Pumped storage power station monitoring system control method and system

Abstract

The invention provides a pumped storage power station monitoring system control method and device. The method comprises the following steps of: acquiring and converging internal state data from a power grid dispatching center, a power station sensor and a system in real time according to configuration information and operation condition requirements of a monitoring system to form a unified and formatted multi-source real-time data stream; constructing a system state non-loop connection tree model in combination with the equipment model, the operation constraint condition and the state transition rule; calculating an optimal control sequence by applying a linear time complexity shortest path algorithm for directed acyclic graph optimization; and according to the optimal control sequence, control instructions arranged according to a time sequence are generated, monitoring system parameters are dynamically adjusted, and the control instructions are issued to an equipment execution mechanism. According to the method, the global optimal control path can be quickly found, and the response speed and the control precision are improved.

Description

Technical Field

[0001] This invention relates to the field of power plant automation control technology, and more specifically, to a control method and system for a pumped storage power plant monitoring system. Background Technology

[0002] As an important component of the power system, pumped storage power stations undertake multiple functions such as peak shaving, valley filling, frequency regulation, phase regulation, and emergency backup. Their operating conditions are complex and variable, which places extremely high demands on the response speed, control accuracy, and stability of the monitoring system.

[0003] Currently, monitoring systems for pumped storage power stations typically employ an architecture based on SCADA (Supervisory Control and Data Acquisition) and PLC (Programmable Logic Controller). For example, some systems use a hierarchical distributed structure, collecting status information from generating units, hydraulic equipment, and electrical equipment through field control units, uploading it to a central monitoring host, which then makes judgments and issues control commands based on preset logic or scheduling instructions. Other systems may integrate more complex model predictive control or expert systems to improve control performance.

[0004] The most relevant prior art to this invention consists of monitoring system methods that attempt to optimize control logic to cope with complex operating conditions and dispatch instructions. These methods typically analyze input grid dispatch instructions, the current operating status of the generating units (e.g., speed, power, temperature, vibration), reservoir water level information, and grid parameters (e.g., frequency, voltage), and generate control sequences based on a pre-defined control strategy library (e.g., containing start-up and shutdown procedures under different operating conditions, power regulation curves, fault handling plans, etc.). They attempt to select appropriate strategies based on the input information and execute them step by step.

[0005] However, existing control methods often have significant technical shortcomings when dealing with multiple operating conditions, rapid switching, and complex scheduling commands (e.g., requiring a switch from pumping operation to full-load power generation in a very short time while satisfying multiple constraints). First, their control decision-making process may be based on a fixed rule base or simple feedback adjustments. For multivariable, strongly coupled, and rapidly time-varying pumped storage systems, it is difficult to quickly find the globally optimal or near-optimal control path, resulting in slow response speeds. Second, due to the failure to adequately predict and optimize the timing and priority of control commands, especially during transient processes of operating condition switching, control accuracy may be low, leading to overshoot, oscillations, or even excessive equipment stress, affecting system stability and equipment lifespan. Summary of the Invention

[0006] The purpose of this invention is to overcome the above-mentioned problems in the prior art and provide a control method and system for a pumped storage power station monitoring system. This system can cope with complex and ever-changing operating conditions and scheduling instructions, quickly find the globally optimal or near-optimal control path, improve response speed, and improve control accuracy and system stability by optimizing the timing and priority of control instructions.

[0007] To achieve the above objectives, the present invention provides a control method for a pumped storage power station monitoring system, comprising: Based on the configuration information of the monitoring system and the operating conditions of the power station, real-time data from the power grid dispatch center, local sensors of the power station and the internal status data of the pumped storage power station monitoring system are collected and aggregated to form a unified formatted multi-source real-time data stream. Based on the unified formatted multi-source real-time data stream, combined with the power plant equipment model, operating constraints, and state transition rules, a system state non-cyclic connection tree model representing all feasible paths from the current state to the target state is constructed. Using the aforementioned system state acyclic connection tree model, and applying a linear time complexity shortest path algorithm optimized for directed acyclic graphs, the optimal path from the current state node to the target state node is calculated, and the optimal control sequence representing the optimal control strategy is obtained. Based on the optimal control sequence, control instructions are generated in a time sequence, and the data acquisition frequency and alarm threshold of the monitoring system are dynamically adjusted. The control instructions are then sent to the equipment actuators of the pumped storage power station monitoring system.

[0008] According to one embodiment of the present invention, based on the monitoring system configuration information and the monitoring power station's operating condition requirements, real-time data from the power grid dispatch center, local sensors at the power station, and the pumped storage power station monitoring system are collected and aggregated to form a unified formatted multi-source real-time data stream, including: The system analyzes the preset configuration information of the monitoring system and the current operating conditions of the monitoring power plant to identify the data items necessary to complete the control task, locate the physical sensors, computing nodes and external data interfaces corresponding to the data items, and obtain the list of data items to be collected and their corresponding specific data sources. Based on the list of data items and their corresponding specific data sources, data requests or subscription services are initiated to the specified data sources through the communication interface of the monitoring system. Raw measurement values ​​or status signals are acquired at preset or dynamically adjusted frequencies to generate raw multi-source heterogeneous data streams containing timestamps. The original multi-source heterogeneous data stream is processed and a precise timestamp is added to it to obtain a processed and timestamped valid data set.

[0009] According to one embodiment of the present invention, based on the monitoring system configuration information and the monitoring power station's operating condition requirements, real-time data from the power grid dispatch center, local sensors at the power station, and the pumped storage power station monitoring system are collected and aggregated to form a unified formatted multi-source real-time data stream, including: Data fusion technology is applied to the cleaned and timestamped valid dataset to fuse multiple measurements of the same physical quantity to obtain fused data. The fused data is formatted and organized according to the unified data model defined within the pumped storage power station monitoring system to construct a unified formatted multi-source real-time data stream.

[0010] According to one embodiment of the present invention, based on the uniformly formatted multi-source real-time data stream, and combined with the power plant equipment model, operating constraints, and state transition rules, a system state acyclic connection tree model representing all feasible paths from the current state to the target state is constructed, including: Using the unified formatted multi-source real-time data stream, as well as the predefined power plant equipment physical model, operating constraints, state transition rules, and received target scheduling instructions, key state variables are extracted from the multi-source real-time data stream to determine the current precise operating state of the pumped storage power station monitoring system, and a state node is constructed to represent the current state of the pumped storage power station monitoring system. The target scheduling instructions in the unified formatted multi-source real-time data stream are parsed, and the target state that the system needs to achieve is determined by combining the device model and constraints, and a clearly defined target state node is constructed.

[0011] According to one embodiment of the present invention, based on the uniformly formatted multi-source real-time data stream, and combined with the power plant equipment model, operating constraints, and state transition rules, a system state acyclic connection tree model representing all feasible paths from the current state to the target state is constructed, including: Based on the clearly defined current state node and target state node, as well as the device model, constraints and state transition rules, a data structure is created to represent the state tree, and the current state node is set as the root node of the tree, and an initial state tree structure containing the root node is constructed. Using the initial state tree structure containing the root node, according to the state transition rules, find all possible one-step transition operations, calculate the new state node it leads to, evaluate the cost of the transition operation, and form a partially expanded non-cyclic connection tree containing the newly added nodes and edges. Based on the non-cyclic connection tree and state transition rules, perform state node expansion iterations to explore all paths to the target state node. Prune paths that cannot reach the target state node or violate key constraints to generate a system state non-cyclic connection tree model containing all feasible paths from the current state node to the target state node.

[0012] According to one embodiment of the present invention, using the system state acyclic connection tree model, a linear time complexity shortest path algorithm optimized for directed acyclic graphs is applied to calculate the optimal path from the current state node to the target state node, and obtain the optimal control sequence representing the optimal control strategy, including: For the system state acyclic connection tree model and the preset global optimization objective, a path cost function is defined according to the optimization objective. The path cost function can accumulate the cost of all edges on the path, and the costs of different dimensions are weighted or combined according to the optimization objective to create a path total cost calculation function. Using the system state acyclic connection tree model, the current state node and the target state node, and the path total cost calculation function, the directed acyclic graph shortest path algorithm is used to calculate the minimum cumulative cost from the current state node to each node, and the predecessor nodes and edges traversed to reach the minimum cumulative cost are recorded. The output includes the minimum cost of each node and the calculation results for the predecessor nodes and edges.

