Quality management method and system for new energy power station

Abstract

The invention provides a quality management method and system for a new energy power station, and relates to the field of power station management, and the method comprises the steps: receiving an element model list and an element number list of a target power storage unit of a target new energy power station; counting a power storage influence attribute time sequence matrix, a power storage unit sample and workload time sequence information; according to the workload time sequence information, performing redundancy life evaluation on the target power storage unit to obtain a redundancy life prediction value; and carrying out temporary quality detection identification or default period quality detection identification on the target power storage unit. The technical problems that in the prior art, new energy power station construction quality detection only depends on a regular monitoring mode, key risk moments are possibly omitted, and potential operation hazards exist are solved, redundant life evaluation is carried out according to service data and working load time sequence information of a power storage unit, targeted quality detection identification is achieved, and the reliability of the system is improved. And the operation risk is effectively prevented.

Description

Technical Field

[0001] This invention relates to the field of power plant management, and in particular to a quality management method and system for new energy power plants. Background Technology

[0002] With the rapid development of the new energy industry, ensuring the construction quality of new energy power plants, as important carriers of clean energy, is particularly crucial. Currently, the management of new energy power plants both domestically and internationally mainly adopts a periodic monitoring approach. This involves inspecting and evaluating various components of the power plant at fixed time intervals to achieve quality management. While this periodic monitoring method can identify problems in power plant operation to some extent, its fixed monitoring cycle cannot be flexibly adjusted according to the actual operating status of the power plant, thus presenting significant limitations. Especially for critical equipment such as energy storage units, their operating status is affected by various factors, such as workload and environmental conditions. These factors can change drastically in a short period, and periodic monitoring cannot promptly capture the potential risks arising from these changes. This leads to the potential omission of critical risk moments between periodic inspections, thus creating hidden dangers for the safe operation of the power plant. Summary of the Invention

[0003] This invention addresses the technical problem that existing technologies rely solely on periodic monitoring for quality inspection of new energy power plant construction, which may lead to the omission of critical risk moments and the existence of operational hazards. It provides a quality management method and system for new energy power plants to solve this problem.

[0004] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: In a first aspect, the present invention provides a quality management method for a new energy power plant, comprising: receiving a component model list and a component quantity list of a target energy storage unit in a target new energy power plant; retrieving service data of the target energy storage unit in the target new energy power plant and statistically analyzing a time-series matrix of energy storage impact attributes; statistically analyzing a sample of energy storage units that satisfy the component model list and the component quantity list, and the workload time-series information in the time-series matrix of energy storage impact attributes; performing a redundancy lifetime assessment on the target energy storage unit based on the workload time-series information to obtain a redundancy lifetime prediction value; when the redundancy lifetime prediction value is less than or equal to a redundancy lifetime threshold, marking the target energy storage unit with a temporary quality inspection; and when the redundancy lifetime prediction value is greater than the redundancy lifetime threshold, marking the target energy storage unit with a default periodic quality inspection.

[0005] Secondly, the present invention provides a quality management system for a new energy power plant, comprising: a data receiving module for receiving a list of component models and a list of component quantities of target energy storage units in a target new energy power plant; a service analysis module for retrieving service data of target energy storage units in the target new energy power plant and statistically analyzing a time-series matrix of energy storage impact attributes; a load statistics module for statistically analyzing the working load time-series information of energy storage unit samples that satisfy the list of component models and the list of component quantities in the time-series matrix of energy storage impact attributes; a lifespan assessment module for performing a redundancy lifespan assessment on the target energy storage units based on the working load time-series information to obtain a redundancy lifespan prediction value; a temporary inspection identification module for temporarily identifying the target energy storage units when the predicted redundancy lifespan value is less than or equal to a redundancy lifespan threshold; and a default identification module for performing a default periodic quality inspection identification on the target energy storage units when the predicted redundancy lifespan value is greater than the redundancy lifespan threshold.

[0006] The beneficial effects of this invention are: By receiving the component model list and component quantity list of the target energy storage units in the target new energy power plant, a basis for component configuration is provided for subsequent analysis; service data of the target energy storage units in the target new energy power plant is retrieved, and the time series matrix of energy storage impact attributes is statistically analyzed to provide data support for evaluating changes in energy storage system performance; samples of energy storage units that meet the component model and component quantity lists are statistically analyzed, and samples of energy storage units with the same configuration are selected from the workload time series information in the energy storage impact attribute time series matrix, and their workload time series characteristics are analyzed to establish a comparable reference benchmark; based on the workload time series information, the redundancy lifetime of the target energy storage units is assessed to obtain the redundancy lifetime prediction value, and the remaining service life of the energy storage units is predicted through workload data analysis to achieve predictive assessment; when the redundancy lifetime prediction value is less than or equal to the redundancy lifetime threshold, the target energy storage unit is marked with a temporary quality inspection indicator to indicate the need for temporary quality inspection and prevent potential failures; when the redundancy lifetime prediction value is greater than the redundancy lifetime threshold, the target energy storage unit is marked with a default periodic quality inspection indicator to ensure that routine maintenance is carried out.

[0007] The above technical solution enables intelligent quality inspection planning for energy storage units in new energy power plants. Different inspection strategies are adopted according to different predicted lifespans to achieve targeted quality inspection and identification, thereby improving inspection efficiency and resource utilization and effectively preventing operational risks of energy storage units. Attached Figure Description

[0008] Figure 1 A flowchart illustrating a quality management method for a new energy power plant provided by the present invention; Figure 2 This invention provides a structural schematic diagram of a quality management system for a new energy power plant.

[0009] In the attached diagram, the components represented by each number are as follows: Data receiving module 11, service analysis module 12, load statistics module 13, life assessment module 14, inspection identification module 15, and default identification module 16. Detailed Implementation

[0010] 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.

[0011] In the description of this invention, the terms "first" and "second" 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" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0012] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0013] Example 1, as Figure 1 As shown, this embodiment of the invention provides a quality management method for new energy power plants, including: S100: Receive the component model list and component quantity list of the target energy storage unit of the target new energy power station.

