Electric power resource delivery evaluation method, electronic equipment and medium

Abstract

The invention discloses an electric power resource release evaluation method, electronic equipment and a medium, and the method comprises the steps: collecting first ecological data of an electric power system in real time, the first ecological data comprising system production data and external environment data; preprocessing the first ecological data to obtain a first power data set, and determining putting effect data according to the first power data set and a trained putting effect prediction model; dividing the first power data set into at least two first classification subsets, performing feature extraction on each first classification subset to obtain a first power feature set, and determining delivery risk data according to the first power feature set and a trained delivery risk prediction model; and according to the putting effect data and the putting risk data, determining the evaluation result of the power resource putting, thereby solving the problems that the evaluation dimension of the putting decision is single and the data sources of the double models are inconsistent, improving the accuracy of the evaluation task, and ensuring the safe and stable operation of the power system.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an evaluation method, electronic device and medium for power resource allocation. Background Technology

[0002] Power resource allocation, as a key technological link, runs through the entire process of power planning, construction, maintenance, and upgrading. With the increasing penetration of renewable energy, the increasing complexity of power grid topology, and the diversification of load demand, the scientific nature of power resource allocation has an increasingly significant impact on the operational efficiency of the power system.

[0003] Currently, the power industry has widely adopted big data technology, and the evaluation tasks mainly focus on two dimensions: deployment effectiveness and deployment risk. That is, by building deployment effectiveness prediction models and deployment risk prediction models, the corresponding evaluation objectives can be achieved.

[0004] However, in existing dual-model schemes, the deployment effect prediction model plays a leading role in decision-making, while the deployment risk prediction model only serves as an auxiliary early warning function, making it difficult to effectively correct deployment decisions. Furthermore, the two models employ independent modeling strategies, resulting in significant differences in data selection. For example, the deployment effect prediction model primarily uses electricity supply and demand data, while the deployment risk prediction model focuses more on equipment status data. This inconsistency in data sources leads to biases in feature representation, weakening the consistency of the two models' prediction results, and consequently making it difficult for the final evaluation results to effectively support investment decisions.

[0005] The aforementioned problems not only easily lead to the waste of power resources through ineffective or excessive allocation, but may even affect the safe and stable operation of the power system. Summary of the Invention

[0006] This invention provides an evaluation method, electronic device, and medium for power resource allocation, which solves the problems of a single evaluation dimension and inconsistent data sources between the two models, improves the accuracy of the evaluation task, and ensures the safe and stable operation of the power system while improving the utilization rate of power resources.

[0007] According to one embodiment of the present invention, an evaluation method for power resource allocation is provided, the method comprising:

[0008] Real-time acquisition of primary ecological data of the power system, which includes system production data and external environmental data;

[0009] The first ecological data is preprocessed to obtain the first power dataset. Based on the first power dataset and the trained deployment effect prediction model, the deployment effect data is determined.

[0010] The first power dataset is divided into at least two first category subsets. Features are extracted from each first category subset to obtain a first power feature set. Based on the first power feature set and the trained deployment risk prediction model, deployment risk data is determined.

[0011] The evaluation results of power resource allocation are determined based on the data on allocation effectiveness and the data on allocation risk.

[0012] According to another embodiment of the present invention, an evaluation device for power resource allocation is provided, the device comprising:

[0013] The first ecological data acquisition module is used to collect the first ecological data of the power system in real time. The first ecological data includes system production data and external environment data.

[0014] The deployment effect data determination module is used to preprocess the first ecological data to obtain the first power dataset, and determine the deployment effect data based on the first power dataset and the trained deployment effect prediction model.

[0015] The deployment risk data determination module is used to divide the first power dataset into at least two first category subsets, extract features from each first category subset to obtain a first power feature set, and determine the deployment risk data based on the first power feature set and the trained deployment risk prediction model.

[0016] The evaluation result determination module is used to determine the evaluation result of power resource allocation based on the allocation effect data and the allocation risk data.

[0017] According to another embodiment of the present invention, an electronic device is provided, the electronic device comprising:

[0018] At least one processor; and

[0019] A memory communicatively connected to the at least one processor; wherein,

[0020] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the power resource allocation assessment method according to any embodiment of the present invention.

[0021] According to another embodiment of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement the power resource allocation evaluation method described in any embodiment of the present invention.

[0022] According to another embodiment of the present invention, a computer program product is provided, including a computer program that, when executed by a processor, implements the power resource allocation evaluation method described in any embodiment of the present invention.

[0023] The technical solution of this invention eliminates data source barriers and solves the problem of inconsistent data sources between the two models by comprehensively collecting ecological data of the power system. Furthermore, by classifying and dividing the power dataset and conducting deep data mining on each subset, the risk identification capability of the deployment risk prediction model is guaranteed. By coupling the deployment effect evaluation model and the deployment risk prediction model, the evaluation task of power resource deployment is completed collaboratively. The two-dimensional evaluation system avoids the misleading influence of a single evaluation dimension on deployment decisions. This invention provides accurate and reliable data support for power resource deployment by improving multiple technical implementations, significantly improving the scientific and rational nature of investment decisions, and ensuring the efficient and safe operation of the power system while increasing the utilization rate of power resources.

[0024] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A flowchart illustrating an evaluation method for power resource allocation according to an embodiment of the present invention;

[0027] Figure 2 A flowchart illustrating another method for evaluating power resource allocation, provided as an embodiment of the present invention;

[0028] Figure 3 This is a schematic diagram of the structure of an evaluation device for power resource allocation provided in one embodiment of the present invention;

[0029] Figure 4 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present invention. Detailed Implementation

[0030] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. 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 should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0032] Figure 1 This is a flowchart illustrating an evaluation method for power resource allocation according to an embodiment of the present invention. This embodiment is applicable to situations where an implementation evaluation of a power resource allocation plan is conducted. The method can be executed by a power resource allocation evaluation device, which can be implemented in hardware and / or software and can be configured in a terminal device. Figure 1 As shown, the method includes:

[0033] S110: Real-time acquisition of primary ecological data from the power system.

[0034] A power system is an organic whole composed of power generation, transmission, transformation, distribution, and consumption. Its core function is to achieve a closed-loop operation of the entire process of electricity production, transmission, distribution, and consumption, while also including auxiliary systems such as control, protection, and communication to ensure the stable operation of this process. Specifically, a power system encompasses various types of power generation equipment, transmission and transformation lines and equipment, distribution networks, electricity terminals, and corresponding dispatch and management systems. For example, power generation equipment includes, but is not limited to, thermal power units, hydropower units, wind farms, and photovoltaic power stations; transmission and transformation lines and equipment include, but are not limited to, high-voltage transmission lines, substations, and transformers; and user terminals include, but are not limited to, industrial terminals and residential terminals.

[0035] Among them, the first ecological data refers to a set of data collected in real time that can comprehensively reflect the operation of the power system and external influencing factors, possessing real-time and full-dimensional coverage.

