Gravity energy storage distributed power generation control method

Abstract

The invention provides a gravity energy storage distributed power generation control method, and relates to the technical field of intelligent control, and the method comprises the steps: calculating a gravitational potential energy value based on pre-stored train quality data and downlink height data collected in real time; the train is controlled to move downwards, and the gravitational potential energy value is converted into mechanical energy; the mechanical energy is received, and the number of power stations is calculated by dividing the total length of the downhill rail by a preset power station spacing value; deploying power station groups distributed at equal intervals based on the number of power stations, and converting the mechanical energy into distributed electric energy according to the receiving strength proportion of each power station; the distributed electric energy is received, the load state of the power grid is calculated in real time, when the load value is higher than a preset threshold value, storage battery capacity calculation is executed, all input electric energy is collected to a total storage battery to be stored, and when the load value is lower than the preset threshold value, grid-connected power calculation is executed, and the input electric energy is connected into the power grid. The peak-valley difference of the power grid can be reduced, and the start-stop frequency of the frequency modulation unit is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control, in particular to a gravity energy storage distributed power generation control method. BACKGROUND

[0002] As a physical energy storage technology, gravity energy storage realizes the mutual conversion of electric energy and potential energy through the vertical displacement of heavy objects (such as trains). The traditional gravity energy storage system usually adopts the following technical scheme: For example, a single power station is set at the bottom of the slope, and the gravity potential energy is converted into electric energy through a mechanical transmission device when the train descends. Therefore, the long-distance transmission loss of mechanical energy is as high as 15%-20%, which may lead to low overall efficiency, and a single point failure will cause the entire system to malfunction, and the reliability is insufficient.

[0003] For example, the number and location of power stations are fixed, which cannot adapt to different lengths of slope tracks. Therefore, the power station is surplus on a short slope, causing resource waste, and the coverage is insufficient on a long slope, leading to insufficient energy collection, uneven mechanical energy distribution, and decreased energy conversion rate in the section far from the power station.

[0004] For another example, the electric energy is directly connected to the power grid or stored in a fixed capacity battery, and there is a lack of dynamic scheduling mechanism. Therefore, when the power grid is connected during the peak period, the power grid fluctuation is aggravated, and the risk of overcharging the battery during the storage period is increased. If a load response mechanism is not established, the power grid regulation function of peak load shifting may not be realized. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a gravity energy storage distributed power generation control method, which can reduce the peak-valley difference of the power grid and reduce the start-stop frequency of the frequency modulation unit.

[0006] To solve the above technical problems, the technical scheme of the present application is as follows: The gravity energy storage distributed power generation control method comprises the following steps: Step 1: based on the pre-stored train mass data and the real-time collected descending height data, the gravity potential energy value is calculated; the train is controlled to move downward, and the gravity potential energy value is converted into mechanical energy; Step 2: receiving the mechanical energy output by step 1, combining the preset power station spacing value, calculating the number of power stations by dividing the total length of the downhill track by the spacing value; based on the number of power stations, deploying an equidistant distributed power station group, and converting the mechanical energy into distributed electric energy according to the receiving intensity proportion of each power station; Step 3: receiving the distributed electric energy output by step 2, and calculating the load state of the power grid in real time: When the load value is higher than the preset threshold value, the battery capacity calculation is performed, and all input electric energy is collected to the total battery storage; When the load value is lower than the preset threshold value, grid-connected power calculation is performed to integrate the input electric energy into the power grid.

[0007] Further, step 1, based on the pre-stored train mass data and the real-time collected descending height data, the gravitational potential energy value is calculated, including: Step 11, access the pre-stored train attribute database, and match the corresponding mass raw data according to the current train number; Step 12, unit uniformization calculation is performed on the output mass raw data to convert it into a standard mass unit value; Step 13, real-time collection of the first elevation data of the train descending starting point and the second elevation data of the current position; Step 14, difference calculation is performed on the output first elevation data and second elevation data to generate a dynamic descending height value; Step 15, receiving the output standard mass unit value and the output dynamic descending height value, performing scalar product calculation, outputting real-time gravitational potential energy value.

[0008] Further, the gravitational potential energy value is converted into mechanical energy by controlling the train descending motion, including: Step 16, collect historical running data set, including train gravitational potential energy value, actual descending speed value and corresponding mechanical energy conversion efficiency value; Step 17, performing normalization calculation on the historical running data set to generate a standardized training set after eliminating the dimensional differences; based on the standardized training set, the energy-speed mapping reference parameters are iteratively calculated by the minimum energy loss optimization algorithm; Step 18, real-time collection of track friction coefficient and environmental wind speed data, combined with the reference parameters to perform dynamic compensation calculation to obtain the compensation result; Step 19, generating target speed value according to the compensation result, and converting it into acceleration control instruction of train driving system; through kinetic energy conversion device, the current gravitational potential energy value is converted into mechanical energy according to the preset mechanical conversion ratio.

[0009] Further, step 2, receiving the mechanical energy output by step 1, combining the preset power station spacing value, calculating the number of power stations by dividing the total length of the downhill track by the spacing value, including: Step 21, receiving mechanical energy data stream, extracting real-time intensity value of mechanical energy; Step 22, calling the pre-set power station spacing value database to obtain the standard spacing value corresponding to the current track; Step 23, real-time measurement of the distance value from the starting point to the end point of the downhill track, combined with the output standard spacing value, calculating and generating the integer value of the number of power stations.

[0010] Further, based on the number of power stations, the power station group is deployed with equidistant distribution, and the mechanical energy is converted into distributed electric energy according to the proportion of the receiving intensity of each power station, including: Step 24, receiving the integer value of the number of power stations output by step 23; Step 25, according to the starting point coordinates and end point coordinates of the downhill track, combining the number of power stations, the linear interpolation algorithm is used to calculate the equidistant deployment position coordinate set of each power station; Step 26, receiving the real-time intensity value of mechanical energy output by step 21, combining the position coordinate set to calculate the mechanical energy receiving proportion coefficient of each power station; Step 27, based on the proportion coefficient set, the total mechanical energy is proportionally distributed to the corresponding power station; Step 28, control each power station to convert the distributed mechanical energy into electric energy through the generator, wherein the electric energy output value is proportional to the input mechanical energy intensity; Step 29, aggregating the electric energy output by all power stations to generate a distributed electric energy data stream.