[0014] Specifically, target new energy power plants refer to power plants that use renewable energy sources such as wind, solar, and tidal energy as energy sources and are equipped with energy storage facilities. Unlike traditional thermal power plants, new energy power plants are environmentally friendly and sustainable, but they also face technical challenges such as unstable energy input and grid peak-shaving requirements. Therefore, energy storage systems play a crucial role. Target energy storage units refer to the functional units in new energy power plants responsible for storing and releasing electrical energy. They can be composed of various energy storage components, such as lithium-ion battery packs, sodium-sulfur battery packs, flow battery packs, and supercapacitors. The quality and performance of energy storage units directly affect the operational stability and power distribution capabilities of the entire new energy power plant, and are an important object of quality inspection in power plant construction.

[0015] During the quality inspection of energy storage units, the system first receives a list of component models and a list of component quantities for the target energy storage units in the target new energy power plant. The component model list records the model information of all components used in the target energy storage unit, including but not limited to battery models, capacitor models, converter models, cooling device models, and control module models. The component quantity list records the quantity information of each component model, reflecting the scale and configuration structure of the energy storage unit. In practical applications, this information can be received in various ways, such as: directly from the construction or operation party of the new energy power plant via a dedicated data interface; extracted from the database of the power plant management system; obtained by scanning and identifying the equipment tags of the energy storage units; or manually entered.

[0016] Combinations of different models and quantities of components result in energy storage units of varying sizes and performance characteristics. These differences directly impact the workload capacity, operational stability, and lifespan of the energy storage units. Therefore, lists of component models and quantities are crucial references for subsequent redundancy life assessments, laying the foundation for accurate quality testing.

[0017] S200: Retrieve the service data of the target energy storage unit of the target new energy power station and compile the time series matrix of energy storage impact attributes.

[0018] Specifically, while receiving the component model list and component quantity list, the system retrieves the service data of the target energy storage units in the target new energy power plant and compiles a time-series matrix of energy storage impact attributes. Service data refers to the records of various operating parameters generated by the energy storage units during actual operation, including but not limited to real-time monitoring data such as charge / discharge cycles, depth of charge / discharge, operating temperature, voltage and current changes, ambient humidity, and power output. This data comprehensively reflects the performance of the energy storage units under different operating states and environmental conditions, serving as the foundational material for quality assessment.

[0019] The energy storage impact attribute time series matrix is ​​a multi-dimensional data structure formed after systematically processing service data. Each column represents an attribute that has a significant impact on energy storage performance (such as temperature, humidity, depth of charge / discharge, etc.), and each row represents a record at a specific point in time. By constructing this time series matrix, the trends of various influencing factors over time and their interrelationships can be clearly shown, providing a data foundation for subsequent energy storage unit life assessment.

[0020] The process of compiling a time-series matrix of energy storage impact attributes involves various data processing techniques, including data cleaning, feature extraction, and time-series analysis. The aim is to extract the most representative and predictive information from massive amounts of raw service data. By using time-series data-based analysis methods, the limitations of traditional periodic inspections are overcome, enabling the capture of dynamic changes in energy storage units throughout their entire lifecycle, thus helping to identify potential quality risks.

[0021] S300: Statistically analyze the energy storage unit samples that satisfy the component model list and the component quantity list, and the workload time series information of the energy storage influence attribute time series matrix.

[0022] Specifically, after obtaining the component model and quantity list of the target energy storage units in the target new energy power plant and constructing the time series matrix of energy storage impact attributes, the workload time series information of energy storage unit samples that meet specific configuration conditions is further statistically analyzed to provide basic information for the redundancy life assessment of the target energy storage units. By analyzing energy storage unit samples with similar configurations, the workload variation patterns in actual operation are obtained, providing a reference basis for the life assessment of the target energy storage units.

[0023] First, using the component model list and component quantity list of the target energy storage unit as screening criteria, samples with the same or highly similar component configurations are selected from historical databases or online monitoring systems. This component-based screening method ensures the comparability of the selected samples with the target energy storage unit in terms of structural and performance characteristics, improving the relevance and accuracy of subsequent analysis results. Next, combined with the obtained energy storage impact attribute time-series matrix, statistical analysis of the working load time-series information of the selected energy storage unit samples is performed. Working load time-series information refers to the actual load status of the energy storage unit at different time points, reflecting the unit's performance under various environmental conditions and usage scenarios. By statistically analyzing the working load time series of multiple similar configuration samples, typical load patterns and trends of specific types of energy storage units can be identified, providing more reliable data support for subsequent redundancy life assessment.

[0024] By using statistical analysis based on actual operating data, it breaks through the limitations of traditional quality testing that relies solely on theoretical models or design parameters. It can more comprehensively reflect the performance status of energy storage units in real working environments, effectively improving the accuracy and predictability of quality testing.

[0025] S400: Based on the workload timing information, perform a redundancy lifetime assessment on the target energy storage unit to obtain a redundancy lifetime prediction value.

[0026] Specifically, after acquiring the workload timing information, a redundancy lifetime assessment is performed on the target energy storage unit, and a redundancy lifetime prediction value is obtained. Redundancy lifetime refers to the remaining service life of the energy storage unit under normal operating performance conditions, and is an indicator for assessing the quality status of the energy storage unit.

[0027] Based on the obtained workload time-series information, combined with the design parameters and theoretical models of the energy storage unit, a specific lifetime prediction algorithm is used to model and analyze the future performance degradation trend of the target energy storage unit. This prediction process not only considers the current state of the energy storage unit but also fully integrates historical load variation patterns and environmental influencing factors, achieving dynamic assessment of redundancy lifetime and obtaining a redundancy lifetime prediction value. The redundancy lifetime prediction value is a quantitative indicator, usually expressed in time units (such as days, months, years) or cycle counts, reflecting the expected duration for which the energy storage unit can continue to operate normally under the current usage mode and environmental conditions.

[0028] By introducing a redundancy life assessment mechanism, the limitations of traditional periodic inspection methods in identifying potential risks have been overcome. This enables a forward-looking assessment of the quality status of energy storage units, providing a basis for timely detection and handling of potential quality issues, and effectively improving the operational safety and stability of new energy power plants.