[0036] In this embodiment, the first ecological data includes system production data and external environmental data. System production data represents operational information generated during various production processes of the power system, while external environmental data represents environmental factors that directly or indirectly affect the power system's system production data.

[0037] For example, system production data includes, but is not limited to, generator output, power generation efficiency, and fuel consumption in the power generation stage; line voltage, current, power loss, and line temperature in the transmission stage; transformer load rate, insulation status, and tap position in the substation stage; distribution network voltage quality, load distribution, and line fault information in the distribution stage; and generator dispatch instructions, load forecast results, and operation mode adjustment records in the dispatching stage. External environmental data includes, but is not limited to, meteorological environmental data, geographical environmental data, social environmental data, and policy environmental data. Among these, meteorological environmental data can affect the power generation of new energy sources, the operating status of lines, and electricity load; geographical environmental data can affect the laying and operational safety of transmission and distribution lines; social environmental data can affect the spatiotemporal distribution of electricity load; and policy environmental data can affect the planning and implementation of power resource allocation.

[0038] This section only provides examples of system production data and external environment data, and does not limit them. You can customize the settings according to your actual needs.

[0039] In one optional embodiment, the system production data includes system operation data and equipment status data, and / or, the external environment data includes intra-domain environment data and cross-domain environment data.

[0040] In this embodiment, the system operation data represents the operating condition information of the power system, and the equipment status data represents the health status information of the physical equipment in the power system.

[0041] In one optional embodiment, the system operation data includes at least one of electrical data, power load data, system fault records, system anomaly records, system efficiency data, and operation records under extreme weather conditions.

[0042] For example, electrical data includes, but is not limited to, node voltage, line current, active power, reactive power, and system frequency; power load data includes, but is not limited to, load peak, load valley, average load, and power demand load curves for different time periods; system fault records include, but are not limited to, fault occurrence time, fault location, fault type, fault duration, fault impact range, and the amount of resources lost due to the fault; system anomaly records include, but are not limited to, information on voltage fluctuation exceeding limits, information on frequency fluctuation exceeding limits, and information on line overload; system efficiency data includes, but is not limited to, transmission loss rate, line utilization rate, and power flow distribution data; and extreme weather includes, but is not limited to, heavy rain, heavy snow, typhoons, high temperatures, and cold waves.

[0043] In one optional embodiment, the device status data includes at least one of the following: device attribute data of physical devices, operation monitoring data, device fault and deterioration data, maintenance data, and spare parts related data.

[0044] For example, equipment attribute data includes, but is not limited to, equipment model, installation time, and design life; operation monitoring data includes, but is not limited to, cumulative operating time, equipment temperature, vibration amplitude, and oil quality parameters; equipment failure and deterioration data includes, but is not limited to, number of failures, failure precursor data, equipment deterioration rate, and equipment deterioration trend; maintenance data includes, but is not limited to, maintenance time, maintenance reasons, maintenance costs, maintenance cycle, and delayed records of maintenance not carried out as planned; and spare parts related data includes, but is not limited to, spare parts inventory data and supply stability. Among these, failure precursor data includes, but is not limited to, dissolved gas content in oil, partial discharge quantity, and insulation performance parameters.

[0045] In this embodiment, the intra-domain environmental data represents feedback information of environmental factors that are directly coupled with the power system, and the cross-domain environmental data represents regulation information of environmental factors that are indirectly related to the power system.

[0046] In one alternative embodiment, the domain environmental data includes at least one of meteorological environmental data, data on the amount of resources that need to be transferred for electricity use, and supply and demand environmental data.

[0047] For example, meteorological environmental data includes, but is not limited to, temperature data of transmission lines, relative humidity data of substation areas, wind data, solar radiation intensity data of photovoltaic power stations, precipitation data, and lightning data. Data on the amount of resources that need to be transferred for electricity use includes, but is not limited to, the amount of resources that need to be transferred for electricity use at different times, the amount of resources that need to be transferred for electricity use collected in real time, the amount of resources that need to be transferred for electricity use in history, the amount of resources that need to be transferred for grid electricity use, and the amount of resources that need to be transferred for commercial electricity use. Supply and demand environmental data includes, but is not limited to, electricity consumption data, power generation data, power generation scale data, power generation forecast data for future periods, and electricity consumption forecast data.

[0048] In one alternative embodiment, cross-domain environmental data includes, but is not limited to, at least one of social environmental data, policy environmental data, energy environmental data, geographical environmental data, and economic environmental data associated with the amount of resources input.

[0049] For example, social environment data includes, but is not limited to, population data, urban planning data, GDP data, and industrial structure data of the region where the data is distributed. Among these, population data includes, but is not limited to, the number of permanent residents, the number of migrants, population growth rate, and population migration trends. Policy environment data includes, but is not limited to, electricity subsidy policies, carbon emission policies, ownership transfer policies, and approval standards for the construction of power facilities. Energy environment data includes, but is not limited to, energy reserve data, energy consumption data, pollutant emission data, energy allocation data, energy storage data, and energy demand trends. Geographic environment data includes, but is not limited to, topographic slope and elevation data, geological data of transmission lines, topographic data of hydropower stations and wind power stations, and vegetation cover data. Economic environment data related to the amount of resources invested includes, but is not limited to, conversion parameters of the amount of resources invested, tax burden parameters, and ownership transfer matching parameters.

[0050] Based on the above embodiments, optionally, the intra-domain environmental data includes electricity consumption data, and the cross-domain environmental data includes electricity consumption environment data associated with electricity consumption. Accordingly, before determining the delivery effect data based on the first power dataset and the trained delivery effect prediction model, the method further includes: acquiring preprocessed electricity consumption data and electricity consumption environment data from the first power dataset; statistically analyzing the electricity consumption data and electricity consumption environment data according to a preset time period to obtain electricity consumption time-series data and electricity consumption environment time-series data; determining the electricity demand trend based on the electricity consumption time-series data and electricity consumption environment time-series data, and merging the electricity demand trend into the first power dataset.

[0051] Specifically, the electricity consumption environment data includes environmental factors related to electricity consumption. For example, the electricity consumption environment data includes environmental factors such as temperature, resident population, urbanization rate, and GDP. For instance, stable high or low temperatures will drive the growth of electricity consumption, an increase in the resident population will drive the growth of residential electricity consumption, a faster growth rate in urbanization will drive the growth of infrastructure electricity consumption, and a faster growth rate in GDP will drive the growth of industrial electricity consumption. This is just an example of electricity consumption environment data, but it is not limited to the example given above.

[0052] Specifically, the electricity consumption data includes real-time electricity consumption and historical electricity consumption. For example, the preset time period is one month, one quarter, or one year, but is not limited to the example scenario.