[0013] According to one embodiment of the present invention, using the system state acyclic connection tree model, a linear time complexity shortest path algorithm optimized for directed acyclic graphs is applied to calculate the optimal path from the current state node to the target state node, and obtain the optimal control sequence representing the optimal control strategy, including: Based on the minimum cost of each node and the calculation results of predecessor nodes and edges, as well as the target state node, starting from the target state node, the path is traversed in reverse according to the recorded predecessor information until the current state node is reached, and the reverse optimal path data is obtained. The reverse optimal path data is rearranged in order, and the path information is organized in the order from the current state node to the target state node to construct the optimal control sequence representing the optimal control strategy.

[0014] According to one embodiment of the present invention, based on the optimal control sequence, control instructions are generated in a time-ordered manner, and the data acquisition frequency and alarm threshold of the monitoring system are dynamically adjusted. The control instructions are then sent to the equipment actuators of the pumped storage power station monitoring system, comprising: Traverse each edge in the optimal control sequence, and parse each state transition into basic control actions with time-series relationships based on the edge attributes, generating a preliminary control action list containing specific operations, parameters, and execution timing. Each basic control action in the preliminary control action list is converted into a standard control instruction that conforms to the communication protocol and instruction format of the underlying execution unit.

[0015] According to one embodiment of the present invention, based on the optimal control sequence, control instructions are generated in a time-ordered manner, and the data acquisition frequency and alarm threshold of the monitoring system are dynamically adjusted. The control instructions are then sent to the equipment actuators of the pumped storage power station monitoring system, comprising: The optimal control sequence, the status of the pumped storage power station monitoring system, and the control objectives are analyzed to evaluate the characteristics of the optimal control sequence and the requirements of the pumped storage power station monitoring system for monitoring accuracy and response speed. The parameters of the monitoring system itself are dynamically adjusted to obtain the updated set of monitoring system operating parameters. The updated monitoring system operating parameters are applied to the monitoring system to match the monitoring behavior with the control tasks.

[0016] The present invention also provides a control device for a pumped storage power station monitoring system, comprising: The data acquisition and fusion module is used to collect and aggregate internal status data from the power grid dispatch center, local sensors of the power station and the pumped storage power station monitoring system in real time according to the configuration information of the monitoring system and the operating conditions of the monitoring power station, construct a unified formatted multi-source real-time data stream, and transmit the unified formatted multi-source real-time data stream to the state tree model construction module. The state tree model construction module is used to receive the unified formatted multi-source real-time data stream, combine the power plant equipment model, operating constraints and state transition rules, construct a system state non-cyclic connection tree model representing all feasible paths from the current state to the target state, and transmit the system state non-cyclic connection tree model to the optimal path calculation module. The optimal path calculation module is used to receive the system state acyclic connection tree model, apply the linear time complexity shortest path algorithm optimized for directed acyclic graphs, calculate the optimal path from the current state node to the target state node, output the optimal control sequence representing the optimal control strategy, and transmit the optimal control sequence to the control command generation and execution module. The control command generation and execution module is used to receive the optimal control sequence, generate control commands arranged in time sequence, dynamically adjust the data acquisition frequency and alarm threshold of the monitoring system, and send the control commands to the equipment execution mechanism of the pumped storage power station monitoring system.

[0017] Compared with the prior art, the technical solution of the present invention has the following beneficial effects: This invention proposes a method for constructing a non-cyclic connection tree model of a system state to represent all feasible transition paths and their costs from the current state to the target state of a pumped storage power station, providing a structured foundation for subsequent optimization calculations. This model comprehensively considers all possible states of the system and their transition paths, which helps in discovering globally optimal or near-optimal control strategies.

[0018] The core of this invention lies in utilizing the acyclic (DAG) characteristics of the constructed model and applying a faster shortest path algorithm to calculate the optimal control sequence. Compared to traditional methods, this approach can find control strategies that satisfy complex constraints and optimization objectives more quickly, thus improving response speed. This algorithm fully leverages the characteristics of DAGs, avoids the complexity of handling loops, and significantly improves computational efficiency. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings used in the embodiments will be briefly described below. These drawings are incorporated in and constitute a part of this specification. They illustrate embodiments conforming to this disclosure and, together with the specification, serve to explain the technical solutions of this disclosure. It should be understood that the following drawings only show some embodiments of this disclosure and should not be considered as limiting the scope. Those skilled in the art can obtain other related drawings based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of a control method for a pumped storage power station monitoring system according to an embodiment of the present invention; Figure 2 This is a detailed flowchart of the steps for real-time acquisition and fusion of multi-source information according to an embodiment of the present invention; Figure 3 This is a detailed flowchart of the steps for constructing a system state non-cyclic connected tree model according to an embodiment of the present invention; Figure 4 This is a detailed flowchart of the steps for calculating the optimal control sequence using the shortest path algorithm according to an embodiment of the present invention; Figure 5 This is a detailed flowchart of the dynamic control instruction generation and execution steps according to an embodiment of the present invention; Figure 6 This is a structural block diagram of a control device for a pumped storage power station monitoring system according to an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. The components of the embodiments of this disclosure described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this disclosure provided in the accompanying drawings is not intended to limit the scope of the claimed disclosure, but merely represents selected embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without inventive effort are within the scope of protection of this disclosure.

[0022] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0023] In this document, the term "and / or" merely describes a relationship, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Furthermore, the term "at least one" in this document means any combination of at least two of any one or more elements. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Example 1 like Figure 1 As shown, according to a first embodiment of the present invention, a control method for a pumped storage power station monitoring system is provided, comprising: Step S1: Based on the monitoring system configuration information and the monitoring power station operation requirements, collect and aggregate the internal status data from the power grid dispatch center, the power station local sensors, and the pumped storage power station monitoring system in real time to form a unified formatted multi-source real-time data stream. This step first identifies and determines which data needs to be collected based on the pre-configured equipment list, measurement point list, and current operational requirements in the monitoring system. For example, in a pumped storage power station, this data may include equipment parameters such as upstream and downstream reservoir water levels, unit operating modes (power generation / pumping / phase adjustment / shutdown), speed, power output / input values, guide vane opening, bearing temperature, and vibration values, as well as instruction information from the power grid dispatch center (such as target power, operating mode switching requirements, etc.). The monitoring system actively initiates requests to these data sources or establishes data subscriptions through various communication protocols (such as OPC, Modbus, IEC 61850, etc.) to obtain real-time data.

[0026] The collected raw data often suffers from inconsistent formats and varying quality. The monitoring system cleans and preprocesses this data, including validity checks (e.g., determining if values ​​are within reasonable ranges), filtering (removing noise interference), and repairing or eliminating bad data points. For multiple measurements of the same parameter (e.g., the temperature of the same bearing from different sensors), the monitoring system applies data fusion techniques (e.g., weighted averaging, Kalman filtering) to improve data reliability. Finally, the monitoring system converts and organizes all processed data according to an internally defined data model, forming a unified data stream that provides a foundation for subsequent decision analysis.

[0027] Step S2: Based on the unified formatted multi-source real-time data stream, combined with the power plant equipment model, operating constraints, and state transition rules, construct a monitoring system state non-cyclic connection tree model representing all feasible paths from the current state to the target state; This step first analyzes the real-time data stream generated in step S1, combining it with predefined equipment models (such as turbine characteristic curves, generator parameters, etc.) and operational constraints (such as maximum power limits, minimum flow requirements, equipment temperature limits, etc.) to accurately determine the current state of the monitoring system. This state can be understood as a multi-dimensional vector containing the current values ​​of all key state variables. Simultaneously, the monitoring system determines the target state it needs to achieve by parsing instructions from the dispatch center or locally set operating plans.

[0028] After determining the starting point (current state) and the ending point (target state), the monitoring system creates an initial state tree data structure, with the current state as the root node. Then, the monitoring system begins exploring possible one-step transition operations from the current state. Examples include "increase guide vane opening by 2%" and "activate the auxiliary cooling monitoring system." For each possible operation, the monitoring system predicts the new state of the monitoring system after executing the operation and evaluates the cost of this transition (such as time consumption, energy loss, and impact on equipment lifespan). If the new state satisfies operational constraints and does not appear in the previous path (to avoid loops), it is added as a child node of the current node.