[0029] S500: When the predicted redundancy lifetime value is less than or equal to the redundancy lifetime threshold, a temporary quality inspection mark is made on the target energy storage unit.

[0030] Specifically, after obtaining the predicted redundancy lifetime value, quality inspection decisions are made based on this value. When the predicted redundancy lifetime value is less than or equal to a preset redundancy lifetime threshold, a temporary quality inspection flag will be issued for the target energy storage unit. The redundancy lifetime threshold is a pre-set critical value based on factors such as the type of energy storage unit, operating environment, and safety requirements, representing a lifespan node where the energy storage unit requires special attention.

[0031] Temporary quality inspection identification is a high-priority inspection scheduling mechanism implemented for energy storage units with low predicted redundancy lifetimes and potential quality risks. Unlike traditional fixed-period inspections, this temporary inspection is dynamically triggered based on data analysis results, enabling timely responses to potential risks arising from changes in energy storage unit performance. Temporary quality inspection identification typically includes information such as inspection time, inspection items, and priority, and is recorded in the power plant management system and distributed to relevant maintenance personnel. This prediction-based inspection identification mechanism represents a shift from periodic inspections to predictive inspections, significantly improving the targeting and efficiency of quality inspections.

[0032] By using temporary quality inspection labels, potential risks can be identified and addressed before serious quality problems occur in energy storage units. This effectively avoids safety hazards and economic losses caused by untimely inspections, providing a more reliable guarantee for the safe operation of new energy power plants. This proactive prevention inspection strategy can effectively improve the overall operational stability of new energy power plants and extend the service life of energy storage equipment.

[0033] S600: When the predicted redundancy lifetime value is greater than the redundancy lifetime threshold, the target energy storage unit is marked with a default periodic quality test.

[0034] Specifically, when the predicted redundancy lifetime value is greater than the preset redundancy lifetime threshold, the target energy storage unit will be marked with a default periodic quality inspection, using conventional inspection frequency and inspection scheme.

[0035] The default periodic quality inspection marker is a routine inspection schedule set for energy storage units that are currently in good condition and have sufficient expected service life. Unlike ad hoc quality inspections, default periodic quality inspections follow a pre-defined inspection plan and are conducted at fixed time intervals, such as quarterly, semi-annual, or annual inspections. The specific cycle can be flexibly set according to the type of energy storage unit and operating environment. The default periodic quality inspection marker also includes information such as inspection time and inspection items, but its priority is usually lower than that of ad hoc quality inspections. This differentiated inspection strategy achieves a reasonable allocation of inspection resources, ensuring that potentially risky energy storage units receive timely attention while avoiding over-inspection of energy storage units in good condition, thus improving overall inspection efficiency.

[0036] By comparing the predicted redundancy lifetime with the redundancy lifetime threshold, intelligent allocation of testing resources is achieved, overcoming the limitations of traditional periodic testing methods. This data-driven adaptive testing mechanism not only improves the accuracy and efficiency of quality testing but also reduces testing costs, enabling targeted quality testing and labeling of energy storage units and effectively preventing operational risks.

[0037] Furthermore, the service data of the target energy storage units of the target new energy power station are retrieved, and the time series matrix of energy storage impact attributes is statistically analyzed, including: S210: Based on the energy storage type of the target energy storage unit, perform frequent mining to obtain the first energy storage influence attribute up to the Nth energy storage influence attribute; S220: From the service data, extract the initial time series information of the first energy storage impact attribute and perform neighborhood hierarchical clustering analysis to obtain the time series information of the first energy storage impact attribute; S230: Until the service data is used to extract the initial time series information of the Nth energy storage impact attribute and perform neighborhood hierarchical clustering analysis, the time series information of the Nth energy storage impact attribute is obtained. S240: Construct the energy storage impact attribute time series matrix based on the first energy storage impact attribute time series information up to the Nth energy storage impact attribute time series information.

[0038] In one feasible implementation, when constructing the time-series matrix of energy storage impact attributes, firstly, based on the target energy storage type of the target energy storage unit (such as lithium-ion batteries, sodium-sulfur batteries, supercapacitors, etc.), targeted frequency mining is performed. Frequency mining is a data mining technique that identifies key attributes that have a significant impact on a specific energy storage type by statistically analyzing the correlation frequency between various attributes and energy storage types. In this way, the most relevant set of attributes can be screened from numerous potential influencing factors, including the first energy storage impact attribute up to the Nth energy storage impact attribute. These screened impact attributes will become the core dimensions of subsequent analysis, ensuring the relevance and efficiency of data processing. After determining the impact attributes, initial time-series information related to the first energy storage impact attribute is extracted from the service data. This information reflects the original values ​​of the attribute at different time points. Subsequently, a neighborhood hierarchical clustering analysis method is applied to this initial time-series information. Based on the temporal continuity and numerical similarity between data points, data points with similar change patterns are grouped into the same category, thereby reducing noise interference and extracting more representative time-series patterns. After this processing, the time-series information of the first energy storage influence attribute is obtained, which can more accurately reflect the pattern of the attribute's change over time.

[0039] Following the processing flow of S220, the same operations are performed sequentially on the second, third, and so on up to the Nth energy storage impact attribute. For each impact attribute, its initial time-series information is extracted from the service data, and then optimized through neighborhood hierarchical clustering analysis to finally obtain the time-series information of that attribute. This sequential iterative processing mechanism ensures that all identified impact attributes can be systematically analyzed, forming a complete set of time-series information. Subsequently, all time-series information from the first energy storage impact attribute to the first energy storage impact attribute time-series information is integrated into a structured time-series matrix, forming the energy storage impact attribute time-series matrix. In this matrix, each column corresponds to a type of energy storage impact attribute, each row corresponds to a time point, and the matrix elements represent the value of a specific impact attribute at a specific time point. This matrix-based data structure allows for a unified expression of the time-series changes of each impact attribute and their interrelationships, facilitating subsequent comprehensive analysis and providing necessary data support for subsequent workload statistics and redundancy life assessment.