[0053] Before determining the electricity demand trend based on the electricity consumption time-series data and the electricity environment time-series data, the method further includes: aligning the statistically obtained electricity consumption time-series data and the electricity environment time-series data in terms of time dimension, and removing data in the electricity consumption time-series data and the electricity environment time-series data that are inconsistent in time dimension.

[0054] Specifically, the electricity demand trend represents the short-term fluctuation pattern or long-term direction of electricity consumption in response to environmental factors. For example, the electricity demand trend can be short-term growth, short-term stability, or short-term decline, or the electricity demand trend can be long-term growth, long-term stability, or long-term decline.

[0055] For example, electricity demand trends can be obtained through environmental-driven modeling or trend fitting algorithms, but are not limited to the given examples.

[0056] Based on the above embodiments, optionally, before determining the electricity demand trend based on the electricity consumption time-series data and the electricity environment time-series data, the method further includes: smoothing the electricity consumption time-series data and the electricity environment time-series data according to the average moving period.

[0057] Taking electricity consumption time-series data as an example, data is selected from the electricity consumption time-series data according to the average moving period. The electricity consumption statistics within the average moving period are averaged to obtain a moving average. Multiple moving averages constitute the smoothed electricity consumption time-series data. For example, the first... moving average Satisfy the following formula:

[0058]

[0059] in, Indicates the average moving period. Indicates the first Electricity consumption within a preset time period.

[0060] The advantage of setting up smoothing is that it reduces the interference of short-term random fluctuations on the trend of electricity demand, highlights the core trend of electricity demand changes with environmental factors, and thus improves the accuracy and stability of electricity demand trends.

[0061] The advantage of setting electricity demand trends is that it strengthens the matching logic between electricity resource allocation and electricity demand. By quantifying the temporal fluctuations and scale changes in electricity demand, the allocation effect prediction model can more objectively evaluate core performance indicators such as resource utilization, supply and demand balance, and system operation losses, thereby improving the accuracy and reliability of the allocation effect prediction model.

[0062] S120. The first ecological data is preprocessed to obtain the first power dataset. Based on the first power dataset and the trained deployment effect prediction model, the deployment effect data is determined.

[0063] The preprocessing process involves a series of standardized and normalized technical operations performed on the first ecological data. In one optional embodiment, preprocessing includes at least one of data cleaning, noise reduction, format conversion, and standardization.

[0064] Data cleaning is a preprocessing operation that verifies and corrects the validity of the primary ecological data. It identifies and handles missing, outlier, duplicate, and logically contradictory values ​​in the primary ecological data. Through methods such as completion, correction, or removal, it eliminates invalid information generated during data collection or transmission, ensuring the accuracy, completeness, and consistency of the data. For example, completion methods can be implemented using interpolation or deep learning models.

[0065] Among them, denoising is a preprocessing operation that separates and suppresses random interference signals mixed in the first ecological data. Random interference signals mostly originate from electromagnetic interference of data acquisition equipment, environmental interference, or data transmission loss. Through filtering, smoothing and other methods, the core trends and effective characteristics of power ecological parameters are retained, and meaningless random fluctuations are eliminated.

[0066] The format conversion operation is a preprocessing operation that converts the storage format, data type, or structure of the first ecological data into a unified standard format. The standardization processing is a preprocessing operation that normalizes or standardizes the power ecological parameters of different magnitudes and dimensions in the first ecological data. For example, it can map the data to a unified numerical range or make it meet a specific statistical distribution through mathematical transformation, so as to eliminate the influence of differences in dimensions and numerical ranges, and meet the input requirements of the deployment effect prediction model.

[0067] Among them, the deployment effect prediction model refers to the model that uses machine learning technology to train by learning the mapping relationship between the power system's power ecological parameters and the actual deployment effect in historical execution scenarios of power resource deployment schemes, and has the ability to quantify the execution effect of power resource deployment schemes.

[0068] For example, the power resource allocation scheme can be applied to power deployment scenarios such as equipment upgrades, capacity expansion construction, power generation technology transformation, new energy integration, and emergency power supply guarantee. The allocation effect prediction model can be a random forest model, a neural network model, a support vector regression model, a gradient boosting tree model, or a multiple linear regression model, but it is not limited to the example scenarios given above.

[0069] Among them, the deployment effect data is a set of indicators of the operational benefits generated by implementing the power resource deployment plan in a real-time power ecosystem scenario.

[0070] In one optional embodiment, the deployment effect data includes economic benefit data and / or social benefit data. The economic benefit data characterizes the indicators relating the power system to the amount of resources after implementing the power resource deployment plan in a real-time power ecosystem scenario. For example, economic benefit data includes the ratio of the power system's output resources to its input resources, the time period for the output resources to reach the input resources, and the input resources corresponding to a unit of power output, but is not limited to the given example.

[0071] Among them, social benefit data characterizes the indicators related to the power system and people's livelihood after the power resource allocation plan is implemented in a real-time power ecosystem scenario. For example, social benefit data includes the power supply reliability improvement rate, the power outage duration reduction rate, and carbon emission reduction. The power supply reliability improvement rate characterizes the degree to which the power resource allocation plan improves the continuous and stable operation of the power system, and the carbon emission reduction characterizes the degree to which the power resource allocation plan contributes to ecological and environmental protection.

[0072] It should be noted that the deployment effect data in this embodiment is only used as data support for evaluating the feasibility, optimization potential and comprehensive value of the power resource deployment plan, so as to adapt to the application scenario requirements such as decision-making, parameter adjustment and effect verification of the power resource deployment plan.

[0073] S130. Divide the first power dataset into at least two first category subsets, extract features from each first category subset to obtain a first power feature set, and determine the deployment risk data based on the first power feature set and the trained deployment risk prediction model.

[0074] Specifically, the first category subset represents a set of homogeneous and low-redundancy power ecological parameters. Each first category subset focuses on a risk event or risk impact dimension, and the characteristic expression of the power ecological parameters within the set is consistent.

[0075] Subset partitioning can effectively solve the problem of poor predictive generalization of global data. It not only avoids cross-dimensional data interference, but also meets the input directionality requirements of the risk prediction model.

[0076] In one optional embodiment, the multiple first category subsets include a first category subset corresponding to system production data and a first category subset corresponding to external environment data. The advantage of this setup is that it clarifies the subject boundary of the data source, enabling the deployment risk prediction model to distinguish between internal and external risks.

[0077] Understandably, based on system production data and external environment data, smaller-granular subset divisions can be made. For example, first-category subsets can be defined corresponding to system operation data, equipment status data, intra-domain environment data, and cross-domain environment data, respectively. Alternatively, first-category subsets can be defined corresponding to electrical data, power load data, and system fault records within the system operation data. Smaller-granular subset divisions enable the deployment risk prediction model to focus on specific risk events. By highlighting the importance of a certain first-category subset, data interference from other risk events can be avoided, thereby improving the prediction accuracy of the deployment risk prediction model for a single risk event.

