Intelligent scheduling method and system for small hydropower station group

Abstract

The invention relates to the technical field of intelligent dispatching of hydropower stations, in particular to an intelligent dispatching method and system for a small hydropower station group. The method comprises the following steps: collecting multi-source operation parameters and unit operation parameters of each operation node in real time in a small hydropower station group, packaging the obtained multi-source operation parameters and unit operation parameters according to a collection period, and uploading the packaged parameters to a centralized control platform; predicting a runoff change trend of a future target time period on the centralized control platform by using the multi-source operation parameters, determining a hydropower station group collaborative scheduling strategy according to the runoff change trend, and calculating a corresponding unit adjustment amount based on the hydropower station group collaborative scheduling strategy and the unit operation parameters; distributing the unit adjusting quantity to each operation node; through runoff trend prediction and group station cooperative scheduling closed-loop control, the adaptive capacity and scheduling stability of the small hydropower station group to runoff fluctuation are enhanced, and the scheduling response capacity and operation stability of the small hydropower station group are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent dispatching technology for hydropower stations, and in particular to an intelligent dispatching method and system for a small hydropower station group. Background Technology

[0002] Existing small hydropower stations are mostly distributed across river basins, characterized by their large number, small scale, and significant differences in operating conditions. They typically rely on manual experience or simple rules for scheduling and control, making it difficult to reflect the dynamic evolution of runoff changes, load fluctuations, and unit operating status in a timely manner. In actual operation, uneven rainfall, unstable inflow, and the coupling of upstream and downstream water levels create significant correlations between power stations in terms of output allocation, water utilization, and regulation rhythm. However, existing scheduling methods often focus on independent control of individual stations, lacking systematic modeling and unified scheduling of the overall collaborative relationship among hydropower station groups. This can easily lead to problems such as localized overload, increased water wastage, or scheduling delays. Furthermore, with the increasing automation and remote operation of small hydropower stations, a large amount of multi-source operating parameters and unit status data are generated during operation. However, this data largely remains at the monitoring level and is not effectively used for predictive analysis and scheduling decisions, resulting in insufficient adaptability of scheduling strategies to future runoff changes. Summary of the Invention

[0003] Therefore, it is necessary to provide an intelligent scheduling method and system for small hydropower station groups to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for intelligent scheduling of small hydropower station groups includes the following steps: Step S1: Collect multi-source operating parameters and unit operating parameters of each operating node in real time in the small hydropower station group, and upload the acquired multi-source operating parameters and unit operating parameters to the centralized control platform after packaging them according to the collection cycle. Step S2: On the centralized control platform, use multi-source operating parameters to predict the runoff change trend for the future target period, determine the coordinated scheduling strategy of the hydropower station group based on the runoff change trend, and calculate the corresponding unit regulation based on the coordinated scheduling strategy of the hydropower station group and the unit operating parameters. Step S3: Distribute the unit adjustment amount to each operating node, execute the adjustment operation according to the unit adjustment amount, and complete the real-time control in the local edge computing node; Step S4: After completing the adjustment operation, collect the adjusted unit operating parameters and upload them to the centralized control platform; the centralized control platform updates the hydropower station group collaborative scheduling strategy according to the uploaded unit operating parameters, and completes the simulation verification of the scheduling strategy in the preset digital twin virtual environment, and redistributes the verified scheduling strategy to each operating node for execution.

[0005] The present invention also provides an intelligent scheduling system for a small hydropower station group, used to execute the intelligent scheduling method for a small hydropower station group as described above. The intelligent scheduling system for a small hydropower station group includes: The data extraction module is used to collect multi-source operating parameters and unit operating parameters of each operating node in a small hydropower station group in real time, and then encapsulate the acquired multi-source operating parameters and unit operating parameters according to the collection cycle and upload them to the centralized control platform. The runoff change trend prediction module is used to predict the runoff change trend in the future target period using multi-source operating parameters on the centralized control platform, determine the coordinated scheduling strategy of the hydropower station group based on the runoff change trend, and calculate the corresponding unit regulation based on the coordinated scheduling strategy of the hydropower station group and the unit operating parameters. The adjustment operation module is used to distribute the unit adjustment amount to each operating node, execute adjustment operations according to the unit adjustment amount, and complete real-time control in the local edge computing node; The scheduling strategy update module is used to collect the adjusted unit operating parameters after the adjustment operation is completed and upload them to the centralized control platform. The centralized control platform updates the hydropower station group collaborative scheduling strategy according to the uploaded unit operating parameters, and completes the simulation verification of the scheduling strategy in the preset digital twin virtual environment. The verified scheduling strategy is then redistributed to each operating node for execution.

[0006] The beneficial effects of this invention are as follows: By constructing a closed-loop scheduling mechanism during the operation of a small hydropower station group—"real-time acquisition of operating parameters—runoff trend prediction—cooperative scheduling calculation—real-time control at the edge—strategy update at the central control platform—digital twin simulation verification"—the scheduling decision-making of the hydropower station group is transformed from static experience-based control to dynamic scheduling control based on real-time operating status and future runoff changes. By uniformly processing multi-source operating parameters and predicting runoff change trends for future target periods at the central control platform, a forward-looking match between scheduling strategies and actual hydrological changes is achieved, reducing the risk of scheduling lag caused by runoff fluctuations. Simultaneously, by introducing local edge computing nodes at each operating node to monitor and automatically correct the unit regulation process in real time, scheduling commands achieve rapid response and adaptive correction capabilities at the execution level, improving the stability and execution accuracy of the unit regulation process. Furthermore, by introducing a digital twin virtual environment at the central control platform, the updated collaborative scheduling strategy of the hydropower station group is simulated and verified, and feasible strategies are selected before being issued for execution, avoiding unreasonable scheduling strategies from directly affecting the actual unit operation, thereby improving the overall safety, coordination, and scheduling reliability of the small hydropower station group. Attached Figure Description

