A method and system for generating remote centralized control plans for cascade hydropower projects

Abstract

This invention relates to the field of hydropower system control technology, and particularly to a method and system for generating remote centralized control plans for cascade hydropower stations. The method includes: acquiring multi-source heterogeneous dynamic data of the target cascade hydropower station; statistically reconstructing the multi-source heterogeneous dynamic data to generate a dynamic feature tensor; identifying risks of the target cascade hydropower station based on the dynamic feature tensor to obtain a risk index sequence of the target cascade hydropower station; dynamically weighting and superimposing a preset set of plan templates based on the risk index sequence to generate an adaptive plan set; and dynamically matching the adaptive plan set based on the real-time operating status of the cascade hydropower remote centralized control system to determine the optimal plan, so as to control the cascade hydropower remote centralized control system based on the optimal plan. This invention achieves real-time adaptive matching of risk warning and emergency plans for cascade hydropower stations through dynamic data fusion and risk quantification analysis, significantly improving the accuracy and timeliness of emergency response.

Description

Technical Field

[0001] This invention relates to the field of hydropower system control technology, and in particular to a method and system for generating remote centralized control plans for cascade hydropower projects. Background Technology

[0002] The remote centralized control system for cascade hydropower stations is a crucial link in achieving the efficient and safe utilization of watershed hydropower resources. To address the potential chain reactions and safety risks caused by unforeseen events such as equipment failures, natural disasters, and cyberattacks, the industry has generally established emergency response plan matching mechanisms to ensure rapid response and mitigate losses in the event of an accident.

[0003] However, the relevant technical solutions have significant limitations in practical applications, hindering the timeliness and effectiveness of emergency response. These limitations are mainly reflected in the difficulty of effectively integrating multi-dimensional dynamic information reflecting the true risk situation of the system, resulting in insufficient and untimely risk perception. Furthermore, the core contingency plan generation and matching mechanism is relatively rigid, relying primarily on preset fixed rules or limited static historical data models, lacking the ability to dynamically adjust based on real-time risk status and system operating environment. This makes the recommended emergency plans often insufficiently adaptable and precise when facing complex, changing, or newly emerging risk scenarios, failing to meet the needs of rapid and accurate emergency response in cascade hydropower remote control systems.

[0004] Therefore, how to break through the bottlenecks of existing technologies and build a technical solution that can deeply integrate multi-source heterogeneous dynamic data, accurately depict dynamic risk situations, and intelligently generate and dynamically match the optimal contingency plan to significantly improve the emergency response speed and handling effect of the remote centralized control system of cascade hydropower stations in complex and ever-changing environments has become a core technical problem that urgently needs to be solved. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method and system for generating remote centralized control plans for cascade hydropower projects.

[0006] A first aspect of the present invention provides a method for generating a remote centralized control plan for cascade hydropower projects, comprising:

[0007] Acquire multi-source heterogeneous dynamic data of the target cascade hydropower station;

[0008] The multi-source heterogeneous dynamic data is statistically reconstructed to generate a dynamic feature tensor, which is used to characterize the temporal evolution and statistical distribution characteristics of the multi-source heterogeneous dynamic data in a structured multidimensional space.

[0009] Based on the dynamic feature tensor, risk identification is performed on the target cascade hydropower station to obtain the risk index sequence of the target cascade hydropower station;

[0010] Based on the risk index sequence, the preset set of contingency plan templates are dynamically weighted and superimposed to generate an adaptive set of contingency plans;

[0011] Based on the real-time operating status of the cascade hydropower remote control system, dynamic matching is performed in the adaptive contingency plan set to determine the optimal contingency plan, so as to control the cascade hydropower remote control system based on the optimal contingency plan.

[0012] A second aspect of the present invention provides a remote centralized control plan generation system for cascade hydropower projects, comprising:

[0013] The data acquisition module is used to acquire multi-source heterogeneous dynamic data of the target cascade hydropower station;

[0014] The statistical reconstruction module is used to perform statistical reconstruction on the multi-source heterogeneous dynamic data and generate a dynamic feature tensor. The dynamic feature tensor is used to characterize the temporal evolution and statistical distribution characteristics of the multi-source heterogeneous dynamic data in a structured multidimensional space.

[0015] The risk identification module is used to identify the risks of the target cascade hydropower station based on the dynamic feature tensor, and obtain the risk index sequence of the target cascade hydropower station.

[0016] The contingency plan generation module is used to dynamically weight and superimpose a preset set of contingency plan templates based on the risk index sequence to generate an adaptive set of contingency plans.

[0017] The contingency plan matching module is used to dynamically match the adaptive contingency plan set based on the real-time operating status of the cascade hydropower remote control system, and determine the optimal contingency plan so as to control the cascade hydropower remote control system based on the optimal contingency plan.

[0018] The beneficial effects of this invention are reflected in:

[0019] First, by integrating multi-source heterogeneous dynamic data and constructing a dynamic feature tensor, a unified and structured representation of complex operational states is achieved, significantly improving the comprehensiveness and timeliness of risk perception. Second, based on a dynamic risk index sequence, contingency plan templates are dynamically weighted and superimposed to generate a highly adaptive set of contingency plans, overcoming the shortcomings of traditional contingency plan libraries that are rigid and have low matching degrees. Finally, by dynamically matching and executing the optimal contingency plan in conjunction with real-time operational status, the accuracy, agility, and pertinence of emergency response are ensured, greatly improving the safety assurance capability and emergency control efficiency of the cascade hydropower remote centralized control system in complex and ever-changing environments. Attached Figure Description

[0020] Figure 1 A flowchart illustrating a method for generating remote centralized control plans for cascade hydropower provided by the present invention;

[0021] Figure 2 This is a flowchart illustrating an adaptive contingency plan generation method provided by the present invention.

[0022] Figure 3 This is a schematic diagram of the structure of a cascade hydropower remote centralized control plan generation system provided by the present invention. Detailed Implementation

[0023] 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 embodiments of the present invention, and not all embodiments. 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.

[0024] The existing technologies and their main problems can be summarized in the following aspects:

[0025] 1. Static contingency plan invocation mechanism based on fixed rules:

[0026] Existing systems generally use preset fixed rules (such as specific fault types or threshold exceeding limits) to trigger corresponding static emergency plans. While this mechanism can achieve automated invocation, it lacks dynamic adjustment capabilities. It cannot adjust the specific content or execution intensity of the plan based on the real-time changing comprehensive risk level (e.g., the combined effects of multiple faults occurring simultaneously or environmental factors) and the current overall operating status of the system, resulting in insufficient adaptability and flexibility of the plan.