[0029] The monitoring system continuously repeats this expansion process until all feasible paths to the target state are found, or a preset search depth limit is reached. During this process, the monitoring system simultaneously performs pruning operations, removing branches of paths that are clearly impossible to reach the target or have excessively high costs, in order to control the size of the search space. Ultimately, the monitoring system constructs a non-cyclic connected tree model containing all feasible paths; this model is essentially a directed acyclic graph (DAG).

[0030] Step S3: Using the state acyclic connection tree model of the monitoring system, apply the shortest path algorithm with linear time complexity optimized for directed acyclic graphs to calculate the optimal path from the current state node to the target state node, and obtain the optimal control sequence representing the optimal control strategy. This step first defines a path cost function based on the control objective (such as shortest response time, highest efficiency, best stability, or a weighted combination of these factors). This function accumulates the multidimensional costs of all edges on the path and converts them into a comprehensive evaluation value. For example, if the primary optimization objective is "shortest time," then the time dimension can be given a higher weight, while other dimensions can be given lower weights.

[0031] With the cost function in place, the monitoring system applies a shortest path algorithm optimized for directed acyclic graphs. Since the model constructed in step S2 is acyclic, the monitoring system can leverage this characteristic by employing a dynamic programming method based on topological sorting. This method has a time complexity of O(V+E), where V is the number of nodes and E is the number of edges. This is more efficient than the general Dijkstra's algorithm (complexity O(E+VlogV)).

[0032] During algorithm execution, the monitoring system calculates the minimum cumulative cost from the current state node to each node in the graph and records the predecessor nodes traversed to achieve that cost. Once the calculation is complete, the monitoring system starts from the target node and, based on the recorded predecessor information, traces backward to find the optimal path from the current state to the target state. Finally, the monitoring system reverses this path to obtain a forward control sequence, which includes all intermediate transformation operations and their parameters required to reach the target state from the current state.

[0033] Step S4: Based on the optimal control sequence, generate control instructions arranged in time sequence, dynamically adjust the data acquisition frequency and alarm threshold of the monitoring system, and send the control instructions to the equipment actuators of the pumped storage power station monitoring system.

[0034] This step first converts the optimal control sequence obtained in step S3 into specific executable instructions. The monitoring system traverses each edge (representing a state transition) in the control sequence, analyzes its attributes, and transforms the abstract state transition into a series of basic control actions, such as "increase the guide vane opening to 40% at a rate of 1% per second" and "adjust the excitation current to 2000A". These basic actions are organized according to strict timing relationships to form a preliminary list of control actions.

[0035] Next, the monitoring system further converts these abstract control actions into standard control instructions that conform to the communication protocols of the underlying devices. For example, "increase guide vane opening" is converted into a Modbus write command for a specific PLC, specifying the target register address and value.

[0036] Meanwhile, the monitoring system will dynamically adjust its own parameters based on the characteristics of the optimal control sequence to match the needs of the current control task. For example, during the rapid power adjustment phase, the sampling frequency of power-related parameters may be temporarily increased; during equipment startup, the alarm thresholds of certain parameters may be appropriately relaxed to avoid unnecessary alarms triggered by normal fluctuations during the transition process.

[0037] Finally, the monitoring system sends control commands to each actuator (such as speed governor, excitation monitoring system, valve drive device, etc.) according to the calculated optimal timing, and monitors the execution status in real time, making fine adjustments as necessary. This dynamic control and monitoring coordination method ensures high precision in the control process and stability of the monitoring system.

[0038] The pumped storage power station monitoring system control method provided in this embodiment processes multi-source information in a structured manner, constructs a non-cyclic connected tree model, and applies an optimized shortest path algorithm to calculate the optimal control strategy, ultimately generating and executing dynamic control commands. This method features fast response speed, high control accuracy, and strong monitoring system stability, effectively addressing the control requirements of pumped storage power stations under complex operating conditions.

[0039] The pumped storage power station monitoring system control method provided in this embodiment processes multi-source information in a structured manner, constructs a non-cyclic connected tree model, and applies an optimized shortest path algorithm to calculate the optimal control strategy, ultimately generating and executing dynamic control commands. This method features fast response speed, high control accuracy, and strong monitoring system stability, effectively addressing the control requirements of pumped storage power stations under complex operating conditions.

[0040] Example 2 According to a second embodiment of the present invention, based on the above-described embodiment 1, a specific implementation method for real-time acquisition and fusion of multi-source information is further described. For example... Figure 2 As shown, based on the monitoring system configuration information and the operational requirements of the power station, real-time data is collected and aggregated from the power grid dispatch center, local sensors at the power station, and the internal status data of the pumped storage power station monitoring system to form a unified, formatted, multi-source real-time data stream, including: Step S1.1: Analyze the preset configuration information of the monitoring system and the current operating conditions of the monitoring power station, identify the data items necessary to complete the control task, locate the physical sensors, computing nodes and external data interfaces corresponding to the data items, and obtain the list of data items to be collected and their corresponding specific data sources. In this embodiment, during the data requirement identification and source location process in step S1.1, the monitoring system identifies the data items necessary to complete the current control task based on the preset configuration information of the monitoring system (e.g., equipment list, measurement point list) and the current operational requirements (defined by power grid dispatch instructions or local operation plans). These data items include power grid frequency, voltage, dispatch power instructions, unit speed, power, guide vane opening, temperature of each bearing, vibration value, water level of upstream and downstream reservoirs, gate opening, and valve status. Then, the monitoring system locates the physical sensors, computing nodes, or external data interfaces (such as the communication interface with the power grid dispatch center) corresponding to these data items.

[0041] Step S1.2: Based on the list of data items and their corresponding specific data source lists, initiate a data request or subscription service to the specified data source through the communication interface of the monitoring system, acquire raw measurement values ​​or status signals at a preset or dynamically adjusted frequency, and generate a raw multi-source heterogeneous data stream containing timestamps; During the real-time data acquisition process in step S1.2, the monitoring system initiates data requests or subscription services to the designated data source (PLC, RTU, sensor, dispatch monitoring system) through the communication interface of the monitoring system (such as OPC, Modbus, IEC 61850, etc.) and acquires raw measurement values ​​or status signals at preset or dynamically adjusted frequencies.

[0042] Step S1.3: Process the original multi-source heterogeneous data stream and uniformly add accurate timestamps to obtain a processed and timestamped valid data set.

[0043] During the data preprocessing and cleaning process in step S1.3, the monitoring system performs validity checks (such as range verification and state consistency verification), filtering (removing noise interference), bad data removal or repair (interpolation based on historical data or data from nearby measurement points) on the received raw data, and uniformly adds accurate timestamps.

[0044] This embodiment ensures the quality and reliability of the data foundation upon which subsequent control decisions rely through this hierarchical data acquisition and processing method.

[0045] Example 3 According to a third embodiment of the present invention, based on the above embodiment 2, the specific implementation method of data fusion and formatting is further described. Based on the monitoring system configuration information and the operating conditions requirements of the power station, real-time data from the power grid dispatch center, local sensors at the power station, and the internal status data of the pumped storage power station monitoring system are collected and aggregated to form a unified formatted multi-source real-time data stream, including: Step S1.4: Apply data fusion technology to the cleaned and timestamped valid data set to fuse multiple measurements of the same physical quantity to obtain fused data; In this embodiment, the data fusion process in step S1.4 can employ techniques such as weighted averaging and Kalman filtering to fuse multiple measurements of the same physical quantity, thereby improving data accuracy and reliability. For example, for the same bearing temperature data obtained from different sensors, different weights can be assigned based on factors such as sensor accuracy, reliability, and data freshness before fusion to obtain a more accurate temperature estimate.

[0046] Step S1.5: Convert and organize the fused data according to the unified data model defined within the pumped storage power station monitoring system to construct a unified formatted multi-source real-time data stream.

[0047] In step S1.5, the monitoring system converts and organizes all data from all sources according to a unified data model defined within the monitoring system. This unified data model ensures that data from different sources and of different types can be processed and used consistently within the monitoring system.

[0048] Through data fusion and formatting, this embodiment can convert multi-source heterogeneous raw data into a high-quality, uniformly formatted data stream, providing a reliable data foundation for subsequent state tree model construction and path planning.