[0007] Figure 1 A flowchart illustrating the steps of an intelligent scheduling method for a small hydropower station group; Figure 2 This is a schematic diagram of the flood warning identification results; Figure 3 A schematic diagram of the intelligent scheduling process for a group of small hydropower stations; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0008] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0009] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0010] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0011] To achieve the above objectives, please refer to Figures 1 to 3 A method for intelligent scheduling of small hydropower station groups includes the following steps: Step S1: Collect multi-source operating parameters and unit operating parameters of each operating node in real time in the small hydropower station group, and upload the acquired multi-source operating parameters and unit operating parameters to the centralized control platform after packaging them according to the collection cycle. Step S2: On the centralized control platform, use multi-source operating parameters to predict the runoff change trend for the future target period, determine the coordinated scheduling strategy of the hydropower station group based on the runoff change trend, and calculate the corresponding unit regulation based on the coordinated scheduling strategy of the hydropower station group and the unit operating parameters. Step S3: Distribute the unit adjustment amount to each operating node, execute the adjustment operation according to the unit adjustment amount, and complete the real-time control in the local edge computing node; Step S4: After completing the adjustment operation, collect the adjusted unit operating parameters and upload them to the centralized control platform; the centralized control platform updates the hydropower station group collaborative scheduling strategy according to the uploaded unit operating parameters, and completes the simulation verification of the scheduling strategy in the preset digital twin virtual environment, and redistributes the verified scheduling strategy to each operating node for execution.

[0012] All specific values ​​involved in this embodiment are exemplary parameters used to clearly illustrate the technical operation process and are not the only limitation of the present invention.

[0013] In one embodiment, the small hydropower station group includes five operating nodes deployed along the same river basin, with each operating node corresponding to one small hydropower station. Within each operating node, sensors are deployed at the intake, water level well, and unit control cabinet to collect multi-source operating parameters, including: inflow rate (sampling accuracy 0.1 m³ / s), upstream water level (sampling accuracy 1 cm), tailrace water level, rainfall, and ambient temperature. Simultaneously, unit operating parameters are read from the unit control system, including current output, active power, guide vane opening, rotational speed, and rated output parameters. Each operating node synchronously collects the above parameters with a 1-second acquisition cycle, and attaches a unified timestamp to each set of collected data. Subsequently, with a 5-second upload cycle, the data from five consecutive acquisition cycles is encapsulated into a single data packet and uploaded to the centralized control platform via a dedicated communication network.

[0014] After receiving data uploaded by each operating node, the centralized control platform aligns the data from different nodes according to timestamps and fills in missing records caused by communication delays, forming a continuous multi-source operating parameter time series. Using a 30-minute sliding time window, features are extracted from historical flow, water level, and rainfall parameters, calculating the average value, maximum variation amplitude, and rate of change within the window, which serve as input features for runoff prediction. Based on these input features, the centralized control platform uses a historical similarity interval matching method to predict the runoff change trend within the next hour and outputs an indicator indicating whether the runoff is increasing or decreasing. After obtaining the runoff change trend, the overall adjustable load range of the hydropower station group is determined by combining the current unit output and adjustable range of each operating node. Subsequently, based on the positional order of each operating node in the watershed, hydraulic correlations are established between the nodes, and the overall load is proportionally allocated to each operating node, generating a coordinated scheduling strategy for the hydropower station group. This coordinated scheduling strategy is then matched and calculated with the unit operating parameters of each node to obtain the corresponding unit adjustment amount.

[0015] The centralized control platform breaks down the calculated unit adjustment quantities according to the operating nodes and sends them to the control devices of the corresponding operating nodes through the scheduling communication channel. Upon receiving the unit adjustment quantity, each operating node's control device, in conjunction with the current unit operating status parameters, converts the adjustment quantity into an executable guide vane opening adjustment value and generates a control command. After the control command is sent to the unit execution unit, the unit completes the parameter adjustment at a set rate. The local edge computing node monitors the unit output and guide vane opening change curves in real time with a monitoring cycle of 200ms and calculates the deviation between adjacent sampling points. When the deviation exceeds a set threshold, the edge computing node automatically corrects the control command and resends it, achieving local real-time control. In one embodiment, after completing the unit adjustment operation, each operating node continues to collect unit operating parameters according to the original acquisition cycle and uploads the adjusted operating parameters to the centralized control platform.

[0016] The centralized control platform compares the newly uploaded data with the data before adjustment, calculates the changes in the operating status of each node, and triggers the update process of the hydropower station group coordinated dispatch strategy when the change in the status of any node exceeds a set threshold. During the strategy update process, the centralized control platform recalculates the load allocation and unit adjustment, and loads the updated dispatch strategy into a preset digital twin virtual environment for simulation. During the simulation, the output and flow matching of the virtual nodes are continuously monitored. When the simulation error is within the allowable range, the dispatch strategy is deemed effective.

[0017] In another embodiment, the small hydropower station group includes eight dispersed operating nodes, each using a different type of generating unit. Each operating node collects multi-source operating parameters at a 2-second acquisition cycle. Flow and water level parameters are read by the SCADA system, while unit operating parameters are directly output by the PLC control system. The collected data is formatted and initial screening for outliers is performed locally. When a parameter exceeds the equipment's allowable range, it is marked as an outlier and the original record is retained. Subsequently, the multi-source operating parameters and unit operating parameters are packaged every 10 seconds and uploaded to the centralized control platform via a public network encrypted channel.