[0027] 2. Risk perception based on a single data source:

[0028] The existing contingency plan matching system suffers from significant limitations in its data foundation. Its automated data collection primarily focuses on electrical equipment operating parameters (such as voltage, current, and frequency), failing to effectively integrate other key dimensions closely related to power plant safety, such as environmental parameters (hydrology, meteorology), human operational factors, and heterogeneous data on cybersecurity. This singular data source and fragmented information leads to an incomplete and inaccurate assessment of the overall system risk, directly impacting the effectiveness of subsequent contingency plan selection.

[0029] 3. Static classification model trained based on historical data:

[0030] Some improvement schemes attempt to apply machine learning algorithms (such as decision trees and support vector machines) to recommend solutions by training on historical data patterns. However, these methods have inherent drawbacks:

[0031] (1) The training samples are highly dependent and the updates are lagging: the model performance is highly dependent on limited, potentially outdated historical sample data.

[0032] (2) Lack of real-time dynamic response: It is difficult to effectively integrate and respond to the dynamic data stream that is generated in real time and reflects the current risk evolution.

[0033] (3) Weak generalization ability: When faced with new risk types not included in the training data or sudden scenarios with complex coupling of multiple factors, the model matching accuracy drops significantly.

[0034] (4) Poor interpretability: The model decision-making process is usually opaque, which is not conducive to operators' understanding and trust in the recommended results of the plan.

[0035] 4. Static and rigid contingency plan database management:

[0036] Existing systems' contingency plan databases are generally simple to build but difficult to maintain.

[0037] (1) Fixed content: The items in the contingency plan are usually static, predefined text or process, lacking the ability to dynamically generate or adjust the content according to real-time risks.

[0038] (2) Inefficient management: The contingency plan database may have problems such as redundancy, incomplete coverage, and unclear classification, which may lead to difficulties in finding contingency plans or content conflicts.

[0039] (3) Lagging updates: The updates and optimizations of the contingency plan database mainly rely on manual intervention and lack automated and intelligent assessment, optimization and synchronization mechanisms, making it difficult to keep up with the rapid changes in the risk situation and the accumulation of system operation experience.

[0040] In view of this, the present invention proposes a method and system for generating remote centralized control plans for cascade hydropower projects to solve the above-mentioned technical problems.

[0041] Example 1:

[0042] Reference manual attached Figure 1 The diagram shows a flowchart of a method for generating remote centralized control plans for cascade hydropower provided by the present invention.

[0043] This invention provides a method for generating remote centralized control plans for cascade hydropower projects, comprising:

[0044] S1: Acquire multi-source heterogeneous dynamic data of the target cascade hydropower station.

[0045] This step aims to build a comprehensive, real-time, and reliable data foundation for subsequent risk identification and contingency plan generation.

[0046] In practice, dynamic data is collected through a distributed sensing network deployed throughout the cascade hydropower stations. This network covers:

[0047] Equipment operating parameter acquisition unit: including generator winding temperature sensor, bearing vibration accelerometer, turbine flow meter, gate opening sensor, bus voltage / current transformer, etc.;

[0048] Environmental status monitoring unit: including reservoir water level radar, dam piezometer, watershed rain gauge, meteorological temperature and humidity sensor, geological disaster monitoring instrument, etc.;

[0049] Human operation and network security unit: integrates the operation command logs of the scheduling master control system, access control and security records, network traffic probes and intrusion detection system alarm information.

[0050] All data is aggregated in real time to the remote control center through a multi-channel concurrent transmission mechanism, forming a four-dimensional heterogeneous data stream covering equipment operation, environmental status, human operation, and network security, namely the aforementioned multi-source heterogeneous dynamic data.

[0051] S2: Perform statistical reconstruction on multi-source heterogeneous dynamic data to generate dynamic feature tensors. The dynamic feature tensors are used to characterize the temporal evolution and statistical distribution characteristics of multi-source heterogeneous dynamic data in a structured multidimensional space.

[0052] In one possible implementation, S2 involves statistically reconstructing the multi-source heterogeneous dynamic data to generate a dynamic feature tensor, specifically including:

[0053] Data preprocessing is performed on multi-source heterogeneous dynamic data to eliminate dynamic anomalies in the multi-source heterogeneous dynamic data, resulting in preprocessed multi-source heterogeneous dynamic data.

[0054] A sliding feature reconstruction mechanism is introduced to reconstruct the sliding features of each type of variable in the preprocessed multi-source heterogeneous dynamic data, generating a time-aligned structured feature set.

[0055] Based on the time axis, each type of variable in the structured feature set is encapsulated as a structured feature tensor to obtain a dynamic feature tensor.

[0056] This step transforms the original dynamic dataset into a structured data representation with temporal consistency and statistical expressive power through temporal feature reconstruction technology. Specifically, it includes the following operations:

[0057] After receiving multi-source heterogeneous dynamic data from the target cascade hydropower station, a state-aware data cleaning algorithm can be used to address typical problems in the sensing link, such as packet loss, duplication, and delay. During algorithm execution, the data acquisition time window is set as follows: For each sensor variable sequence of multi-source heterogeneous dynamic data x ( t First, continuous detection and anomaly removal are performed. If the conditions are met... If the probability of its consecutive occurrence is less than 1, then it is considered that there is a mutation anomaly at that point. The value is marked as an isolated outlier and removed, and missing values ​​are filled in by linear interpolation to generate preprocessed multi-source heterogeneous dynamic data.

[0058] Among them, parameters The threshold for the transition criterion, Both the continuous anomaly tolerance threshold and the threshold value can be calculated based on the equipment's historical operating statistics.

[0059] In one possible implementation, a sliding feature reconstruction mechanism is introduced to reconstruct the features of each type of variable in the preprocessed multi-source heterogeneous dynamic data, generating a time-aligned structured feature set, specifically including:

[0060] Within a preset sliding window, the set of statistics corresponding to each type of variable in the preprocessed multi-source heterogeneous dynamic data is calculated. The set of statistics includes the mean and weighted variance.

[0061] Based on the set of statistics, the statistics of each type of variable in multi-source heterogeneous dynamic data at each time step are encapsulated as binary structure features, resulting in a set of binary structure features.

[0062] Temporally assemble the binary structural features in the binary structural feature set to obtain a temporally aligned structured feature set.

[0063] To maintain the continuity of data features and the sensitivity to changes in system state, this invention introduces a sliding feature reconstruction mechanism based on local structural statistics to avoid the loss of trend information caused by traditional extreme value normalization methods.