[0078] In another alternative embodiment, the multiple first category subsets include a first category subset corresponding to real-time dynamic data and a first category subset corresponding to basic static data. For example, real-time dynamic data includes electrical data, power load data, and meteorological environmental data, etc., while basic static data includes equipment attribute data, policy environment data, spare parts basic data, and geographical environment data, etc.

[0079] The advantage of this setup is that it achieves subset division from two dimensions: real-time volatility risk and long-term inherent risk, enabling the risk prediction model to have the ability to identify static and dynamic risks in a hierarchical manner.

[0080] Specifically, since the feature representation of the first classification subset is consistent, features can be extracted from some or all of the power ecological parameters in the first classification subset. The power classification features in the first power feature set represent the comprehensive feature information corresponding to some or all of the power ecological parameters in the first classification subset.

[0081] In one alternative embodiment, the first power feature set includes at least two power classification features among system failure rate, fluctuation rate of power resources, rate of change of power consumption trend, and degree of constraint of policy influence.

[0082] In one specific embodiment, the system failure rate is determined based on the device failure rate corresponding to at least one physical device, and the first category subset corresponding to the system failure rate includes at least the number of failures and the cumulative running time of the physical devices.

[0083] For example, equipment failure rate Satisfy the following formula:

[0084]

[0085] in, Indicates the number of failures. This indicates the cumulative running time.

[0086] In an optional embodiment, the first category subset corresponding to the system failure rate further includes at least one of the following: fault precursor data, maintenance data, system anomaly records, and equipment degradation data.

[0087] In one embodiment, for example, a baseline failure rate is determined based on the number of failures and cumulative uptime; a correction factor is determined based on failure precursor data; a correction coefficient is determined based on maintenance data, system anomaly records, and equipment degradation data; and the equipment failure rate is determined based on the baseline failure rate, the correction factor, and the correction coefficient.

[0088] Specifically, the greater the fault risk represented by the fault precursor data, the larger the correction factor. For example, the greater the dissolved gas content in the oil, the greater the partial discharge, or the greater the insulation performance parameters, the larger the correction factor. Similarly, the greater the correction coefficient, the greater the repetition of maintenance causes in the maintenance data, the longer the lag time of unplanned maintenance, the higher the temperature exceedance in the system anomaly records, the more abnormal the vibration amplitude, and the faster the rate of oil quality deterioration in the equipment deterioration data.

[0089] For example, equipment failure rate Satisfy the following formula:

[0090]

[0091] in, Indicates the baseline failure rate. This represents the correction factor. This represents the correction factor.

[0092] In one specific embodiment, the fluctuation rate of electricity consumption is determined based on the real-time fluctuation rate of electricity consumption and / or the fluctuation rate of electricity consumption in at least one time period. The first category subset corresponding to the fluctuation rate of electricity consumption includes the amount of electricity consumption that needs to be transferred in real time, the amount of electricity consumption that needs to be transferred in at least one historical time period, and / or the amount of electricity consumption that needs to be transferred in at least two time periods.

[0093] For example, the volatility of electricity consumption. Satisfy the following formula:

[0094]

[0095] Specifically, when the fluctuation rate of electricity consumption is the real-time fluctuation rate of resource quantity, This represents the amount of resources that need to be transferred for electricity consumption, collected in real time. This represents the average amount of resources required to transfer for electricity consumption over multiple historical time periods. When the volatility of electricity resource consumption is the time-period resource volatility, This indicates the amount of resources that need to be transferred for electricity consumption during the current time period. This represents the average amount of resources that need to be transferred for electricity consumption over multiple time periods.

[0096] For example, the electricity demand trend can be quantified by determining the rate of change of electricity demand trend based on the electricity demand trend at the current data collection time and the average of the electricity demand trends at multiple historical data collection times.

[0097] Specifically, the policy impact constraint degree represents the degree of impact of changes in environmental factors in policy environment data on the power system. The policy impact constraint degree is determined based on changes in at least one environmental factor.

[0098] In one embodiment, for example, for each environmental factor in the policy environment data, a range-based quantitative system is established for that environmental factor. Based on the change information of environmental factors in the first subset, the influence degree of each environmental factor is determined. Based on the influence degree of at least one environmental factor, the policy influence constraint degree is determined. For instance, the greater the reduction in electricity subsidies under the electricity subsidy policy, the greater the influence degree of the electricity subsidy policy; the greater the carbon quota deficit rate under the carbon emission policy, the greater the influence degree of the carbon emission policy; and the influence degree of the power facility construction approval regulations can be used as a coefficient to adjust the policy influence constraint degree. For example, if the approval time in the power facility construction approval regulations is extended, the proportion of the policy influence constraint degree increases.

[0099] Among them, the deployment risk prediction model refers to the model that uses machine learning or statistical analysis techniques to learn the mapping relationship between the power classification characteristics of the power system and the actual deployment risk in historical execution scenarios of power resource deployment schemes. It has the ability to quantify the probability of one or more risk events occurring.

[0100] For example, the risk prediction model can be a logistic regression model, a gradient boosting decision tree, an attention model, a Bayesian network model, or a support vector machine, but is not limited to the given example.

[0101] Among them, the deployment risk data is a set of probabilities of risk events occurring after the deployment plan of power resources is executed in a real-time power ecosystem scenario. It is used to reflect the degree of risk exposure of the deployment plan of power resources in the real power ecosystem scenario.

[0102] Based on the above embodiments, the method may optionally further include: outputting warning information to indicate the risk event when the probability of the risk event occurring is greater than a probability threshold.

[0103] Based on the above embodiments, optionally, the deployment risk prediction model is optimized and updated in the following manner: according to the deployment effect prediction model, sensitivity analysis is performed on the power ecological parameters in the first power dataset to obtain at least one sensitive ecological parameter; at least one first category subset to which the at least one sensitive ecological parameter belongs is obtained, and the feature weights of the power classification features corresponding to the first category subset in the deployment risk prediction model are optimized and adjusted.

[0104] Sensitivity analysis, in particular, uses the controlled variable method in the delivery effect prediction model to determine the degree of influence of each power ecological parameter in the first power dataset on the delivery effect data. Specifically, one power ecological parameter is perturbed while the remaining power ecological parameters in the first power dataset are fixed. The delivery effect prediction model is run repeatedly and the change range of the delivery effect data is recorded. By quantifying the correlation between the perturbation range of the power ecological parameter and the fluctuation range of the delivery effect data, the sensitivity of the power ecological parameter to the delivery effect data is determined.

[0105] For example, for each power ecological parameter, the power ecological parameter is adjusted within the historical fluctuation range of the power ecological parameter, and the adjusted power ecological parameter is used as the input variable of the deployment effect prediction model to obtain the deployment effect data output by the deployment effect prediction model. Based on the adjustment range of the power ecological parameter and the fluctuation range of the deployment effect data, the sensitivity coefficient corresponding to the power ecological parameter is determined. If the sensitivity coefficient is greater than 1, the power ecological parameter is regarded as a sensitive ecological parameter.

