Distributed resource electric quantity trusted measurement evaluation method based on block chain technology

Abstract

A distributed resource electric quantity trusted measurement evaluation method based on a block chain technology relates to the technical field of electric energy quality monitoring, and specifically comprises the following steps: acquiring real-time data of distributed power supply equipment; preprocessing the obtained real-time data, and then uploading the data to a block chain node; the data receiver verifies the data of the block chain node by adopting a Byzantine fault-tolerant consensus; if verification fails, data acquisition, data preprocessing, uploading to the block chain node and data verification are performed again; if the verification is passed, the data of the block chain node is encrypted and transmitted to the electric quantity evaluation model through a transmission protocol; performing trusted measurement evaluation on the data by the electric quantity evaluation model to obtain a trusted measurement score; converting the trusted measurement score into a trusted measurement grade; corresponding processing strategies are executed according to different trusted measurement levels; the method is used for solving the problems that traditional evaluation training is high in cost and an evaluation result is inaccurate.

Description

Technical Field

[0001] The present invention relates to the technical field of power quality monitoring, and in particular to a distributed resource electricity quantity trustworthy measurement evaluation method based on blockchain technology. Background Art

[0002] With the transformation of my country's energy development strategy, the proportion of distributed resources such as solar photovoltaic panels, small wind turbines, and energy storage systems in the power system is rapidly increasing. Distributed resources are gradually developing towards a multi-energy complementary and integrated energy system, and will become the mainstay of my country's energy structure in the future. However, distributed resources currently face challenges such as grid connection difficulties, low power reliability, difficulty in data collection, and uneven data quality. Researching the power reliability of distributed resources and developing power reliability measurement standards are of great significance to improving power system reliability and enhancing the grid's ability to accommodate distributed resources.

[0003] Electricity reliability refers to the authenticity, accuracy, and reliability of electricity data in the power system. It is a key indicator for measuring the trustworthiness of electricity data and its acceptance and reliance by relevant stakeholders such as power companies, users, and regulators. Electricity reliability primarily assesses the ability of distributed resources to reliably provide electricity within a specific time period. Its level directly impacts the operational efficiency of the power system, the fairness of market transactions, and user trust in electricity services. Assessing the reliability of distributed resource electricity requires not only considering the traditional accuracy of electricity data calculation but also comprehensively considering the impact of various uncertainties on electricity data.

[0004] Assessing the reliability of electricity typically involves multiple aspects, including data sources, collection methods, transmission processes, processing algorithms, and data quality assessment. Distributed resources like wind and photovoltaic power are typically highly affected by weather, resulting in fluctuating output. Trustworthiness metrics, using historical data and probabilistic models, quantify the impact of risks like extreme weather and equipment failures on electricity output by measuring the reliability of the "actually available electricity." This converts the intermittent output of distributed resources into the equivalent stable output of traditional thermal power units, aiding grid planning and scheduling. For example, the "reliable electricity" of 100MW of photovoltaic power may be equivalent to 30MW of thermal power.

[0005] Currently, there are two main approaches for assessing the reliability of distributed resource electricity. One is a machine learning-based approach, which only requires input of time-series electricity characteristics such as mean and variance, and then calculates the output electricity data reliability score using algorithms such as LightGBM. However, this method has high training costs and relies on high-quality labeled data. Moreover, the black-box nature of the model leads to poor interpretability. The other is standardization. For example, the US NERC regulates data quality through standards such as PRC-002-2, the European ENTSO-E promotes data sharing, and China has formulated the "Electricity Data Quality Verification and Assessment Standard" to improve electricity reliability by standardizing electricity data. However, in practical applications, this type of approach often faces problems such as uneven standard enforcement and a lack of cross-regional coordination mechanisms.

[0006] Therefore, we propose a method with accurate evaluation results and low training cost. Summary of the Invention

[0007] The purpose of the present invention is to provide a distributed resource power credibility measurement evaluation method based on blockchain technology, which is used to solve the problems of high training cost and inaccurate evaluation results in traditional evaluation.