[0040] Furthermore, based on the energy storage type of the target energy storage unit, frequent mining is performed to obtain the first energy storage impact attribute up to the Nth energy storage impact attribute, including: S211: Obtain the set of influencing attributes to be filtered; S212: Traverse the set of influencing attributes to be screened and count the frequency set that co-occurs with the energy storage type; S213: Based on the frequency set, filter the set of first-level influence attributes that are greater than or equal to the frequency threshold from the set of influence attributes to be filtered; S214: Traverse the set of first-level influence attributes to perform support evaluation and obtain the support set; S215: Based on the support set, filter the first energy storage influence attributes that are greater than or equal to the support threshold from the first-level influence attribute set up to the Nth energy storage influence attribute.

[0041] In a preferred embodiment, during frequency mining, the first step is to obtain a set of attributes that may affect energy storage performance, i.e., the set of attributes to be screened. This set of attributes covers various factors that may affect the performance of the energy storage unit, such as ambient temperature, humidity, charging and discharging current, voltage fluctuations, cycle count, and workload. These attributes can be obtained from various sources, including professional literature, equipment manuals, historical operating records, and expert experience. The obtained set of attributes to be screened forms the basis for subsequent precise screening, ensuring that no potential key influencing factors are overlooked during the analysis. After obtaining the set of attributes to be screened, each influencing attribute in the set is iterated over, and the frequency of its co-occurrence with a specific energy storage type is statistically analyzed. Specifically, a large number of operating records involving this energy storage type are extracted from the historical database, and the frequency of each influencing attribute appearing in these records and having a significant impact on energy storage performance is analyzed. This statistical analysis, based on the principle of association rule mining, can quantitatively describe the correlation strength between each attribute and the energy storage type, forming a frequency set. Each element in this set corresponds to an influencing attribute and its frequency of occurrence, providing a numerical basis for subsequent screening.

[0042] Next, based on the obtained frequency set, a frequency threshold is introduced as a screening criterion. Attributes with frequencies greater than or equal to the frequency threshold are selected from the set of influencing attributes to be screened, forming a primary influencing attribute set. The frequency threshold is a preset critical value that can be flexibly adjusted according to actual application scenarios and accuracy requirements. This frequency-based initial screening mechanism effectively filters out attributes with low correlation to energy storage types, retaining the most relevant influencing factors. Compared to the original set of influencing attributes to be screened, the primary influencing attribute set is significantly smaller in size, but its information value is more concentrated, laying the foundation for subsequent fine-tuning. Subsequently, support evaluation is performed on each attribute in the primary influencing attribute set. Support is a more precise evaluation indicator than frequency, considering not only the frequency of attribute occurrence but also the actual impact of attribute changes on energy storage performance. By analyzing historical data, the support value of each primary influencing attribute can be calculated, forming a support set. This support-based evaluation mechanism can more comprehensively reflect the actual importance of influencing attributes, avoiding the one-sidedness that may result from relying solely on frequency. Subsequently, based on the obtained support set, a support threshold was introduced as a second-round screening criterion. Attributes with support values ​​greater than or equal to the support threshold were further selected from the primary influencing attribute set, ultimately determining the influencing attributes from the first energy storage attribute to the Nth energy storage attribute. These ultimately selected influencing attributes not only co-occur frequently with specific energy storage types but also have a significant practical impact on energy storage performance, forming the core dimension for subsequently establishing the time-series matrix of energy storage influencing attributes.

[0043] Through the above two-level screening mechanism, the set of most representative and predictive impact attributes can be effectively identified, laying the foundation for accurately assessing the redundancy life of energy storage units.

[0044] Furthermore, support is evaluated by traversing the set of first-level influence attributes to obtain a support set, including: S2141: Collect several historical energy storage data of the energy storage type, wherein any one historical energy storage data includes a set of primary influence attribute record values ​​and energy storage record values; S2142: Calculate the same attribute deviation modulus of the aforementioned energy storage historical data to obtain multiple sets of first-level influence attribute record value deviations and multiple energy storage record value deviations; S2143: Obtain the set of first-level influencing attribute deviation thresholds; S2144: From the set of deviations of the multiple first-level influence attribute record values, extract the deviations of multiple first-level influence attribute record values ​​where only the deviation of the first influence attribute is greater than the first influence attribute deviation threshold. S2145: Based on the deviation of the recorded values ​​of the multiple first influencing attributes, extract the energy storage record value deviation associated with the multiple first influencing attributes from the deviation of the multiple energy storage record values; S2146: Calculate the ratio of the deviation of the energy storage record value associated with the plurality of first influence attributes to the deviation of the record value of the plurality of first influence attributes, then perform mode statistics to obtain the support of the first influence attributes, and add it to the support set.

[0045] In a preferred embodiment, when obtaining the support set, firstly, several historical energy storage data related to the target energy storage type are collected from a historical database or online monitoring system. These historical energy storage data have clear structural characteristics; each data point contains two key pieces of information: a set of primary influence attribute records and a stored energy record value. The primary influence attribute record set records the specific values ​​of the selected primary influence attributes at a given moment, such as an ambient temperature of 25°C, humidity of 60%, and charging current of 5A. The stored energy record value represents the actual energy storage state of the energy storage unit at the corresponding moment, typically expressed as a capacity percentage or in units of actual stored electricity. This paired data structure, containing cause (influence attributes) and result (stored energy), provides the necessary conditions for subsequent analysis of the correlation between each attribute and energy storage performance. After obtaining several historical energy storage data points, the original data are converted into a more comparable deviation form through the same attribute deviation modulus calculation. Specifically, first, baseline values ​​(such as average or nominal values) for each attribute and energy storage are determined. Then, the deviations of each attribute value and energy storage value in each historical energy storage data point from the corresponding baseline values ​​are calculated, and their absolute values ​​(modulo values) are taken. After this processing, the original historical energy storage data are transformed into multiple sets of deviations for primary influencing attribute records and multiple deviations for energy storage records. This deviation-based analysis method can more intuitively reflect the correspondence between changes in each influencing attribute and changes in energy storage, eliminating the interference of differences in dimensions and numerical ranges between different attributes, and improving the accuracy and comparability of subsequent analyses.

