A method and system for robust evaluation of energy efficiency of a thermal power plant

Abstract

The application discloses a kind of thermal power plant energy efficiency robust evaluation method and system, it is related to energy efficiency evaluation technical field.The method includes: acquisition thermal power plant operating unit multidimensional operating parameter data, constructs time sequence characteristics capable of reflecting operating state change;Based on the time sequence characteristics determine the dynamic weight of multiple energy efficiency indicators;According to dynamic weight construction weight disturbance set, and under different disturbance conditions, calculate energy efficiency closeness set;Energy efficiency closeness set is carried out robust fusion processing, and generates comprehensive energy efficiency evaluation value;According to the comprehensive energy efficiency evaluation value determines energy efficiency grade.The application is by introducing time sequence characteristics modeling, dynamic weight adjustment, multiple disturbance closeness calculation and robust fusion processing in energy efficiency evaluation process, so that energy efficiency evaluation result can keep consistency and reliability under complex working condition and uncertain condition, and based on historical data constructs energy efficiency grade system, improves the applicability and generalizability of energy efficiency analysis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy efficiency evaluation, and in particular to a method and system for robust evaluation of energy efficiency of a thermal power plant. BACKGROUND

[0002] During long-term operation of a thermal power plant, its energy efficiency level is influenced by various factors such as equipment state, load variation, fuel quality, environmental conditions, and operation management strategy. Various operating parameters exhibit obvious trends, fluctuations, and coupling over time, and it is difficult to comprehensively reflect the overall energy efficiency status by relying on a single indicator. In application scenarios such as operation optimization, horizontal benchmarking, and operation and maintenance management of thermal power units, a multi-index comprehensive evaluation method is needed to determine the energy efficiency level of the operating unit.

[0003] Existing energy efficiency evaluation methods usually rely on fixed weights or models mainly based on linear weighting and static closeness degree calculation. Such methods cannot reflect the differences in importance of indicators caused by changes in operating state over time, and are also difficult to cope with weight sensitivity problems caused by load switching, parameter fluctuations, or external disturbances. In cases where operation data is unstable or indicators fluctuate greatly, fixed weight models are prone to evaluation result deviation, and lack consistency and repeatability. In complex operating scenarios, there is significant correlation between multiple energy efficiency indicators, and the closeness degree results under different disturbance conditions often exhibit significant differences. If there is a lack of robust processing mechanism for multiple disturbance situations, the comprehensive evaluation result is easily affected by local anomalies or short-term fluctuations, and it is difficult to truly reflect the overall energy efficiency level of the unit. At the same time, most existing energy efficiency grade division methods rely on empirical rules or manually set thresholds, and lack a systematic grade construction method based on historical operation data distribution characteristics, resulting in a lack of uniformity and comparability in grade division.

[0004] In the above background, how to extract features from multi-dimensional operating parameters that can depict the time variation law, how to dynamically adjust the importance of each energy efficiency indicator according to the change in operating state, how to obtain stable and reliable comprehensive energy efficiency evaluation results under multiple disturbance conditions, and how to construct an energy efficiency grade system with a data basis, have become problems that need to be solved in current energy efficiency evaluation technology. SUMMARY

[0005] The present application aims to provide a method and system for robust evaluation of energy efficiency of a thermal power plant, to solve the problem of insufficient adaptation to time series changes, indicator uncertainty, and disturbance sensitivity of existing energy efficiency evaluation methods in handling complex operating conditions, and to overcome limitations such as unstable traditional evaluation results, lack of basis for weight setting, and non-uniform grade division. The present application improves the response capability of the energy efficiency evaluation process to changes in operating state, the robustness to abnormal fluctuations, and the utilization degree of historical data characteristics, making the comprehensive energy efficiency evaluation result more reliable, consistent, and with engineering applicability.

[0006] To achieve the above object, the present application is realized by the following technical solutions:

[0007] In one aspect, the present application provides a method for robust evaluation of energy efficiency of a thermal power plant, comprising:

[0008] Step 1: Collecting multi-dimensional operation parameter data of a running unit of the thermal power plant, and performing time series feature extraction to obtain time series features capable of reflecting changes in the running state;

[0009] Step 2: Based on the time series features, determining dynamic weights of a plurality of energy efficiency indicators varying with the running state;

[0010] Step 3: Constructing a weight disturbance set according to the dynamic weights, and calculating a set of energy efficiency closeness degrees of the running unit under each weight disturbance condition;

[0011] Step 4: Robustly fusing the set of energy efficiency closeness degrees to obtain a comprehensive energy efficiency evaluation value of the running unit;

[0012] Step 5: Determining an energy efficiency grade of the running unit according to the comprehensive energy efficiency evaluation value.

[0013] As a preferred scheme of the present application, the time series feature extraction comprises:

[0014] Performing trend analysis on the change sequence of the multi-dimensional operation parameter data in a continuous evaluation period to obtain long-term change features;

[0015] Based on the long-term change features, performing fluctuation analysis on the change sequence of the multi-dimensional operation parameter data in a short time scale to obtain short-term fluctuation features;

[0016] According to the change rate and amplitude characteristics of the multi-dimensional operation parameter data, dividing the multi-dimensional operation parameter data into intervals, and obtaining corresponding long-term change features and short-term fluctuation features in each interval; taking the long-term change features and short-term fluctuation features obtained in each interval as interval change features;

[0017] For interval segments that do not meet the change continuity condition, identifying abnormal segments through interval difference analysis and eliminating them;

[0018] Based on the time series correlation calculation results of the multi-dimensional operation parameter data, obtaining the cooperative relationship features between the operation parameters;

[0019] Aligning the long-term change features, short-term fluctuation features, interval change features and cooperative relationship features obtained in different intervals according to consistency constraint rules;

[0020] According to a feature combination rule, the long-term change feature, the short-term fluctuation feature, the interval change feature and the synergy relationship feature after the alignment processing are combined to generate a time sequence feature for dynamic weight determination.