[0106] The advantage of this setup is that the linkage mechanism of adjusting deployment risk through feedback on deployment results ensures that the weighting of deployment risk is precisely matched with the impact of power ecological parameters on deployment results. This strengthens the consistency of the two evaluation models in identifying sensitive ecological parameters. This coupled adjustment can filter out redundant interference from non-sensitive ecological parameters, optimize the feature synergy efficiency of the two models, and further improve the scientificity and reliability of power resource deployment decisions.

[0107] S140. Based on the deployment effect data and the deployment risk data, determine the evaluation result of the power resource deployment.

[0108] Specifically, the deployment effect data and the deployment risk data are normalized, and the evaluation results of power resource deployment are determined based on the normalized deployment effect data and the deployment risk data.

[0109] Specifically, the evaluation results characterize the evaluation score, evaluation classification, evaluation decision, or scheme evaluation optimization of the power resource allocation scheme in a real-time power ecological scenario.

[0110] In one optional embodiment, determining the evaluation result of power resource allocation based on the normalized allocation effect data and the allocation risk data includes: determining allocation effect indicators based on the normalized allocation effect data, determining allocation risk indicators based on the normalized allocation risk data, and determining the evaluation result of power resource allocation based on the allocation effect indicators and the allocation risk indicators.

[0111] In one specific embodiment, the product of the deployment risk index and the risk aversion coefficient is obtained, and the difference between the deployment effect index and the product result is used as the evaluation score for power resource deployment. A higher risk aversion coefficient indicates that the decision-maker is more sensitive to risk.

[0112] In another specific embodiment, based on the application scenario of the power resource allocation plan, the weights corresponding to the allocation effectiveness index and the allocation risk index are determined respectively. Combining these weights, a weighted sum is applied to the allocation effectiveness index and the allocation risk index to obtain the evaluation score for power resource allocation. For example, when the application scenario is centralized residential power supply, the weight of the allocation effectiveness index is less than the weight of the allocation risk index; when the application scenario is industrial production, the weight of the allocation effectiveness index is equal to the weight of the allocation risk index, but this is not limited to the example scenarios.

[0113] In another specific embodiment, the target quadrants of the four-quadrant matrix are determined for the deployment effectiveness indicators and deployment risk indicators, and these target quadrants are used as the evaluation classification for power resource deployment. The four quadrants of the four-quadrant matrix represent high efficiency and low risk, high efficiency and high risk, low efficiency and low risk, and low efficiency and high risk, respectively.

[0114] In another specific embodiment, the evaluation category of power resource allocation is determined based on the allocation effectiveness index and allocation risk index, and the evaluation decision for power resource allocation is determined based on the evaluation category. For example, taking the evaluation category as the target quadrant, the evaluation decision corresponding to high efficiency and low risk is immediate execution; the evaluation decision corresponding to high efficiency and high risk is execution under the premise of additional risk control requirements; the evaluation decision corresponding to low efficiency and low risk is pending; and the evaluation decision corresponding to low efficiency and high risk is no execution.

[0115] In another specific embodiment, an evaluation category or score for power resource allocation is determined based on allocation effectiveness indicators and allocation risk indicators. Based on this evaluation category or score, an optimization plan for power resource allocation is then determined. For example, taking the evaluation score as an example: when the score is 80-100, the plan optimization is deemed unnecessary, and the allocation capacity or allocation area in the power resource allocation plan can be expanded; when the score is 60-79, the plan optimization involves adjusting the power resource allocation plan by combining the deductions from the allocation effectiveness data and allocation risk data; when the score is 0-59, the plan optimization involves plan reconstruction, such as changing the allocation area or modifying the time schedule.

[0116] Specifically, the allocation plan for power resources can be optimized and adjusted by combining sensitive ecological parameters. For example, when the sensitive ecological parameter is the equipment aging rate, the allocation plan can include the amount of reserved resources for equipment replacement. When the sensitive ecological parameter is the amount of resources that need to be transferred for electricity use, the validity period of the power ownership transfer can be adjusted in the allocation plan. This is not limited to the examples given.

[0117] The technical solution of this embodiment eliminates data source barriers and solves the problem of inconsistent data sources between the two models by comprehensively collecting ecological data of the power system. Furthermore, by classifying and dividing the power dataset and conducting in-depth data mining on each subset, the risk identification capability of the deployment risk prediction model is guaranteed. By coupling the deployment effect evaluation model and the deployment risk prediction model, the evaluation task of power resource deployment is completed collaboratively. The two-dimensional evaluation system avoids the misleading influence of a single evaluation dimension on deployment decisions. This embodiment of the invention improves multiple technical implementations, providing accurate and reliable data support for power resource deployment, significantly improving the scientific and rational nature of investment decisions, and ensuring the efficient and safe operation of the power system while increasing the utilization rate of power resources.

[0118] Figure 2 This is a flowchart illustrating another method for evaluating power resource allocation according to an embodiment of the present invention. This embodiment further refines the training method of the allocation effect prediction model in the above embodiment. Figure 2 As shown, the method includes:

[0119] S210. Collect the second ecological data of the power system during its historical operating cycle.

[0120] The historical operating cycle refers to a specific time interval in which the power system has completed continuous operation. For example, the historical operating cycle can be the past year, three full quarters, or five typical load cycles, but it is not limited to the example scenario.

[0121] Specifically, the second ecological data in this embodiment corresponds to or is similar to the first ecological data in the above embodiments, and will not be described again in this embodiment.

[0122] S220. The second ecological data is preprocessed to obtain the second power dataset, and the second power dataset is divided into a training dataset, a test dataset, and a validation dataset.

[0123] In one optional embodiment, the preprocessing process for the second ecological data is the same as the preprocessing process for the first ecological data in the above embodiments.

[0124] In another optional embodiment, the preprocessing of the second ecological data further includes a data filtering operation, and the first ecological data in the following steps is the same as or similar to the second ecological data obtained after the data filtering operation.

[0125] Specifically, the data screening operation uses a significant correlation with both the deployment effect and the deployment risk as the screening criterion. By setting correlation thresholds for both deployment effect and deployment risk through quantitative correlation analysis, power ecological parameters that are simultaneously correlated with both deployment effect and deployment risk are extracted from the second ecological data. Redundant ecological parameters that are correlated in only one dimension or not correlated in both dimensions are eliminated. For example, the correlation analysis uses Pearson correlation coefficient, mutual information entropy or rank correlation analysis, but is not limited to the given example.

[0126] Based on the above embodiments, the method may optionally further include: randomly sampling the second ecological data obtained after the data filtering operation, recalculating the correlation coefficient, in order to verify the validity of the finally obtained second ecological data.