[0018] The centralized control platform uses a 15-minute basic analysis cycle to perform short-term trend analysis on multi-source operating parameters. It uses three consecutive analysis cycles as prediction windows. When the flow change direction is consistent and the change amplitude exceeds a preset threshold within two consecutive prediction windows, the runoff change trend is considered stable for the future target period. Based on this, a time-segmented load allocation table is generated according to the correspondence between the available water volume and the rated output of each unit at each operating node. The adjustment amplitude of each unit within the target period is then calculated as the unit adjustment amount. When the edge computing nodes of each operating node execute adjustment operations, they simultaneously record intermediate state parameters during the adjustment process. When a unit response speed is detected to be lower than a preset response threshold, the edge computing nodes dynamically scale the adjustment step size and update the control commands to ensure the smoothness of the adjustment process. The digital twin virtual environment is built based on the topology and unit parameters of the actual hydropower station group. After the centralized control platform passes simulation verification, the verified scheduling strategy is redistributed to each operating node, and the control device executes the corresponding unit adjustments to enter the next scheduling cycle.

[0019] Please refer to [link / reference needed] for further information. Figure 2 The image shows a factory workshop severely submerged by turbid green water. By collecting data in real time and uploading it to a centralized control platform, future runoff trends can be accurately predicted, and collaborative scheduling strategies can be formulated. Its core lies in the collaboration between the centralized control platform and local edge nodes. The centralized control end is responsible for macroscopic prediction and optimization, while the edge end performs rapid real-time control and deviation self-correction, thereby achieving precise scheduling and safe management of hydropower and avoiding accidents caused by sudden changes in water conditions.

[0020] Preferably, step S2 includes: In the centralized control platform, features are extracted from the received multi-source operating parameters, and a dynamic input sample sequence for runoff prediction is constructed based on a sliding time window; Predict runoff trends for future target time periods using dynamic input sample sequences; Determine the future operational status and dispatchable range of each operational node based on runoff change trends, as a measure of operational status changes; Calculate the load distribution of the hydropower station group based on the trend of runoff change and the changes in operating status; The coordinated scheduling strategy for hydropower station groups is determined based on the load allocation of each operating node. The computer adjusts the amount of data based on the coordinated scheduling strategy of the hydropower station group and the operating parameters of the generating units.

[0021] In this embodiment, after receiving multi-source operating parameters uploaded by each operating node, the centralized control platform performs a unified preprocessing operation on the parameters. This involves synchronizing and aligning the data uploaded by different operating nodes according to their timestamps, interpolating and completing data segments with time gaps, and removing duplicate data records to form a continuous time series of multi-source operating parameters. After preprocessing, the centralized control platform performs feature extraction processing on the multi-source operating parameters. Historical operating parameters are divided into segments using a fixed-length sliding time window, with the sliding window length set to 30 minutes and the window movement step size set to 5 minutes. Within each sliding window, the average value, maximum variation amplitude, and rate of change of parameters such as flow rate, water level, and rainfall are calculated as the corresponding feature values ​​for that window.

[0022] The feature values ​​extracted from each sliding window are arranged in chronological order to construct a dynamic input sample sequence for runoff prediction. Based on this dynamic input sample sequence, the centralized control platform analyzes the characteristic change trends of adjacent time windows. When the sign of the flow rate change rate remains consistent across multiple consecutive time windows and the change amplitude exceeds a preset threshold, it is determined that the runoff in the future target period shows a stable upward or downward trend, and the runoff change trend result is output. After obtaining the runoff change trend for the future target period, the centralized control platform combines the current unit operating parameters of each operating node to predict and analyze the operating status of each operating node in the target period. Based on the runoff change trend, the platform estimates the inflow conditions that each operating node can obtain, and combines the rated output parameters of the unit with the current operating load to determine the dispatchable range of each operating node in the target period, as the change in operating status.

[0023] The centralized control platform calculates the load allocation at the hydropower station group level based on runoff variation trends and changes in the operating status of each operating node. During the calculation, the dispatchable range of each operating node is used as a constraint. A basic load allocation value is generated according to the ratio between available water volume and rated unit output. When the basic load allocation value of a certain operating node exceeds its dispatchable upper limit, the excess portion is redistributed according to the available proportions of other operating nodes, thus forming a group-level load allocation that satisfies the overall constraints.

[0024] After obtaining the group-level load allocation, the centralized control platform further determines the coordinated scheduling strategy of the hydropower station group for each operating node. Based on the spatial order of each operating node in the basin, the hydraulic correlation between the nodes is established. Combined with the distribution of the load allocation in the time dimension, the power transfer ratio between adjacent operating nodes is corrected. On this basis, the target output plan of each operating node in the target time period is determined, forming the coordinated scheduling strategy of the hydropower station group. The centralized control platform matches and calculates the coordinated scheduling strategy of the hydropower station group with the unit operating parameters of each operating node to obtain the unit adjustment amount of each unit in the target time period, which serves as the input basis for subsequent adjustment operations.

[0025] Preferably, the process of extracting features from the received multi-source operating parameters in the centralized control platform and constructing a dynamic input sample sequence for runoff prediction based on a sliding time window includes: In the centralized control platform, multi-source operating parameters are synchronized according to the acquisition timestamp, records with discontinuous time are filtered out, missing records are filled in, duplicate records are removed, and pre-processed multi-source operating parameters are generated. For continuous parameters in multi-source operating parameters, time segments are divided according to a fixed-length sliding window, and the parameter sequence in each window is used as a set of input samples. Extract the mean, minimum, maximum, and rate of change as feature values ​​for each input sample group; The extracted feature values ​​are arranged in chronological order into a continuous sample set, which serves as the dynamic input sample sequence.