[0021] As a preferred scheme of the present application, the dynamic weight of the plurality of energy efficiency indexes changing with the operation state is determined based on the time sequence feature, and the dynamic weight determination comprises:

[0022] An operation state feature vector of the operation unit is generated by taking the time sequence feature as an input;

[0023] The operation state contribution of each energy efficiency index is calculated based on the operation state feature vector;

[0024] The historical discriminant coefficient of each energy efficiency index is calculated by taking the energy efficiency index data in the historical operation sample as an input;

[0025] The uncertainty factor of each energy efficiency index is calculated by taking the prediction residual or fluctuation data of the energy efficiency index as an input;

[0026] The non-normalized weight contribution of the energy efficiency index is calculated according to the following formula:

[0027] ;

[0028] Wherein: is the operation unit number; is the energy efficiency index number; is the evaluation time; is the operation state contribution of the energy efficiency index ; is the historical discriminant coefficient of the energy efficiency index ; is the uncertainty factor of the energy efficiency index ; is the uncertainty penalty coefficient; is the non-normalized weight contribution of the energy efficiency index ;

[0029] The dynamic weight of the energy efficiency index is calculated according to the following formula:

[0030] ;

[0031] Wherein: is the dynamic weight of the energy efficiency index ; is the number of energy efficiency indexes; is the sequence number of the energy efficiency index summation; indicates the operation unit number at the evaluation time , the first Un-normalized weight contribution calculated by the energy efficiency index

[0032] As a preferred scheme of the present application, constructing the weight disturbance set according to the dynamic weight comprises:

[0033] determining an adaptive disturbance amplitude for generating the disturbance based on the running state stability parameter of the running unit at the evaluation moment, and performing forward disturbance and reverse disturbance on the dynamic weight of each energy efficiency index according to the adaptive disturbance amplitude and taking the dynamic weight as a reference, to obtain a plurality of initial weight disturbance vectors;

[0034] performing disturbance direction checking on the initial weight disturbance vectors according to the correlation constraint rules between the energy efficiency indexes, eliminating the disturbance vectors that do not satisfy the correlation constraint rules, to obtain a candidate weight disturbance vector set;

[0035] performing normalization processing on the candidate weight disturbance vector set according to the weight non-negative constraint and the weight summation constraint, to obtain a weight disturbance vector set that satisfies the constraint conditions;

[0036] taking each weight disturbance vector in the weight disturbance vector set as input to respectively calculate the energy efficiency closeness of the running unit under the corresponding disturbance weight, to generate an energy efficiency closeness set.

[0037] As a preferred scheme of the present application, the step of performing robust fusion processing on the energy efficiency closeness set is as follows:

[0038] performing numerical consistency checking on each energy efficiency closeness result in the energy efficiency closeness set, and eliminating the energy efficiency closeness results that do not satisfy the numerical consistency condition;

[0039] performing weighted synthesis on the remaining energy efficiency closeness results according to a preset fusion weight rule, to obtain a preliminary fusion result;

[0040] performing deviation correction processing on the preliminary fusion result according to the distribution characteristics of the energy efficiency closeness set, to obtain a corrected fusion result;

[0041] taking the corrected fusion result as the comprehensive energy efficiency evaluation value of the running unit.

[0042] As a preferred scheme of the present application, the robust fusion processing on the energy efficiency closeness set comprises constructing a robust fusion function based on the energy efficiency closeness set, and generating a corrected fusion result according to the robust fusion function, and the robust fusion function comprises:

[0043] ;

[0044] wherein: is the preliminary fusion result; is the number of weight disturbance vectors; In order to assess the moment The set of weighted perturbation vectors obtained below is the first One energy efficiency closeness result; In order to target the The fusion weights are calculated from the energy efficiency proximity results; the deviation correction process includes constructing a correction compensation function based on the deviation distribution of the energy efficiency proximity set, and applying the correction compensation function to the preliminary fusion results to obtain the corrected fusion results. The formula for the correction compensation function is:

[0045] ;

[0046] in: This is the deviation compensation amount; This is the deviation amplification factor calculated based on the distribution characteristics of the energy efficiency proximity set; the corrected fusion result is... express.

[0047] As a preferred embodiment of the present invention, before performing robust fusion processing on the energy efficiency proximity set, the energy efficiency proximity set is subjected to perturbation sensitivity screening through the following operation:

[0048] Based on the correspondence between the weighted perturbation vector and the energy efficiency closeness result, the perturbation sensitivity of each energy efficiency closeness result is calculated; all energy efficiency closeness results are sorted according to the perturbation sensitivity; and energy efficiency closeness results with perturbation sensitivity exceeding the preset threshold range are removed from the energy efficiency closeness set.

[0049] As a preferred embodiment of the present invention, in the process of determining the energy efficiency level of the operating unit based on the comprehensive energy efficiency evaluation value, the energy efficiency level is determined through the following operations:

[0050] A set of energy efficiency level thresholds is constructed based on the comprehensive energy efficiency evaluation values ​​of multiple historical operating units;

[0051] The overall energy efficiency evaluation value is compared with the threshold values ​​of each level in the set of energy efficiency level thresholds;

[0052] Based on the comparison results, select the energy efficiency level corresponding to the level threshold that matches the comprehensive energy efficiency evaluation value.

[0053] As a preferred embodiment of the present invention, the process of constructing the energy efficiency level threshold set includes the following operations:

[0054] The statistical data set for describing the distribution characteristics of the comprehensive energy efficiency evaluation value is generated based on the comprehensive energy efficiency evaluation values of a plurality of historical running units; the plurality of historical running units are clustered according to the statistical data set to obtain a plurality of energy efficiency level categories; the range of the comprehensive energy efficiency evaluation value in each energy efficiency level category is taken as an energy efficiency level threshold range, and the plurality of energy efficiency level threshold ranges form an energy efficiency level threshold set.

[0055] In another aspect, the present application also provides a thermal power plant energy efficiency robust evaluation system, which is applied to a thermal power plant energy efficiency robust evaluation method as described above, and the system comprises:

[0056] a time sequence feature extraction module, which is used for collecting multi-dimensional running parameter data of a running unit of a thermal power plant and performing time sequence feature extraction to obtain time sequence features capable of reflecting running state changes;

[0057] a dynamic weight determination module, which is used for determining dynamic weights of a plurality of energy efficiency indexes based on the time sequence features;

[0058] a weight disturbance closeness module, which is used for constructing a weight disturbance set according to the dynamic weights and calculating a set of energy efficiency closeness of the running unit under a plurality of weight disturbance conditions;

[0059] a robust fusion module, which is used for performing robust fusion processing on the set of energy efficiency closeness to obtain a comprehensive energy efficiency evaluation value of the running unit;

[0060] an energy efficiency level determination module, which is used for determining an energy efficiency level of the running unit according to the comprehensive energy efficiency evaluation value.