[0127] The advantage of this setup is that, based on the technical coupling relationship of the entire campaign process, it locks in the ecological parameter information that simultaneously affects campaign performance and campaign risk, ensuring that the second ecological data obtained after data filtering has the ability to support both campaign performance and campaign risk.

[0128] S230. Train the deployment effect prediction model based on the training dataset, test dataset and validation dataset.

[0129] The training dataset is used for parameter learning and pattern fitting of the campaign performance prediction model, the validation dataset is used for hyperparameter tuning and overfitting monitoring during the training process of the campaign performance prediction model, and the test dataset is used to evaluate the final performance and generalization ability of the campaign performance prediction model after training.

[0130] In one optional embodiment, the network architecture of the campaign performance prediction model is a multi-layer feedforward neural network. For example, the input variables of the campaign performance prediction model... It can be represented as ,in, This indicates the number of power ecological parameters in the input variables. The model for predicting the deployment effect satisfies the following formula:

[0131]

[0132] in, In the model for predicting the effectiveness of campaign deployment, the first... The feature vector output by each network layer, when hour, For campaign performance data, This represents the total number of layers in the neural network. This represents the activation function. Indicates the first The weight matrix of each network layer Indicates the first The bias vectors of each network layer.

[0133] During the training of the campaign performance prediction model, a closed-loop logic of "forward propagation - error calculation - backpropagation" is followed. The loss function value is minimized by adjusting the weight matrix and bias vector of the network layers. The loss function value quantifies the difference between the predicted and actual values ​​of the campaign performance prediction model. Specifically, during forward propagation, the predicted value is output based on the input variables, the current weight matrix, and the current bias vector. During error calculation, the loss function value is calculated based on the predicted and actual values. Finally, the loss function value is propagated from the output layer back to the input layer through the backpropagation algorithm, iteratively updating the weight matrix and bias vector of each network layer to gradually reduce the loss function value.

[0134] For example, the loss function value Satisfy the following formula:

[0135]

[0136] in, This represents the sample size of the training dataset. Indicates the actual value. This represents the predicted value.

[0137] Specifically, the model training process and the model validation process are carried out alternately until a preset stopping condition is met, such as stopping the model training process when the loss function value corresponding to the validation dataset no longer decreases, in order to avoid the occurrence of overfitting.

[0138] If overfitting occurs during model training, regularization methods can be used to optimize the model parameters of the power delivery effect prediction model. This limits the range of values ​​for the parameter weights corresponding to the power ecological parameters, thereby reducing the complexity of the model and improving its generalization ability. The regularization method can be L1 or L2 regularization. For example, L1 regularization satisfies the following formula:

[0139]

[0140] L2 regularization satisfies the following formula:

[0141]

[0142] in, Represents the regularization coefficient. Represents the first in the input variables The parameter weights corresponding to each power ecological parameter.

[0143] In an optional embodiment, the deployment risk prediction model is trained as follows: the second power dataset is divided into at least two second category subsets, and features are extracted from each second category subset to obtain a second power feature set. The second power feature set is then divided into k second feature subsets. The k-1 second feature subsets are used as training datasets, and the deployment risk prediction model is trained based on the training datasets. The first performance index of the deployment risk prediction model obtained in the current training process is determined based on the remaining 1 second feature subset. The step of using the k-1 second feature subsets as training datasets is iteratively executed k times, and the k first performance indices are statistically analyzed to obtain training performance indices. When the training performance index meets the training termination condition, the deployment risk prediction model obtained in the current training process is taken as the completed deployment risk prediction model.

[0144] In this embodiment, the remaining one subset of the second feature in each of the k training processes is different, where k ≥ 2. For example, training performance metrics include, but are not limited to, statistical values ​​corresponding to mean squared error, recall, accuracy, and F1 score.

[0145] In one optional embodiment, the network architecture of the risk prediction model is a multiple linear regression network, and the initial feature weights of the power classification features are obtained through expert scoring.

[0146] For example, the risk prediction model for deployment satisfies the following formula:

[0147]

[0148] in, Indicates the probability of a risk event occurring. Represents the intercept term. Represents the second set of power characteristics. Electricity classification features, Indicates the first The feature weights corresponding to each electricity classification feature This indicates the error term.

[0149] S240: Real-time acquisition of primary ecological data from the power system.

[0150] S250. The first ecological data is preprocessed to obtain the first power dataset. Based on the first power dataset and the trained deployment effect prediction model, the deployment effect data is determined.

[0151] S240-S250 in this embodiment are the same as those in the above embodiment. Figure 1 The S110-S120 shown are the same or similar, and will not be described again in this embodiment.

[0152] Based on the above embodiments, optionally, the delivery effect prediction model is optimized and updated in the following manner: determining the data difference index between the first power dataset and the second power dataset; when the data difference index is less than the difference index threshold, fine-tuning the delivery effect prediction model based on the first power dataset; when the data difference index is greater than or equal to the difference index threshold, retraining the delivery effect prediction model based on the first power dataset and at least a portion of the second power dataset.

[0153] Among them, the data difference index is an indicator used to quantify the degree of deviation, distribution consistency or information overlap between the first power dataset and the second power dataset. For example, the data difference index includes, but is not limited to, mean deviation rate, standard deviation ratio, KL divergence and parameter overlap, but is limited to the example situation.

[0154] The threshold for the difference indicator is set based on historical data. Historical data provides an empirical range of data changes. By analyzing historical data, a reasonable threshold can be determined. The setting of the threshold needs to consider the stability, accuracy, and sensitivity to data changes of the campaign performance prediction model. If the difference indicator threshold is too high, the campaign performance prediction model will not be sensitive to data changes, resulting in untimely model optimization and updates. If the difference indicator threshold is too low, the campaign performance prediction model will be frequently optimized and updated, thus affecting the stability of the campaign performance prediction model.

[0155] For example, during fine-tuning training, a small learning rate optimization algorithm is used to update the weight matrix and bias vector. The small learning rate ensures that the weight matrix and bias vector are adjusted within a narrow range, maintaining the stability of the delivery effect prediction model and adapting to minor changes. During retraining, a new second power dataset is constructed from the first power dataset and at least a portion of the second power dataset. If the data change index exceeds a threshold, indicating significant data change, the model is retrained. A new training set is formed using the new data and a portion of the old data, preserving the historical patterns of the old power data. After re-initializing the weight matrix and bias vector, the initialized delivery effect prediction model is trained. This allows the delivery effect prediction model to adapt to power scenarios with significant changes in power ecological parameters, thereby improving the accuracy and adaptability of the delivery effect prediction model.

[0156] S260. Divide the first power dataset into at least two first category subsets, extract features from each first category subset to obtain a first power feature set, and determine the deployment risk data based on the first power feature set and the trained deployment risk prediction model.

[0157] S260 in this embodiment is the same as in the above embodiment. Figure 1 The S130 shown is the same or similar, and will not be described again in this embodiment.

