information (the standard range of historical atmospheric information is, for example, but not limited to, the average value ± 2σ), or the duration of the deviation is greater than a preset period (e.g., more than one week), the processing unit 30 then performs an optimization procedure to optimize the correlation data, such as dynamically correcting the solar power generation trend line, to generate optimized correlation data to improve the accuracy of solar power generation prediction.

[0040] In some embodiments, the machine learning anomaly detection mode further includes: using a long short-term memory (LSTM) or isolation forest model to detect whether there are anomalies in real-time atmospheric information. For example, if the weather pattern for the month should be sunny, but real-time atmospheric information detects a cloud and rain band and thunderstorms that last for half a month, then an anomaly is detected.

[0041] In some embodiments, the intelligent solar power dispatching system based on atmospheric information of the present invention further includes: an alarm unit and an electronic information unit, wherein the alarm unit is signal-connected to the processing unit 30 and the electronic information unit respectively. When the alarm unit receives a warning notification that the real-time atmospheric information deviates from the historical atmospheric information, it immediately notifies the electronic information device (e.g., a smartphone or computer) so that relevant monitoring personnel can be informed and take corresponding actions.

[0042] In some embodiments, the optimization procedure includes, but is not limited to, retraining the correlation data using a random forest regression or long short-term memory (LSTM) time series model, or adjusting the proportion of data used to establish the prediction of solar power generation (e.g., increasing the proportion of cloud thickness on the impact of solar power generation in the case of thunderstorms lasting up to a month), so that the solar power generation prediction trend provided by the prediction model is closer to the actual solar power generation corresponding to atmospheric changes in the current observation area.

[0043] The energy scheduling algorithm performed by the processing unit 30 includes, but is not limited to, machine learning and reinforcement learning (RL) algorithms. In some embodiments, when solar power generation typically peaks during the hot summer months, coinciding with school summer vacation, solar power generation faces the risk of overload. In this case, the portion of the electricity originally allocated to schools for residential use in the energy scheduling algorithm needs to be diverted to other uses or converted to other forms of energy storage to address the aforementioned risk of solar power overload. In addition to considering the deviation value, the scheduling algorithm for energy dispatch also takes into account the overload or the gap in solar power generation below the threshold. It also calculates and predicts when the excess electricity needs to be stored in solar batteries or converted to hydrogen production through water electrolysis (for example, based on historical atmospheric information and real-time atmospheric information, it is predicted that the next month will be sunny and may result in high solar power generation. Therefore, the processing unit 30 generates an energy dispatch instruction to instruct the energy storage and release unit 40 to schedule the energy to reduce the high solar power generation load on the microgrid) or to release the stored backup energy for use.

[0044] Referring again to Figures 2 and 3, the energy storage and release unit 40 is signal-connected to the processing unit 30, wherein the energy storage and release unit 40 stores solar power according to the optimized correlation data and the energy dispatch command. In some embodiments, the energy storage and release unit 40 includes, but is not limited to, a water electrolyzer, a solar battery (e.g., but not limited to LiFePO4), or a pumped hydro storage device. Preferably, the energy storage and release unit 40 is a water electrolyzer, wherein, in order to reduce the high cost required for solar batteries, the use of a water electrolyzer improves the expensive hardware cost of solar power storage and provides a storage form in which excess solar power is converted into hydrogen through water electrolysis.

[0045] Effect verification of intelligent scheduling system.

[0046] Please refer to Figure 4, which shows the performance verification results of one embodiment of the present invention. The data shows the relationship between cloud optical depth (OD) and solar panel power generation (watts (W)). It is evident that the greater the cloud optical depth, the smaller the power generation of the solar panel. In other words, when the sun is blocked by clouds, the power generation of the solar panel will decrease accordingly. Since atmospheric data observed by satellite cannot be used for real-time prediction of the timing and thickness of cloud formations, active real-time sensing devices such as LiDAR are required to obtain relevant dynamic information.

[0047] Referring again to Figure 4, the power generation of solar panels is inversely proportional to the optical thickness of clouds, showing that cloud shading has a crucial impact on power generation. Only by using active sensing devices (such as LiDAR) to measure atmospheric information in real time can we more accurately grasp the actual power generation status of solar energy, thereby improving the accuracy of subsequent power dispatch. In addition, the cloud drift speed is related to when it affects power, which is also crucial for subsequent Automatic Frequency Control (AFC). These are all information that LiDAR can measure and provide in real time.