[0049] Example 4 According to a fourth embodiment of the present invention, based on the above embodiments, a specific implementation method for determining the current state and the target state is further described. For example... Figure 3 As shown, based on the unified formatted multi-source real-time data stream, combined with the power plant equipment model, operational constraints, and state transition rules, a non-cyclic connection tree model of the monitoring system state, representing all feasible paths from the current state to the target state, is constructed, including: Step S2.1: Using the unified formatted multi-source real-time data stream, as well as the predefined power station equipment physical model, operating constraints, state transition rules, and received target scheduling instructions, extract key state variables from the multi-source real-time data stream, determine the current precise operating state of the pumped storage power station monitoring system, and construct a state node to represent the current state of the pumped storage power station monitoring system. In this embodiment, during step S2.1 of determining the current state, the monitoring system extracts a set of predefined key state variables (KSVs) from the real-time data stream. These variables are sufficient to describe the macroscopic state of the power plant's current operation. For example, KSVs may include: unit operating mode (generation / pumping / shutdown / phase adjustment), current output / input power, unit speed, guide vane opening, upstream reservoir water level, downstream reservoir water level, key equipment temperature, and relevant switch status. The values ​​of these KSVs collectively constitute the current state node S. current It can be imagined as a multidimensional vector.

[0050] Step S2.2: Parse the target scheduling instructions in the unified formatted multi-source real-time data stream, combine the device model and constraints, determine the target state that the monitoring system needs to achieve, and construct a clearly defined target state node.

[0051] In step S2.2, during the determination of the target state, the monitoring system parses the target scheduling instruction. For example, the instruction might be "achieve a power generation capacity of 300MW within 10 minutes." Combining the equipment model and operational constraints, the monitoring system determines the feasibility of this instruction and transforms it into a specific target state description. Target state node S targetSimilarly, it is defined by a set of KSVs values, some of which are directly required by the instruction (e.g., power = 300MW, mode = power generation), while other values ​​are calculated based on the model and constraints (e.g., to achieve 300MW of power generation, the guide vane opening should be within a certain range and the rotational speed should be the rated speed, based on the efficiency curve and head). If the instruction includes time constraints, these will also serve as important constraints for subsequent path searches.

[0052] In this way, this embodiment can accurately define the start and end points of the control problem, laying the foundation for subsequent path search.

[0053] Example 5 According to a fifth embodiment of the present invention, based on the above embodiment 4, a specific implementation method for the construction and expansion of the state tree is further described. Based on the uniformly formatted multi-source real-time data stream, combined with the power plant equipment model, operating constraints, and state transition rules, a non-cyclic connection tree model of the monitoring system state, representing all feasible paths from the current state to the target state, is constructed, including: Step S2.3: Based on the clearly defined current state node and target state node, as well as the device model, constraints and state transition rules, create a data structure to represent the state tree, and set the current state node as the root node of the tree, and construct an initial state tree structure containing the root node; In this embodiment, during step S2.3 of initializing the state tree, the monitoring system creates an appropriate data structure (such as an adjacency list or a list of node objects) to represent the state tree, and sets the current state node S... current Set it as the root node of the tree.

[0054] Step S2.4: Using the initial state tree structure containing the root node, according to the state transition rules, find all possible one-step transformation operations, calculate the new state node caused by them, evaluate the cost of the transformation operation, and form a partially expanded non-cyclic connection tree containing the newly added nodes and edges. During the exploration and expansion process in step S2.4, the monitoring system never fully expanded any nodes (initially S). current Starting from the state transition rules, the system searches for all possible one-step transition operations. For each valid transition operation (which does not violate operational constraints), the monitoring system calculates the resulting new state node S. next And assess the costs of this conversion operation (such as time consumption, energy loss, stability impact, etc.). The monitoring system ensures S next The node S has not appeared on the path from the root node to the current node to ensure the non-cyclic property of the tree. next Add it as a child node to the current node and record the transformation operation and its cost.

[0055] Step S2.5: Perform state node expansion iteration according to the non-cyclic connection tree and state transition rules, explore all paths that can reach the target state node, prune paths that cannot reach the target state node or violate key constraints, and generate a monitoring system state non-cyclic connection tree model containing all feasible paths from the current state node to the target state node.

[0056] During the tree completion and pruning process in step S2.5, the monitoring system repeats the expansion process in step S2.4 until the termination conditions are met (such as finding the target path, reaching the search depth limit, etc.). Then, paths that cannot reach the target or violate constraints are pruned to optimize the state tree structure.

[0057] By using this state-space exploration method with a systematic monitoring approach, this embodiment can construct an optimized state-acyclic connection tree model containing all feasible paths, providing a complete search space for subsequent optimal path calculation.

[0058] According to a sixth embodiment of the present invention, based on the above embodiments, the specific implementation of the path cost function definition and the shortest path algorithm execution is further described. For example... Figure 4 As shown, using the state acyclic connection tree model of the monitoring system, and applying a linear time complexity shortest path algorithm optimized for directed acyclic graphs, the optimal path from the current state node to the target state node is calculated, obtaining the optimal control sequence representing the optimal control strategy, including: Step S3.1: For the non-cyclic connection tree model of the monitoring system state and the preset global optimization objective, define a path cost function according to the optimization objective. The path cost function can accumulate the cost of all edges on the path, and weight or combine the costs of different dimensions according to the optimization objective to create a path total cost calculation function. In the operation and control of pumped storage power stations, multiple optimization objectives typically need to be considered, such as response speed, energy efficiency, equipment stress, and stability. The core task of step S3.1 is to mathematically integrate these potentially conflicting objectives into a unified evaluation criterion, namely the path cost function. This function will be used to compare the merits of different control paths, providing a metric for the shortest path algorithm.

[0059] First, the monitoring system determines the weights of each dimension based on preset global optimization goals. For example, if grid dispatch requires rapid response, the weight of the time dimension will be set higher; if economic operation is prioritized, the weight of the energy consumption dimension will be higher; if equipment has just undergone maintenance, stability may be emphasized, and the weight of stability indicators will be increased accordingly. These weights can be preset based on expert experience or dynamically adjusted according to the current grid status or specific dispatch instructions.

[0060] Next, the monitoring system defines multi-dimensional cost metrics for each edge (i.e., each state transition operation) in the state transition tree. Typical cost dimensions include: Time cost: The estimated time required to complete this state transition, which may be calculated based on factors such as equipment characteristics and the range of operation. For example, adjusting the guide vane opening from 30% to 40% can be calculated to take approximately 10 seconds based on the guide vane adjustment rate limit (e.g., a maximum of 1% per second).

[0061] Energy costs: The anticipated energy loss or additional consumption during the conversion process. Examples include reduced efficiency due to rapid adjustments and the power consumption of starting auxiliary equipment.

[0062] Stability Cost: The potential impact of this transition on the stability of the monitoring system is typically represented as a penalty value. For example, rapid and significant adjustments may lead to water hammer or oscillations in the power monitoring system, resulting in a higher stability cost.

[0063] Equipment stress cost: The potential impact of operation on equipment lifespan, which may take into account factors such as temperature change rate and number of start-stop cycles.

[0064] With the multidimensional costs of the edges and the weights of each dimension, the total path cost calculation function is usually in the form of a weighted sum: TotalCost(Path) = w1×Sum(Time Cost) + w2×Sum(Energy Cost) + w3×Sum(Stability Cost) + w4×Sum(Equipment Stress Cost) Where w1, w2, w3, and w4 are the weight coefficients of the corresponding dimensions, and Sum represents the sum of the costs of all edges on the path in that dimension.

[0065] In some cases, monitoring systems may employ more complex combinations, such as introducing nonlinear relationships or threshold constraints (for example, assigning an extremely high penalty value if any operation on a path causes the rate of temperature change to exceed a threshold, effectively excluding that path).

[0066] In this way, the monitoring system creates a path total cost calculation function that can comprehensively evaluate the advantages and disadvantages of different control paths, providing a clear optimization objective for the shortest path algorithm in step S3.2.

[0067] Step S3.2: Using the state acyclic connection tree model of the monitoring system, the current state node and the target state node, and the total path cost calculation function, the shortest path algorithm for directed acyclic graphs is used to calculate the minimum cumulative cost from the current state node to each node, and the predecessor nodes and edges traversed to reach the minimum cumulative cost are recorded. The output includes the minimum cost of each node and the calculation results for the predecessor nodes and edges.