[0026] In one embodiment, after receiving multi-source operating parameters uploaded by each operating node, the centralized control platform performs unified sorting and synchronization processing on the data from different operating nodes and different parameter sources based on the acquisition timestamps. For data records with discontinuous timestamps, linear interpolation is performed to fill in the gaps according to the numerical change trend of adjacent valid sampling points. For parameter records uploaded repeatedly at the same timestamp, only the earliest data record is retained, and the rest are discarded, thereby generating preprocessed multi-source operating parameters that are continuous in time and do not repeat. After completing the preprocessing, for the parameters with continuous change characteristics among the multi-source operating parameters, including flow rate, water level, and rainfall, The time series is divided according to a fixed-length sliding time window, with the sliding window length set to 30 minutes and the window movement step size set to 5 minutes, so that the time intervals between adjacent windows overlap. The time series of parameters corresponding to each sliding window is used as a set of input samples. Then, within each set of input samples, the average, minimum, and maximum values ​​of each parameter within the window, as well as the rate of change between adjacent sampling points, are calculated to characterize the overall level and trend of the parameter within the time interval. The feature values ​​extracted from each sliding window are arranged in chronological order to form a continuous set of feature samples.

[0027] In another embodiment, to address the issue of differing acquisition frequencies for multi-source operating parameters, the centralized control platform first resamples data from different parameter sources according to a unified time benchmark. Using the resampled timestamps as the synchronization basis, data records with time jumps or abnormal sampling intervals are marked and removed. Short-term missing data segments are filled in using forward padding, thus obtaining a multi-source operating parameter sequence with a consistent time structure. Based on this, only continuous parameters such as flow rate and water level are processed using a sliding window. The sliding window length is set to 15 minutes, and the window step size is set to 3 minutes to improve the response sensitivity to short-term runoff changes. The parameter sequence within each window's coverage period is used as an independent input sample. During feature extraction, in addition to calculating the average, minimum, and maximum values ​​of each parameter within the window, the ratio of the parameter difference at the start and end times of the window to the window duration is used to obtain the parameter change rate, reflecting the intensity of change within that time segment. Subsequently, the feature values ​​corresponding to each time window are concatenated in chronological order to form a time-continuous sample set, which is used to construct a dynamic input sample sequence for runoff prediction.

[0028] Preferably, predicting runoff change trends for future target periods using dynamic input sample sequences includes: The generated dynamic input sample sequence is divided into multiple time segments in chronological order; Calculate the rate of change sequence between adjacent time segments, and mark time segments with a rate of change exceeding a threshold as feature evolution segments; Identify the continuous change direction of characteristic evolution segments and determine the short-term runoff change trend; According to the preset prediction period length, rolling calculations are performed on continuous feature evolution segments to obtain the change sequence of continuous prediction periods; By comparing the change series of continuous forecast periods with the short-term runoff change trend, the overall upward or downward direction of runoff change can be determined, and the runoff change trend for the future target period can be output.

[0029] In one embodiment, after obtaining the constructed dynamic input sample sequence, the centralized control platform divides the sample sequence into several adjacent time segments according to the time sequence corresponding to the samples. Each time segment corresponds to a feature sample extracted by a sliding window. Subsequently, the centralized control platform calculates the change amplitude of the corresponding parameter features within adjacent time segments to obtain a change rate sequence reflecting the change intensity between adjacent time segments. Time segments with change rates exceeding a preset threshold are marked as feature evolution segments to distinguish time intervals with significant runoff changes. Based on this, directional analysis is performed on consecutively appearing feature evolution segments. When the change direction within multiple feature evolution segments remains consistent, it is determined that the runoff within that time range exhibits stability. After determining a short-term upward or downward trend, the centralized control platform performs rolling calculations on continuous feature evolution segments according to a preset prediction period length. That is, while keeping the prediction period length unchanged, it sequentially extracts continuous time intervals covering the prediction period and calculates the cumulative changes in runoff-related features within each time interval, thereby forming a sequence of changes for continuous prediction periods. The centralized control platform compares and analyzes this sequence of changes with the aforementioned short-term runoff change trend. When the direction of change in multiple consecutive prediction periods is consistent with the direction of the short-term trend and the magnitude of change reaches the set conditions, it determines that the overall direction of runoff change in the future target period is upward or downward, and outputs the runoff change trend result for the future target period accordingly.

[0030] Preferably, rolling calculations are performed on continuous feature evolution segments according to a preset prediction period length to obtain a sequence of changes over the continuous prediction period, including: When performing rolling calculations, the prediction period length is kept constant, and the window step size is set to be an integer multiple of the prediction period. Extract a time interval of the same length as the prediction period from the characteristic evolution segment, and calculate the cumulative change of runoff parameters within the interval; The cumulative changes between adjacent time intervals are calculated sequentially according to the window step size to form a set of periodic changes; Arrange the set of periodic changes to generate a sequence of changes for continuous prediction periods.

[0031] In one embodiment, when performing rolling calculations on continuous feature evolution segments according to a preset prediction cycle length to obtain a sequence of changes in continuous prediction cycles, a prediction cycle length for runoff trend prediction is preset in the centralized control platform. This prediction cycle length corresponds to the time scale required for assessing future runoff changes. During the rolling calculation process, the prediction cycle length remains constant, and the window step size is set based on this prediction cycle length, ensuring that the window step size and the prediction cycle length are integer multiples of each other, thereby ensuring that adjacent calculation windows advance regularly on the time axis. Subsequently, multiple time intervals are sequentially extracted from the identified continuous feature evolution segments according to the prediction cycle length, with the start and end times of each time interval specified. The current scrolling position is determined, and each time interval contains complete runoff parameter time series data. For each selected time interval, based on the starting and ending values ​​of the runoff parameters within that time interval, the cumulative change in runoff within the corresponding prediction period is calculated. The cumulative change is used to characterize the overall change amplitude of runoff within the prediction period. After completing the calculation of the cumulative change for a single interval, the scrolling window is moved forward along the time axis according to the window step size, and the calculation of the cumulative change is repeated for adjacent time intervals, thereby obtaining multiple periodic change values ​​arranged in chronological order. The obtained periodic change values ​​corresponding to each prediction period are arranged in sequence to form a change value sequence reflecting the change of runoff within consecutive prediction periods.