[0011] Further, step 3, receiving the distributed electric energy output by step 2, real-time calculating the load state of power grid, including: Step 31, receiving the distributed electric energy data stream output by step 29, extracting the total power input value of power grid at the current time; Step 32, collecting the historical load data set of power grid, including the time series data of power consumption, seasonal factor and industrial activity index; Step 33, based on the historical load data set of power grid, the benchmark load value at the current time is calculated through the time series prediction model; Step 34, calculating the real-time load deviation value by subtracting the benchmark load value from the total power input value of power grid; Step 35, performing sliding average filtering calculation on the load deviation value to generate a smooth load value after eliminating instantaneous fluctuations; Step 36, comparing the smooth load value with the preset threshold interval to output the load state label of power grid.

[0012] Further, when the load value is higher than the preset threshold value, the battery capacity calculation is performed, and all input electric energy is collected to the total battery storage, including: Receiving the "peak period" load state label output by step 36; Calling the real-time state data set of battery pack, including the residual capacity value, the maximum safe charging power value and the temperature value; Based on the output state data and the total power input value of power grid in step 31, the charging power distribution value of each battery unit is calculated through the dynamic capacity distribution algorithm; Control the electric energy collection system to store the input electric energy to the corresponding battery unit according to the output distribution value; Real-time verification of the capacity growth rate of the stored battery pack after storage deviates from the expected value by ≤5%, and the output power storage completion state.

[0013] Further, when the load value is lower than the preset threshold, perform grid-connected power calculation to input the power into the power grid, including: Receiving the "low valley period" load state label output by step 36; Collecting real-time parameters of the power grid, including voltage value, frequency value and phase angle; Based on the input total power value and the power grid parameters of step 31, calculate the grid-connected power set value; Adjust the inverter output to synchronize the power parameters with the power grid parameters, and perform grid-connected operation according to the set value; Monitoring the power grid frequency fluctuation after grid connection ≤0.5Hz, outputting the grid connection success state.

[0014] The above scheme of the present application at least includes the following beneficial effects: Based on the pre-stored train mass data and the real-time collected descending height data, the gravitational potential energy is calculated, which can accurately match the actual running state and improve the energy conversion efficiency; by controlling the train descending movement to realize the conversion of potential energy into mechanical energy, the real-time and dynamic response ability of energy utilization is ensured.

[0015] According to the distance value between power stations, equidistantly distributed power station groups are calculated and deployed, which can adapt to different terrain and power grid layout requirements, realize flexible planning of distributed power generation, and optimize energy distribution strategy to avoid local overload or energy waste.

[0016] Real-time monitoring of the load state of the power grid, when the load is higher than the threshold, the power is collected to the battery storage, which can relieve the power supply pressure of the power grid and avoid overload; when the load is lower than the threshold, perform grid-connected power calculation to realize timely consumption of the power and improve the stability of the power grid operation and the power utilization rate.

[0017] The calculation method combining pre-stored data and real-time collected data can reduce the influence of data error on the system and enhance the operation reliability; the deployment mode of the distributed power station group can disperse the construction and maintenance cost, and at the same time, through the intelligent switching of energy storage and grid connection, the energy loss is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 It is the flowchart of the gravity energy storage distributed power generation control method provided by the embodiment of the present application.

[0019] Figure 2 It is the flowchart of calculating the gravitational potential energy value based on the pre-stored train mass data and the real-time collected descending height data of the present application.

[0020] Figure 3 This is a schematic diagram of the process of controlling the downward movement of the train and converting gravitational potential energy into mechanical energy according to the present invention. Detailed Implementation

[0021] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0022] like Figure 1 As shown, embodiments of the present invention propose a gravity energy storage distributed generation control method, including: Step 1: Calculate the gravitational potential energy value based on the pre-stored train mass data and the real-time collected downhill altitude data; control the train's downhill motion to convert the gravitational potential energy value into mechanical energy; Step 2: Receive the mechanical energy output from Step 1, and calculate the number of power stations by dividing the total length of the downhill track by the preset power station spacing value; deploy an equidistantly distributed power station group based on the number of power stations, and convert the mechanical energy into distributed electrical energy according to the receiving intensity ratio of each power station. Step 3: Receive the distributed energy output from Step 2 and calculate the grid load status in real time. When the load value is higher than the preset threshold, the battery capacity is calculated and all input electrical energy is collected into the total battery storage. When the load value is lower than the preset threshold, the grid-connected power calculation is performed, and the input electrical energy is connected to the grid.

[0023] In the embodiment of the present application, the gravitational potential energy value is accurately calculated by pre-stored train mass data and real-time collected descending height data, and the gravitational potential energy is converted into mechanical energy by controlling the train descending movement. This method can maximize the use of gravitational potential energy, significantly improve the energy conversion efficiency compared with the traditional energy conversion method, and reduce energy loss. The number of power stations is calculated in combination with the pre-set power station spacing value, and the equidistantly distributed power station group is deployed, so that the power generation facilities can be reasonably arranged according to the terrain and energy transmission demand; at the same time, the mechanical energy is converted into distributed electric energy according to the receiving intensity of each power station, which balances the work load of each power station, improves the spatial adaptability of the power generation system, and reduces the equipment maintenance cost. The real-time load state of the power grid is calculated, when the load value is higher than the pre-set threshold value, the electric energy is collected to the total storage battery to relieve the power supply pressure of the power grid; when the load value is lower than the pre-set threshold value, the electric energy is merged into the power grid to fully utilize the excess electric energy, realize the dynamic balance of power supply and demand, and ensure the stable operation of the power grid. The control method forms a complete and efficient power generation system through accurate energy conversion, reasonable power station layout and intelligent power grid response, improves the system reliability, reduces the economic loss caused by energy waste and unreasonable operation of equipment, and reduces the energy storage and transmission cost.