[0046] Then, a corresponding deviation threshold is set for each primary impact attribute, forming a set of deviation thresholds for primary impact attributes. These deviation thresholds serve as reference standards for judging whether changes in each attribute are significant; only when an attribute change exceeds a specific threshold is the change considered meaningful. The setting of deviation thresholds can be based on professional knowledge, equipment specifications, or statistical analysis results, with different threshold standards applied to different attributes. For example, ±5℃ might be set as the deviation threshold for temperature, while ±10% might be set for humidity. By introducing these specialized threshold standards, attribute changes with small fluctuations and limited impact on energy storage performance can be effectively filtered out, focusing on analyzing key changes that truly have a significant impact. Subsequently, data screening is performed on the primary impact attributes. From the set of deviation values ​​for multiple primary impact attribute records, records where only the deviation of the primary impact attribute exceeds the corresponding threshold, while the deviations of other attributes do not exceed their respective thresholds, are extracted. This screening condition ensures that the primary impact attribute is the only significant change factor in the selected records, eliminating the possibility of interference from other attributes and creating a clean data environment for subsequent analysis of the independent support of the primary impact attribute. By controlling other variables to remain relatively stable, the influence of a single variable is highlighted, thereby achieving an accurate assessment of the support of each attribute.

[0047] Based on the obtained deviations of multiple primary impact attribute recorded values, the energy storage deviations corresponding to these primary impact attribute recorded value deviations are further extracted from the multiple energy storage recorded value deviations, termed primary impact attribute-associated energy storage recorded value deviations. These paired deviation data establish a direct correspondence between changes in primary impact attributes and changes in energy storage, providing fundamental data for calculating support. This association extraction process ensures the accuracy of causal correspondence, avoids interference from irrelevant data, and improves the reliability of subsequent support calculations. Subsequently, the ratio of each pair of primary impact attribute recorded value deviations to its corresponding primary impact attribute-associated energy storage recorded value deviation is calculated. These ratios reflect the degree of energy storage change caused by a unit attribute change, providing a direct quantitative expression of the attribute's influence strength. Afterward, the mode of all calculated ratios is statistically analyzed, identifying the most frequently occurring value among these ratios and using it as the support of the primary impact attribute, adding it to the support set. Using the mode rather than the mean as the statistical method for support effectively reduces the interference of outlier data and improves the stability and representativeness of the support assessment. This support calculation method based on ratios and the mode achieves accurate quantification of the importance of impact attributes, providing a basis for subsequent final selection.

[0048] Through the above steps, a systematic support assessment can be conducted on the primary influence attribute and even all primary influence attributes, forming a complete support set, which lays the foundation for the final determination of key influence attributes.

[0049] Furthermore, statistical analysis is performed on the energy storage unit samples that satisfy the component model list and the component quantity list, including the workload time-series information of the energy storage impact attribute time-series matrix, including: S310: Using the component model list and the component quantity list as constraints on the unit size, select the energy storage unit sample set; S320: Using the energy storage capacity as the workload indicator and the energy storage impact attribute time series matrix as the environmental constraint, collect several workload time series records of the energy storage unit sample set; S330: Perform simultaneous load mode evaluation on the several workload time-series record data to obtain the workload time-series information.

[0050] In one feasible implementation, firstly, a sample set of energy storage units is selected using a list of component models and a list of component quantities as constraints on unit size. This allows for the filtering of reference samples with similar configurations to the target energy storage unit based on component composition information. Specifically, using the obtained list of component models and a list of component quantities as constraints on unit size, energy storage units with highly similar component compositions are selected from historical databases or online monitoring systems to form the energy storage unit sample set. This component composition-based filtering mechanism ensures the consistency of the selected samples with the target energy storage unit in terms of physical structure and basic performance, providing a reliable reference standard for subsequent load analysis. In practice, a certain degree of matching tolerance can be allowed; for example, the number of components can fluctuate within ±5%, or functionally equivalent alternative components can be accepted to increase the number of available samples and improve the stability of statistical results. This sample filtering mechanism ensures the relevance and reference value of subsequent workload analysis.

[0051] After determining the sample set of energy storage units, the workload data of these samples during actual operation was further collected. First, the energy storage capacity was selected as the core indicator of workload, as this indicator directly reflects the actual energy storage status and usage intensity of the energy storage units. Simultaneously, a constructed time-series matrix of energy storage influence attributes was introduced as an environmental constraint. This means that when collecting workload data, the actual values ​​of various influence attributes were considered to ensure that the collected load data has clear environmental background information. Based on these settings, several workload time-series records were extracted from the operating records of each sample energy storage unit. Each workload time-series record contains a complete sequence of energy storage capacity changes over time and the corresponding environmental attribute information. This data collection method, combined with environmental constraints, ensures the comprehensiveness and environmental adaptability of the workload analysis, providing more accurate data support for subsequent load statistics and lifetime assessment.

[0052] Subsequently, representative workload time-series information was extracted from the workload time-series records. Specifically, the collected workload time-series records were first aligned by time points, and then the load value at each time point was evaluated using the mode, i.e., identifying the load level that occurred most frequently at that moment. This statistical method based on the mode rather than the mean effectively filters out the interference of outlier data and extracts the most representative workload patterns. After this time-series mode evaluation, standardized workload time-series information was obtained, which reflects the workload variation patterns of a specific configuration of energy storage units under typical environmental conditions, providing a core reference for subsequent redundancy lifetime assessment. Compared with single-sample analysis, this multi-sample statistical method has higher stability and representativeness, effectively reducing the impact of random factors and improving the accuracy of redundancy lifetime prediction.

[0053] Furthermore, based on the workload timing information, a redundancy lifetime assessment is performed on the target energy storage unit to obtain a predicted redundancy lifetime value, including: S410: The workload timing information includes the first time-domain workload up to the Mth time-domain workload; S420: Based on the first time domain workload and rated life, perform a redundancy life assessment on the target energy storage unit to obtain the first time domain redundancy life prediction value. S430: Based on the second time-domain workload and the first time-domain redundancy lifetime prediction value, perform a redundancy lifetime assessment on the target energy storage unit to obtain the second time-domain redundancy lifetime prediction value. S440: Until the target energy storage unit is evaluated for redundancy lifetime based on the Mth time domain workload and the M-1th time domain redundancy lifetime prediction value, the redundancy lifetime prediction value is obtained.