[0068] Having defined the total path cost calculation function in step S3.1, the task of step S3.2 is to find the optimal path from the current state to the target state in the state space. The key here is to utilize the acyclic property of the state tree (i.e., a directed acyclic graph, DAG) and employ a more efficient algorithm to calculate the shortest path.

[0069] First, the monitoring system performs a topological sort on the state tree model. Topological sorting arranges all nodes in a directed acyclic graph into a linear sequence, such that for any directed edge (u, v) in the graph, node u always appears before node v in the sequence. For the state transitions of a pumped storage power station, this means sorting them according to the logical order of the state transitions. For example, "opening the inlet valve" must precede "adjusting the guide vane opening," and therefore will appear first in the topological sequence.

[0070] After completing the topology sorting, the monitoring system uses dynamic programming to calculate the shortest path. The core idea of ​​this method is: if the shortest path from the starting point to node u is known, then the shortest path from the starting point to any successor node v of u is the minimum of the sum of the shortest path from the starting point to u and the cost of the edge from u to v. The specific algorithm steps are as follows: Initialization: Set the shortest path cost of the starting point (current state node) to 0, and set the shortest path cost of all other nodes to infinity.

[0071] Process each node u sequentially according to the topological sorting order: For each successor node v of u, check if the path from u to v is shorter than the currently known shortest path to v.

[0072] If so, update the shortest path cost of v and record u as the predecessor node of v.

[0073] For example, suppose the currently processed node u represents the state of "guide vane opening 20%, power 100MW", and its shortest path cost has been calculated to be 30 (possibly the result weighted by the cost over 30 seconds). Node u has a successor node v representing "guide vane opening 30%, power 150MW", and the edge cost from u to v is 15. At this time, the monitoring system will check whether "30+15=45" is less than the shortest path cost currently recorded for v. If so, the shortest path cost of v is updated to 45, and u is recorded as the predecessor node of v.

[0074] This dynamic programming method based on topological sorting has a time complexity of O(V+E), where V is the number of nodes in the state tree and E is the number of edges. Compared to the general Dijkstra algorithm (complexity O(E+VlogV)), it is more efficient when dealing with directed acyclic graphs, especially when the state space is large, where the performance improvement brought by this optimization is very significant.

[0075] In certain complex situations, such as when the state space is particularly large or when heuristic search needs to be considered, the monitoring system may employ an AGI optimized for DAG. Algorithm variant. This method combines the efficiency of dynamic programming with the guidance of heuristic search, allowing for faster finding of near-optimal solutions.

[0076] After the algorithm is executed, the monitoring system receives the following output: The minimum cumulative cost from the current state node to every node in the graph (especially the target state node). The predecessor nodes and corresponding edges (i.e., state transition operations) that achieve these minimum costs. This information forms the basis for step S3.3, which involves backtracking to generate the optimal control sequence.

[0077] Example 7 According to the seventh embodiment of the present invention, based on the above embodiment 6, a specific implementation method for generating the optimal control sequence is further described. Using the state acyclic connection tree model of the monitoring system, and applying a linear time complexity shortest path algorithm optimized for directed acyclic graphs, the optimal path from the current state node to the target state node is calculated to obtain the optimal control sequence representing the optimal control strategy, including: Step S3.3: Based on the minimum cost of each node and the calculation results of predecessor nodes and edges, as well as the target state node, starting from the target state node, perform a reverse traversal of the path according to the recorded predecessor information until the current state node is reached, and obtain the reverse optimal path data. Step S3.3 is a path reconstruction process, the purpose of which is to find the complete optimal path from the current state to the target state based on the shortest path information calculated in step S3.2. This process uses a reverse tracing method, tracing back from the endpoint to the starting point, and then adjusting the order to obtain an executable control sequence.

[0078] First, the monitoring system locates the target status node (S). target In step S3.2, the monitoring system has calculated the minimum cumulative cost from the current state node to each node in the graph (including the target state node) and recorded the information of the predecessor nodes and edges that reach that cost. This predecessor information forms a "pointer tree" that indicates the path back to the starting point from any node (if reachable).

[0079] Starting from the target state node, the monitoring system searches for its predecessor node record. For example, if the target state is "power generation mode, power 300MW", its predecessor might be "power generation mode, power 280MW". The monitoring system obtains this predecessor node and records the edge from the predecessor to the current node (i.e., the state transition operation, such as "increase power generation by 20MW") and its attributes (such as execution time, method, etc.).

[0080] Then, the monitoring system continues to search for the predecessor of this predecessor node, repeating the above process until it traces back to the starting point (current state node S). current At each step, the monitoring system records information about nodes and edges, forming a reverse path data structure. This data structure contains all intermediate states and state transition operations from the endpoint to the starting point.

[0081] In practical applications of pumped storage power stations, this reverse traversal process may involve state changes in multiple dimensions. For example, the path from "pumping mode, power -250MW, high water level" to "power generation mode, power 300MW, slightly low water level" may include several intermediate states: first reducing pumping power, then shutting down, switching operating direction, starting power generation mode, and gradually increasing power generation, etc. Each state transition has specific operational requirements and execution conditions.

[0082] By performing this reverse traversal, the monitoring system can accurately reconstruct the optimal control path, ensuring that every transformation step required to achieve the goal is recorded, thus laying the foundation for the next step of forward rearrangement.

[0083] Step S3.4: Rearrange the reverse optimal path data in order, organize the path information in the order from the current state node to the target state node, and construct the optimal control sequence representing the optimal control strategy.

[0084] The reverse optimal path data obtained in step S3.3 is a backtracking from the target state node to the current state node, in the reverse order of execution. The task of step S3.4 is to convert this reverse path into a control sequence arranged according to the actual execution order, so that the subsequent monitoring system can execute these control operations in the correct order.

[0085] Reordering is a relatively simple process. The monitoring system reverses the sequence of nodes and edges in the reverse path to obtain the forward path from the current state to the target state. In this forward path, nodes represent the states that the monitoring system will experience during the execution of control, and edges represent the control operations that need to be performed from one state to the next.

[0086] During the rearrangement process, the monitoring system retains all attribute information for each state transition operation (edge), such as execution method, parameter settings, and cost estimates. This information is crucial for the generation of subsequent specific control commands. For example, an edge might represent "increasing the guide vane opening from 25% to 35%", and its attributes might include "adjustment rate of 1% per second", "estimated time of 10 seconds", and "energy cost of x".

[0087] After rearrangement, the monitoring system obtains a complete optimal control sequence, which can be represented in the following form: [(S current , Op1, S1), (S1, Op2, S2), ..., (Sn-1, Opn, S target )] Among them, S current S1, S2, ..., Sn-1 are the initial states, and S is the intermediate state. target Op1, Op2, ..., Opn are the target states, and Op1, Op2, ..., Opn are the corresponding control operations.

[0088] This control sequence contains all the operations required for the monitoring system to reach the target state from the current state, arranged in chronological order, with detailed execution information accompanying each operation. This sequence is the direct input to the subsequent step S4 (generating and executing dynamic control instructions).

[0089] By rationally defining the cost function, efficiently executing the shortest path algorithm, accurately backtracking the path, and rearranging its order, this embodiment can find a control sequence that satisfies multi-objective optimization in a complex state space, providing decision support for the precise control of pumped storage power stations. This method is particularly suitable for handling scenarios with complex operating conditions and high control requirements, such as rapid response to grid dispatch commands and smooth switching of operating modes.

[0090] Example 8 According to the eighth embodiment of the present invention, based on the above embodiments, a specific implementation method for control instruction parsing and generation is further described. For example... Figure 5 As shown, based on the optimal control sequence, control commands are generated in a time-ordered manner, and the data acquisition frequency and alarm threshold of the monitoring system are dynamically adjusted. The control commands are then sent to the equipment actuators of the pumped storage power station monitoring system, including: Step S4.1: Traverse each edge in the optimal control sequence, and parse each state transformation into basic control actions with time-series relationships according to the attributes of the edge, generating a preliminary control action list containing specific operations, parameters and execution sequence. The core task of step S4.1 is to transform the abstract control sequence output in step S3 into a more specific, executable list of control actions. This is the first step in the transformation from control strategy to actual operation, concretizing the high-level state transition description into a series of atomic-level operation instructions.

[0091] First, the monitoring system iterates through each edge in the optimal control sequence in chronological order. Each edge represents a state transition, such as "increasing power from 180MW to 220MW" or "switching from shutdown to generation." These abstract state transitions need to be parsed into basic control actions that the actual equipment can perform.