[0024] In a preferred embodiment of the present application, step 1, based on pre-stored train mass data and real-time collected descending height data, the gravitational potential energy value is calculated, comprising: Step 11, accessing the pre-stored train attribute database, matching the corresponding mass raw data according to the current train number; Step 12, unit uniformization calculation is performed on the output mass raw data to convert it into a standard mass unit value; Step 13, real-time collection of the first elevation data of the train descending starting point and the second elevation data of the current position; Step 14, difference calculation is performed on the output first elevation data and second elevation data to generate a dynamic descending height value; Step 15, receiving the output standard mass unit value and the output dynamic descending height value, performing scalar product calculation, outputting the real-time gravitational potential energy value.

[0025] In the embodiment of the present application, the train attribute database stores the mass raw data corresponding to different numbered trains, and when the gravitational potential energy of a certain train needs to be calculated, the system will search in the database according to the current train number. For example, the train number "001" in the database corresponds to the mass raw data "50 tons", and the system can obtain the mass raw data of "50 tons" by matching the train number "001".

[0026] The unit of the acquired quality raw data may be different, such as "tons" or "kilograms", and the system will perform unit unification calculation according to the standard quality unit (such as kilograms). If the quality raw data is "50 tons", because 1 ton is equal to 1000 kilograms, "50 tons" is multiplied by 1000 to convert the value into the standard quality unit value of "50000 kilograms".

[0027] In the above step 13, the first elevation data of the starting point of the train descending is collected in real time by using an elevation measuring device (such as high-precision GPS or altitude sensor) when the train starts to descend, and the altitude of the train at the initial position is recorded. The second elevation data of the current position of the train is continuously collected as the train descends, and the real-time altitude of the train during the descending process is continuously obtained.

[0028] In the above step 14, the first elevation data and the second elevation data are calculated by difference. Assuming that the first elevation data is 1000 meters and the second elevation data is 800 meters, 1000 meters is subtracted from 800 meters to obtain 200 meters, which is the dynamic descending height value of the train from the starting point to the current position.

[0029] In the above step 15, the standard quality unit value after unit unification processing and the dynamic descending height value calculated are calculated according to the logic of scalar multiplication. For example, the standard quality unit value is 50000 kilograms, and the dynamic descending height value is 200 meters. By multiplying the two, the real-time gravitational potential energy value is obtained.

[0030] In the above step 11, the quality raw data is accurately matched by the train number to avoid data calling errors. In the above step 12, the unit unification calculation is performed to eliminate the data confusion caused by different units and ensure that the quality data is accurate and available. In the above steps 13 to 15, the elevation data is collected in real time, the dynamic descending height is calculated, and the standard quality unit value is calculated by scalar multiplication. The gravitational potential energy value can be dynamically updated during the descending process of the train, the energy state of the train at different positions can be accurately reflected, and the real-time and accuracy of the gravitational potential energy calculation are improved. The above accurate data processing and calculation process makes the gravitational potential energy calculation result more reliable.

[0031] In a preferred embodiment of the present application, the descending movement of the train is controlled, and the gravitational potential energy value is converted into mechanical energy, comprising: Step 16, collect historical running data set, including train gravitational potential energy value, actual descending speed value and corresponding mechanical energy conversion efficiency value; Step 17, perform normalization calculation on the historical running data set to generate a standardized training set after eliminating the dimensional difference; based on the standardized training set, the energy-speed mapping reference parameters are iteratively calculated by a minimum energy loss optimization algorithm; Step 18, real-time acquisition of track friction coefficient and environmental wind speed data, combined with the reference parameters to perform dynamic compensation calculation to obtain the compensation result; Step 19, generating target speed value according to the compensation result, and converting it into acceleration control instruction of train driving system; converting the current gravitational potential energy value into mechanical energy according to the preset mechanical conversion ratio through kinetic energy conversion device.

[0032] In the embodiment of the present application, the above-mentioned step 16 retrieves the records of the train in different running stages from the historical operation database, collects the gravitational potential energy value of the train, such as the gravitational potential energy of a certain value at a certain position in a certain operation; the actual descending speed value, such as the running speed of the train at that time is how many kilometers per hour; and the corresponding mechanical energy conversion efficiency value, that is, the proportion of gravitational potential energy converted into mechanical energy in this state, and integrates these data to form a historical operation data set.

[0033] The above-mentioned step 17, since the gravitational potential energy value, descending speed value and mechanical energy conversion efficiency value in the historical operation data set are of different dimensions, in order to facilitate analysis and processing, the system adopts normalization method to map the data to a specific interval, for example, all data are mapped between 0-1, eliminating the dimensional difference, generating a standardized training set, then using the minimum energy loss optimization algorithm to analyze the standardized training set, trying different parameter combinations through multiple iterations, finding the parameters that can minimize energy loss, and finally determining the energy-speed mapping reference parameters.

[0034] The above-mentioned step 18, through the sensor installed on the track, real-time acquisition of track friction coefficient data, understanding the influence of track surface roughness on train operation; using anemometer to collect environmental wind speed data, mastering the force of wind on train operation, combining the track friction coefficient and environmental wind speed data collected with the reference parameters obtained in step 17, calculating according to the pre-set compensation rules, considering the influence of friction and wind speed on train operation and energy conversion, and obtaining the dynamic compensation result.

[0035] The above-mentioned step 19, according to the compensation result of step 18, combining the energy-speed mapping relationship, determining the target speed value that the train should reach at present, and then converting the target speed value into acceleration control instruction that can be recognized by the train driving system, and sending it to the train driving system to control the train to run at the target speed, at the same time, using the kinetic energy conversion device, converting the currently calculated gravitational potential energy value into mechanical energy according to the preset mechanical conversion ratio.