[0054] In one feasible implementation, the workload time-series information includes workload in the first time domain up to workload in the Mth time domain. For example, the obtained workload time-series information is divided into time domains, decomposing continuous time-series data into multiple discrete time periods, from workload in the first time domain up to workload in the Mth time domain. The time domain division can be based on different division criteria according to the operating characteristics of the energy storage unit. For example, it can be divided according to natural time units (such as days, weeks, months), or according to the working cycles of the energy storage unit (such as the number of charge and discharge cycles), or according to the change nodes of energy storage impact attributes. The workload data in each time domain reflects the actual working status and usage intensity of the energy storage unit in that specific time period. This time-domain processing method enables the system to accurately capture the dynamic characteristics of workload changes over time, laying the foundation for achieving high-precision, progressive life assessment.

[0055] Then, using the rated lifespan of the energy storage unit as an initial reference benchmark, and combining it with the first time-domain workload, a specific lifespan assessment model is applied to calculate the predicted redundant lifespan of the energy storage unit at the end of the first time domain. The rated lifespan, typically provided by the manufacturer, represents the expected service life of the energy storage unit under standard operating conditions and usage modes. However, actual operating environments and load patterns often differ significantly from standard test conditions, leading to variations in the actual lifespan consumption rate of the energy storage unit. By comparing the first time-domain workload with the rated lifespan, the actual lifespan consumption of the energy storage unit in the first time domain stage can be assessed, and the remaining lifespan at the end of this stage can be predicted. This assessment method based on actual load overcomes the limitations of traditional time-based calculations and more realistically reflects the performance degradation process of the energy storage unit.

[0056] Subsequently, the progressive stage of redundancy lifetime assessment begins. Unlike step S420, the initial rated lifetime is no longer used as the benchmark. Instead, the first time-domain redundancy lifetime prediction value calculated in step S420 is used as the new starting point. Combined with the workload data of the second time domain, the redundancy lifetime prediction value at the end of the second time domain is calculated, i.e., the second time-domain redundancy lifetime prediction value. This progressive assessment mechanism fully considers the cumulative effect and nonlinear characteristics of lifetime consumption, and can accurately reflect the dynamic evolution of energy storage unit performance with changes in usage time and load. At the same time, since each assessment is based on the latest redundancy lifetime status and the actual load of the current time domain, it can sensitively capture inflection points and abnormal fluctuations in energy storage unit performance, providing a basis for early warning of quality risks. Following the same assessment principle, the workload data from the third time domain to the final Mth time domain are processed sequentially. For each time domain, the redundancy lifetime prediction value of the previous time domain is used as the assessment starting point, and combined with the workload situation of the current time domain, the latest redundancy lifetime prediction value at the end of that time domain is calculated. Through this chain-iterative evaluation method, the redundancy lifetime prediction value at the end of the Mth time domain is finally obtained, which is the current redundancy lifetime status of the target energy storage unit.

[0057] Through multi-time-domain progressive evaluation, the changes in the working status and performance evolution trajectory of the energy storage unit throughout its service life are fully considered, enabling high-precision dynamic prediction of redundancy life. This provides a basis for subsequent quality inspection decisions and effectively improves the foresight and accuracy of quality management of new energy power plants.

[0058] Furthermore, based on the first time-domain workload and rated lifespan, a redundancy lifespan assessment is performed on the target energy storage unit to obtain a first time-domain redundancy lifespan prediction value, including: S421: Using the component model list and the component quantity list as constraints, collect multiple sets of data, wherein any set of the multiple sets of data includes workload record data, initial life set value record data and a label identifying redundant life. S422: Configure a threshold for the number of decision trees, using the number of component models as a constraint; S423: Based on the threshold number of decision trees, perform random forest training in combination with the multiple sets of data to obtain a redundancy lifetime estimator, and bind the redundancy lifetime estimator to the target energy storage unit.

[0059] In a preferred embodiment, firstly, multiple sets of data for storage and energy storage units with similar configurations are collected from a historical database, using a list of component models and a list of component quantities as filtering criteria. This component-based filtering mechanism ensures the relevance and reference value of the collected data, providing a high-quality sample foundation for subsequent model training. The collected data sets have a clear three-element structure: workload records reflect the load state changes of the storage and energy storage units during actual operation; initial life setpoint records record the original design life parameters of each storage and energy storage unit; and labels identifying redundancy life record the remaining life value actually measured or empirically estimated under specific load conditions. This complete data structure, including input features (workload and initial life) and output target (redundancy life), provides the necessary training samples for building a predictive model.

[0060] Then, using the number of component types in the target energy storage unit as a reference, the threshold for the number of decision trees in the random forest model is configured. The number of decision trees is one of the core parameters of the random forest algorithm, directly affecting the model's complexity, generalization ability, and computational efficiency. Generally, the more component types and the more complex the structure of the energy storage unit, the more diverse the performance influencing factors become, thus requiring more decision trees to capture these complex relationships. By dynamically configuring decision tree parameters based on the number of component types, an adaptive match between model complexity and the structural complexity of the energy storage unit is achieved. This avoids overfitting of models in simple systems while ensuring the expressive power of models in complex systems, improving the accuracy and applicability of redundancy lifetime assessment.

[0061] Subsequently, based on the configured threshold for the number of decision trees, the random forest algorithm was applied to train the model using multiple sets of collected data. Random forest is an ensemble learning method that effectively handles high-dimensional features, nonlinear relationships, and noisy data by constructing multiple decision trees and averaging their predictions, making it highly suitable for performance prediction of complex systems such as energy storage units. During training, workload records and initial lifetime setpoint records were used as input features, and labels identifying redundant lifetimes were used as output targets. Through iterative optimization, a redundant lifetime estimator with high prediction accuracy was finally obtained. Instead of using a general model, the trained redundant lifetime estimator was bound to a specific target energy storage unit. This targeted model binding mechanism considers the individual differences of different energy storage units, achieving a high degree of customization in redundant lifetime assessment. This significantly improves the accuracy and reliability of the prediction results, providing a solid foundation for subsequent quality inspection decisions.