[0008] The present invention is achieved through the following technical solutions:

[0009] The distributed resource power trust evaluation method based on blockchain technology specifically includes:

[0010] Obtain real-time data of distributed power supply equipment;

[0011] Preprocess the acquired real-time data and then upload it to the blockchain node;

[0012] The data receiver uses the Byzantine fault-tolerant consensus to verify the data of the blockchain node;

[0013] If the verification fails, the data will be acquired again, pre-processed, uploaded to the blockchain node, and verified;

[0014] If the verification is successful, the data of the blockchain node is encrypted and transmitted to the power evaluation model through the transmission protocol;

[0015] The power evaluation model performs a credibility evaluation on the data to obtain a credibility score;

[0016] Converting the credibility metric scores into credibility metric levels;

[0017] Execute corresponding disposal strategies according to different trust measurement levels.

[0018] Furthermore, the power evaluation model includes data quality dimension, system operation dimension, user behavior dimension and environmental factor dimension;

[0019] The data quality dimensions include the acquisition accuracy deviation rate index, the transmission packet loss rate index, the time series continuity index and the outlier ratio index;

[0020] The system operation dimensions include load fluctuation coefficient index, equipment reliability index index, network loss rate index and voltage qualification rate index;

[0021] The user behavior dimensions include a power consumption pattern similarity index, a demand deviation index, and an abnormal power consumption probability index;

[0022] The environmental factor dimensions include a temperature sensitivity coefficient index, a humidity impact index, and a seasonal adjustment factor index.

[0023] Furthermore, the calculation formula of the acquisition accuracy deviation rate index is:

[0024]

[0025] In the formula is the sample mean; μ is the population mean; σ is the sample standard deviation; n is the sample size;

[0026] The calculation formula of the transmission packet loss rate indicator is:

[0027]

[0028] Where λ is the average packet loss rate, and k is the actual number of packet losses;

[0029] The calculation formula of the temporal continuity index is:

[0030]

[0031] Where ρ is the autocorrelation coefficient, is the sample mean, n is the total length of the time series, k is the lag order, X t is the observed value of the time series at time t;

[0032] The specific steps for calculating the outlier ratio index are as follows:

[0033] Outlier value range:

[0034] Q3+1.5IQR<x<Q1-1.5IQR

[0035] Obtain the range of non-outlier values ​​based on outlier values:

[0036] Q3+1.5IQR≥X≥Q1-1.5IQR;

[0037] Calculate the proportion of non-outliers:

[0038]

[0039] Calculate the proportion of outliers:

[0040] x η =1-X η ;

[0041] Among them, x represents an outlier, x η represents the proportion of outliers; X represents non-outliers, X η Indicates the proportion of non-outliers; Q1 and Q3 are the lower quartile and upper quartile of the data respectively; IQR = Q3-Q1 represents the range of the middle 50% of the data.

[0042] Furthermore, the calculation formula of the load fluctuation coefficient index is:

[0043]

[0044] Where, σ is the standard deviation of load data; μ is the mean of load data;

[0045] The calculation formula of the equipment reliability index is:

[0046]

[0047] Where η is the scale parameter; β is the shape parameter, and t is the equipment operation time;

[0048] The calculation formula of the network loss rate index is:

[0049] y=β0+β1x1+β2x2+...+β n x n +ε;

[0050] Where β0 is the intercept term; β i is the regression coefficient, reflecting the marginal impact of each independent variable on the network loss rate; x i is the influencing factor (independent variable); ε is a random error variable that is independent of each other and obeys the same normal distribution N(0,σ);

[0051] The calculation formula of the voltage qualification rate index is:

[0052]

[0053] Where μ is the process mean; σ is the process standard deviation; (USL, LSL) are the upper and lower limits of voltage compliance.