[0036] The present application can fully utilize the historical data to mine the energy conversion rule, optimize the energy-speed mapping relationship, reduce the energy loss, and improve the efficiency of the conversion of gravitational potential energy into mechanical energy by collecting historical operation data sets and performing normalization processing, and determining the benchmark parameters by combining the minimum energy loss optimization algorithm; Real-time collection of track friction coefficient and environmental wind speed data and dynamic compensation calculation consider the influence of various factors in the actual running environment, making the train operation control and energy conversion more in line with the actual situation; According to the compensation result, the target speed value is generated and converted into acceleration control instruction, which can accurately control the downward movement of the train, ensure the train to run at the optimal speed, and then ensure the gravitational potential energy to be converted into mechanical energy at a predetermined proportion, and improve the stability and accuracy of the whole system operation.

[0037] In a preferred embodiment of the present application, step 2, receiving mechanical energy output by step 1, combining a preset power station spacing value, calculating the number of power stations by dividing the total length of the downhill track by the spacing value, including: Step 21, receiving mechanical energy data stream, extracting real-time intensity value of mechanical energy; Step 22, calling a preset power station spacing value database to obtain a standard spacing value corresponding to the current track; Step 23, real-time measurement of the distance value from the starting point to the ending point of the downhill track, combining the output standard spacing value to calculate and generate an integer value of the number of power stations.

[0038] In the embodiment of the present application, step 21, the mechanical energy data stream output by step 1 is received through the data interface, which contains the changing mechanical energy information during the train descending process, and the system filters and extracts the key data reflecting the current mechanical energy size from the data stream, that is, the real-time intensity value of mechanical energy. For example, the data stream contains the time stamp, conversion state and other information of mechanical energy, and the system extracts only the numerical part representing the intensity of mechanical energy through specific data recognition rules.

[0039] Step 22, access the preset power station spacing value database, which stores the standard spacing values corresponding to different track types or running scenarios. The system retrieves in the database according to the track identification or running scenario information of the current train, for example, if the current track is a mountain heavy haul track, the system finds the matching record in the database and obtains the standard spacing value corresponding to the track, such as "set a power station every 2 kilometers".

[0040] The step 23 is used to measure the actual distance value from the start point to the end point of the downhill track in real time by using the distance measuring device (such as laser range finder or satellite positioning-based distance measuring system) installed along the track, and the obtained distance value is compared with the standard interval value obtained in the step 22. Assuming that the total length of the downhill track is 10 kilometers and the standard interval value is 2 kilometers, the calculation result is 5 by dividing 10 by 2, and the final generated integer value of the power station number is 5 after the rounding processing.

[0041] The present application can flexibly adjust the layout of the power station according to the actual track situation by measuring the length of the downhill track in real time and combining the standard interval value of the corresponding track to calculate the number of power stations, avoid excessive dense or sparse power stations, and make the power station distribution more suitable for energy transmission and conversion needs; the real-time intensity value of the mechanical energy can be extracted, the actual situation of the mechanical energy can be fully considered when calculating the number of power stations, the energy can be more reasonably distributed, the energy transmission loss can be reduced, and the energy utilization efficiency of the entire power generation system can be improved; based on the preset power station interval value database, the corresponding standard interval value can be quickly obtained according to different track types and running scenes, so that the system can adapt to various complex environments; at the same time, the database can be updated and maintained at any time, which facilitates the addition of new track types or the adjustment of the standard interval value, and enhances the scalability of the system.

[0042] In a preferred embodiment of the present application, the power station group is deployed in equidistant distribution based on the number of power stations, and the mechanical energy is converted into distributed electrical energy according to the receiving intensity proportion of each power station, comprising: Step 24, receiving the integer value of the number of power stations output by the step 23; Step 25, calculating the equidistant deployment position coordinate set of each power station by linear interpolation algorithm according to the start point coordinate and the end point coordinate of the downhill track and combining the number of power stations; Step 26, receiving the real-time intensity value of the mechanical energy output by the step 21, and calculating the mechanical energy receiving proportion coefficient of each power station according to the position coordinate set; Step 27, distributing the total mechanical energy to the corresponding power station according to the proportion coefficient set; Step 28, controlling each power station to convert the distributed mechanical energy into electrical energy through the generator, wherein the electrical energy output value is proportional to the input mechanical energy intensity; Step 29, aggregating the electrical energy output by all power stations to generate a distributed electrical energy data stream.

[0043] In the embodiment of the present application, the step 24 receives the integer value of the number of power stations calculated by the step 23 through the data transmission channel, which clearly indicates the total number of power stations to be deployed on the current downhill track.

[0044] The step 25 obtains the start point coordinate and the end point coordinate of the downhill track, which determines the spatial range of the downhill track. Based on the integer value of the number of power stations, the total length of the downhill track is equally divided according to the number of power stations by using a linear interpolation algorithm. For example, if the start point coordinate of the downhill track is (0, 0) and the end point coordinate is (10, 0), and the number of power stations is 5, then the total length of 10 is evenly divided into 4 segments, each with a length of 2.5, and the deployment position coordinates of each power station are calculated in turn to generate an equidistant deployment position coordinate set [(2.5, 0), (5, 0), (7.5, 0)].

[0045] The step 26 receives the real-time mechanical energy intensity value output by the step 21, combines the position coordinate set obtained by the step 25, and comprehensively considers the position of each power station on the track and the attenuation of mechanical energy in the transmission process to calculate the mechanical energy receiving proportion coefficient of each power station. For example, the receiving proportion coefficient of a power station close to the mechanical energy input point may be larger, and the receiving proportion coefficient of a power station far away may be relatively smaller. By analyzing the relationship between mechanical energy distribution and power station position, the receiving proportion coefficient of each power station is determined.

[0046] The step 27 distributes the total mechanical energy to each power station according to the proportion coefficient set calculated by the step 26. Assuming that the total mechanical energy is 100 units and the proportion coefficient of a certain power station is 0.2, the mechanical energy allocated to the power station is 100 x 0.2 = 20 units. Similarly, the total mechanical energy is reasonably distributed to the corresponding power station.