[0061] The present application has the following beneficial effects: the present application sequentially performs time sequence feature extraction, dynamic weight determination, weight disturbance construction and closeness calculation, and robust fusion processing, so that the energy efficiency evaluation process can reflect the time variation law of the running unit under different working conditions, and still maintain high evaluation reliability when there is an uncertain disturbance. The time sequence feature extraction enables the weight calculation to fully utilize the trend information, fluctuation characteristics and correlation of the running data, thereby ensuring that the importance of different energy efficiency indexes dynamically changes with the running state. The weight disturbance set constructed based on the dynamic weights covers a plurality of disturbance conditions that may occur in the running state, so that the generated closeness set can comprehensively describe the response of the running result to the weight change. The robust fusion processing performs consistency checking and deviation correction on the closeness result, so that the comprehensive energy efficiency evaluation value remains stable under complex running conditions, thereby effectively solving the defects of the traditional fixed weight mode, i.e., the evaluation result is easily affected by fluctuations and lacks stability. BRIEF DESCRIPTION OF DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort based on these drawings. Among them:

[0063] Figure 1 The method flowchart of the present application;

[0064] Figure 2 The system modular structure schematic diagram of the present application. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solutions and advantages of the embodiments of the present application more clear, the following will combine the drawings of the embodiments of the present application to make a clear and complete description of the technical solutions of the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all the embodiments. Based on the described embodiments of the present application, all other embodiments obtained by those skilled in the art belong to the scope of protection of the present application.

[0066] As Figure 1 shown in the present application, in one embodiment, the thermal power plant energy efficiency robust evaluation method first performs the collection and time sequence feature construction of operation data. The real-time operation data of the unit can be generally obtained by the unit DCS system, thermal monitoring system or online energy efficiency analysis platform. In the present embodiment, the selected operation parameters can include main steam temperature, main steam pressure, feed water temperature, load power, coal consumption index, exhaust gas temperature and other multi-dimensional operation parameters, which are used to represent the operation change trend of the unit under different working conditions.

[0067] Step 1: Collecting multi-dimensional operation parameter data of the thermal power plant operation unit, and performing time sequence feature extraction to obtain time sequence features reflecting the change of operation state; in order to obtain multi-level time sequence features reflecting the operation law, the present embodiment is based on the data in the continuous monitoring period, and the time sequence features are constructed in the following way.

[0068] The time sequence feature extraction includes:

[0069] Performing trend analysis on the change sequence of multi-dimensional operation parameter data in the continuous evaluation period to obtain long-term change characteristics;

[0070] The trend analysis can be realized by calculating the moving average sequence, local linear fitting value or monotonic change trend, etc. so that the performance change of the unit in a long time scale gradually appears, which is convenient for identifying the long-term law such as load change, equipment aging or operation quality improvement.

[0071] Based on the long-term trend, based on the long-term change characteristics, the fluctuation analysis is performed on the change sequence of the multi-dimensional operating parameter data on a short time scale to obtain a short-term fluctuation characteristic; the short-term fluctuation characteristic can be extracted by monitoring the change amplitude, fluctuation frequency or fluctuation rate in a short window range, reflecting the transient disturbance situation of the unit in actual operation.

[0072] According to the change rate and amplitude characteristics of the multi-dimensional operating parameter data, the multi-dimensional operating parameter data is divided into intervals, and the corresponding long-term change characteristics and short-term fluctuation characteristics are obtained in the divided intervals; the long-term change characteristics and the short-term fluctuation characteristics obtained in each interval are taken as interval change characteristics; interval division can make the feature calculation in different operating stages more targeted, such as extracting features in low load section, high load section or stable operation section, making the time sequence representation more accurate.

[0073] Further, for the interval section that does not satisfy the change continuity condition, the abnormal section is identified and removed through interval difference analysis; this is to avoid the influence of abnormal intervals on subsequent feature quality due to instrument failure, instantaneous jump, data missing, etc., so that the time sequence features formed finally are more stable and reliable. Interval difference analysis can be realized by comparing the mean difference, slope difference or fluctuation amplitude difference of adjacent intervals, and those skilled in the art can identify abnormal sections with discontinuous changes accordingly.

[0074] Based on the time sequence correlation calculation result of the multi-dimensional operating parameter data, the cooperative relationship characteristics between the operating parameters are obtained; the cooperative relationship characteristics are used to describe the dynamic association between different energy efficiency indicators, such as the coupling relationship between the boiler efficiency and the exhaust gas temperature, the cooperative change between the main steam pressure and the main steam temperature, etc., which is helpful to form a more comprehensive state representation.

[0075] After the feature construction is completed, the multi-class features need to be uniformly processed. Therefore, the long-term change characteristics, short-term fluctuation characteristics, interval change characteristics and cooperative relationship characteristics obtained in different intervals are aligned according to the consistency constraint rule; the alignment processing usually includes time window synchronization, feature dimension unification or scale normalization, so that different types of features can be used under the unified time sequence reference.

[0076] According to the feature combination rule, the long-term change characteristics, short-term fluctuation characteristics, interval change characteristics and cooperative relationship characteristics after alignment processing are combined to generate time sequence features for dynamic weight determination. The generated time sequence features comprehensively reflect the time change information of the operating unit in multiple dimensions, and can be directly used for dynamic weight calculation and disturbance analysis in the subsequent steps, ensuring that the entire energy efficiency evaluation process has sufficient data basis and time sequence expression ability.

[0077] After the construction of the timing characteristics is completed, the embodiment determines the dynamic importance of the plurality of energy efficiency indicators under the current operating state based on the timing characteristics. To this end, the embodiment comprehensively considers the operating state contribution, the historical data differentiation degree, and the uncertainty factors in operation in a multi-source information fusion manner to form dynamic weights that can be adjusted with the change of the operating state.

[0078] Step 2: Based on the timing characteristics, the dynamic weights of the plurality of energy efficiency indicators that change with the operating state are determined.