[0062] After obtaining the redundancy lifetime estimator, it is used to implement the progressive redundancy lifetime prediction in steps S410 to S440. By binding the redundancy lifetime estimator to the target energy storage unit, the first time-domain workload and rated lifetime data can be input to obtain the first time-domain redundancy lifetime prediction value; the second time-domain workload and the redundancy lifetime prediction value of the previous time domain can be input to obtain the second time-domain redundancy lifetime prediction value; and so on, until the final redundancy lifetime prediction value is obtained. This machine learning-based progressive evaluation mechanism overcomes the limitations of the traditional linear decay model and can accurately capture the nonlinear performance changes of the energy storage unit throughout its entire life cycle, providing accurate and reliable data support for subsequent quality inspection decisions.

[0063] Example 2, as Figure 2 As shown, based on the same inventive concept as the quality management method for a new energy power plant provided in Embodiment 1, this embodiment of the invention also provides a quality management system for a new energy power plant, including: Data receiving module 11 is used to receive the component model list and component quantity list of the target energy storage unit of the target new energy power station; Service analysis module 12 is used to retrieve service data of target energy storage units in target new energy power plants and to statistically analyze the time series matrix of energy storage impact attributes. Load statistics module 13 is used to count the working load time sequence information of the energy storage unit sample that meets the component model list and the component quantity list in the energy storage influence attribute time sequence matrix; The life assessment module 14 is used to perform a redundancy life assessment on the target energy storage unit based on the workload timing information, and obtain a redundancy life prediction value. The temporary inspection identification module 15 is used to temporarily identify the target energy storage unit when the predicted redundancy lifetime value is less than or equal to the redundancy lifetime threshold. The default identification module 16 is used to perform a default periodic quality detection identification on the target energy storage unit when the predicted redundancy lifetime value is greater than the redundancy lifetime threshold.

[0064] Furthermore, the service analysis module 12 includes the following execution steps: Based on the energy storage type of the target energy storage unit, perform frequent mining to obtain the first energy storage impact attribute up to the Nth energy storage impact attribute; From the service data, the initial time-series information of the first energy storage impact attribute is extracted and subjected to neighborhood hierarchical clustering analysis to obtain the time-series information of the first energy storage impact attribute. Until the service data is used to extract the initial time series information of the Nth energy storage impact attribute and perform neighborhood hierarchical clustering analysis, the time series information of the Nth energy storage impact attribute is obtained. The energy storage impact attribute time series matrix is ​​constructed based on the first energy storage impact attribute time series information up to the Nth energy storage impact attribute time series information.

[0065] Furthermore, the service analysis module 12 also includes the following execution steps: Obtain the set of influencing attributes to be filtered; Iterate through the set of influencing attributes to be screened and count the frequency set of attributes that co-occur with the energy storage type. Based on the frequency set, a set of first-level influence attributes that are greater than or equal to the frequency threshold is selected from the set of influence attributes to be screened; The support set is obtained by traversing the set of first-level influence attributes. Based on the support set, the first energy storage impact attribute that is greater than or equal to the support threshold is selected from the first-level impact attribute set up up to the Nth energy storage impact attribute.

[0066] Furthermore, the service analysis module 12 also includes the following execution steps: Collect several historical energy storage data of the aforementioned energy storage type, wherein any one piece of historical energy storage data includes a set of primary influence attribute record values ​​and an energy storage record value; The same attribute deviation modulus is calculated for the aforementioned historical energy storage data to obtain multiple sets of deviations for primary influence attribute records and multiple deviations for energy storage records. Obtain the set of first-level influence attribute deviation thresholds; From the set of deviations of the multiple first-level influence attribute records, extract the deviations of multiple first-level influence attribute records where only the deviation of the first-level influence attribute is greater than the first-level influence attribute deviation threshold. Based on the deviations of the recorded values ​​of the multiple first influencing attributes, the deviations of the recorded values ​​of the multiple energy storage attributes are extracted from the deviations of the recorded values ​​of the multiple first influencing attributes; Calculate the ratio of the deviation of the energy storage record value associated with the multiple first influence attributes to the deviation of the record value of the multiple first influence attributes, then perform mode statistics to obtain the support of the first influence attributes, and add it to the support set.

[0067] Furthermore, the load statistics module 13 includes the following execution steps: Using the component model list and the component quantity list as constraints on the unit size, a sample set of energy storage units is selected; Using the energy storage capacity as the workload indicator and the time series matrix of the energy storage impact attributes as the environmental constraint, several workload time series records of the energy storage unit sample set are collected. Simultaneous load mode evaluation is performed on the aforementioned workload time-series records to obtain the workload time-series information.

[0068] Furthermore, the life assessment module 14 includes the following execution steps: The workload timing information includes the first time-domain workload up to the Mth time-domain workload; Based on the first time-domain workload and rated life, the redundancy life of the target energy storage unit is evaluated to obtain the first time-domain redundancy life prediction value. Based on the second time-domain workload and the first time-domain redundancy lifetime prediction value, the redundancy lifetime of the target energy storage unit is evaluated to obtain the second time-domain redundancy lifetime prediction value. The redundancy lifetime of the target energy storage unit is evaluated based on the Mth time-domain workload and the M-1th time-domain redundancy lifetime prediction value to obtain the redundancy lifetime prediction value.

[0069] Furthermore, the life assessment module 14 also includes the following execution steps: Constrained by the component model list and the component quantity list, multiple sets of data are collected, wherein any one of the multiple sets of data includes workload record data, initial life set value record data and a tag identifying redundant life. Configure a threshold for the number of decision trees, using the number of component models as a constraint. Based on the threshold number of decision trees, a random forest is trained using the multiple sets of data to obtain a redundancy lifetime estimator, which is then bound to the target energy storage unit.