[0047] The step 28 sends control instructions to each power station to start the generator in the power station. The generator converts the allocated mechanical energy into electrical energy according to a fixed energy conversion relationship according to the input mechanical energy intensity. The greater the input mechanical energy intensity, the higher the electrical energy output value, which ensures the stability and proportionality of energy conversion.

[0048] The step 29 collects the electrical energy output by all power stations through the power transmission line. The electrical energy with different voltages and currents output by each power station is processed through voltage adjustment and current integration to form a unified format of distributed electrical energy data stream.

[0049] The application calculates the deployment position by using a linear interpolation algorithm according to downhill track coordinates and the number of power stations, ensures equidistant distribution of the power stations, optimizes the spatial layout of the power stations on the track, and improves the spatial utilization rate of energy collection and conversion; combines the real-time intensity value of mechanical energy and the position of the power station to calculate a receiving proportion coefficient, then proportionally distributes the mechanical energy, so that the energy received by each power station is more in line with actual demand, reduces energy transmission loss, and improves the rationality and efficiency of energy distribution; controls each power station to proportionally convert electrical energy according to the intensity of mechanical energy, and aggregates the output electrical energy, thereby ensuring the stability and consistency of the electrical energy output and improving the quality of distributed electrical energy; from the deployment of the power station to the series of operations of energy distribution and conversion, a complete and efficient power generation process is formed, and the operation efficiency of the entire gravity energy storage distributed power generation system is improved.

[0050] In a preferred embodiment of the application, step 3, receiving the distributed electrical energy output by step 2, calculates the real-time load state of the power grid, including: Step 31, receiving the distributed electrical energy data stream output by step 29, extracts the total power input value of the power grid at the current time; Step 32, collecting a historical load data set of the power grid, including time series data of power consumption, seasonal factors and industrial activity indexes; Step 33, based on the historical load data set of the power grid, calculating the reference load value at the current time through a time series prediction model; Step 34, calculating the real-time load deviation value by subtracting the reference load value from the total power input value of the power grid; Step 35, performing a moving average filter calculation on the load deviation value to eliminate transient fluctuations and generate a smoothed load value; Step 36, comparing the smoothed load value with a preset threshold interval to output a power grid load state label.

[0051] In the embodiment of the application, step 31 above receives the distributed electrical energy data stream output by step 29 through a power data transmission channel, which contains various information of the electrical energy output by each power station. The data items related to the power input of the power grid are filtered from the data stream, and the total power input value of the power grid at the current time is extracted through specific data recognition rules. This value represents the total power size of the distributed electrical energy input into the power grid at this moment.

[0052] Step 32 above extracts minute-level or hour-level power consumption data from the past 3-5 years from the power dispatching system database, such as the total power consumption of the power grid every hour from 2022 to 2024; cleans the data, eliminates abnormal values (such as sudden data caused by equipment failure), and fills in missing values through linear interpolation to ensure data continuity; aggregates the data according to the time dimension to form a time series indexed by "date-time", such as recording the power consumption at 8:00, 9:00, etc. every day.

[0053] Seasonal factor quantification method: Divide the whole year into 4 seasons (spring, summer, autumn, winter); calculate the electricity consumption characteristic coefficient of each season, for example, statistics the average electricity consumption of 14:00-16:00 every day in summer (June-August) in the past 3 years, and compare it with the average electricity consumption of the same day and time period in the whole year, to get the electricity consumption increase coefficient in summer high temperature period (such as summer air conditioning load makes the electricity consumption in this period increase by 25%).

[0054] Generate a standardized factor for each season, such as spring factor 1.05 (slightly higher than the annual average), summer factor 1.2 (high temperature electricity peak), winter factor 1.15 (heating electricity peak), which is used to correct the seasonal fluctuations in the prediction model.

[0055] Industrial activity index construction: Collect macroeconomic indicators, such as monthly industrial added value (from the statistics bureau), manufacturing PMI index, and key industry (such as steel, chemical) production rate data; convert industrial added value to relative growth rate, PMI index to expansion / contraction index with 50 as the critical point, and production rate to comprehensive value by industry weight.

[0056] Establish the mapping relationship between industrial activity and electricity load, for example, when the manufacturing PMI index of a certain region rises from 52 to 55, the historical data shows that the industrial electricity consumption increases by 8%-10% synchronously, according to which the influence coefficient of industrial activity index is generated.

[0057] Step 33 above, using the method of "basic time series model + feature engineering", such as LSTM (long short-term memory network) combined with seasonal decomposition (such as STL decomposition), or Prophet model (Facebook open source time series model, which supports seasonal factors and holiday effects by default).

[0058] Model input feature design: Time features, hour-of-day (0-23), day-of-week (Monday to Sunday), month (1-12), whether it is a holiday, etc.

[0059] External features, seasonal factors (such as summer factor 1.2) extracted in step 32, industrial activity index (such as current month industrial added value relative +3%), weather data (temperature, humidity, optional).

[0060] Historical load features, 24-hour, 7-day, and 12-month electricity consumption lag values (such as t-1, t-24, t-168 time point load).

[0061] Map the numerical features such as electricity consumption, industrial activity index to the interval [0, 1], and keep the original proportion of the seasonal factors (such as 1.05, 1.2); Convert day-of-week to one-hot encoding (7-dimensional vector) and month to 12-dimensional vector; Divide the historical data (such as 2022-2024) into training set and test set in the ratio of 8:2, use the training set for model parameter optimization, and use the test set for evaluation of prediction accuracy.

[0062] Model training process, taking LSTM model as an example: Define network structure, set 2 layers of LSTM layer (128 neurons each), 1 layer of fully connected layer for output of benchmark load value, and use ReLU as activation function.

[0063] Loss function setting, use mean square error (MSE) or mean absolute error (MAE) to measure the deviation between predicted value and actual load.