[0079] To achieve the above-mentioned purpose, the embodiment generates dynamic weights through the following process.

[0080] The dynamic weights of the plurality of energy efficiency indicators that change with the operating state are determined based on the timing characteristics, and include:

[0081] The operating state feature vector of the operating unit is generated by taking the timing characteristics as input. The operating state feature vector can be regarded as a high-dimensional description of the overall operating condition of the current unit, including long-term features reflecting trend changes, short-term fluctuation features reflecting disturbance characteristics, interval change features, and collaborative relationship features among multiple parameters. The embodiment uniformly converts these features into a state vector through feature mapping for subsequent quantitative analysis.

[0082] The operating state contribution of each energy efficiency indicator is calculated based on the operating state feature vector. The operating state contribution is used to describe the sensitivity and influence degree of each energy efficiency indicator on the comprehensive energy efficiency under the current operating condition. For example, when the boiler efficiency changes significantly under a certain operating condition segment and the exhaust gas temperature changes weakly, the operating state contribution can reflect the difference, so that the obtained weight is more consistent with the actual operating condition.

[0083] The historical differentiation degree coefficient of each energy efficiency indicator is calculated by taking the energy efficiency indicator data in the historical operating sample as input. The historical differentiation degree coefficient can reflect the ability of a certain energy efficiency indicator to distinguish different energy efficiency levels in the long-term operation. Generally speaking, the greater the change range and the more obvious the contribution to the differentiation of the comprehensive energy efficiency in the historical data, the higher the historical differentiation degree coefficient.

[0084] The uncertainty factor of each energy efficiency indicator is calculated by taking the prediction residual or fluctuation data of the energy efficiency indicator as input.

[0085] The uncertainty factor is used to reflect the reliability of the indicator in actual operation. For example, the instability caused by sensor noise, data missing, environmental fluctuation, etc. can be quantified by the uncertainty factor, so that indicators with high uncertainty are inhibited in the maximum weight calculation.

[0086] After the above quantification is completed, the non-normalized weight contribution is calculated according to the formula.

[0087] The unnormalized weight contribution of the energy efficiency indicator is calculated according to the following formula:

[0088] ;

[0089] Wherein: is the running unit number, indicating the specific unit or unit subsystem for which the energy efficiency evaluation is performed; is the energy efficiency indicator number, indicating the number of the energy efficiency indicator, such as boiler efficiency, coal consumption for power supply, main steam temperature, etc.; is the evaluation time, i.e. the corresponding time point when the energy efficiency evaluation is performed; is the running state contribution of the energy efficiency indicator , describing the influence degree of the energy efficiency indicator on the energy efficiency change under the current running state; is the historical differentiation coefficient of the energy efficiency indicator , reflecting the ability of the indicator to distinguish different energy efficiency levels in historical data; is the uncertainty factor of the energy efficiency indicator , describing the stability and reliability of the indicator in the current running; is the uncertainty penalty coefficient, The introduction of the uncertainty penalty coefficient makes the unnormalized weight contribution of the indicator with higher uncertainty smaller, thereby enhancing the stability of the overall weight under disturbance; is the unnormalized weight contribution of the energy efficiency indicator ;

[0090] In a specific implementation, the unnormalized weight contribution of the energy efficiency indicator can be calculated by the following steps: 1) for each energy efficiency indicator, first obtain the running state contribution of the indicator according to the running state feature vector, for example, the deviation of the indicator at the current evaluation time from its historical average value or reference value can be normalized to make the running state contribution within a predetermined range; 2) according to the historical running samples, the historical values of each energy efficiency indicator are counted, the ability of the indicator to distinguish different energy efficiency levels is calculated, and the historical differentiation coefficient is obtained by normalization, so that the higher the differentiation degree, the larger the historical differentiation coefficient; 3) according to the fluctuation or prediction residual of the energy efficiency indicator in a short time window, the uncertainty of each energy efficiency indicator is measured, and the uncertainty factor is obtained by normalization, so that the greater the fluctuation or the greater the prediction residual, the greater the uncertainty factor; 4) for each energy efficiency indicator, first multiply the running state contribution of the indicator by the historical differentiation coefficient to obtain an intermediate quantity, then divide the intermediate quantity by "1 plus the product of the uncertainty penalty coefficient and the uncertainty factor", thereby obtaining the unnormalized weight contribution of the indicator, wherein the uncertainty penalty coefficient is a pre-set constant, used to control the suppression degree of the uncertainty factor on the unnormalized weight contribution.

[0091] The overall structure of the formula embodies the calculation logic of contribution enhancement and fluctuation suppression, i.e., the running state contribution amount and the historical distinction degree are jointly enhanced, and the uncertainty factor suppresses the influence of the unstable indicator through the denominator part, so that the maximum weight has adaptability and robustness.

[0092] After obtaining the non-normalized weight contribution amount, the embodiment further generates the dynamic weight through normalization.

[0093] The dynamic weight of the energy efficiency indicator is calculated according to the following formula:

[0094] ;

[0095] Wherein: is the dynamic weight of the energy efficiency indicator , indicating the relative importance of the energy efficiency indicator at the evaluation moment ; is the number of energy efficiency indicators; is the sequence number of the summation of energy efficiency indicators; indicates the non-normalized weight contribution amount of the running unit at the evaluation moment , calculated for the th energy efficiency indicator.

[0096] In a specific implementation manner, the dynamic weight of the energy efficiency indicator can be obtained by normalizing the non-normalized weight contribution amount, specifically including: 1) at a given evaluation moment, first sum all the non-normalized weight contribution amounts of the energy efficiency indicators to obtain a total sum; 2) for each energy efficiency indicator, divide the non-normalized weight contribution amount of the indicator by the above total sum to obtain the dynamic weight of the indicator; 3) through the above normalization processing, it can be ensured that the dynamic weights of all energy efficiency indicators are non-negative, and the sum of the dynamic weights of all energy efficiency indicators at the same evaluation moment is equal to 1, thereby meeting the mathematical constraints of the comprehensive energy efficiency evaluation model.

[0097] The normalization process ensures that the sum of all weights is 1, meets the mathematical constraint condition of the energy efficiency evaluation model, and can make the dynamic weights comparable under different running states.