[0070] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0071] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0072] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0073] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0074] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0075] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0076] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A quality management method for a new energy power plant, characterized in that, include: Receive the component model list and component quantity list of the target energy storage unit of the target new energy power station; Retrieve the service data of the target energy storage units of the target new energy power station and compile the time series matrix of energy storage impact attributes; Collect samples of energy storage units that meet the component model list and the component quantity list, and analyze the workload time-series information of the energy storage impact attribute time-series matrix. Based on the workload timing information, the redundancy lifetime of the target energy storage unit is assessed to obtain a redundancy lifetime prediction value. When the predicted redundancy lifetime value is less than or equal to the redundancy lifetime threshold, the target energy storage unit is temporarily marked for quality inspection. When the predicted redundancy lifetime value is greater than the redundancy lifetime threshold, the target energy storage unit is marked with a default periodic quality inspection.

2. The method as described in claim 1, characterized in that, Retrieve the service data of the target energy storage units in the target new energy power plant, and statistically analyze the time series matrix of energy storage impact attributes, including: Based on the energy storage type of the target energy storage unit, perform frequent mining to obtain the first energy storage impact attribute up to the Nth energy storage impact attribute; From the service data, the initial time-series information of the first energy storage impact attribute is extracted and subjected to neighborhood hierarchical clustering analysis to obtain the time-series information of the first energy storage impact attribute. Until the service data is used to extract the initial time series information of the Nth energy storage impact attribute and perform neighborhood hierarchical clustering analysis, the time series information of the Nth energy storage impact attribute is obtained. The energy storage impact attribute time series matrix is ​​constructed based on the first energy storage impact attribute time series information up to the Nth energy storage impact attribute time series information.

3. The method as described in claim 2, characterized in that, Based on the energy storage type of the target energy storage unit, frequent mining is performed to obtain the first energy storage impact attribute up to the Nth energy storage impact attribute, including: Obtain the set of influencing attributes to be filtered; Iterate through the set of influencing attributes to be screened and count the frequency set of attributes that co-occur with the energy storage type. Based on the frequency set, a set of first-level influence attributes that are greater than or equal to the frequency threshold is selected from the set of influence attributes to be screened; The support set is obtained by traversing the set of first-level influence attributes. Based on the support set, the first energy storage impact attribute that is greater than or equal to the support threshold is selected from the first-level impact attribute set up up to the Nth energy storage impact attribute.

4. The method as described in claim 3, characterized in that, The support set is obtained by traversing the set of primary influence attributes to perform support evaluation, including: Collect several historical energy storage data of the aforementioned energy storage type, wherein any one piece of historical energy storage data includes a set of primary influence attribute record values ​​and an energy storage record value; The same attribute deviation modulus is calculated for the aforementioned historical energy storage data to obtain multiple sets of deviations for primary influence attribute records and multiple deviations for energy storage records. Obtain the set of first-level influence attribute deviation thresholds; From the set of deviations of the multiple first-level influence attribute records, extract the deviations of multiple first-level influence attribute records where only the deviation of the first-level influence attribute is greater than the first-level influence attribute deviation threshold. Based on the deviations of the recorded values ​​of the multiple first influencing attributes, the deviations of the recorded values ​​of the multiple energy storage attributes are extracted from the deviations of the recorded values ​​of the multiple first influencing attributes; Calculate the ratio of the deviation of the energy storage record value associated with the multiple first influence attributes to the deviation of the record value of the multiple first influence attributes, then perform mode statistics to obtain the support of the first influence attributes, and add it to the support set.

5. The method as described in claim 1, characterized in that, The sample of energy storage units that meet the requirements of the component model list and the component quantity list is analyzed, and the workload time-series information of the energy storage impact attribute time-series matrix includes: Using the component model list and the component quantity list as constraints on the unit size, a sample set of energy storage units is selected; Using the energy storage capacity as the workload indicator and the time series matrix of the energy storage impact attributes as the environmental constraint, several workload time series records of the energy storage unit sample set are collected. Simultaneous load mode evaluation is performed on the aforementioned workload time-series records to obtain the workload time-series information.

6. The method as described in claim 1, characterized in that, Based on the workload timing information, a redundancy lifetime assessment is performed on the target energy storage unit to obtain a predicted redundancy lifetime value, including: The workload timing information includes the first time-domain workload up to the Mth time-domain workload; Based on the first time-domain workload and rated life, the redundancy life of the target energy storage unit is evaluated to obtain the first time-domain redundancy life prediction value. Based on the second time-domain workload and the first time-domain redundancy lifetime prediction value, the redundancy lifetime of the target energy storage unit is evaluated to obtain the second time-domain redundancy lifetime prediction value. The redundancy lifetime is assessed based on the Mth time-domain workload and the (M-1)th time-domain redundancy lifetime prediction value to obtain the redundancy lifetime prediction value.

7. The method as described in claim 6, characterized in that, Based on the first time-domain workload and rated lifespan, a redundancy lifetime assessment is performed on the target energy storage unit to obtain a first time-domain redundancy lifetime prediction value, including: Constrained by the component model list and the component quantity list, multiple sets of data are collected, wherein any one of the multiple sets of data includes workload record data, initial life set value record data and a tag identifying redundant life. Configure a threshold for the number of decision trees, using the number of component models as a constraint. Based on the threshold number of decision trees, a random forest is trained using the multiple sets of data to obtain a redundancy lifetime estimator, which is then bound to the target energy storage unit.

8. A quality management system for a new energy power plant, characterized in that, For implementing the method as described in any one of claims 1 to 7, comprising: The data receiving module is used to receive the component model list and component quantity list of the target energy storage unit of the target new energy power station; The service analysis module is used to retrieve the service data of the target energy storage units of the target new energy power station and to statistically analyze the time series matrix of energy storage impact attributes. The load statistics module is used to statistically analyze the working load time sequence information of the energy storage unit samples that meet the component model list and the component quantity list in the energy storage influence attribute time sequence matrix. The life assessment module is used to perform a redundancy life assessment on the target energy storage unit based on the workload timing information, and obtain a redundancy life prediction value. The temporary inspection identification module is used to temporarily identify the target energy storage unit when the predicted redundancy lifetime value is less than or equal to the redundancy lifetime threshold. The default identification module is used to perform a default periodic quality detection identification on the target energy storage unit when the predicted redundancy lifetime value is greater than the redundancy lifetime threshold.

Images