[0092] Taking the typical control of a pumped storage power station as an example, when the monitoring system encounters an edge indicating "switching from shutdown to generation, target power 150MW", it needs to decompose this high-level instruction into a series of ordered basic actions, such as: Close the unit maintenance isolation door; Start the hydraulic device and wait for the hydraulic pressure to build up (estimated 10 seconds). Open the main inlet valve to 100% (at a rate of 5% / second); Wait for confirmation signal that the inlet valve is fully open; Open the guide vanes to 5% of their initial opening; Wait for the unit speed to reach 95% of the rated value; Close the main circuit breaker of the generating unit; Establish excitation (set voltage regulation target); According to the power adjustment curve, the guide vane opening is gradually adjusted to 25% (at a rate of 0.5% / second). Monitor the power output until it stabilizes at 150MW; During this analysis process, the monitoring system will consider many factors: Equipment physical characteristics and limitations: for example, the maximum rate of guide vane adjustment, the acceleration characteristics of the turbine, etc.

[0093] Safety constraints and protection requirements: For example, the rotational speed must be within a safe range before closing the circuit.

[0094] Temporal dependency: Some operations can only be performed after other operations have been completed.

[0095] Parallelism possibility: Some operations can be performed simultaneously, such as other preparatory work being carried out while waiting for oil pressure to build up.

[0096] For each basic control action, the monitoring system will define it in detail: Operation targets: such as guide vanes, inlet valves, circuit breakers and other specific equipment.

[0097] Operation type: such as turn on, turn off, adjust to a specific value, change at a specific rate, etc.

[0098] Target parameters: such as target opening degree, target pressure value, etc.

[0099] Execution conditions: such as confirmation of completion of the preceding action, or reaching a threshold for a specific parameter.

[0100] Estimated execution time: The estimated completion time calculated based on equipment characteristics and target parameters.

[0101] Verification conditions: Condition checks to confirm that the action has been completed correctly.

[0102] This parsing process ensures that high-level control strategies can be translated into specific instructions understandable at the device level, while maintaining operational logic and security. Ultimately, the monitoring system generates a preliminary list containing all necessary basic control actions, arranged in a strict temporal sequence, preparing for subsequent conversion into standard control instructions.

[0103] Step S4.2: Convert each basic control action in the preliminary control action list into a standard control instruction that conforms to the communication protocol and instruction format of the underlying execution unit.

[0104] The task of step S4.2 is to convert the list of basic control actions generated in step S4.1 into standard control instructions that can actually be executed. These instructions must conform to the communication protocol and instruction format requirements of the underlying execution device. This is the final conversion step from logical operation description to physical control signals.

[0105] In modern pumped-storage power plants, the control and monitoring system typically adopts a hierarchical distributed architecture, including a central monitoring system, process control stations, and various programmable logic controllers (PLCs), remote terminal units (RTUs), and intelligent actuators. These devices may come from different manufacturers and support different communication protocols and command formats. Step S4.2 needs to address this heterogeneity to ensure that control commands are correctly transmitted and executed.

[0106] First, the monitoring system establishes a mapping relationship, associating basic control actions with corresponding execution units. For example, the operation of "adjusting the guide vane opening" may need to be mapped to the guide vane control PLC; the operation of "closing the circuit breaker" may need to be mapped to the switchgear control unit. This mapping is usually based on the monitoring system configuration database, which stores equipment topology and communication parameter information.

[0107] Then, for each basic control action, the monitoring system converts it into specific standard control instructions based on the characteristics of the target execution unit: If the target is a PLC that supports the Modbus protocol, the monitoring system will generate the appropriate Modbus function code and register address / value combination. For example, "adjust guide vane opening to 30%" is converted to "write value 3000 (representing 30.00%) to register 4001".

[0108] If the target supports the IEC 61850 standard, the monitoring system will generate an MMS (Manufacturing Message Specification) message that conforms to the standard, such as "Write the value true to the LLN0.CSWI1.Pos object" to operate a switch.

[0109] For devices with proprietary protocols, the monitoring system will call a predefined protocol conversion module to convert standardized control requests into a format that the specific device can understand.

[0110] During the conversion process, the monitoring system will also handle the following key issues: Command parameter unit conversion: Ensure that the unit of the value conforms to the requirements of the target device, such as converting percentages to raw values ​​or engineering units to the device's internal units.

[0111] Command splitting and combination: Some complex operations may need to be broken down into multiple atomic commands, or multiple related operations may need to be combined into a batch command.

[0112] Command priority assignment: Assign appropriate priorities to different control commands to ensure that critical operations can be executed in a timely manner.

[0113] Security check code generation: Adds necessary verification information to control commands to ensure data integrity during transmission.

[0114] Execution timestamp: Adds an execution timestamp to the instruction, supporting scheduled execution or sequence control.

[0115] After the conversion is complete, each basic control action becomes one or more standardized control instructions. These instructions contain all necessary elements such as target device address, opcode, parameter value, timestamp, and verification information, and can be directly sent to the corresponding execution unit through the communication network.

[0116] This precise instruction generation mechanism ensures that abstract control strategies can be accurately mapped to physical device operations, minimizing errors that may be introduced during human interpretation and conversion, and improving the accuracy and reliability of the control and monitoring system.

[0117] Example 9 According to the ninth embodiment of the present invention, based on the above embodiment 8, a specific implementation method for dynamically adjusting the parameters of the monitoring system is further described. Based on the optimal control sequence, control instructions arranged in a time sequence are generated, and the data acquisition frequency and alarm threshold of the monitoring system are dynamically adjusted. The control instructions are then sent to the equipment actuators of the pumped storage power station monitoring system, including: Step S4.3: Analyze the optimal control sequence, the status of the pumped storage power station monitoring system, and the control objectives; evaluate the characteristics of the optimal control sequence and the requirements of the pumped storage power station monitoring system for monitoring accuracy and response speed; dynamically adjust the parameters of the monitoring system itself; and obtain the updated set of monitoring system operating parameters. The core innovation of step S4.3 lies in treating the configuration parameters of the monitoring system as dynamic variables rather than fixed constants, thereby achieving coordinated optimization between monitoring behavior and control tasks. Different control tasks and operating conditions place different requirements on the monitoring system. By dynamically adjusting the monitoring parameters, the monitoring system can operate in the way most suitable for the current task, improving overall performance and reliability.

[0118] First, the monitoring system performs feature analysis on the optimal control sequence to identify key control stages and characteristics. For example, a control sequence may contain different types of stages: rapid power adjustment stage, equipment startup stage, steady-state operation stage, etc. Each stage has different requirements for the monitoring system. The rapid power adjustment phase requires a higher data acquisition frequency to accurately track the change process, but may tolerate some alarm thresholds with greater leniency (because parameters may fluctuate significantly in a short period of time).

[0119] During the equipment startup phase, it may be necessary to focus on the trend of specific parameters (such as the rate of increase of bearing temperature) rather than the absolute value.

[0120] During steady-state operation, stricter alarm thresholds are needed to detect anomalies as early as possible, but the data collection frequency may not need to be too high.

[0121] The monitoring system also considers the current status and resource availability, such as network bandwidth usage, data processing load, and storage space. When resources are limited, the monitoring system will prioritize ensuring the monitoring quality of critical parameters.

[0122] Based on these analyses, the monitoring parameters that the monitoring system dynamically adjusts mainly include: Data acquisition frequency: Different sampling periods are set for different devices and parameters. For example, during unit startup, the sampling frequency of speed and vibration may be temporarily increased from the standard 1 time per second to 10 times per second to capture rapidly changing details.

[0123] Alarm thresholds: Adjust the upper and lower limits and rate of change limits for various parameters. For example, during the rapid opening of the guide vanes, the alarm limit for water pressure fluctuations may be temporarily relaxed, while maintaining strict monitoring of the bearing temperature.

[0124] Data filtering parameters: Adjust the parameters of the filtering algorithm, such as the moving average window size and outlier judgment criteria, to balance sensitivity and anti-interference ability.

[0125] Status assessment rules: Modify the rules used to comprehensively assess the status of equipment or monitoring systems, such as weight adjustments and decision logic.

[0126] Data storage strategy: Adjust the storage granularity and retention period of different types of data, for example, record data of critical processes at a higher resolution.