[0048] In summary, the intelligent scheduling system based on atmospheric information from LiDAR telemetry of the present invention monitors and compares real-time atmospheric information from LiDAR telemetry with historical atmospheric information. If a deviation is detected, the real-time atmospheric information is used as one of the bases for whether to optimize the historical solar panel power generation efficiency information. This allows the predicted solar power generation to be closer to the actual power generation under the current atmospheric information changes, with the assistance of reliable real-time atmospheric information. With the assistance of real-time atmospheric information monitoring, when the solar power generation prediction model predicts that the solar power generation will be overloaded or below the threshold in a specific future period, the processing module performs an energy scheduling algorithm and generates an energy scheduling command to allow the energy storage and release unit 40 to provide corresponding backup processing (e.g., the overloaded solar power is used for water electrolysis to produce hydrogen for storage, or the energy storage and release unit 40 releases the reserved power), thereby improving the resilience, stability, and safety of residential electricity use in the microgrid. Therefore, the invention's purpose can indeed be achieved.

[0049] However, the above description is merely a preferred embodiment of the present invention and should not be construed as limiting the scope of the present invention. Any simple equivalent changes and modifications made in accordance with the claims and description of the present invention are still within the scope of the present invention. Furthermore, no embodiment or claim of the present invention needs to achieve all the objectives, advantages, or features disclosed in the present invention. [Simplified Explanation of the Diagram]

[0050] Figure 1 is a schematic diagram of the effect of atmospheric environment on the power output of solar panels, which describes the intermittent and random characteristics of solar power generation energy.

[0051] Figure 2 is a schematic diagram of a system according to one embodiment of the present invention.

[0052] Figure 3 is a schematic diagram of a method according to one embodiment of the present invention.

[0053] Figure 4 shows the performance verification results of one embodiment of the present invention. [Biomaterial Storage]

[0055] None

Claims

1. An intelligent dispatching system based on atmospheric information from LiDAR telemetry, comprising: a historical database storing correlation data, wherein the correlation data is related to historical atmospheric information and historical solar panel power generation efficiency information; at least one LiDAR unit telemetry of at least one real-time atmospheric information; a processing unit signal-connected to the historical database and the at least one LiDAR unit, wherein when a deviation between the real-time atmospheric information and the historical atmospheric information exceeds a standard range, the processing unit optimizes the correlation data to generate optimized correlation data; furthermore, the processing unit performs an energy dispatching scheduling algorithm based on the deviation to generate an energy dispatching command; and an energy storage and release unit signal-connected to the processing unit, wherein the energy storage and release unit stores and / or releases solar power based on the optimized correlation data and the energy dispatching command. The historical atmospheric information includes: cloud cover, cloud drift path, cloud optical thickness, cloud height, wind direction, wind speed, humidity or temperature; or the real-time atmospheric information includes: cloud cover, cloud drift path, cloud optical thickness, cloud height, wind direction, wind speed, humidity or temperature.

2. The intelligent scheduling system based on atmospheric information from light-based telemetry as described in claim 1, wherein the relevant data is obtained through machine learning.

3. The intelligent scheduling system based on atmospheric information from light-based telemetry as described in claim 1, wherein the atmospheric information is atmospheric information remotely measured by the light-based unit within a radius of 1 to 10 km.

4. The scheduling system based on atmospheric information via LiDAR telemetry as described in claim 1, wherein the types of deviation values ​​include: cloud coverage deviation, cloud drift path deviation, cloud optical thickness deviation, cloud height deviation, wind direction deviation, wind speed deviation, humidity deviation, and temperature deviation.

5. The scheduling system based on light-based telemetry atmospheric information as described in claim 1, wherein the standard range refers to the average value ± standard deviation of the real-time atmospheric information or the historical atmospheric information.

6. The dispatching system based on atmospheric information via LiDAR telemetry as described in claim 1, wherein the energy storage and release unit comprises: a water electrolyzer, a solar battery, or a pumped storage device.

7. A method for intelligent dispatching, comprising: Step 1: using a historical database to store correlation data, wherein the correlation data is related to historical atmospheric information and historical solar panel power generation efficiency information; Step 2: using at least one light-emitting unit to remotely measure at least one real-time atmospheric information; Step 3: using a processing unit to analyze the difference between the real-time atmospheric information and the historical atmospheric information, wherein if a deviation value between the real-time atmospheric information and the historical atmospheric information is greater than a standard range, the processing unit optimizes the correlation data to generate optimized correlation data, and furthermore, the processing unit performs an energy dispatching scheduling algorithm based on the deviation value to generate an energy dispatching command; and Step 4: using an energy storage and release unit, wherein the energy storage and release unit stores and / or releases solar power based on the optimized correlation data and the energy dispatching command; The historical atmospheric information includes: cloud cover, cloud drift path, cloud optical thickness, cloud height, wind direction, wind speed, humidity or temperature; or the real-time atmospheric information includes: cloud cover, cloud drift path, cloud optical thickness, cloud height, wind direction, wind speed, humidity or temperature.

8. The intelligent scheduling method as described in claim 7, wherein the correlation data is obtained through machine learning.

9. The intelligent scheduling method as described in claim 7, wherein the atmospheric information is atmospheric information remotely measured by the light-receiving unit within a radius of 1 to 10 km.

Images