[0064] Specifically, let the length of the sliding window be... The current time is In the interval Internally defined in t Exponential decay plus or minus the mean at time 1 With weighted variance They are respectively:

[0065]

[0066]

[0067] in, It is an exponential decay factor used to control the influence of historical data from a distant time period on the current statistic, reflecting the characteristic of "high weight for recent data", that is, giving higher weight to recent data. As a measure of the degree of fluctuation of variables, it can capture the difference between short-term drastic changes or steady states in time-varying systems, and is often used in anomaly detection and fluctuation modeling.

[0068] Calculate μ(t) and for each variable separately. This process forms a time-aligned set of structured features, which are then stacked along the time axis to generate a three-dimensional structure tensor. ,in This represents the total number of sampling times. This represents the number of sensed variables.

[0069] This step significantly improves the accuracy of risk identification and the adaptability of contingency plan matching for cascade hydropower stations through statistical reconstruction and dynamic feature tensor generation of multi-source heterogeneous dynamic data. Specifically, a dynamic anomaly elimination mechanism effectively removes sensor noise and isolated outliers, ensuring data quality; the sliding feature reconstruction uses exponentially decaying weighted mean and variance to quantify short-term fluctuations while preserving temporal trends, avoiding information loss caused by traditional normalization; and the three-dimensional dynamic feature tensor uniformly encapsulates multivariate statistical features, forming a spatiotemporally coupled structured expression, providing a high-dimensional data foundation for risk factor quantification. Ultimately, this addresses the root causes of data noise interference and insufficient feature representation, making subsequent risk identification results more reliable and contingency plan generation more accurately adapted to dynamic risk situations.

[0070] S3: Based on the dynamic feature tensor, risk identification is performed on the target cascade hydropower station to obtain the risk index sequence of the target cascade hydropower station.

[0071] In one possible implementation, step S3 involves risk identification of the target cascade hydropower stations based on the dynamic feature tensor, resulting in a risk index sequence for the target cascade hydropower stations. Specifically, this includes:

[0072] Based on a preset set of risk factors, risk factors are extracted from the dynamic feature tensor to obtain the target set of risk factors.

[0073] Extract the subset of sensor variable features associated with each type of risk factor in the target risk factor set to obtain a set of risk feature vectors;

[0074] Based on the risk feature vector set, a weighted nonlinear transformation is performed on each type of risk factor in the target risk factor set to obtain a set of risk activity scalar values.

[0075] Based on the set of risk activity scalar values ​​and the preset core indicator data of each type of factor in the target risk factor set, a standardized calculation is performed to obtain a set of comprehensive risk indices; the preset core indicator data includes historical frequency, current activity and potential impact intensity.

[0076] The risk index sequence is obtained by performing time-series assembly processing on the comprehensive risk index vector in the comprehensive risk index set.

[0077] This step uses dynamic feature tensors as input to identify and quantify four pre-defined risk factors for cascade hydropower stations through a unified feature extraction framework:

[0078] Environmental risk factors: Based on the statistical characteristics of hydrology (water level, flow rate) and meteorology (rainfall, wind speed), a flood probability model and an extreme weather impact index are constructed;

[0079] Equipment-related risk factors: Analyze the time-varying patterns of unit vibration, temperature, and electrical parameters to quantify the risks of mechanical fatigue and electrical failure;

[0080] Operational risk factors: analyze the timing anomalies in the dispatch instruction logs and combine them with access control system data to pinpoint the probability of erroneous operations;

[0081] Information security risk factors: Corresponding to the variance of network traffic probes and the frequency of intrusion alarms, an attack threat level model is constructed.

[0082] Of course, the set of risk factors preset in this invention may also include other categories of risk factors, and this invention does not specifically limit them.

[0083] Specifically, let there be a total Potential risk factors are denoted as follows: Each type of risk factor is associated with a set of characteristic variables. Let the first type be... The original variable vector associated with risk factors is:

[0084] ,

[0085] in, Indicates the first i The number of variable dimensions corresponding to risk factors.

[0086] To accurately identify the activity level of risk factors in the current state, a risk activity function is constructed. The definition is as follows:

[0087]

[0088] in, Indicates the first i Class of risks j The weighting coefficients of each variable are used to represent the degree of contribution of that variable to the overall risk. Indicates the first Class of risks For each variable, a risk sensitivity function is selected based on the physical meaning and volatility characteristics of the variable, using a linear, hyperbolic tangent, or ReLU function, and nonlinear enhancement or threshold activation processing is applied. Indicates time Time of the first Class of Riskj The collected values ​​of each variable. This function constitutes a functional expression of the activity level of the risk factor at the current moment, and forms the basis for subsequent quantification.

[0089] To achieve a unified measurement of risk across categories, a risk index is introduced. S i As a quantification result, it represents the first The risk index for risk categories is calculated using a model that comprehensively considers three core indicators: the historical frequency, current activity, and potential impact strength of risk factors. Specifically, the aforementioned preset core indicator data includes the historical frequency, current activity, and potential impact strength of each risk category. The definitions are as follows:

[0090]

[0091] in, Indicates the first i The normalization coefficients of risk factors ensure that all risk indices are on the same scale. This indicates the current activity level of a risk factor, specifically calculated using the risk activity function. f i ( t ) within the time window [ t - w +1, t The average value within ] reflects the first i The active trend of risk factors over a period of time (rather than the instantaneous state at a single moment). F i Indicates the first i The frequency of a risk factor in historical records, i.e., the historical frequency mentioned above, can be determined by the frequency of the first occurrence of a risk factor within a statistical observation period. i The proportion of the number of times a risk factor is triggered relative to the total number of observations is calculated, reflecting the frequency of this type of risk in history; L i To indicate the first i The potential loss value of risk factors, i.e., the potential impact strength mentioned above. L i Indicates the first i The expected impact of such risks on system operation is determined based on historical incident databases, expert experience, or operational simulation assessments. , , The weighting coefficients for the three indicators mentioned above are respectively, satisfying... This is used to express the relative importance of each risk quantification element in the comprehensive assessment.

[0092] Specifically, regarding the determination and acquisition method of the potential impact intensity, the actual loss data (such as economic losses, downtime, etc.) of similar risks in historical records can be statistically analyzed, and the average impact level can be obtained through fitting or weighted calculation as a basic reference; or experts in the hydropower field can classify or score the risk consequences according to the characteristics of the power station (such as scale and importance), and incorporate them into the calculation after quantification; or the system response after the risk occurs can be simulated through a system simulation platform (such as a digital twin model) to assess the impact on core functions (power generation, flood control, etc.) and output simulation calculation values. This invention does not make specific limitations on this.