[0032] Preferably, comparing the change series of continuous forecast periods with the short-term runoff change trend to determine the overall upward or downward direction of runoff change, and outputting the runoff change trend for the future target period includes: Compare the change series of continuous forecast periods with the short-term runoff change trend; When making comparisons, the changes within the same forecast period and the short-term trends of the corresponding time intervals are selected for matching. Calculate the difference sequence of continuously predicted periodic changes and determine whether the sign of the difference is consistent with the direction of the short-term change trend. When the change in two consecutive forecast periods is consistent with the short-term trend and the magnitude of the change exceeds the threshold, the overall runoff change direction is determined to be either upward or downward. Based on the determined overall direction of change, the runoff change trend for the future target period is output.

[0033] In one embodiment, after determining the change sequence and short-term runoff change trend for the continuous forecast period, the centralized control platform compares and analyzes the change sequence and short-term runoff change trend for the continuous forecast period to determine the overall upward or downward direction of runoff change. Using the forecast period as a unified time scale, the platform matches the change in each period of the continuous forecast period change sequence with the identified short-term runoff change trend within its corresponding time interval, ensuring that the change and short-term change trend being compared originate from the same time range. After time alignment, the platform performs difference calculation on the change in two adjacent forecast periods within the continuous forecast period to obtain a difference sequence reflecting the acceleration or deceleration characteristics of runoff change, and then determines the difference based on the positive or negative sign of each difference. The direction of runoff change within the corresponding prediction period is determined. Subsequently, the direction of the difference is compared with the direction of the short-term runoff change trend within the corresponding time interval. When the direction of change represented by the difference sign is consistent with the direction indicated by the short-term runoff change trend, the direction of runoff change within the prediction period is considered stable. Further, when the change in two consecutive prediction periods is consistent with the direction of the corresponding short-term runoff change trend, and the absolute value of the change or the difference exceeds the preset change amplitude threshold, the centralized control platform determines that the runoff change exhibits a continuous characteristic in time, and determines whether the overall runoff change direction is upward or downward. Based on the determined overall change direction, the runoff change trend within the future target period is output as the runoff prediction result.

[0034] Preferably, the calculation of the load allocation at the hydropower station group level based on future runoff trends and changes in operating status includes: By comparing the runoff change trend in the future target period with the changes in the operating status of each operating node, the allocable water volume range of each operating node in the target period is determined. Based on the available water volume in the allocable water volume range and the rated output parameters of the unit operating parameters in the node, the basic load allocation value of the node is calculated according to the proportional correspondence between the two. The base load allocation value is weighted and adjusted based on the demand differences among the operating nodes. When the weighted adjusted load allocation value exceeds the adjustable limit of a node, it is limited to the maximum available load of that node, and the excess portion is redistributed to other operating nodes according to the current load ratio of each node. The corrected load allocation values ​​are arranged in chronological order to generate a hydropower station group-level load allocation sequence for the target time period.

[0035] In one embodiment, after obtaining the runoff change trend and the operational status changes of each operating node for a future target period, the centralized control platform compares and analyzes the runoff change trend with the corresponding operational status changes of each operating node to determine the allocatable water volume range that each operating node can participate in scheduling during the target period. This allocatable water volume range is jointly defined by the node's current available reservoir capacity, water level constraints, and operational status changes. After determining the allocatable water volume range for each operating node, the centralized control platform uses the available water volume within this range as a water volume-side constraint. Simultaneously, it combines this with the rated output parameters of the generating units in each operating node to establish a proportional relationship between available water volume and rated output, thereby calculating the allocatable water volume range for each operating node during the target period. The platform obtains the basic load allocation value. After obtaining the basic load allocation value, the centralized control platform further performs weighted correction processing on the basic load allocation value based on the differences in load demand, power supply priority, or regulation capability of each operating node on the grid side, so that the load allocation result can reflect the actual scheduling needs between nodes. When the weighted correction load allocation value exceeds the adjustable upper limit of a certain operating node, the centralized control platform limits the load allocation value of that node to its maximum available load, and redistributes the excess part to other operating nodes with regulation margin according to the current load ratio of each operating node. The load allocation values ​​of each operating node after limitation and redistribution are arranged in chronological order within the target time period to form a load allocation sequence at the hydropower station group level.

[0036] Of particular importance is that the coordinated scheduling strategy for hydropower station groups, based on load allocation, includes: Before determining the scheduling strategy, read the load allocation and unit operation constraint parameters corresponding to each operating node; Based on the time series of load distribution, a corresponding analysis is conducted on the available output and water level regulation capacity of each operating node; Establish a hydraulic correlation matrix for the operating nodes according to their location in the watershed, and determine the water level transfer relationship between the nodes; By combining the hydraulic correlation matrix with the time-period distribution of load allocation, the coordinated power transfer ratio between nodes is calculated. Based on the coordinated power transfer ratio and the node output limit, determine the target power scheduling value for each node; When the target power scheduling value exceeds the node's adjustable limit, the excess portion will be allocated to adjacent nodes according to the coordination ratio. The target power scheduling values ​​of each node are arranged in time sequence to form a coordinated scheduling strategy for the hydropower station group, which serves as the input basis for calculating the unit regulation.