[0064] Optimizer selection, use Adam optimizer, set initial learning rate to 0.001, and reduce learning rate by 0.9 after every 50 iterations.

[0065] Iterative training, input training set features (time features + seasonal factors + industrial activity index + historical load), get predicted value through forward propagation, update weights through back propagation, until loss function converges (such as MSE less than 50kW 2 ).

[0066] Model fusion of seasonal factors and industrial activity index: In the forward propagation of the model, the seasonal factor is integrated into the prediction result as a multiplication factor, for example, the base predicted load is 400kW, and the summer factor is 1.2, then the corrected predicted value is 400×1.2=480kW.

[0067] Industrial activity index as addition or multiplication bias, such as industrial activity index increases by 1 unit, predicted load increases by 20kW (based on linear relationship of historical data statistics).

[0068] Model evaluation and tuning: Use test set to calculate evaluation indicators such as root mean square error (RMSE), mean absolute percentage error (MAPE), and RMSE is not more than 5% of the average actual load, MAPE<8%.

[0069] Dynamic adjustment of hyperparameters, if the model predicts summer load too low, increase the weight of seasonal factor; if the industrial activity index is not significant, adjust the dimension of its input features or introduce interaction terms (such as industrial index × seasonal factor).

[0070] The step 34 calculates the difference between the total power input value extracted in step 31 and the reference load value calculated in step 33; for example, if the total power input value is 500 kW and the reference load value is 400 kW, 100 kW is obtained by subtracting 400 kW from 500 kW, which is the real-time load deviation value at the current time, reflecting the difference between the actual input power and the expected load.

[0071] The step 35 calculates the average value of the load deviation values in the window, replaces the middle value in the window with the average value, and continuously calculates the average value as the window slides forward over time, eliminating transient fluctuations in the data and making the load value more smooth and stable, ultimately generating a smooth load value.

[0072] The step 36 compares the smooth load value generated in step 35 with the pre-set threshold interval, which includes an upper limit value and a lower limit value; if the smooth load value is greater than the upper limit value, the output grid load status label is "high load"; if the smooth load value is less than the lower limit value, the output label is "low load"; if the smooth load value is between the upper and lower limit values, the output label is "normal load".

[0073] The present application can accurately obtain the real-time load status of the power grid by extracting the total power input value, combining historical load data to predict the reference load value, and performing difference calculation and filtering processing, effectively eliminating data fluctuation interference and improving the accuracy of load monitoring; comparing the smooth load value with the pre-set threshold to output the load status label, so that the system can timely and accurately judge the load status of the power grid, and take appropriate measures such as energy storage or grid connection operation when the load is too high or too low, to ensure stable operation of the power grid and avoid power failure caused by abnormal load. Accurate load status of the power grid helps to reasonably allocate power resources. In high load, the energy is stored, and in low load, the energy is connected to the power grid, improving the utilization efficiency of power resources, reducing the cost of power generation, and achieving dynamic balance of power supply and demand.

[0074] In a preferred embodiment of the present application, when the load value is higher than the pre-set threshold, the battery capacity calculation is performed to collect all the input energy to the total battery storage, including: Receiving the "peak period" load status label output by step 36; Calling the real-time status data set of the battery pack, including the remaining capacity value, the maximum safe charging power value and the temperature value; Based on the output status data and the total power input value in step 31, the charging power distribution value of each battery unit is calculated by a dynamic capacity distribution algorithm; The control power collection system stores the input electric energy to the corresponding battery unit according to the distribution value of the output; The real-time verification is performed on the capacity growth rate of the stored battery pack, and the deviation from the expected value is less than or equal to 5%, and the output electric energy storage completion state is output.

[0075] In the embodiment of the present application, the output result of step 36 is monitored in real time through the data transmission channel. When the grid load state is calculated and judged as "peak period", step 36 outputs the corresponding "peak period" load state label. Once the system receives this label information, it will immediately trigger the subsequent battery capacity calculation and electric energy storage process, which serves as a signal to start the energy storage operation.

[0076] The state monitoring database of the battery pack is accessed, which updates the parameters of the battery pack in real time. The remaining capacity value, i.e. the amount of electric energy that the current battery pack can still store, the maximum safe charging power value, i.e. the maximum power allowed to be input under the premise of ensuring the safety of the battery, and the temperature value, which reflects the current working temperature of the battery, are extracted and output synchronously.

[0077] The dynamic capacity allocation algorithm first analyzes the remaining capacity value of each battery unit. The units with high remaining capacity will be allocated relatively high charging power to be quickly filled. The units with low remaining capacity will be allocated lower power to prevent overcharging. In addition, the algorithm will refer to the maximum safe charging power value to ensure that the power allocated to each unit does not exceed its safe threshold. Furthermore, the temperature value will also be taken into account. If the temperature of a certain battery unit is too high, its allocated power will be appropriately reduced to avoid affecting the battery life or causing safety hazards due to temperature rise. Finally, the total power value from step 31 is divided according to the allocation strategy of each unit to obtain the charging power allocation value of each battery unit.

[0078] The calculated charging power allocation values of each battery unit are converted into control instructions and sent to the electric energy collection system. After receiving the instructions, the electric energy collection system adjusts the current, voltage and other parameters in the circuit to control the flow and size of the input electric energy. According to the allocation value, the corresponding power electric energy is accurately delivered to the corresponding battery unit to realize the orderly storage of electric energy.

[0079] During the electric energy storage process, the capacity change of the battery pack is continuously monitored, and the capacity growth rate, i.e. the increase proportion of the capacity of the battery pack per unit time, is calculated in real time. At the same time, according to the previous allocation strategy and input power, the expected capacity growth rate is calculated. The real-time calculated capacity growth rate is compared with the expected value. If the deviation is within 5%, it means that the electric energy storage process meets the expectation, and the system outputs the electric energy storage completion state. If the deviation exceeds 5%, the charging power allocation value is adjusted to correct the storage process.