[0158] Based on the above embodiments, optionally, the deployment risk prediction model is optimized and updated in the following manner: In response to the optimization instruction for the deployment risk prediction model, the first power feature set is divided into k first sub-feature sets; based on the k first sub-feature sets, second performance indicators of the deployment risk prediction model are determined respectively, and verification performance indicators are obtained by statistically analyzing the k second performance indicators; when the verification performance indicators meet the model optimization conditions, the hyperparameters of the deployment risk prediction model are adjusted according to the verification performance indicators; based on the k first sub-feature sets, the adjusted deployment risk prediction model is cross-validated and trained to obtain the trained deployment risk prediction model.

[0159] For example, the conditions for generating optimization instructions include risk optimization cycles and structural changes in the power system, but are not limited to the given examples.

[0160] For example, validation performance metrics include, but are not limited to, statistical values ​​corresponding to mean squared error, recall, coefficient of determination, accuracy, and F1 score.

[0161] The advantage of this setup is that it ensures the accuracy and generalization ability of the risk prediction model, reduces the occurrence of missed or misjudged risk events, and thus ensures the reliability of the assessment results of power resource allocation under the ever-changing power scenarios.

[0162] S270. Based on the deployment effect data and the deployment risk data, determine the evaluation result of the power resource deployment.

[0163] S270 in this embodiment is the same as that in the above embodiment. Figure 1 The S140 shown is the same or similar, and will not be described again in this embodiment.

[0164] The technical solution of this embodiment effectively solves the problem of insufficient prediction accuracy of power delivery effect and risk by training the power delivery effect prediction model and the power delivery risk prediction model with comprehensive collection of ecological data of the power system. At the same time, this embodiment realizes the continuous optimization and updating of the power delivery effect prediction model and the power delivery risk prediction model by monitoring changes in ecological data, which improves the generalization ability of power resource assessment tasks to adapt to the ever-changing power ecological environment and ensures the safe and efficient operation of the power system in a sustainable manner.

[0165] The following are embodiments of the power resource allocation assessment device provided in this invention. This device and the power resource allocation assessment method described in the above embodiments belong to the same inventive concept. For details not described in detail in the embodiments of the power resource allocation assessment device, please refer to the content of the power resource allocation assessment method in the above embodiments.

[0166] Figure 3 This is a schematic diagram of the structure of an evaluation device for power resource allocation according to an embodiment of the present invention. Figure 3 As shown, the device includes: a first ecological data acquisition module 310, a deployment effect data determination module 320, a deployment risk data determination module 330, and an evaluation result determination module 340.

[0167] The first ecological data acquisition module 310 is used to collect the first ecological data of the power system in real time. The first ecological data includes system production data and external environment data.

[0168] The deployment effect data determination module 320 is used to preprocess the first ecological data to obtain the first power dataset, and determine the deployment effect data based on the first power dataset and the trained deployment effect prediction model.

[0169] The deployment risk data determination module 330 is used to divide the first power dataset into at least two first category subsets, extract features from each first category subset to obtain a first power feature set, and determine the deployment risk data based on the first power feature set and the trained deployment risk prediction model.

[0170] The evaluation result determination module 340 is used to determine the evaluation result of power resource allocation based on the allocation effect data and the allocation risk data.

[0171] The technical solution of this embodiment eliminates data source barriers and solves the problem of inconsistent data sources between the two models by comprehensively collecting ecological data of the power system. Furthermore, by classifying and dividing the power dataset and conducting in-depth data mining on each subset, the risk identification capability of the deployment risk prediction model is guaranteed. By coupling the deployment effect evaluation model and the deployment risk prediction model, the evaluation task of power resource deployment is completed collaboratively. The two-dimensional evaluation system avoids the misleading influence of a single evaluation dimension on deployment decisions. This embodiment of the invention improves multiple technical implementations, providing accurate and reliable data support for power resource deployment, significantly improving the scientific and rational nature of investment decisions, and ensuring the efficient and safe operation of the power system while increasing the utilization rate of power resources.

[0172] In an optional embodiment, the device further includes:

[0173] The first model optimization module is used to perform sensitivity analysis on the power ecological parameters in the first power dataset according to the deployment effect prediction model to obtain at least one sensitive ecological parameter.

[0174] Obtain at least one first category subset to which the at least one sensitive ecological parameter belongs, and optimize and adjust the feature weights of the power classification features corresponding to the first category subset in the deployment risk prediction model.

[0175] In an optional embodiment, the system production data includes system operation data and equipment status data, wherein the system operation data represents the operating condition information of the power system, and the equipment status data represents the health status information of the physical equipment in the power system;

[0176] And / or, the external environment data includes intra-domain environment data and cross-domain environment data, wherein the intra-domain environment data represents feedback information of environmental factors that are directly coupled with the power system, and the cross-domain environment data represents regulation information of environmental factors that are indirectly related to the power system.

[0177] In an optional embodiment, the intra-domain environmental data includes electricity consumption data, and the cross-domain environmental data includes electricity consumption environment data associated with electricity consumption. Accordingly, the device further includes:

[0178] The electricity demand trend determination module is used to obtain the electricity consumption data and electricity environment data after centralized preprocessing of the first power data set.

[0179] According to a preset time period, the electricity consumption data and the electricity environment data are statistically analyzed to obtain electricity consumption time-series data and electricity environment time-series data, respectively.

[0180] Based on the electricity consumption time-series data and the electricity environment time-series data, the electricity demand trend is determined, and the electricity demand trend is merged into the first power dataset.

[0181] In an optional embodiment, the device further includes:

[0182] The deployment effect prediction model training module is used to collect the second ecological data of the power system during its historical operating cycle;

[0183] The second ecological data is preprocessed to obtain the second power dataset, which is then divided into a training dataset, a test dataset, and a validation dataset.

[0184] The deployment effect prediction model is trained based on the training dataset, test dataset, and validation dataset.

[0185] In an optional embodiment, the device further includes:

[0186] The second model optimization module is used to determine the data difference index between the first power dataset and the second power dataset.

[0187] If the data difference index is less than the difference index threshold, the delivery effect prediction model is fine-tuned and trained based on the first power dataset.

[0188] If the data difference index is greater than or equal to the difference index threshold, the delivery effect prediction model is retrained based on the first power dataset and at least a portion of the second power dataset.

[0189] In an optional embodiment, the device further includes:

[0190] The deployment risk prediction model training module is used to divide the second power dataset into at least two second classification subsets, extract features from each second classification subset to obtain a second power feature set, and divide the second power feature set into k second feature subsets;

[0191] Use k-1 subsets of the second features as the training dataset, train the deployment risk prediction model based on the training dataset, and determine the first performance index of the deployment risk prediction model obtained in the current training process based on the remaining 1 subset of the second features.