[0093] Furthermore, in order to integrate the originally scattered single-moment risk indices into time-correlated sequence data to intuitively present the changes in the activity level of various risk factors over time (such as the continuous rise of equipment-related risk indices and sudden fluctuations in environmental risk indices), and thus provide a structured time-series basis for identifying risk evolution patterns and dynamically matching emergency plans, this invention performs time-series assembly processing on the comprehensive risk index vectors in the comprehensive risk index set. The comprehensive risk index vectors at different time points are systematically integrated along the time dimension to form a continuous and traceable risk evolution sequence, i.e., the risk index sequence. The specific processing method is as follows:

[0094] 1. Time Dimension Alignment: Based on a preset unified timestamp (e.g., milliseconds, seconds), the comprehensive risk index vector calculated at each time point (e.g., ...) is aligned. t At time 1 [ S 1( t 1), S 2( t 1),..., S m ( t m )], t At time 2 [ S 1( t 2), S 2( t 2),..., S m ( t m The vectors are sorted in chronological order to ensure that each vector corresponds to a unique time point and to avoid time discrepancies.

[0095] 2. Sequence Structure Construction: Arrange the sorted comprehensive risk index vectors vertically along the time axis to form a two-dimensional time series matrix. For example, if within the time window [ t 0, t N [Collected internally] N Given a vector at time +1, the risk index sequence can be represented as:

[0096] [ S 1( t 0), S 2( t 0),..., S m ( t 0)], / / t Risk index vector at time 0

[0097] [ S 1( t 1), S 2( t 1),...,S m ( t 1)], / / t Risk index vector at time 1 ...

[0098] [ S 1( t N ), S 2( t N ),..., S m ( t N )] / / t N Risk index vector at any moment

[0099] Each row corresponds to the composite index of all risk factors at a given point in time, and each column corresponds to the index change trend of a certain type of risk factor throughout the entire time window.

[0100] 3. Handling of temporal continuity: If there are missing time points (such as data acquisition failure at a certain moment), the missing vector is supplemented by linear interpolation or the mean of nearby time points to ensure the continuity of the sequence in the time dimension and avoid the impact of data gaps on subsequent analysis of risk evolution trends.

[0101] 4. Dynamic window association: The assembled time series is bound to a preset time sliding window (such as a 5-minute or 10-minute window) so that the series can reflect both the real-time risk status (the vector at the current moment) and the short-term evolution trend of the risk (the vector change within the window), providing dynamic input for subsequent contingency plan matching based on time series features.

[0102] Thus, through the above methods, m The comprehensive risk index set of each risk factor is output in time dimension to form a risk index sequence. S ( t )=[ S 1( t ),S 2( t ) , ... ,S m ( t This serves as the triggering basis for subsequent intelligent contingency plan generation and the calculation basis for contingency plan matching. Among them, S ( t )express t A collection of comprehensive risk indices at any given moment. S 1( t This represents the composite risk index of the first type of risk factor. S 2( t This represents the composite risk index of the second type of risk factor. S 3( t ) indicates the first m The comprehensive risk index for similar risk factors, and so on. This sequence has time dependence and variable explanatory power, and can dynamically reflect the main risk types and weight structure currently faced by the system.

[0103] In summary, this module, by constructing a risk-driven structural activation model and combining historical contingency plan templates with real-time risk characteristics, achieves dynamic assembly, parameter reconstruction, and execution evolution of response strategies. Its key feature is the use of a risk index to construct a contingency plan structure index mapping and the introduction of a feedback-driven structural weight fine-tuning mechanism. This enables the contingency plan generation to possess continuous self-adaptation and online optimization capabilities, demonstrating significant adaptability and improved response efficiency in complex and ever-changing cascade hydropower remote centralized control systems.

[0104] S4: Based on the risk index sequence, the preset set of contingency plan templates are dynamically weighted and superimposed to generate an adaptive set of contingency plans.

[0105] Intelligent contingency plan generation is the core link in realizing the transformation of "risk index-driven - response strategy construction" in this invention. Its role is to dynamically construct a set of remote centralized control contingency plans for cascade hydropower that accurately corresponds to the current system risk situation based on the risk factor identification results and quantitative indicators output by the preceding modules.

[0106] Reference manual attached Figure 2 The diagram shows a flowchart of an adaptive contingency plan generation method provided by the present invention.

[0107] In one possible implementation, step S4 involves dynamically weighting and superimposing a preset set of contingency plan templates based on a risk index sequence to generate an adaptive contingency plan set. This specifically includes sub-steps S401 to S403:

[0108] S401: Generate a risk response impact vector based on a preset set of response weights and a risk index sequence;

[0109] S402: The risk response impact vector is mapped to the contingency plan structure factor vector through the structure activation mapping function;

[0110] S403: Based on the contingency plan structure factor vector, the preset contingency plan template set is dynamically weighted and superimposed to generate an adaptive contingency plan set.

[0111] After obtaining the risk factor set R 1 ,R 2 , ... ,R m and its corresponding risk index sequence S 1( t ) ,S 2( t ) , ... ,S m ( t By further modeling and calculation, static contingency plan templates can be transformed into multi-response path schemes with variable structures and adaptive parameters, thus realizing the transition from data-driven to strategy generation.

[0112] Specifically, for any given time t The total number of identified risk factors is m Risk index series ,in S i ( t ) indicates the first i Class of risks t The quantitative activity intensity at any given time. To capture the nonlinear weighting relationship of the contribution of each risk to the contingency plan structure, a risk response influence vector is first constructed. Its definition is:

[0113]

[0114] in For the first i The response weights corresponding to different risk types satisfy the normalization conditions. This weight represents the system's level of attention or control preference for various risk factors, and can be obtained through training based on historical execution priorities. This vector is input into the pre-plan structure activation mapping function. Generate a contingency plan structure index vector z ( t ):

[0115]

[0116] in The strategy structure weight matrix, For bias terms, It is a non-linear activation function (either Sigmoid or Softmax can be used). This is a contingency plan structure factor vector, representing the configuration weights of the stress response under the current system state. Each dimension of this vector represents the activation intensity of a contingency plan component in the overall response scheme, used to guide subsequent strategy assembly and parameter customization.

[0117] exist z ( t Once determined, the system enters the contingency plan generation stage, and further dynamically weights and superimposes the preset contingency plan template set based on the contingency plan structure factor vector to generate an adaptive contingency plan set.

[0118] In one possible implementation, step S403 involves dynamically weighting and superimposing a preset set of contingency plan templates based on the contingency plan structure factor vector to generate an adaptive contingency plan set, specifically including:

[0119] The structural factor vector of the plan is normalized to obtain the structural factor vector of the plan after normalization.