[0037] In one embodiment, before determining the coordinated scheduling strategy of the hydropower station group based on the load allocation, the centralized control platform reads the load allocation corresponding to each operating node within the target time period and the operating constraint parameters of the units in that operating node. The operating constraint parameters include at least the unit output limit, output ramp-up constraint, and water level regulation limit. After reading, the centralized control platform performs corresponding analysis on the available output capacity and water level regulation capacity of each operating node within the corresponding time period based on the time series of the load allocation to determine whether the node has the regulation conditions to complete the allocated load in different time segments. Subsequently, the centralized control platform establishes a hydraulic correlation matrix according to the upstream and downstream positions of each operating node in the same watershed, and clarifies the water level transmission relationship between nodes and its impact through this hydraulic correlation matrix. Based on this, and combining the hydraulic correlation matrix with the time distribution of load allocation within the target period, the coordinated power transfer ratio between operating nodes is calculated so that the output change of upstream nodes can match the adjustment capacity of downstream nodes. Then, based on the coordinated power transfer ratio and the output limit of each operating node, the target power scheduling value of each operating node within the target period is calculated. When the target power scheduling value of a certain operating node exceeds its adjustable limit, the centralized control platform will allocate the excess part to the adjacent operating nodes with hydraulic correlation according to the coordinated power transfer ratio to ensure that the scheduling result meets the node constraints. The target power scheduling values ​​determined by each operating node within the target period are arranged in chronological order to form a coordinated scheduling strategy for the hydropower station group.

[0038] Preferably, step S3 includes: The unit adjustment parameters generated by the centralized control platform are distributed to the built-in control devices of each operating node; After receiving the corresponding unit adjustment, each control device calculates the executable adjustment execution value based on the current unit operating status parameters, and generates control commands for executing the adjustment accordingly. The control device sends control commands to the unit's execution unit, which then adjusts the unit's parameters in real time according to the control commands. The real-time change curve of the unit parameters is monitored by a preset local edge computing node. When the detected parameter deviation exceeds the set threshold, the control command is automatically corrected and the corrected command is resent to the corresponding execution unit. All adjustment data generated during the adjustment process is processed by the local edge computing node and then transmitted back to the centralized control platform.

[0039] In one embodiment, after generating the unit adjustment quantity, the centralized control platform breaks down the unit adjustment quantity according to the operating nodes and distributes it to the control devices built into each operating node through a communication link. Upon receiving the corresponding unit adjustment quantity, each control device reads the current operating status parameters of the unit, including real-time output, water level status, and unit start / stop status. Using these operating status parameters as constraints, it performs an executability judgment on the unit adjustment quantity, thereby calculating the adjustment execution value that satisfies the current operating status. After determining the adjustment execution value, each control device generates a control command for executing the adjustment based on the adjustment execution value and sends the control command to the unit. The group's execution unit adjusts the unit output, turbine guide vane opening, or speed parameters in real time according to control commands. During the adjustment process, a preset local edge computing node continuously monitors the real-time change curve of the unit parameters and calculates the parameter deviation between adjacent sampling times. When the parameter deviation exceeds a set threshold, the edge computing node automatically corrects the current control command based on the deviation direction and magnitude, and reissues the corrected control command to the corresponding execution unit. At the same time, the unit operating parameters, control commands, and correction records generated during the adjustment process are processed by the local edge computing node and uniformly transmitted back to the centralized control platform.

[0040] Preferably, the real-time change curves of unit parameters are monitored using preset local edge computing nodes. When the detected parameter deviation exceeds a set threshold, the control commands are automatically corrected, and the corrected commands are reissued to the corresponding execution units, including: The real-time change curves of unit parameters are continuously monitored using designated local edge computing nodes, and the unit's operating parameters are collected during the monitoring process. The difference in operating parameters at adjacent sampling times is calculated to generate the unit operating deviation. When the operating deviation exceeds the set threshold, the automatic correction process of the adjustment command is triggered, and corresponding updated operating parameters are generated based on the deviation. The local edge computing node calculates adjustment correction values ​​based on the updated operating parameters and generates updated adjustment instructions; The updated adjustment command is sent to the corresponding control device, which then re-executes the unit adjustment operation.

[0041] In one embodiment, a preset local edge computing node continuously monitors the real-time change curves of unit parameters during the unit regulation execution process, and periodically collects the unit's operating parameters as real-time monitoring data during the monitoring process. After obtaining the continuously sampled operating parameters, the local edge computing node calculates the difference between the operating parameters corresponding to adjacent sampling times, thereby generating a unit operating deviation reflecting the instantaneous changes in the unit. When the unit operating deviation exceeds a preset threshold, the local edge computing node triggers an automatic correction process for the regulation command, and generates updated operating parameters corresponding to the current deviation state based on the operating deviation. The local edge computing node recalculates the current regulation execution state based on the updated operating parameters to obtain a regulation correction value for correcting the unit regulation process, and generates an updated regulation command based on the regulation correction value. The updated regulation command is then sent to the corresponding control device, which re-executes the unit regulation operation according to the updated regulation command to bring the unit operating state back to the expected regulation range.

[0042] Of particular importance, step S4 includes: After completing the adjustment operation, the adjusted unit operating parameters are collected and uploaded to the centralized control platform; The running parameters are collected according to the preset sampling period, and the running parameter dataset is generated after format verification and missing data filling. After receiving the operating parameters, the centralized control platform calculates the changes in the operating status of each operating node. When the changes exceed the set threshold, it triggers an update to the coordinated scheduling strategy of the hydropower station group. During the update process, the load allocation and unit regulation are recalculated, a new scheduling strategy is generated, and simulation verification is performed in a digital twin virtual environment. During simulation, monitor the matching status of virtual node output power and traffic, and determine the verified scheduling strategy when the scheduling error is within the allowable threshold range; The verified scheduling strategy is redistributed to each operating node, and the control device executes unit adjustments to achieve closed-loop update and rescheduling of the strategy.