[0098] Through the above steps, the embodiment realizes a dynamic weight determination mechanism based on multi-source information, so that the importance of the energy efficiency indicator can adaptively change with the running state, thereby providing a basis for subsequent weight disturbance construction and robust evaluation.

[0099] After obtaining the dynamic weights of the plurality of energy efficiency indicators, the embodiment performs perturbation construction on the dynamic weights to simulate energy efficiency evaluation results under different running fluctuations, data uncertainties or changes in indicator sensitivity. By generating a plurality of sets of perturbed weights and calculating corresponding energy efficiency closeness, the robustness of the comprehensive evaluation results to running perturbations can be improved.

[0100] Step 3: constructing a set of weight perturbations according to the dynamic weights, and calculating a set of energy efficiency closeness of the running unit under each weight perturbation condition;

[0101] In the embodiment, the set of weight perturbations is constructed as follows.

[0102] Constructing a set of weight perturbations according to the dynamic weights comprises:

[0103] Determining an adaptive perturbation amplitude for generating perturbations based on the running state stability parameter of the running unit at the evaluation time, and perturbing the dynamic weights of each energy efficiency indicator in a positive direction and a reverse direction according to the adaptive perturbation amplitude based on the dynamic weights, to obtain a plurality of initial weight perturbation vectors;

[0104] In specific implementation, the running state stability parameter can be quantified by a short-term fluctuation amplitude, a load change rate or a coefficient of variation of multi-dimensional parameters. The worse the stability, the greater the perturbation amplitude can be, so that the final evaluation can better reflect the real running uncertainty.

[0105] Perturbing the initial weight perturbation vectors according to the relevance constraint rules between the energy efficiency indicators, eliminating the perturbation vectors that do not satisfy the relevance constraint rules, and obtaining a set of candidate weight perturbation vectors;

[0106] In engineering practice, there are natural correlations between different energy efficiency indicators, for example, an increase in boiler efficiency is usually accompanied by a decrease in flue gas temperature, and other indicators may change simultaneously when the coal consumption for power supply increases. Therefore, the embodiment constructs the relevance constraint through feature correlation (such as Pearson coefficient, mutual information, etc.).

[0107] For the perturbation vectors that violate the trend of index correlation, whether the perturbation direction is consistent with the index cooperative relationship is judged, and the perturbation vectors are eliminated to prevent the introduction of physically unreasonable perturbation weights.

[0108] Further, the set of candidate weight perturbation vectors is normalized according to the non-negative weight constraint and the weight summation constraint to obtain a set of weight perturbation vectors that satisfy the constraint conditions;

[0109] Physical properties of weights require:

[0110] Each weight cannot be negative (non-negativity constraint)

[0111] The sum of the weights of all energy efficiency indicators must be 1 (normalization constraint)

[0112] The embodiment realizes the correction of the disturbance weight by linear normalization, so that each set of disturbance weights meets the mathematical conditions that can be used for closeness calculation.

[0113] Each weight disturbance vector in the set of weight disturbance vectors is input to calculate the energy efficiency closeness of the running unit under the corresponding disturbance weight, and an energy efficiency closeness set is generated.

[0114] For each set of disturbance weights, closeness calculation is performed on the basis of the normalized indicator value (such as TOPSIS positive ideal point, negative ideal point distance, etc.). By traversing all disturbance weight vectors, multiple closeness results are obtained.

[0115] The finally formed energy efficiency closeness set covers evaluation results under multiple disturbance conditions, and can be used as an important input for subsequent robust fusion processing, thereby improving the stability and reliability of the final evaluation result in the actual running environment.

[0116] After the energy efficiency closeness set is calculated by multiple sets of weight disturbance vectors, the embodiment further performs robust fusion processing on the set. Through consistency checking, weighted fusion and deviation correction of the closeness result, the comprehensive energy efficiency evaluation value finally obtained can remain stable under different disturbance conditions, and the influence of local outliers can be reduced.

[0117] Step 4: performing robust fusion processing on the energy efficiency closeness set to obtain a comprehensive energy efficiency evaluation value of the running unit;

[0118] To realize robust fusion, the embodiment is processed in the following manner.

[0119] Numerical consistency checking is performed on each energy efficiency closeness result in the energy efficiency closeness set, and energy efficiency closeness results that do not meet the numerical consistency condition are removed. The energy efficiency closeness set may contain outliers caused by extreme disturbance, input data fluctuation or error amplification. The embodiment can use the median deviation, box interval or distribution density method to identify the closeness results deviating from the main distribution area, and remove them from the set to ensure that the subsequent fusion process has a relatively stable basis.

[0120] The remaining energy efficiency closeness results are weighted and synthesized according to the preset fusion weight rule to obtain a preliminary fusion result. The fusion weight can be related to the disturbance amplitude, closeness fluctuation degree or disturbance direction consistency, or can be set according to the running state stability parameter. The embodiment obtains a preliminary fusion value under multiple disturbance conditions through the weighting process, so that the evaluation result can reflect the main distribution trend.

[0121] The preliminary fusion result is subjected to deviation correction processing according to the distribution characteristics of the energy efficiency closeness set, to obtain a corrected fusion result;

[0122] The corrected fusion result is taken as the comprehensive energy efficiency evaluation value of the running unit.

[0123] The robust fusion processing of the energy efficiency closeness set includes constructing a robust fusion function based on the energy efficiency closeness set, and generating a corrected fusion result according to the robust fusion function. Since the closeness calculation under each perturbation weight may present slight systematic deviation, the embodiment analyzes the average deviation amount, deviation amplification coefficient or distribution center position of the closeness set, constructs a deviation compensation amount and corrects the preliminary fusion result, so that the final result is closer to the stable center value of each perturbation result.

[0124] The corrected fusion result is taken as the comprehensive energy efficiency evaluation value of the running unit.

[0125] In further specific implementation, the robust fusion processing of the energy efficiency closeness set includes constructing a robust fusion function based on the energy efficiency closeness set, and generating a corrected fusion result according to the robust fusion function, and the robust fusion function includes:

[0126] ;

[0127] Wherein: is the preliminary fusion result; is the number of weight perturbation vectors; is the energy efficiency closeness result obtained by the weight perturbation vector set at the evaluation moment ; is the energy efficiency closeness result obtained by the weight perturbation vector set at the evaluation moment ; is the fusion weight calculated for the energy efficiency closeness result ;

[0128] The preliminary fusion result is obtained by weighted linear combination of each closeness result, and the contribution of the closeness result under different perturbation weights to the comprehensive energy efficiency evaluation value is embodied by the fusion weight .