[0120] Obtain the structure activation weight matrix corresponding to the structural factor vector of the pre-planned structure after structural normalization;

[0121] Based on the structural activation weight matrix, the combination of preset plan templates is dynamically scheduled to obtain the dynamic combination of the preset plan template set;

[0122] The preset contingency plan template set is dynamically weighted and superimposed using a dynamic combination method to obtain an adaptive contingency plan set.

[0123] Let the set of contingency plan templates be ,in T j Indicates the first j Each strategy template includes information such as response logic, execution modules, control parameter interfaces, and scheduling control processes. The generated plan structure is composed of multiple templates, and their combination is determined by structure activation values. z ( t Scheduling is performed. To ensure the rationality and stability of structure generation, the system introduces structure normalization processing to ensure that the template call strength meets the following requirements:

[0124]

[0125] Based on this, each of the final generated plans P ( t This can be represented as a weighted sum of template sets:

[0126]

[0127] in P (t This is the final centralized control plan. For the first A basic template, the contents of which may include voltage regulation schemes, load switching strategies, node control logic, personnel response paths, etc. Although the representation is a linear combination in symbolism, in actual execution it is a module-level assembly, that is, selecting a template according to the activation intensity and loading its parameter configuration to construct the control process.

[0128] In one possible implementation, after dynamically weighting and superimposing the preset plan template set based on the plan structure factor vector to generate an adaptive plan set, the method further includes:

[0129] Get the function to evaluate the execution effect;

[0130] The effectiveness of the adaptive contingency plan set is evaluated based on the performance evaluation function to obtain the performance evaluation results.

[0131] The effectiveness evaluation results include the execution performance indicators of each plan in the adaptive contingency plan set during the execution process. The execution performance indicators include response time, control stability, and degree of goal achievement.

[0132] Obtain historical execution performance data;

[0133] Based on historical execution performance data and performance evaluation results, the structure activation weight matrix corresponding to the structural factor vector of the contingency plan is updated by gradient so as to adaptively adjust the set of adaptive contingency plans.

[0134] To improve the adaptability and execution quality of the generated contingency plan, the system introduces an evolutionary mechanism based on historical feedback after the structure is generated. A historical execution performance evaluation function is defined. This measures indicators such as response time, control stability, and goal achievement in the actual implementation of the current contingency plan. The system records past... k Effectiveness of contingency plan implementation within the wheel E ( P ( t - k +1)) , ... ,E ( P ( t )), and on the activation matrix W Perform gradient-based fine-tuning:

[0135]

[0136] in For learning rate, This represents the gradient information of the evaluation function with respect to the structure matrix. This mechanism enables continuous adaptive adjustment of the plan structure weights, allowing the system to continuously enhance effective strategies and automatically weaken ineffective structures during long-term operation, thereby achieving self-evolution and optimization of the adaptive plan set.

[0137] S5: Based on the real-time operating status of the cascade hydropower remote control system, dynamic matching is performed in the adaptive contingency plan set to determine the optimal contingency plan, so as to control the cascade hydropower remote control system based on the optimal contingency plan.

[0138] As the terminal response scheduling component of the system of this invention, the automatic matching and optimization of contingency plans is responsible for accurately matching the constructed contingency plans with the current operating status of the cascade hydropower remote centralized control system and dynamically executing them. Its core value lies in achieving a high degree of adaptability and effect optimization between the contingency plans and the real-time control environment.

[0139] This step is based on the set of contingency plans. P ( t (and combined with the state vector obtained by the system operation monitoring module) By constructing a dynamic matching scoring function and a response path selection mechanism, the optimal strategy path is selected and its execution parameters are refined to improve the immediacy of the regulation effect and the control precision.

[0140] In one possible implementation, the optimal contingency plan is determined by dynamically matching the real-time operating status of the cascade hydropower remote centralized control system within an adaptive contingency plan set, specifically including:

[0141] Construct a dynamic matching scoring function;

[0142] By using a dynamic matching scoring function, the matching degree between each plan in the adaptive plan set and the real-time operating status is calculated to obtain the matching degree set.

[0143] Based on preset matching rules and matching degree sets, candidate emergency response plans are determined;

[0144] Gradient fine-tuning is performed on the key control parameter set of candidate emergency plans to optimize the preset objective function value and obtain the optimal plan.

[0145] In terms of specific design, The set of state vectors corresponding to the system's operating state at any given time is This set of state vectors contains the target state vectors corresponding to system operating parameters such as current reservoir water levels, unit load status, communication delays, and main control strategy indicators. The adaptive contingency plan set is... Each contingency plan All include a structurally adjustable, parameter-weighted response strategy, which can specifically include execution structure (such as control logic topology, scheduling steps), parameter configuration (such as adjusting thresholds, execution timing), and historical execution records. This is to achieve... P ( t The optimal plan is selected from the options and further refined through parameter tuning, introducing a dynamic matching scoring function. The calculation of the dynamic matching scoring function needs to be based on and The correlation characteristics are quantified item by item using three indicators:

[0146] Dynamic matching scoring function Defined as:

[0147]

[0148] in, Contingency Plan The topological deviation index between the execution structure and the current system control logic. This represents the error prediction of the contingency plan strategy for key system indicators (such as water level and load) under the current state. This represents the average response effectiveness loss of the contingency plan under similar historical conditions, i.e., the historical response loss. Assign importance weights to each indicator to satisfy... This function evaluates the fit between each candidate solution and the current system state; a lower matching score indicates a better solution.

[0149] Specifically, the calculation methods for the above indicators include:

[0150] 1. Topological Deviation Index :

[0151] Comparison Plan The execution structure must be consistent with the topology of the current system control logic. For example, if the current system control logic is "tiered joint scheduling priority", while the contingency plan... If the execution structure is "independent adjustment of a single plant", then the deviation is quantified by a topology difference algorithm (such as node matching degree calculation based on graph theory). The topology deviation index... The larger the value, the worse the structural adaptability.

[0152] 2. Error prediction amount :

[0153] Based on real-time running status Prediction and contingency plan The degree of deviation from key system indicators (water level, load, etc.) after execution. For example, based on the current reservoir water level. With contingency plan The adjustment target value is calculated using an error prediction model (such as linear regression or LSTM) to determine the deviation. The smaller the value, the higher the precision of the plan's control over the target.

[0154] 3. Historical response losses :

[0155] Retrieve historical data and current status Extracting solutions for similar scenarios (through cosine similarity or Euclidean distance matching). The average loss is calculated by weighted averaging of performance data in similar scenarios (such as response latency and adjustment overshoot). It reflects the historical reliability of the contingency plan.