[0043] In one embodiment, after completing the unit adjustment operation, each operating node collects the adjusted unit operating parameters according to a preset sampling period and uploads the collected operating parameters to the centralized control platform. During the upload process, the operating parameters are first validated locally, and records that do not meet the data format requirements are removed. At the same time, records with missing timestamps or missing parameters are supplemented to generate a complete and consistent operating parameter dataset. After receiving the operating parameter dataset, the centralized control platform compares the current operating status of each operating node with the corresponding operating status of the previous scheduling cycle to calculate the change in operating status of each operating node. When the change in operating status of any operating node exceeds a set threshold, the hydropower station group collaborative scheduling strategy update process is triggered. During the update process, the centralized control platform recalculates the load allocation at the hydropower station group level based on the latest operating parameter data, and regenerates the corresponding unit adjustment quantities accordingly, thereby forming a new hydropower station group collaborative scheduling strategy. Subsequently, the new scheduling strategy is loaded into the preset digital twin virtual environment, and corresponding scheduling simulations are performed on each virtual operating node in the virtual environment. During the simulation operation, the matching status between the output power and the corresponding flow of each virtual node is continuously monitored. When the monitored scheduling error is within the allowable threshold range, it is determined that the scheduling strategy has passed the simulation verification. The verified scheduling strategy is then redistributed to each operating node, and the control devices in each operating node execute unit adjustments according to the new scheduling strategy, realizing the closed-loop update and rescheduling of the scheduling strategy.

[0044] The present invention also provides an intelligent scheduling system for a small hydropower station group, used to execute the intelligent scheduling method for a small hydropower station group as described above. The intelligent scheduling system for a small hydropower station group includes: The data extraction module is used to collect multi-source operating parameters and unit operating parameters of each operating node in a small hydropower station group in real time, and then encapsulate the acquired multi-source operating parameters and unit operating parameters according to the collection cycle and upload them to the centralized control platform. The runoff change trend prediction module is used to predict the runoff change trend in the future target period using multi-source operating parameters on the centralized control platform, determine the coordinated scheduling strategy of the hydropower station group based on the runoff change trend, and calculate the corresponding unit regulation based on the coordinated scheduling strategy of the hydropower station group and the unit operating parameters. The adjustment operation module is used to distribute the unit adjustment amount to each operating node, execute adjustment operations according to the unit adjustment amount, and complete real-time control in the local edge computing node; The scheduling strategy update module is used to collect the adjusted unit operating parameters after the adjustment operation is completed and upload them to the centralized control platform. The centralized control platform updates the hydropower station group collaborative scheduling strategy according to the uploaded unit operating parameters, and completes the simulation verification of the scheduling strategy in the preset digital twin virtual environment. The verified scheduling strategy is then redistributed to each operating node for execution.

[0045] Please refer to [link / reference needed] for further information. Figure 3 The system begins with the data extraction module, responsible for collecting multi-source monitoring data such as rainfall and dam safety information and uploading it to the centralized control platform. The platform's core functions include constructing dynamic input sample sequences, predicting future runoff trends, and completing load allocation through rolling calculations and weighted corrections, thereby generating a collaborative scheduling strategy. This strategy is distributed to field IoT terminals via the regulation operation module, where local edge nodes receive and execute real-time control, while also possessing monitoring and self-correcting capabilities. Regulation data generated during execution is transmitted back to the centralized control platform to drive updates to the scheduling strategy. Finally, the new strategy is simulated and verified in a digital twin virtual environment, thus forming a complete technical system integrating intelligent prediction, precise scheduling, real-time control, and feedback optimization.

[0046] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for intelligent scheduling of a small hydropower station group, characterized in that, Includes the following steps: Step S1: Collect multi-source operating parameters and unit operating parameters of each operating node in real time in the small hydropower station group, and upload the acquired multi-source operating parameters and unit operating parameters to the centralized control platform after packaging them according to the collection cycle. Step S2: On the centralized control platform, use multi-source operating parameters to predict the runoff change trend for the future target period, determine the coordinated scheduling strategy of the hydropower station group based on the runoff change trend, and calculate the corresponding unit regulation based on the coordinated scheduling strategy of the hydropower station group and the unit operating parameters. Step S3: Distribute the unit adjustment amount to each operating node, execute the adjustment operation according to the unit adjustment amount, and complete the real-time control in the local edge computing node; Step S4: After completing the adjustment operation, collect the adjusted unit operating parameters and upload them to the centralized control platform; the centralized control platform updates the hydropower station group collaborative scheduling strategy according to the uploaded unit operating parameters, and completes the simulation verification of the scheduling strategy in the preset digital twin virtual environment, and redistributes the verified scheduling strategy to each operating node for execution.

2. The intelligent scheduling method for small hydropower station groups according to claim 1, characterized in that, Step S2 includes: In the centralized control platform, features are extracted from the received multi-source operating parameters, and a dynamic input sample sequence for runoff prediction is constructed based on a sliding time window; Predict runoff trends for future target time periods using dynamic input sample sequences; Determine the future operational status and dispatchable range of each operational node based on runoff change trends, as a measure of operational status changes; Calculate the load distribution of the hydropower station group based on the trend of runoff change and the changes in operating status; The coordinated scheduling strategy for hydropower station groups is determined based on the load allocation of each operating node. The computer adjusts the amount of data based on the coordinated scheduling strategy of the hydropower station group and the operating parameters of the generating units.

3. The intelligent scheduling method for small hydropower station groups according to claim 2, characterized in that, The centralized control platform performs feature extraction on the received multi-source operating parameters and constructs a dynamic input sample sequence for runoff prediction based on a sliding time window, including: In the centralized control platform, multi-source operating parameters are synchronized according to the acquisition timestamp, records with discontinuous time are filtered out, missing records are filled in, duplicate records are removed, and pre-processed multi-source operating parameters are generated. For continuous parameters in multi-source operating parameters, time segments are divided according to a fixed-length sliding window, and the parameter sequence in each window is used as a set of input samples. Extract the mean, minimum, maximum, and rate of change as feature values ​​for each input sample group; The extracted feature values ​​are arranged in chronological order into a continuous sample set, which serves as the dynamic input sample sequence.