[0129] In a specific implementation, the specific implementation of the robust fusion function can include: 1) for each energy efficiency closeness result in the energy efficiency closeness set, determining a fusion weight according to its corresponding weight perturbation amplitude, perturbation sensitivity or numerical fluctuation degree, and giving a greater fusion weight to the closeness result which is less sensitive to perturbation and has better numerical stability; 2) normalizing all fusion weights so that the sum of the fusion weights is equal to 1; 3) multiplying each energy efficiency closeness result by its corresponding normalized fusion weight, and summing all the products to obtain a preliminary fusion result of the energy efficiency closeness set; and 4) taking the preliminary fusion result as the basic result output by the robust fusion function, and providing input for subsequent bias correction.

[0130] The bias correction process includes constructing a correction compensation function based on the bias distribution of the energy efficiency closeness set, and applying the correction compensation function to the preliminary fusion result to obtain a corrected fusion result, and the correction compensation function formula is:

[0131] ;

[0132] Wherein: is a bias compensation amount; is a bias amplification coefficient calculated based on the distribution characteristics of the energy efficiency closeness set; and the corrected fusion result is represented by .

[0133] In a specific implementation, the bias correction process can include the following steps: 1) calculating the arithmetic mean of all energy efficiency closeness results in the energy efficiency closeness set that pass the consistency check to obtain a closeness mean; 2) calculating the median or other robust statistics of the above closeness results to obtain a closeness median; 3) calculating the difference between the closeness mean and the closeness median, and multiplying the difference by a bias amplification coefficient to obtain a bias compensation amount, wherein the bias amplification coefficient is a coefficient preset or adaptively determined according to the distribution characteristics, and is used to adjust the compensation strength; and 4) applying the bias compensation amount to the preliminary fusion result, for example, using the preliminary fusion result minus the bias compensation amount to obtain the corrected fusion result, and taking the corrected fusion result as the final comprehensive energy efficiency evaluation value.

[0134] The bias compensation amount is determined by statistically analyzing the overall bias of the closeness set, which helps to eliminate systematic bias caused by perturbation direction bias or uneven distribution, and makes the final evaluation result more robust and reliable.

[0135] Before robust fusion, the energy efficiency closeness set is subjected to perturbation sensitivity screening by the following operations:

[0136] Based on the correspondence between the weight perturbation vector and the energy efficiency closeness result, the perturbation sensitivity of each energy efficiency closeness result is calculated; all energy efficiency closeness results are sorted according to the perturbation sensitivity; and the energy efficiency closeness result whose perturbation sensitivity exceeds a preset threshold range is removed from the energy efficiency closeness set. In this embodiment, the closeness results are sorted according to the sensitivity from large to small, and the closeness result whose sensitivity exceeds the preset threshold is removed from the closeness set, so that the final fusion result is more stable.

[0137] In an optional embodiment, the "calculating the perturbation sensitivity of each energy efficiency closeness result" can be achieved by comparing the deviation between the closeness result under the corresponding perturbation weight and the overall level of the closeness set. For example, the average level of all energy efficiency closeness results can be determined first, and then the sensitivity of the closeness result to the change of the weight perturbation can be obtained according to the relationship between the deviation of each closeness result from the average level and the corresponding weight perturbation amplitude. The greater the deviation or the closeness result that changes significantly under a smaller perturbation amplitude, the higher the perturbation sensitivity.

[0138] After obtaining the comprehensive energy efficiency evaluation value of the running unit, the energy efficiency grade of the running unit is determined according to the evaluation value in this embodiment. The energy efficiency grade is used to represent the level interval of the unit operation efficiency in the industry or the whole plant range from a global perspective, which can be used for horizontal comparison, equipment evaluation and operation optimization, etc. To ensure the objectivity and generalizability of the energy efficiency grade, the energy efficiency grade threshold set is constructed by the statistical law of historical data.

[0139] Step 5: determining the energy efficiency grade of the running unit according to the comprehensive energy efficiency evaluation value.

[0140] To realize the reasonable determination of the energy efficiency grade, the following method is used in this embodiment.

[0141] In the process of determining the energy efficiency grade of the running unit according to the comprehensive energy efficiency evaluation value, the energy efficiency grade is determined by the following operations:

[0142] The energy efficiency grade threshold set is constructed according to the comprehensive energy efficiency evaluation value of the plurality of historical running units. The historical running units can include the running records of the same type of unit at different time periods, or the historical energy efficiency data of different units under stable running conditions. In this embodiment, the generalizable grade boundary is constructed by analyzing the distribution of these historical comprehensive evaluation values.

[0143] The comprehensive energy efficiency evaluation value is compared with each grade threshold in the energy efficiency grade threshold set; the comparison is usually performed in the order from high to low or from low to high, the interval range into which the current comprehensive energy efficiency evaluation value falls is determined, and the belonging grade is determined.

[0144] According to the comparison result, an energy efficiency grade corresponding to a grade threshold value matched with the comprehensive energy efficiency evaluation value is selected. The final energy efficiency grade can adopt a grade system such as A, B, C, D, or a more detailed multi-level classification system. The embodiment does not limit the number of grades, and can be automatically formed according to historical distribution.

[0145] The process of constructing the set of energy efficiency grade thresholds includes the following operations:

[0146] A statistical data set for describing the distribution characteristics of the comprehensive energy efficiency evaluation value is generated based on the comprehensive energy efficiency evaluation values of the plurality of historical running units. The statistical data set can include mean, variance, density distribution, quantile and other information of the evaluation value, for describing the overall distribution form of the comprehensive energy efficiency evaluation value.

[0147] According to the statistical data set, the plurality of historical running units are clustered to obtain a plurality of energy efficiency grade categories. The clustering method can include K-means, density clustering or hierarchical clustering, etc., for automatically dividing different energy efficiency intervals, so that the grade boundaries are not dependent on human experience, but are naturally formed by data distribution.