[0156] Further employing the weighted summation formula (in =1, weight is dynamically adjusted according to the current control priority of the system), for each plan Calculate the score.

[0157] Then iterate through the adaptive contingency plan set. All scores are aggregated to form a matching score set. , ,..., Each element corresponds to a matching degree between a contingency plan and the real-time status, providing a quantitative basis for subsequent screening of candidate contingency plans.

[0158] After obtaining the matching score set (i.e., completing the matching score assessment), candidate emergency response plans are determined based on the preset matching rules and the matching score set. Specifically, the preset matching rules can be to select the plan with the lowest matching score. As the optimal execution strategy at the current moment, the optimization objective can be defined as:

[0159]

[0160] That is, select the solution with the lowest matching score from all candidate solutions. As a candidate emergency response plan for the current situation.

[0161] It is worth noting that the gradient fine-tuning of the key control parameter set of the candidate emergency plan in the above implementation method is consistent with the gradient fine-tuning of the key control parameter set of the optimal plan after controlling the cascade hydropower remote centralized control system based on the optimal plan. The specific method will be elaborated in the following text and will not be repeated here.

[0162] In one possible implementation, after controlling the cascade hydropower remote centralized control system based on the optimal plan, the system further includes:

[0163] Obtain the actual control indicators corresponding to the optimal contingency plan;

[0164] Based on expected and actual control indicators, calculate the performance loss corresponding to the optimal plan.

[0165] Calculate the gradient of the performance loss with respect to the weight vector in the dynamic matching scoring function;

[0166] The weight vector is dynamically updated based on the gradient.

[0167] To further improve the quality of plan implementation, this module also introduces a local parameter optimization process to optimize the plan. control parameter set Gradient fine-tuning is performed to optimize the objective function value as much as possible while keeping the original response logic unchanged. The optimization iterative update form is as follows:

[0168]

[0169] in To optimize step size, This indicates that the matching function matches the parameters. The partial derivatives. This process is performed in a limited number of iterations while meeting the time limits for real-time message processing and execution control, to ensure that parameter adjustment is completed within an acceptable time window, thereby improving execution efficiency and accuracy.

[0170] In addition, to improve the long-term stability of the system, this module also introduces a dynamic weight update mechanism, that is, based on the pre-planned... In actual implementation, the weights are dynamically adjusted. The value is adjusted to better align with the control objectives of the current operating phase of the system.

[0171] Suppose that the actual control indicators perform as follows within a certain operating cycle: y ( t If this is the case, then a performance loss function is introduced. ,in, yref This represents the desired control indicator. Weight updates are based on the following rules:

[0172]

[0173] in Indicates the update step size. This represents the gradient of the loss function with respect to the weights, used to guide the matching mechanism to adjust towards minimizing performance loss, in order to adapt to the long-term evolution of the system.

[0174] In summary, this implementation method, by constructing a risk-driven response matching evaluation mechanism and parameter fine-tuning process, not only achieves precise adaptation between the contingency plan and the system state, but also endows the strategy execution with dynamic self-optimization capabilities. It breaks through the bottleneck of traditional static selection of response plans and parameter solidification, and has significant value in improving dynamic adaptability and execution efficiency in cascade hydropower remote centralized control systems.

[0175] Example 2

[0176] Reference manual attached Figure 3 The diagram shows a schematic of the structure of a cascade hydropower remote centralized control plan generation system provided by the present invention.

[0177] An embodiment of the present invention provides a remote centralized control plan generation system 20 for cascade hydropower projects, comprising:

[0178] Data acquisition module 201 is used to acquire multi-source heterogeneous dynamic data of the target cascade hydropower station;

[0179] The statistical reconstruction module 202 is used to perform statistical reconstruction on the multi-source heterogeneous dynamic data and generate a dynamic feature tensor. The dynamic feature tensor is used to characterize the temporal evolution and statistical distribution characteristics of the multi-source heterogeneous dynamic data in a structured multidimensional space.

[0180] The risk identification module 203 is used to identify the risks of the target cascade hydropower station based on the dynamic feature tensor, and obtain the risk index sequence of the target cascade hydropower station.

[0181] The contingency plan generation module 204 is used to dynamically weight and superimpose a preset set of contingency plan templates based on the risk index sequence to generate an adaptive contingency plan set.

[0182] The contingency plan matching module 205 is used to dynamically match the adaptive contingency plan set based on the real-time operating status of the cascade hydropower remote centralized control system, and determine the optimal contingency plan so as to control the cascade hydropower remote centralized control system based on the optimal contingency plan.

[0183] In one possible implementation, the plan generation module 204 is specifically used for:

[0184] Based on the preset set of response weights and the risk index sequence, a risk response impact vector is generated;

[0185] The risk response impact vector is mapped to a contingency plan structure factor vector using a structure activation mapping function.

[0186] Based on the aforementioned contingency plan structure factor vector, the preset contingency plan template set is dynamically weighted and superimposed to generate an adaptive contingency plan set.

[0187] In one possible implementation, the plan generation module 204 is specifically used for:

[0188] The proposed structural factor vector is subjected to structural normalization to obtain the structural normalized proposed structural factor vector.

[0189] Obtain the structure activation weight matrix corresponding to the structural factor vector of the pre-planned structure after the structure normalization process;

[0190] Based on the structure activation weight matrix, the combination of preset plan templates is dynamically scheduled to obtain the dynamic combination of the preset plan template set;

[0191] Based on the aforementioned dynamic combination method, the preset plan template set is dynamically weighted and superimposed to obtain an adaptive plan set.

[0192] In one possible implementation, the cascade hydropower remote centralized control plan generation system 20 further includes a first weight update module 206, which is specifically used for:

[0193] After dynamically weighting and superimposing the preset plan template set based on the plan structure factor vector to generate an adaptive plan set, the execution effect evaluation function is obtained.

[0194] The effectiveness of the adaptive contingency plan set is evaluated according to the execution effectiveness evaluation function to obtain the effectiveness evaluation result;

[0195] The effect evaluation results include the execution performance indicators of each plan in the adaptive plan set during the execution process, and the execution performance indicators include response time, control stability, and goal achievement degree;

[0196] Obtain historical execution performance data;

[0197] Based on the historical execution effect data and the effect evaluation results, the structure activation weight matrix corresponding to the plan structure factor vector is updated by gradient so as to adaptively adjust the adaptive plan set.

[0198] In one possible implementation, the pre-plan matching module 205 is specifically used for:

[0199] Construct a dynamic matching scoring function;

[0200] The matching degree between each plan in the adaptive plan set and the real-time operating status is calculated using the dynamic matching scoring function to obtain a matching degree set.