[0127] Display refresh rate: Adjust the update frequency of the operator interface to ensure that sufficient and timely information is provided during critical control phases.

[0128] This dynamic adjustment is not arbitrary, but based on a pre-defined adjustment strategy library and equipment characteristic model. The monitoring system applies corresponding monitoring parameter adjustment templates based on the identified control sequence characteristics and makes fine adjustments according to the actual situation.

[0129] Finally, the monitoring system generates an updated set of operating parameters, which includes all parameters that need to be adjusted and their new values, preparing for the parameter application in step S4.4. This dynamic parameter adjustment ensures that the monitoring system operates in a manner best suited to the current control task, improving data quality and reliability while optimizing the use of monitoring system resources.

[0130] Step S4.4: Apply the updated monitoring system operating parameters to the monitoring system to match the monitoring behavior with the control task.

[0131] The task of step S4.4 is to apply the updated set of monitoring system operating parameters generated in step S4.3 to the various components of the monitoring system and ensure that the changes to these parameters take effect, so that the behavior of the monitoring system matches the current control task. This is the execution phase of monitoring and control co-optimization.

[0132] First, the monitoring system distributes the updated parameters to the corresponding target components based on the system's architecture and component distribution. In a typical pumped storage power station monitoring system, these components might include: Data Acquisition Unit (DAU): Responsible for acquiring and initially processing field signals.

[0133] Control station (CS): Responsible for executing control logic and issuing control commands.

[0134] Historical Data Server (HDS): Responsible for long-term data storage and trend analysis.

[0135] Human-Machine Interface (HMI): Provides an interactive interface for operators.

[0136] Alarm Management and Monitoring System (AMS): Responsible for generating, processing, and displaying alarms.

[0137] Parameter distribution needs to consider the characteristics and communication methods of different components. For example: For modern DAUs that support dynamic parameter configuration, the monitoring system may send new sampling frequency configurations directly via standard protocols such as OPC UA.

[0138] For some traditional devices, parameter adjustments may need to be achieved indirectly through an intermediary proxy service or configuration file updates.

[0139] For distributed components, the monitoring system needs to ensure the synchronization of parameter changes to avoid configuration inconsistencies between different components.

[0140] When application parameters change, the monitoring system will take a series of measures to ensure a smooth transition and operational safety: Timing of Changes: Apply changes at appropriate times to avoid data interruptions during critical control moments. For example, it might be necessary to wait until the current sampling period is complete before adjusting the sampling frequency.

[0141] Smooth transition strategy: For certain parameters (such as filter window size), gradual adjustment rather than abrupt changes may be adopted to avoid abrupt changes in data quality.

[0142] Change verification: After applying the parameters, the monitoring system will verify whether the change has taken effect successfully, for example, by checking whether the new data points arrive at the expected frequency.

[0143] Rollback mechanism: Saves the configuration before parameter changes, and can quickly restore the original settings if a problem is found.

[0144] Operator notification: Provide monitoring system operators with information on parameter changes, including which parameters were changed, the reasons for the changes, and the expected impact.

[0145] After the parameters are applied, the behavior of the monitoring system will change accordingly, for example: The data for key parameters is updated more frequently, providing a more detailed observation of the process.

[0146] The alarm thresholds for certain parameters are temporarily relaxed to avoid false alarms during transient processes.

[0147] Data filtering algorithms are made more sensitive or smoother to meet the needs of the current control phase.

[0148] The improved resolution of historical data records provides more detailed information for subsequent analysis.

[0149] Such dynamic adjustments to monitoring parameters are typically temporary and tailored to the specific characteristics of the control task. When the control sequence completes or enters a new phase, the monitoring system will readjust the parameters according to new requirements, or revert to the standard configuration after the task is fully completed.

[0150] By closely coordinating monitoring system parameters with control tasks, this invention achieves a deep integration of control and monitoring, an aspect often overlooked in traditional methods. This collaborative optimization significantly improves the monitoring system's perception and response capabilities, providing more reliable data support and more timely anomaly detection capabilities for the precise operation control of pumped storage power stations.

[0151] The monitoring system can dynamically adjust the monitoring parameters while issuing control commands, enabling it to control and monitor simultaneously in the way most suitable for the current task. This avoids the problem of "sometimes oversensitive and sometimes not sensitive enough" caused by static and fixed monitoring configurations in traditional monitoring systems, and improves the overall flexibility and adaptability of the monitoring system.

[0152] Example 10 like Figure 6 As shown, according to the tenth embodiment of the present invention, a control device for a pumped storage power station monitoring system is provided, characterized in that it includes: The data acquisition and fusion module 10 is used to collect and aggregate internal status data from the power grid dispatch center, local sensors of the power station and the pumped storage power station monitoring system in real time according to the monitoring system configuration information and the operating conditions of the monitoring power station, construct a unified formatted multi-source real-time data stream, and transmit the unified formatted multi-source real-time data stream to the state tree model construction module. The state tree model construction module 20 is used to receive the unified formatted multi-source real-time data stream, combine the power plant equipment model, operating constraints and state transition rules, construct a system state non-cyclic connection tree model representing all feasible paths from the current state to the target state, and transmit the system state non-cyclic connection tree model to the optimal path calculation module. The optimal path calculation module 30 is used to receive the system state acyclic connection tree model, apply the linear time complexity shortest path algorithm optimized for directed acyclic graphs, calculate the optimal path from the current state node to the target state node, output the optimal control sequence representing the optimal control strategy, and transmit the optimal control sequence to the control command generation and execution module. The control command generation and execution module 40 is used to receive the optimal control sequence, generate control commands arranged in time sequence, dynamically adjust the data acquisition frequency and alarm threshold of the monitoring system, and send the control commands to the equipment execution mechanism of the pumped storage power station monitoring system.

[0153] In this embodiment, the data acquisition and fusion module 10 is responsible for performing the functions of step S1 described in embodiments 1-3, including data demand identification and source location, starting real-time data acquisition, data preprocessing and cleaning, data fusion and formatting, and finally outputting a unified formatted multi-source real-time data stream.

[0154] The state tree model construction module 20 is responsible for completing the functions of step S2 described in embodiments 4-5, including determining the current state and the target state, initializing the state tree, exploring and expanding state transition paths, completing and pruning the state tree, and finally constructing a non-cyclic connection tree model of the system state.

[0155] The optimal path calculation module 30 is responsible for performing the functions of step S3 described in Examples 6-7, including defining the path evaluation cost function, executing the faster shortest path algorithm, backtracking to generate the optimal control sequence, and finally outputting a sequence of nodes and edges representing the optimal control strategy.

[0156] The control command generation and execution module 40 is responsible for performing the functions of step S4 described in embodiments 8-9, including parsing the optimal control sequence into an action list, generating standard control commands, dynamically adjusting monitoring system parameters, issuing commands and updating monitoring parameters, and ultimately achieving precise control of the pumped storage power station.

[0157] These four modules are interconnected via data flow to form a complete control process. The entire device can be deployed on the central processing unit of the pumped storage power station's monitoring system or on a dedicated control server.

[0158] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0159] It should be noted that those skilled in the art can make various modifications and variations to this invention without departing from the spirit and scope of this invention. If such modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include such modifications and variations.

[0160] This disclosure also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program performs the steps of a pumped-storage power station monitoring system control method described in the above-described method embodiments. The storage medium can be either volatile or non-volatile computer-readable storage.

[0161] In addition, this disclosure also provides a computer program product, which stores a computer program. When the computer program is run by a processor, it executes the steps of any of the pumped storage power station monitoring system control methods described above. For details, please refer to the above method embodiments, which will not be repeated here.

[0162] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium, which can be a volatile or non-volatile computer-readable storage medium. In another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0163] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices and apparatuses described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. In the several embodiments provided in this disclosure, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0164] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0165] In addition, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0166] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0167] Finally, it should be noted that the above-described embodiments are merely specific implementations of this disclosure, used to illustrate the technical solutions of this disclosure, and not to limit it. The protection scope of this disclosure is not limited thereto. Although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this disclosure; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this disclosure, and should all be covered within the protection scope of this disclosure. Therefore, the protection scope of this disclosure should be determined by the protection scope of the claims.