[0148] The range of the comprehensive energy efficiency evaluation value in each energy efficiency grade category is taken as an energy efficiency grade threshold range, and the plurality of energy efficiency grade threshold ranges form the set of energy efficiency grade thresholds.

[0149] Each category obtained by clustering corresponds to an evaluation value interval range, for example:

[0150] Grade A: the evaluation value is located in the high value interval; Grade B: the evaluation value is located in the high-middle interval; Grade C: the evaluation value is located in the middle-low interval; and Grade D: located in the low value interval.

[0151] These interval ranges constitute the final set of energy efficiency grade thresholds for subsequent grade judgment.

[0152] In specific applications, the embodiment first constructs the set of grade thresholds using historical data, and can update the thresholds regularly according to new running data, so that the grade division always reflects the change of the overall running level of the unit. A larger data sample can improve the stability of the grade boundary, so that the final energy efficiency grade has both scientific basis and high engineering applicability.

[0153] In actual platforms, the running personnel do not need to manually set the thresholds, and the system automatically constructs the grade set through the statistical analysis and clustering mechanism provided by the present application, and automatically judges according to the comprehensive energy efficiency evaluation value, which is convenient for realizing automatic energy efficiency evaluation of a large number of running units.

[0154] In another embodiment of the present application, a system for implementing the above-mentioned robust energy efficiency evaluation method of thermal power plants is provided. The system can be deployed in an energy efficiency monitoring platform or an operation analysis system of a thermal power plant, and can automatically perform the energy efficiency evaluation process and output the comprehensive energy efficiency grade through real-time data processing of multi-dimensional operation parameters. The system provided in this embodiment corresponds to the above-mentioned method steps one by one, so as to ensure the integrity and consistency of the evaluation process.

[0155] As shown in Figure 2 Another embodiment of the present application provides a robust energy efficiency evaluation system of a thermal power plant, which is applied to the above-mentioned robust energy efficiency evaluation method of a thermal power plant. The system comprises:

[0156] a time sequence feature extraction module, configured to collect multi-dimensional operation parameter data of a running unit of a thermal power plant and perform time sequence feature extraction to obtain time sequence features capable of reflecting changes in the running state;

[0157] a dynamic weight determination module, configured to determine dynamic weights of a plurality of energy efficiency indexes based on the time sequence features;

[0158] a weight disturbance closeness module, configured to construct a weight disturbance set according to the dynamic weights and calculate a set of energy efficiency closenesses of the running unit under a plurality of weight disturbance conditions;

[0159] a robust fusion module, configured to perform robust fusion processing on the set of energy efficiency closenesses to obtain a comprehensive energy efficiency evaluation value of the running unit;

[0160] an energy efficiency grade determination module, configured to determine an energy efficiency grade of the running unit according to the comprehensive energy efficiency evaluation value.

[0161] In specific deployment, the system can be connected to a unit control system through a data acquisition interface to update the operation parameters in a periodic or real-time manner. The modules sequentially perform corresponding functions through internal data channels, corresponding to the function processes of steps 1 to 5 of the method, and finally output the energy efficiency grade of the current running unit.

[0162] The system can be integrated into an online monitoring platform or run as an independent software module, and has good engineering implementability and generalizability.

[0163] In summary, the application constructs a set of energy efficiency robust evaluation system with complete structure and closed loop process, which can be deeply combined with the operation monitoring platform of the thermal power plant, and realize the automatic execution from data acquisition, feature processing, weight modeling to energy efficiency evaluation and grade determination. The system can adapt to the needs of different unit types and different operation scenarios, has good engineering deployment ability and expansion ability, and is conducive to popularization and application in the operation optimization, energy efficiency comparison, equipment management and energy saving scheduling of the thermal power plant. Through the unified evaluation process and the expandable analysis framework, the application can form a sustainable energy efficiency evaluation system, and provide a solid technical foundation for the digital operation management and fine energy efficiency improvement of the thermal power plant.

[0164] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example" or "some examples" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the specification and the features of the different embodiments or examples without contradiction.

[0165] Any process or method descriptions in flow diagrams or otherwise described herein can be understood as representing code modules, segments, or portions of code that include one or more executable instructions for implementing specific logic functions or other processes. Also, the preferred embodiments of the present application can be implemented as a temporary program, a state machine, or as one or more application specific integrated circuits (ASICs) or in some other substrate memory device.