[0201] Based on the preset matching rules and the matching degree set, candidate emergency plans are determined;

[0202] The key control parameter set of the candidate emergency response plan is fine-tuned in a gradient to optimize the preset objective function value and obtain the optimal plan.

[0203] In one possible implementation, the cascade hydropower remote centralized control plan generation system 20 further includes a second weight update module 207, which is specifically used for:

[0204] After controlling the cascade hydropower remote centralized control system based on the optimal plan, the actual control indicators corresponding to the optimal plan are obtained;

[0205] Based on the expected indicators and the actual control indicators, calculate the performance loss corresponding to the optimal plan;

[0206] Calculate the gradient of the performance loss with respect to the weight vector in the dynamic matching scoring function;

[0207] The weight vector is dynamically updated based on the gradient.

[0208] In one possible implementation, the risk identification module 203 is specifically used for:

[0209] Based on a preset set of risk factors, risk factors are extracted from the dynamic feature tensor to obtain a target set of risk factors;

[0210] Extract the subset of sensor variable features associated with each type of risk factor in the target risk factor set to obtain a risk feature vector set;

[0211] Based on the set of risk feature vectors, a weighted nonlinear transformation is performed on each type of risk factor in the set of target risk factors to obtain a set of risk activity scalar values.

[0212] Based on the set of risk activity scalar values ​​and the preset core indicator data of each type of factor in the target risk factor set, a standardized calculation is performed to obtain a set of comprehensive risk indices; the preset core indicator data includes historical frequency, current activity, and potential impact intensity.

[0213] The risk index vectors in the comprehensive risk index set are subjected to time-series assembly processing to obtain the risk index sequence.

[0214] In one possible implementation, the statistical reconstruction module 202 is specifically used for:

[0215] The multi-source heterogeneous dynamic data is preprocessed to eliminate dynamic anomalies in the multi-source heterogeneous dynamic data, resulting in preprocessed multi-source heterogeneous dynamic data.

[0216] A sliding feature reconstruction mechanism is introduced to reconstruct the sliding features of each type of variable in the preprocessed multi-source heterogeneous dynamic data, generating a time-aligned structured feature set.

[0217] Based on the time axis, each type of variable in the structured feature set is encapsulated as a structured feature tensor to obtain a dynamic feature tensor.

[0218] In one possible implementation, the statistical reconstruction module 202 is specifically used for:

[0219] Within a preset sliding window, the set of statistics corresponding to each type of variable in the preprocessed multi-source heterogeneous dynamic data is calculated, and the set of statistics includes the mean and weighted variance.

[0220] Based on the set of statistics, the statistics of each type of variable in the multi-source heterogeneous dynamic data at each time step are encapsulated into binary structure features to obtain a set of binary structure features.

[0221] The binary structural features in the binary structural feature set are temporally assembled to obtain a temporally aligned structured feature set.

[0222] The present invention provides a cascade hydropower remote centralized control plan generation system 20, which can realize the steps and effects of the cascade hydropower remote centralized control plan generation method in embodiment 1. To avoid repetition, the present invention will not repeat them.

[0223] The beneficial effects of this invention are reflected in:

[0224] First, by integrating multi-source heterogeneous dynamic data and constructing a dynamic feature tensor, a unified and structured representation of complex operational states is achieved, significantly improving the comprehensiveness and timeliness of risk perception. Second, based on a dynamic risk index sequence, contingency plan templates are dynamically weighted and superimposed to generate a highly adaptive set of contingency plans, overcoming the shortcomings of traditional contingency plan libraries that are rigid and have low matching degrees. Finally, by dynamically matching and executing the optimal contingency plan in conjunction with real-time operational status, the accuracy, agility, and pertinence of emergency response are ensured, greatly improving the safety assurance capability and emergency control efficiency of the cascade hydropower remote centralized control system in complex and ever-changing environments.

[0225] In the description of the embodiments of the present invention, it should be understood that the terms "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "center," "top," "bottom," "top," "bottom," "inner," "outer," "inner side," and "outer side," etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. "Inner side" refers to the interior or enclosed area or space. "Outer perimeter" refers to the area surrounding a specific component or specific area.

[0226] In the description of embodiments of the present invention, the terms "first," "second," "third," and "fourth" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first," "second," "third," or "fourth" may explicitly or implicitly include one or more of that feature. In the description of the present invention, unless otherwise stated, "a plurality of" means two or more.

[0227] In the description of the embodiments of the present invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," "joining," and "assembly" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention based on the specific circumstances.

[0228] In the description of embodiments of the present invention, specific features, structures, materials or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0229] In the description of the embodiments of the present invention, it should be understood that "-" and "~" represent a range of two numerical values, and this range includes the endpoints. For example, "AB" represents a range greater than or equal to A and less than or equal to B. "A~B" represents a range greater than or equal to A and less than or equal to B.

[0230] In the description of embodiments of the present invention, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0231] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for generating remote centralized control plans for cascade hydropower projects, characterized in that, include: Acquire multi-source heterogeneous dynamic data of the target cascade hydropower station; The multi-source heterogeneous dynamic data is statistically reconstructed to generate a dynamic feature tensor, which is used to characterize the temporal evolution and statistical distribution characteristics of the multi-source heterogeneous dynamic data in a structured multidimensional space. Based on the dynamic feature tensor, risk identification is performed on the target cascade hydropower station to obtain a risk index sequence of the target cascade hydropower station; wherein, the risk index sequence is obtained by time-series assembly processing of the intermediate results of the risk identification. Based on the risk index sequence, the preset set of contingency plan templates are dynamically weighted and superimposed to generate an adaptive set of contingency plans; Based on the real-time operating status of the cascade hydropower remote centralized control system, the system is dynamically matched in the set of adaptive contingency plans to determine the optimal contingency plan, so as to control the cascade hydropower remote centralized control system based on the optimal contingency plan; The step of statistically reconstructing the multi-source heterogeneous dynamic data to generate a dynamic feature tensor includes: A sliding feature reconstruction mechanism is introduced to reconstruct the sliding features of each type of variable in the multi-source heterogeneous dynamic data, generating a time-aligned structured feature set. Based on the time axis, each type of variable in the structured feature set is encapsulated as a structured feature tensor to obtain a dynamic feature tensor; The step of dynamically weighting and superimposing a preset set of contingency plan templates based on the risk index sequence to generate an adaptive contingency plan set includes: Based on the preset set of response weights and the risk index sequence, a risk response impact vector is generated; The risk response impact vector is mapped to a contingency plan structure factor vector using a structure activation mapping function. Based on the aforementioned contingency plan structure factor vector, the preset contingency plan template set is dynamically weighted and superimposed to generate an adaptive contingency plan set. Based on the aforementioned contingency plan structure factor vector, the preset contingency plan template set is dynamically weighted and superimposed to generate an adaptive contingency plan set, including: The proposed structural factor vector is subjected to structural normalization to obtain the structural normalized proposed structural factor vector. Obtain the structure activation weight matrix corresponding to the structural factor vector of the pre-planned structure after the structure normalization process; Based on the structure activation weight matrix, the combination of preset plan templates is dynamically scheduled to obtain the dynamic combination of the preset plan template set; Based on the aforementioned dynamic combination method, the preset plan template set is dynamically weighted and superimposed to obtain an adaptive plan set.