[0192] The step of using k-1 subsets of second features as training dataset is executed k times iteratively, and the training performance index is obtained by statistically analyzing k first performance indicators.

[0193] If the training performance index meets the training termination condition, the deployment risk prediction model obtained in the current training process will be used as the deployment risk prediction model after training is completed.

[0194] In this case, the remaining one second feature subset is different in each of the k training processes, where k ≥ 2.

[0195] In an optional embodiment, the device further includes:

[0196] The third model optimization module is used to respond to the optimization instruction for the deployment risk prediction model, divide the first power feature set into k first sub-feature sets, determine the second performance index of the deployment risk prediction model according to the k first sub-feature sets, and statistically obtain the verification performance index by performing statistics on the k second performance indexes.

[0197] If the verification performance indicators meet the model optimization conditions, adjust the hyperparameters of the deployment risk prediction model according to the verification performance indicators.

[0198] Based on the k first sub-feature sets, the adjusted deployment risk prediction model is cross-validated and trained to obtain the completed deployment risk prediction model.

[0199] The power resource allocation assessment device provided in this embodiment of the invention can execute the power resource allocation assessment method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0200] Figure 4 This is a schematic diagram of an electronic device provided according to one embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0201] like Figure 4 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor 11. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0202] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information or data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0203] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as the power resource allocation evaluation method provided in the above embodiments.

[0204] In some embodiments, the power resource allocation assessment method provided in the above embodiments can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the power resource allocation assessment method described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the power resource allocation assessment method by any other suitable means (e.g., by means of firmware).

[0205] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication unit 19, or installed from storage unit 18, or installed from ROM 12. When the computer program is executed by processor 11, it performs the functions defined in the methods of the embodiments of the present invention.

[0206] Various embodiments of the systems and techniques described above herein can be implemented in the following systems or combinations thereof: digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard parts (ASSPs), system-on-chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0207] Computer programs used to implement the power resource allocation assessment method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are performed. The computer programs can be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0208] In the context of this application, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable storage medium. Examples of machine-readable storage media include, based on an electrical connection of at least one wire, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0209] To provide interaction with a user, the systems and techniques described herein can be implemented on a terminal device having: a display device for displaying information to the user (e.g., a cathode-ray tube (CRT) or liquid crystal display (LCD) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the terminal device. Other types of devices can also provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0210] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0211] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system. It addresses the shortcomings of traditional physical hosts and Virtual Private Server (VPS) services, such as high management difficulty and weak business scalability.

[0212] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0213] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for evaluating the allocation of electricity resources, characterized in that, include: Real-time acquisition of primary ecological data of the power system, which includes system production data and external environmental data; The first ecological data is preprocessed to obtain the first power dataset. Based on the first power dataset and the trained deployment effect prediction model, the deployment effect data is determined. The first power dataset is divided into at least two first category subsets. Features are extracted from each first category subset to obtain a first power feature set. Based on the first power feature set and the trained deployment risk prediction model, deployment risk data is determined. The evaluation results of power resource allocation are determined based on the data on allocation effectiveness and the data on allocation risk.

2. The method according to claim 1, characterized in that, The deployment risk prediction model is optimized and updated in the following ways: Based on the aforementioned deployment effect prediction model, sensitivity analysis is performed on the power ecological parameters in the first power dataset to obtain at least one sensitive ecological parameter. Obtain at least one first category subset to which the at least one sensitive ecological parameter belongs, and optimize and adjust the feature weights of the power classification features corresponding to the first category subset in the deployment risk prediction model.

3. The method according to claim 1, characterized in that, The system production data includes system operation data and equipment status data. The system operation data represents the operating condition information of the power system, and the equipment status data represents the health status information of the physical equipment in the power system. And / or, the external environment data includes intra-domain environment data and cross-domain environment data, wherein the intra-domain environment data represents feedback information of environmental factors that are directly coupled with the power system, and the cross-domain environment data represents regulation information of environmental factors that are indirectly related to the power system.

4. The method according to claim 3, characterized in that, The intra-domain environmental data includes electricity consumption data, and the cross-domain environmental data includes electricity consumption environment data associated with electricity consumption. Accordingly, before determining the delivery effect data based on the first electricity dataset and the trained delivery effect prediction model, the method further includes: Obtain the preprocessed electricity consumption data and electricity environment data from the first power dataset; According to a preset time period, the electricity consumption data and the electricity environment data are statistically analyzed to obtain electricity consumption time-series data and electricity environment time-series data, respectively. Based on the electricity consumption time-series data and the electricity environment time-series data, the electricity demand trend is determined, and the electricity demand trend is merged into the first power dataset.

5. The method according to claim 1, characterized in that, The delivery effect prediction model is trained in the following way: Collect the second ecological data of the power system during its historical operating cycle; The second ecological data is preprocessed to obtain the second power dataset, which is then divided into a training dataset, a test dataset, and a validation dataset. The deployment effect prediction model is trained based on the training dataset, test dataset, and validation dataset.

6. The method according to claim 5, characterized in that, The delivery effectiveness prediction model is optimized and updated in the following ways: Determine the data difference index between the first power dataset and the second power dataset; If the data difference index is less than the difference index threshold, the delivery effect prediction model is fine-tuned and trained based on the first power dataset. If the data difference index is greater than or equal to the difference index threshold, the delivery effect prediction model is retrained based on the first power dataset and at least a portion of the second power dataset.

7. The method according to claim 5, characterized in that, The deployment risk prediction model is trained in the following manner; The second power dataset is divided into at least two second category subsets, and features are extracted from each second category subset to obtain a second power feature set. The second power feature set is then divided into k second feature subsets. Use k-1 subsets of the second features as the training dataset, train the deployment risk prediction model based on the training dataset, and determine the first performance index of the deployment risk prediction model obtained in the current training process based on the remaining 1 subset of the second features. The step of using k-1 subsets of second features as training dataset is executed k times iteratively, and the training performance index is obtained by statistically analyzing k first performance indicators. If the training performance index meets the training termination condition, the deployment risk prediction model obtained in the current training process will be used as the deployment risk prediction model after training is completed. In this case, the remaining one second feature subset is different in each of the k training processes, where k ≥ 2.

8. The method according to claim 7, characterized in that, The deployment risk prediction model is optimized and updated in the following ways: In response to the optimization instruction for the deployment risk prediction model, the first power feature set is divided into k first sub-feature sets. Based on the k first sub-feature sets, the second performance index of the deployment risk prediction model is determined respectively, and the verification performance index is obtained by statistically analyzing the k second performance indexes. If the verification performance indicators meet the model optimization conditions, adjust the hyperparameters of the deployment risk prediction model according to the verification performance indicators. Based on the k first sub-feature sets, the adjusted deployment risk prediction model is cross-validated and trained to obtain the completed deployment risk prediction model.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the evaluation method for power resource allocation as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the evaluation method for power resource allocation as described in any one of claims 1-8.

Images