[0080] The application avoids hidden dangers such as overcharging and overheating by dynamically allocating the capacity by referring to the residual capacity, the maximum safe charging power and the temperature value of the battery pack, ensures that the battery works in a safe state, and prolongs the service life of the battery; the charging power is allocated according to the actual state of each battery unit, so that the battery pack can quickly and evenly complete charging, fully utilizes the excess power during the peak period of the power grid, and improves the overall efficiency of the energy storage system; the deviation of the capacity growth rate from the expected value is verified in real time, and the result is adjusted to ensure the accuracy and stability of the energy storage process.

[0081] In a preferred embodiment of the application, when the load value is lower than the preset threshold, the grid-connected power calculation is performed, and the input power is connected to the power grid, including: receiving the "low valley period" load state label output by step 36; collecting real-time parameters of the power grid, including voltage value, frequency value and phase angle; calculating the grid-connected power set value based on the input total power value of step 31 and the power grid parameters; adjusting the output of the inverter to synchronize the power parameters with the power grid parameters, and performing the grid-connected operation according to the set value; monitoring the frequency fluctuation of the power grid after grid connection ≤0.5Hz, and outputting the grid connection success state.

[0082] In the embodiment of the application, the output result of step 36 is continuously monitored through the data interface, and when the load state of the power grid is calculated and determined as "low valley period", step 36 will output the corresponding state label. Once the system captures the "low valley period" label, the grid-connected power calculation module is activated immediately, and the power grid connection process is started, which is the key signal to trigger the grid-connected operation.

[0083] The power monitoring device (such as multifunctional power meter, phase measurement unit PMU) installed at the grid-connected node is used to collect three key parameters of the power grid in real time: the AC voltage value (such as the real-time fluctuation value of the rated voltage of 220V or 10kV) is obtained through the voltage transformer; the frequency of the power grid is monitored through the frequency sensor (the standard frequency of the power grid in China is 50Hz, and the real-time collected value may fluctuate between 49.9Hz-50.1Hz); the phase angle of voltage and current is calculated through the phase detection circuit (reflecting the power factor state of the power grid), which are synchronously transmitted to the grid-connected control unit.

[0084] The grid-connected control unit first analyzes the input total power value of step 31 (i.e. the total distributed power to be connected to the grid at present), and performs adaptive calculation in combination with the collected grid parameters. For example, when the grid voltage is low, the grid-connected power is appropriately reduced to avoid further drop of the grid voltage; when the frequency is close to the standard value of 50 Hz, the full grid-connected power can be connected. At the same time, the active / reactive power ratio of the grid-connected power is adjusted according to the phase angle to ensure that the power factor of the grid after being connected is above 0.95, and finally the grid-connected power setting value meeting the grid access specification is generated (for example, the 100 kW input power is adjusted to a combination of 95 kW active power + 20 kvar reactive power).

[0085] The grid-connected control system sends adjustment instructions to the inverter to adjust the output characteristics of the inverter through a PID (proportional-integral-derivative) control algorithm: first, synchronize the voltage amplitude so that the output voltage of the inverter deviates from the grid voltage by ≤5%; second, synchronize the frequency by using phase-locked loop (PLL) technology to lock the output frequency of the inverter to the grid frequency (error ≤0.05 Hz); and finally, synchronize the phase angle so that the phase difference between the output voltage of the inverter and the grid voltage is ≤5 degrees. When all three parameters are synchronized, the grid-connected breaker is turned on according to the grid-connected power setting value, and the power injection operation is performed.

[0086] After the grid-connected operation is completed, the system continuously monitors the frequency change of the grid, and the sampling period is set to 100 ms (i.e. 10 frequency data are collected per second). The real-time collected frequency value is compared with the standard value of 50 Hz, and the fluctuation amplitude is calculated. If the frequency fluctuation is ≤0.5 Hz (e.g. in the interval of 49.5 Hz-50.5 Hz) in the last 10 sampling periods, it is determined that the grid-connected operation does not significantly affect the stability of the grid, and the "grid-connected successfully" state is output; if the fluctuation exceeds the threshold, the system automatically adjusts the grid-connected power or temporarily disconnects the grid-connected loop for fault diagnosis.

[0087] By real-time collection of the grid voltage, frequency, and phase angle and dynamic adjustment of the inverter parameters, the synchronization of the electrical parameters when the distributed power is connected to the grid is ensured, the grid oscillation caused by the phase difference or frequency deviation is avoided, the frequency fluctuation caused by the grid connection is controlled within 0.5 Hz, and the stability requirement of the national standard for grid-connected power is met. Based on the real-time grid parameters, the grid-connected power setting value is calculated, the active / reactive power ratio is adaptively adjusted, the power factor of the grid after being connected is kept above 0.95, the reactive power loss is reduced, the power quality is improved, and the voltage deviation or harmonic pollution caused by the grid connection is avoided. During the low load period of the grid, the grid-connected operation is automatically performed to inject the excess power generated by the distributed power generation system into the grid in time, and the power waste is avoided. Through the dynamic power matching algorithm, more than 98% of the input power can be effectively connected to the grid, the efficiency is improved by 12% compared with the traditional fixed power grid-connected mode, and the grid frequency fluctuation is continuously monitored after being connected to the grid. When an abnormality occurs, a quick response (response time ≤200 ms) is made to prevent the fault from being enlarged by adjusting the grid-connected power or disconnecting the connection.

[0088] The above description is merely exemplary of some embodiments of the present disclosure and of the principles thereof. It is to be understood that the present disclosure is not limited to the above-described technical solutions, and the scope of the application involved in the embodiments of the present disclosure is not limited to the technical solutions formed by the specific combinations of the technical features described above. Meanwhile, the present disclosure should also cover other technical solutions formed by the arbitrary combinations of the technical features described above or equivalent features thereof without departing from the above-described inventive concept. For example, the technical solutions formed by the mutual replacement of the above-described features and the technical features disclosed in the embodiments of the present disclosure (but not limited to) having similar functions.