2. The method according to claim 1, characterized in that, After dynamically weighting and superimposing the preset plan template set based on the plan structure factor vector to generate an adaptive plan set, the method further includes: Get the function to evaluate the execution effect; The effectiveness of the adaptive contingency plan set is evaluated according to the execution effectiveness evaluation function to obtain the effectiveness evaluation result; The effect evaluation results include the execution performance indicators of each plan in the adaptive plan set during the execution process, and the execution performance indicators include response time, control stability, and goal achievement degree; Obtain historical execution performance data; Based on the historical execution effect data and the effect evaluation results, the structure activation weight matrix corresponding to the plan structure factor vector is updated by gradient so as to adaptively adjust the adaptive plan set.

3. The method according to claim 1, characterized in that, The real-time operating status based on the cascade hydropower remote centralized control system is dynamically matched with the adaptive contingency plan set to determine the optimal contingency plan, including: Construct a dynamic matching scoring function; The matching degree between each plan in the adaptive plan set and the real-time operating status is calculated using the dynamic matching scoring function to obtain a matching degree set. Based on the preset matching rules and the matching degree set, candidate emergency plans are determined; The key control parameter set of the candidate emergency response plan is fine-tuned in a gradient to optimize the preset objective function value and obtain the optimal plan.

4. The method according to claim 3, characterized in that, After controlling the cascade hydropower remote centralized control system based on the optimal plan, the system further includes: Obtain the actual control indicators corresponding to the optimal plan; Based on the expected indicators and the actual control indicators, calculate the performance loss corresponding to the optimal plan; Calculate the gradient of the performance loss with respect to the weight vector in the dynamic matching scoring function; The weight vector is dynamically updated based on the gradient.

5. The method according to claim 1, characterized in that, The step of identifying risks of the target cascade hydropower stations based on the dynamic feature tensor to obtain a risk index sequence of the target cascade hydropower stations includes: Based on a preset set of risk factors, risk factors are extracted from the dynamic feature tensor to obtain a target set of risk factors; Extract the subset of sensor variable features associated with each type of risk factor in the target risk factor set to obtain a risk feature vector set; Based on the set of risk feature vectors, a weighted nonlinear transformation is performed on each type of risk factor in the set of target risk factors to obtain a set of risk activity scalar values. Based on the set of risk activity scalar values ​​and the preset core indicator data of each type of factor in the target risk factor set, standardized calculations are performed to obtain a set of comprehensive risk indices; the preset core indicator data includes historical frequency, current activity, and potential impact intensity. The risk index vectors in the comprehensive risk index set are subjected to time-series assembly processing to obtain the risk index sequence.

6. The method according to claim 1, characterized in that, Before introducing the sliding feature reconstruction mechanism to perform sliding feature reconstruction on each type of variable in the multi-source heterogeneous dynamic data, the following is also included: Data preprocessing is performed on the multi-source heterogeneous dynamic data to eliminate dynamic anomalies in the multi-source heterogeneous dynamic data.

7. The method according to claim 6, characterized in that, The introduced sliding feature reconstruction mechanism reconstructs the features of each type of variable in the multi-source heterogeneous dynamic data, generating a time-aligned structured feature set, including: Within a preset sliding window, calculate the set of statistics corresponding to each type of variable in the multi-source heterogeneous dynamic data, wherein the set of statistics includes the mean and weighted variance. Based on the set of statistics, the statistics of each type of variable in the multi-source heterogeneous dynamic data at each time step are encapsulated into binary structure features to obtain a set of binary structure features. The binary structural features in the binary structural feature set are temporally assembled to obtain a temporally aligned structured feature set.

8. A remote centralized control plan generation system for cascade hydropower projects, characterized in that, include: The data acquisition module is used to acquire multi-source heterogeneous dynamic data of the target cascade hydropower station; The statistical reconstruction module is used to perform statistical reconstruction on the multi-source heterogeneous dynamic data and generate a dynamic feature tensor. The dynamic feature tensor is used to characterize the temporal evolution and statistical distribution characteristics of the multi-source heterogeneous dynamic data in a structured multidimensional space. The risk identification module is used to identify the risks of the target cascade hydropower station based on the dynamic feature tensor, and obtain the risk index sequence of the target cascade hydropower station; wherein, the risk index sequence is obtained by performing time-series assembly processing on the intermediate results of the risk identification. The contingency plan generation module is used to dynamically weight and superimpose a preset set of contingency plan templates based on the risk index sequence to generate an adaptive set of contingency plans. The contingency plan matching module is used to dynamically match the adaptive contingency plan set based on the real-time operating status of the cascade hydropower remote centralized control system, and determine the optimal contingency plan so as to control the cascade hydropower remote centralized control system based on the optimal contingency plan; The statistical reconstruction module is specifically used for: A sliding feature reconstruction mechanism is introduced to reconstruct the sliding features of each type of variable in the multi-source heterogeneous dynamic data, generating a time-aligned structured feature set. Based on the time axis, each type of variable in the structured feature set is encapsulated as a structured feature tensor to obtain a dynamic feature tensor; The contingency plan generation module is specifically used for: Based on the preset set of response weights and the risk index sequence, a risk response impact vector is generated; The risk response impact vector is mapped to a contingency plan structure factor vector using a structure activation mapping function. Based on the aforementioned contingency plan structure factor vector, the preset contingency plan template set is dynamically weighted and superimposed to generate an adaptive contingency plan set. The contingency plan generation module is specifically used for: The proposed structural factor vector is subjected to structural normalization to obtain the structural normalized proposed structural factor vector. Obtain the structure activation weight matrix corresponding to the structural factor vector of the pre-planned structure after the structure normalization process; Based on the structure activation weight matrix, the combination of preset plan templates is dynamically scheduled to obtain the dynamic combination of the preset plan template set; Based on the aforementioned dynamic combination method, the preset plan template set is dynamically weighted and superimposed to obtain an adaptive plan set.

Images