4. The intelligent scheduling method for small hydropower station groups according to claim 2, characterized in that, Predicting runoff trends for future target time periods using dynamic input sample sequences includes: The generated dynamic input sample sequence is divided into multiple time segments in chronological order; Calculate the rate of change sequence between adjacent time segments, and mark time segments with a rate of change exceeding a threshold as feature evolution segments; Identify the continuous change direction of characteristic evolution segments and determine the short-term runoff change trend; According to the preset prediction period length, rolling calculations are performed on continuous feature evolution segments to obtain the change sequence of continuous prediction periods; By comparing the change series of continuous forecast periods with the short-term runoff change trend, the overall upward or downward direction of runoff change can be determined, and the runoff change trend for the future target period can be output.

5. The intelligent scheduling method for small hydropower station groups according to claim 4, characterized in that, According to the preset prediction period length, rolling calculations are performed on continuous feature evolution segments to obtain the change sequence of continuous prediction periods, including: When performing rolling calculations, the prediction period length is kept constant, and the window step size is set to be an integer multiple of the prediction period. Extract a time interval of the same length as the prediction period from the characteristic evolution segment, and calculate the cumulative change of runoff parameters within the interval; The cumulative changes between adjacent time intervals are calculated sequentially according to the window step size to form a set of periodic changes; Arrange the set of periodic changes to generate a sequence of changes for continuous prediction periods.

6. The intelligent scheduling method for small hydropower station groups according to claim 4, characterized in that, By comparing the change series of continuous forecast periods with short-term runoff change trends, the overall upward or downward direction of runoff change is determined, and the runoff change trend for future target periods is output, including: Compare the change series of continuous forecast periods with the short-term runoff change trend; When making comparisons, the changes within the same forecast period and the short-term trends of the corresponding time intervals are selected for matching. Calculate the difference sequence of continuously predicted periodic changes and determine whether the sign of the difference is consistent with the direction of the short-term change trend. When the change in two consecutive forecast periods is consistent with the short-term trend and the magnitude of the change exceeds the threshold, the overall runoff change direction is determined to be either upward or downward. Based on the determined overall direction of change, the runoff change trend for the future target period is output.

7. The intelligent scheduling method for small hydropower station groups according to claim 2, characterized in that, The calculation of load allocation at the hydropower station group level based on future runoff trends and changes in operating status includes: By comparing the runoff change trend in the future target period with the changes in the operating status of each operating node, the allocable water volume range of each operating node in the target period is determined. Based on the available water volume in the allocable water volume range and the rated output parameters of the unit operating parameters in the node, the basic load allocation value of the node is calculated according to the proportional correspondence between the two. The base load allocation value is weighted and adjusted based on the demand differences among the operating nodes. When the weighted adjusted load allocation value exceeds the adjustable limit of a node, it is limited to the maximum available load of that node, and the excess portion is redistributed to other operating nodes according to the current load ratio of each node. The corrected load allocation values ​​are arranged in chronological order to generate a hydropower station group-level load allocation sequence for the target time period.

8. The intelligent scheduling method for small hydropower station groups according to claim 1, characterized in that, Step S3 includes: The unit adjustment parameters generated by the centralized control platform are distributed to the built-in control devices of each operating node; After receiving the corresponding unit adjustment, each control device calculates the executable adjustment execution value based on the current unit operating status parameters, and generates control commands for executing the adjustment accordingly. The control device sends control commands to the unit's execution unit, which then adjusts the unit's parameters in real time according to the control commands. The real-time change curve of the unit parameters is monitored by a preset local edge computing node. When the detected parameter deviation exceeds the set threshold, the control command is automatically corrected and the corrected command is resent to the corresponding execution unit. All adjustment data generated during the adjustment process is processed by the local edge computing node and then transmitted back to the centralized control platform.

9. The intelligent scheduling method for small hydropower station groups according to claim 8, characterized in that, The system utilizes pre-defined local edge computing nodes to monitor the real-time changes in unit parameters. When the detected parameter deviation exceeds a set threshold, the control commands are automatically corrected, and the corrected commands are reissued to the corresponding execution units, including: The real-time change curves of unit parameters are continuously monitored using designated local edge computing nodes, and the unit's operating parameters are collected during the monitoring process. The difference in operating parameters at adjacent sampling times is calculated to generate the unit operating deviation. When the operating deviation exceeds the set threshold, the automatic correction process of the adjustment command is triggered, and corresponding updated operating parameters are generated based on the deviation. The local edge computing node calculates adjustment correction values ​​based on the updated operating parameters and generates updated adjustment instructions; The updated adjustment command is sent to the corresponding control device, which then re-executes the unit adjustment operation.

10. A smart dispatching system for a small hydropower station group, characterized in that, For executing the intelligent scheduling method for a small hydropower station group as described in claim 1, the intelligent scheduling system for the small hydropower station group includes: The data extraction module is used to collect multi-source operating parameters and unit operating parameters of each operating node in a small hydropower station group in real time, and then encapsulate the acquired multi-source operating parameters and unit operating parameters according to the collection cycle and upload them to the centralized control platform. The runoff change trend prediction module is used to predict the runoff change trend in the future target period using multi-source operating parameters on the centralized control platform, determine the coordinated scheduling strategy of the hydropower station group based on the runoff change trend, and calculate the corresponding unit regulation based on the coordinated scheduling strategy of the hydropower station group and the unit operating parameters. The adjustment operation module is used to distribute the unit adjustment amount to each operating node, execute adjustment operations according to the unit adjustment amount, and complete real-time control in the local edge computing node; The scheduling strategy update module is used to collect the adjusted unit operating parameters after the adjustment operation is completed and upload them to the centralized control platform. The centralized control platform updates the hydropower station group collaborative scheduling strategy according to the uploaded unit operating parameters, and completes the simulation verification of the scheduling strategy in the preset digital twin virtual environment. The verified scheduling strategy is then redistributed to each operating node for execution.

Images