Claims

1. A control method of a pumped storage power station monitoring system, characterized by, Comprising: According to the monitoring system configuration information and the monitoring power station operating condition demand, real-time acquisition and convergence of internal state data from the power grid dispatching center, the power station local sensor and the pumped storage power station monitoring system form a unified formatted multi-source real-time data stream; According to the unified formatted multi-source real-time data stream, combined with the power station equipment model, the operation constraint condition and the state transition rule, a system state non-loop connection tree model representing all feasible paths from the current state to the target state is constructed; Using the system state non-loop connection tree model, a linear time complexity shortest path algorithm for directed acyclic graph optimization is applied to calculate the optimal path from the current state node to the target state node, and an optimal control sequence representing the optimal control strategy is obtained; According to the optimal control sequence, control instructions arranged in time sequence are generated, the data acquisition frequency and alarm threshold of the monitoring system are dynamically adjusted, and the control instructions are issued to the device execution mechanism of the pumped storage power station monitoring system.

2. The pumped storage power plant monitoring system control method according to claim 1, characterized by, According to the monitoring system configuration information and the monitoring power station operating condition demand, real-time acquisition and convergence of internal state data from the power grid dispatching center, the power station local sensor and the pumped storage power station monitoring system form a unified formatted multi-source real-time data stream, comprising: The configuration information of the monitoring system and the current monitoring power station operating condition demand are analyzed, the data items necessary for completing the control task are identified, the physical sensors, calculation nodes and external data interfaces corresponding to the data items are located, and the data item list and the corresponding specific data source list are obtained; According to the data item list and the corresponding specific data source list, data requests or subscription services are initiated to the specified data sources through the communication interface of the monitoring system, and original measurement values or state signals are obtained according to the preset or dynamically adjusted frequency, to produce original multi-source heterogeneous data streams containing time stamps; The original multi-source heterogeneous data streams are processed and unified time stamps are added to obtain the effective data set processed and marked with time stamps.

3. The pumped storage power plant monitoring system control method according to claim 2, characterized by, According to the monitoring system configuration information and the monitoring power station operating condition demand, real-time acquisition and convergence of internal state data from the power grid dispatching center, the power station local sensor and the pumped storage power station monitoring system form a unified formatted multi-source real-time data stream, comprising: The data fusion technology is applied to the effective data set cleaned and marked with time stamps, and multiple measurement values of the same physical quantity are fused to obtain the fused data; The fused data is format-converted and organized according to the unified data model defined in the pumped storage power station monitoring system to construct a unified formatted multi-source real-time data stream.

4. The pumped storage power plant monitoring system control method according to claim 3, characterized by, According to the unified formatted multi-source real-time data stream, combined with the power station equipment model, the operation constraint condition and the state transition rule, a system state non-loop connection tree model representing all feasible paths from the current state to the target state is constructed, comprising: The unified formatted multi-source real-time data stream, and a predefined power station device physical model, operation constraint condition, state transition rule and received target scheduling instruction are used to extract key state variables from the multi-source real-time data stream, determine the accurate operation state of the pumped storage power station monitoring system at present, and construct a state node for representing the current state of the pumped storage power station monitoring system; The target scheduling instruction in the unified formatted multi-source real-time data stream is parsed, and the target state to be reached by the system is determined in combination with the device model and constraint, and a definitely defined target state node is constructed.

5. The pumped storage power plant monitoring system control method according to claim 4, characterized by, According to the unified formatted multi-source real-time data stream, in combination with the power station device model, operation constraint condition and state transition rule, a system state acyclic connection tree model representing all feasible paths from the current state to the target state is constructed, including: According to the definitely defined current state node and target state node, and the device model, constraint condition and state transition rule, a data structure is created to represent the state tree, and the current state node is set as the root node of the tree, and an initial state tree structure containing the root node is constructed; According to the state transition rule, all possible one-step conversion operations are found by using the initial state tree structure containing the root node, the new state nodes caused by the conversion operations are calculated, the cost of the conversion operation is evaluated, and a partially expanded acyclic connection tree containing new nodes and edges is formed; According to the acyclic connection tree and state transition rule, state node expansion iteration is performed, paths reaching the target state node are explored, paths unable to reach the target state node or violating key constraints are pruned, and a system state acyclic connection tree model containing all feasible paths from the current state node to the target state node is generated.

6. The pumped storage power plant monitoring system control method according to claim 5, characterized by, According to the system state acyclic connection tree model, a linear time complexity shortest path algorithm optimized for directed acyclic graph is applied to calculate the optimal path from the current state node to the target state node, and an optimal control sequence representing the optimal control strategy is obtained, including: According to the system state acyclic connection tree model and preset global optimization target, a path cost function is defined according to the optimization target, the path cost function can accumulate the cost of all edges on the path, and the cost in different dimensions is weighted or combined according to the optimization target, and a path total cost calculation function is created; According to the system state acyclic connection tree model, current state node and target state node, and the path total cost calculation function, the directed acyclic graph shortest path algorithm is used to calculate the minimum cumulative cost from the current state node to each node, and the predecessor nodes and edges passed through to reach the minimum cumulative cost are recorded, and the calculation results containing the minimum cost of each node and the predecessor nodes and edges are output.

7. The pumped storage power plant monitoring system control method according to claim 6, characterized by, According to the system state acyclic connection tree model, a linear time complexity shortest path algorithm optimized for directed acyclic graph is applied to calculate the optimal path from the current state node to the target state node, and an optimal control sequence representing the optimal control strategy is obtained, including: According to the calculation results of the minimum cost of each node and the predecessor node and edge, and the target state node, starting from the target state node, the reverse optimal path data is obtained by performing reverse traversal according to the recorded predecessor information until the current state node is reached; The reverse optimal path data is sequentially rearranged to organize path information in the order from the current state node to the target state node, and an optimal control sequence representing an optimal control strategy is constructed.

8. The pumped storage power plant monitoring system control method according to claim 7, characterized by, According to the optimal control sequence, control instructions arranged in time sequence are generated, the data acquisition frequency and alarm threshold of the monitoring system are dynamically adjusted, and the control instructions are issued to the device actuator of the pumped storage power station monitoring system, including: Each edge in the optimal control sequence is traversed, each state transition is parsed into a basic control action with a time sequence relationship according to the properties of the edge, and a preliminary control action list containing specific operations, parameters and execution time sequences is generated; Each basic control action in the preliminary control action list is converted into a standard control instruction conforming to the communication protocol and instruction format of the underlying execution unit.

9. The pumped storage power plant monitoring system control method according to claim 8, characterized by, According to the optimal control sequence, control instructions arranged in time sequence are generated, the data acquisition frequency and alarm threshold of the monitoring system are dynamically adjusted, and the control instructions are issued to the device actuator of the pumped storage power station monitoring system, including: The optimal control sequence, the pumped storage power station monitoring system state and the control target are analyzed, the characteristics of the optimal control sequence and the demand of the pumped storage power station monitoring system for monitoring accuracy and response speed are evaluated, the parameters of the monitoring system itself are dynamically adjusted, and an updated monitoring system operating parameter set is obtained; The updated monitoring system operating parameters are applied to the monitoring system to match the monitoring behavior with the control task.

10. A pumped storage power station monitoring system control device, characterized in that, a data acquisition and fusion module for collecting and aggregating internal state data from the power grid dispatching center, the local sensors of the power station and the pumped storage power station monitoring system in real time according to the monitoring system configuration information and the monitoring power station operating condition demand, constructing a unified formatted multi-source real-time data stream, and transmitting the unified formatted multi-source real-time data stream to the state tree model construction module; a state tree model construction module for receiving the unified formatted multi-source real-time data stream, combining the power station equipment model, the operating constraint condition and the state transition rule, constructing a system state acyclic connection tree model representing all feasible paths from the current state to the target state, and transmitting the system state acyclic connection tree model to the optimal path calculation module; an optimal path calculation module for receiving the system state acyclic connection tree model, applying a linear time complexity shortest path algorithm optimized for directed acyclic graph, calculating the optimal path from the current state node to the target state node, outputting an optimal control sequence representing an optimal control strategy, and transmitting the optimal control sequence to the control instruction generation and execution module; The control instruction generation and execution module is configured to receive the optimal control sequence, generate control instructions arranged in time sequence, dynamically adjust the data acquisition frequency and alarm threshold of the monitoring system, and issue the control instructions to the device execution mechanism of the pumped storage power station monitoring system.

Images