Claims

1. A method of controlling a gravity energy storage distributed power generation, characterized by, The method comprises: Step 1, based on the pre-stored train mass data and the real-time collected descending height data, the gravitational potential energy value is calculated; the train descending movement is controlled, and the gravitational potential energy value is converted into mechanical energy; Step 2, receiving the mechanical energy output by step 1, combining the preset power station spacing value, the number of power stations is calculated by dividing the total length of the downhill track by the spacing value; based on the number of power stations, the power station group is deployed at equal intervals, and the mechanical energy is converted into distributed electric energy according to the receiving intensity proportion of each power station; Step 3, receiving the distributed electric energy output by step 2, the real-time calculation of the power grid load state is performed: When the load value is higher than the preset threshold value, the battery capacity calculation is performed, and all input electric energy is collected to the total battery storage; When the load value is lower than the preset threshold value, the grid-connected power calculation is performed, and the input electric energy is connected to the power grid.

2. The method of claim 1, wherein, Step 1, based on the pre-stored train mass data and the real-time collected descending height data, the gravitational potential energy value is calculated, comprising: Step 11, accessing the pre-stored train attribute database, and matching the corresponding mass original data according to the current train number; Step 12, unit uniformization calculation is performed on the output mass original data, and the standard mass unit value is converted; Step 13, real-time collection of the first elevation data of the train descending starting point and the second elevation data of the current position; Step 14, difference value calculation is performed on the output first elevation data and second elevation data to generate dynamic descending height value; Step 15, receiving the output standard mass unit value and the output dynamic descending height value, performing scalar product calculation, and outputting real-time gravitational potential energy value.

3. The method of claim 2, wherein, Control the train descending movement, convert the gravitational potential energy value into mechanical energy, comprising: Step 16, collecting historical running data set, including train gravitational potential energy value, actual descending speed value and corresponding mechanical energy conversion efficiency value; Step 17, performing normalization calculation on the historical running data set to generate a standardized training set after eliminating the dimensional difference; based on the standardized training set, the energy-speed mapping reference parameter is iteratively calculated by the minimum energy loss optimization algorithm; Step 18, real-time collection of track friction coefficient and environmental wind speed data, combined with the reference parameter to perform dynamic compensation calculation to obtain the compensation result; Step 19, generating target speed value according to the compensation result, and converting it into acceleration control instruction of train driving system; through kinetic energy conversion device, the current gravitational potential energy value is converted into mechanical energy according to the preset mechanical conversion proportion.

4. The method of claim 3, wherein, Step 2, receiving the mechanical energy output by step 1, combining the preset power station spacing value, the number of power stations is calculated by dividing the total length of the downhill track by the spacing value, comprising: Step 21, receiving mechanical energy data stream, extracting real-time intensity value of mechanical energy; Step 22, calling the pre-set power station spacing value database to obtain the standard spacing value corresponding to the current track; Step 23, real-time measurement of the distance value from the starting point to the end point of the downhill track, combined with the output standard spacing value, the integer value of the number of power stations is calculated and generated.

5. The method of claim 4, wherein, Based on the number of power stations, the power station group is deployed at equal intervals, and the mechanical energy is converted into distributed electric energy according to the receiving intensity proportion of each power station, comprising: Step 24, receiving the integer value of the number of power stations output by step 23; Step 25, according to the starting point coordinates and end point coordinates of the downhill track, combined with the number value of the power station, the linear interpolation algorithm is used to calculate the equidistant deployment position coordinate set of each power station; Step 26, receive the real-time mechanical energy intensity value output by step 21, and calculate the mechanical energy receiving proportion coefficient of each power station combined with the position coordinate set; Step 27, based on the proportion coefficient set, the total mechanical energy is proportionally distributed to the corresponding power station; Step 28, control each power station to convert the distributed mechanical energy into electric energy through the generator, wherein the electric energy output value is proportional to the input mechanical energy intensity; Step 29, aggregate the electric energy output by all power stations to generate a distributed electric energy data stream.

6. The method of claim 5, wherein, Step 3, receive the distributed electric energy output by step 2, and calculate the real-time load state of the power grid, including: Step 31, receive the distributed electric energy data stream output by step 29, and extract the total power input value at the current time; Step 32, collect the historical load data set of the power grid, including time series data of power consumption, seasonal factors and industrial activity index; Step 33, based on the historical load data set of the power grid, calculate the reference load value at the current time through the time series prediction model; Step 34, calculate the real-time load deviation value by subtracting the reference load value from the total power input value; Step 35, perform sliding average filtering calculation on the load deviation value to generate a smooth load value after eliminating transient fluctuations; Step 36, compare the smooth load value with the preset threshold interval, and output the power grid load state label.

7. The method of claim 6, wherein, When the load value is higher than the preset threshold, perform battery capacity calculation to collect all input electric energy to the total battery storage, including: Receive the "peak period" load state label output by step 36; Call the real-time state data set of the battery pack, including the remaining capacity value, the maximum safe charging power value and the temperature value; Based on the output state data and the total power input value of step 31, calculate the charging power distribution value of each battery unit through the dynamic capacity distribution algorithm; Control the electric energy collection system to store the input electric energy to the corresponding battery unit according to the output distribution value; Real-time verify the capacity growth rate of the battery pack after storage and the expected value deviation ≤ 5%, output the electric energy storage completion state.

8. The method of claim 7, wherein, When the load value is lower than the preset threshold, perform grid-connected power calculation to input electric energy into the power grid, including: Receive the "trough period" load state label output by step 36; Collect real-time parameters of the power grid, including voltage value, frequency value and phase angle; Based on the input total power value of step 31 and the grid parameters, calculate the grid-connected power set value; Adjust the output of the inverter to synchronize the electric energy parameters with the grid parameters, and perform grid-connected operation according to the set value; Monitor the power grid frequency fluctuation ≤ 0.5 Hz after grid connection, and output the grid-connected success state.

Images