[0166] The above description is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of various changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A robust energy efficiency assessment method for thermal power plants, characterized in that, The method includes: Multidimensional operating parameter data of thermal power plant operating units are collected and time-series features are extracted to obtain time-series features that reflect changes in operating status. Based on the time-series features, dynamic weights of multiple energy efficiency indicators that change with operating status are determined. A weighted perturbation set is constructed according to the dynamic weights, and the energy efficiency proximity set of the operating unit is calculated under each weight perturbation condition. The energy efficiency proximity set is robustly fused to obtain the comprehensive energy efficiency evaluation value of the operating unit. The energy efficiency level of the operating unit is determined based on the comprehensive energy efficiency evaluation value. The determination of dynamic weights for multiple energy efficiency indicators that change with operating status based on the time-series characteristics includes: The operating state feature vector of the operating unit is generated using the aforementioned time-series features as input; The operational state contribution of each energy efficiency index is calculated based on the operational state feature vector. The historical discrimination coefficient of each energy efficiency indicator is calculated using energy efficiency index data from historical operating samples as input. The uncertainty factor of each energy efficiency index is calculated using the prediction residuals or fluctuation data of the energy efficiency index as input. The unnormalized weighted contribution of energy efficiency indicators is calculated using the following formula: ； in: The serial number of the operating unit; Energy efficiency index number; For the evaluation time; Energy efficiency indicators The contribution of the operating status; Energy efficiency indicators Historical discrimination coefficient; Energy efficiency indicators Uncertainty factors; This represents the uncertainty penalty coefficient. Energy efficiency indicators The unnormalized weight contribution; The dynamic weights of energy efficiency indicators are calculated using the following formula: ； in: Energy efficiency indicators Dynamic weights; The number of energy efficiency indicators; The sequence number for summing energy efficiency indicators; Indicates the running unit number At the assessment time Next, to the first The unnormalized weighted contribution of each energy efficiency indicator calculated; Constructing a weight perturbation set based on the dynamic weights includes: Based on the stability parameters of the operating state of the operating unit at the evaluation time, the adaptive perturbation amplitude for generating the perturbation is determined, and the dynamic weights of each energy efficiency index are perturbed in the positive and negative directions according to the adaptive perturbation amplitude based on the dynamic weights, to obtain multiple initial weight perturbation vectors. The initial weight perturbation vector is perturbed according to the correlation constraint rules between energy efficiency indicators. Perturbation vectors that do not meet the correlation constraint rules are eliminated to obtain a set of candidate weight perturbation vectors. The candidate weight perturbation vector set is normalized according to the weight non-negativity constraint and the weight summation constraint to obtain the weight perturbation vector set that satisfies the constraint conditions. Using each weight perturbation vector in the set of weight perturbation vectors as input, calculate the energy efficiency closeness of the running unit under the corresponding perturbation weight, and generate an energy efficiency closeness set; Robust fusion processing of the energy efficiency proximity set includes constructing a robust fusion function based on the energy efficiency proximity set, and generating a corrected fusion result according to the robust fusion function. The robust fusion function includes: ； in: This is the preliminary fusion result; This represents the number of weight perturbation vectors; In order to assess the moment The set of weighted perturbation vectors obtained below is the first One energy efficiency closeness result; In order to target the The fusion weights are calculated from the energy efficiency proximity results; the deviation correction process includes constructing a correction compensation function based on the deviation distribution of the energy efficiency proximity set, and applying the correction compensation function to the preliminary fusion results to obtain the corrected fusion results. The formula for the correction compensation function is: ； in: This is the deviation compensation amount; This is the deviation amplification factor calculated based on the distribution characteristics of the energy efficiency proximity set; the corrected fusion result is... express; Before performing robust fusion processing on the energy efficiency similarity set, the energy efficiency similarity set is subjected to perturbation sensitivity screening through the following operation: Based on the correspondence between the weighted perturbation vector and the energy efficiency closeness result, the perturbation sensitivity of each energy efficiency closeness result is calculated; all energy efficiency closeness results are sorted according to the perturbation sensitivity; and energy efficiency closeness results with perturbation sensitivity exceeding the preset threshold range are removed from the energy efficiency closeness set.

2. The method for robust energy efficiency assessment of thermal power plants according to claim 1, characterized in that, The temporal feature extraction includes: Trend analysis is performed on the change sequence of multidimensional operating parameter data within a continuous evaluation period to obtain long-term change characteristics; Based on the aforementioned long-term variation characteristics, fluctuation analysis is performed on the change sequence of multidimensional operating parameter data on a short time scale to obtain short-term fluctuation characteristics. The multidimensional operating parameter data is divided into intervals based on the rate of change and amplitude characteristics of the multidimensional operating parameter data, and the corresponding long-term change characteristics and short-term fluctuation characteristics are obtained in the intervals respectively; the long-term change characteristics and short-term fluctuation characteristics obtained in each interval are used as the interval change characteristics. For interval segments that do not meet the condition of continuity of change, abnormal segments are identified and removed through interval difference analysis; The collaborative relationship characteristics between operating parameters are obtained based on the time series correlation calculation results of multidimensional operating parameter data; Long-term change characteristics, short-term fluctuation characteristics, interval change characteristics, and collaborative relationship characteristics obtained from different intervals are aligned according to consistency constraint rules. According to the feature combination rules, the long-term change features, short-term fluctuation features, interval change features and cooperative relationship features after alignment are combined to generate time-series features for dynamic weight determination.

3. The method for robust energy efficiency assessment of thermal power plants according to claim 1, characterized in that, The steps for robustly fusing the energy efficiency similarity set are as follows: Numerical consistency verification is performed on each energy efficiency closeness result in the energy efficiency closeness set, and energy efficiency closeness results that do not meet the numerical consistency conditions are removed. The remaining energy efficiency similarity results are weighted and synthesized according to the preset fusion weighting rules to obtain the preliminary fusion results; The preliminary fusion results are subjected to deviation correction processing based on the distribution characteristics of the energy efficiency closeness set to obtain the corrected fusion results; The corrected fusion result is used as the comprehensive energy efficiency evaluation value of the operating unit.

4. The method for robust energy efficiency assessment of thermal power plants according to claim 1, characterized in that, In determining the energy efficiency level of the operating unit based on the comprehensive energy efficiency evaluation value, the energy efficiency level is determined through the following operations: A set of energy efficiency level thresholds is constructed based on the comprehensive energy efficiency evaluation values ​​of multiple historical operating units; The overall energy efficiency evaluation value is compared with the threshold values ​​of each level in the set of energy efficiency level thresholds; Based on the comparison results, select the energy efficiency level corresponding to the level threshold that matches the comprehensive energy efficiency evaluation value.

5. The method for robust energy efficiency assessment of thermal power plants according to claim 4, characterized in that, The process of constructing the energy efficiency level threshold set includes the following operations: A statistical data set describing the distribution characteristics of the comprehensive energy efficiency evaluation values ​​is generated based on the comprehensive energy efficiency evaluation values ​​of multiple historical operating units; the multiple historical operating units are clustered according to the statistical data set to obtain multiple energy efficiency level categories; the range of comprehensive energy efficiency evaluation values ​​in each energy efficiency level category is used as the energy efficiency level threshold range, and multiple energy efficiency level threshold ranges are combined into an energy efficiency level threshold set.

6. A robust energy efficiency assessment system for thermal power plants, applied to the robust energy efficiency assessment method for thermal power plants as described in any one of claims 1-5, characterized in that, The system includes: The time-series feature extraction module is used to collect multi-dimensional operating parameter data of thermal power plant operating units and extract time-series features to obtain time-series features that can reflect changes in operating status. A dynamic weight determination module is used to determine the dynamic weights of multiple energy efficiency indicators based on the time-series characteristics. The weighted perturbation proximity module is used to construct a weighted perturbation set based on the dynamic weights, and to calculate the energy efficiency proximity set of the operating unit under multiple weighted perturbation scenarios. The robust fusion module is used to robustly fuse the energy efficiency similarity set to obtain the comprehensive energy efficiency evaluation value of the operating unit; The energy efficiency rating determination module is used to determine the energy efficiency rating of the operating unit based on the comprehensive energy efficiency evaluation value.

Images