[0054] Furthermore, the calculation formula of the power consumption pattern similarity index is:

[0055]

[0056] Where x i represents the feature vector of the i-th user, μ k represents the center of the kth cluster, r ik is an indicator variable, if x i ∈k is 1, otherwise it is 0;

[0057] The calculation formula of the demand deviation index indicator is:

[0058]

[0059] Where (x-μ) represents the difference vector of two data points in each dimension; Σ -1 Represents the inverse matrix of the covariance matrix of the data set;

[0060] The calculation formula of the abnormal power consumption probability index is:

[0061]

[0062] Where β represents the weight vector; β0 represents the bias term, and X is the electricity consumption feature vector.

[0063] Furthermore, the calculation formula of the temperature sensitivity coefficient index is:

[0064]

[0065] The variable x i is the i-th temperature observation; variable y i is the i-th electricity consumption observation value, and are the means of variables x and y respectively;

[0066] The calculation formula of the humidity impact index is:

[0067] Y=XB T +E;

[0068] Where X represents the score matrix of the independent variable x, B represents the score matrix of the dependent variable y, and E represents the residual matrix;

[0069] The calculation formula of the seasonal adjustment factor index is:

[0070]

[0071] Where, T t Represents the “trend-cycle” component; S t represents the seasonal component; k represents the window length; ω i Represents the symmetric weight coefficient; I i Represents irregular components; εt Represents the random error term in the ARI MA model.

[0072] Furthermore, the power evaluation model performs a credibility evaluation on the data to obtain a credibility score, specifically in the following steps:

[0073] The weight ratio of the four dimensions in the power evaluation model is defined as 4:3:2:1;

[0074] The indicator weight allocation algorithm is used to allocate weights to the indicators of each dimension;

[0075] Calculate the value of the indicator in each dimension and normalize it. Multiply the normalized value by the corresponding indicator weight and add them together to obtain the dimension value.

[0076] Multiply the value of each dimension by the corresponding dimension weight and add them together to obtain the credibility measurement score.

[0077] Furthermore, the indicator weight allocation algorithm is specifically as follows:

[0078] The weighted regression model is used to define the indicator weights:

[0079]

[0080] Among them, ω i is the weight of the corresponding indicator, S i Score historical credibility; E i is the environmental stability factor; γ is the attenuation coefficient;

[0081] The weighted least squares method is used to correct the weight distribution results:

[0082]

[0083] where y i is the original calculation result; x ik is the independent variable; β is the regression coefficient.

[0084] Furthermore, the credibility measurement levels include level A, level B, level C and level D, where the scoring range of level A is (credibility score ≥ 90%); the scoring range of level B is (75% ≤ credibility score ≤ 90%); the scoring range of level C is (60% ≤ credibility score ≤ 75%); and the scoring range of level D is (credibility score < 60%).

[0085] Furthermore, the treatment strategy corresponding to the level A is that no correction is required; the treatment strategy corresponding to the level B is that local noise exists and correction is required; the treatment strategy corresponding to the level C is that re-examination is required; and the treatment strategy corresponding to the level D is direct elimination.

[0086] The technical solution of the present invention has at least the following advantages and beneficial effects:

[0087] The present invention discloses a distributed resource electricity credibility measurement evaluation method based on blockchain technology. By transmitting data through blockchain technology, the security of data transmission can be improved, thereby ensuring the integrity of the data. In addition, through the credibility measurement scoring model, the distributed resource electricity can be integrated to conduct a comprehensive evaluation, thereby improving the accuracy of the distributed resource electricity credibility measurement evaluation.

[0088] In addition, the trust metric scoring model adopts a combination of dimensions and indicators, which can reduce the complexity of the model and thus reduce the training cost of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 A schematic flow chart of a method of the present invention is shown;

[0090] Figure 2 A schematic diagram of the system structure of the present invention;

[0091] Figure 3 It is a structural schematic diagram of the electronic device in the present invention. DETAILED DESCRIPTION

[0092] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0093] Example 1

[0094] like Figure 1 The distributed resource power trust measurement evaluation method based on blockchain technology shown in the figure specifically includes:

[0095] Obtain real-time data of distributed power supply equipment;

[0096] The distributed power supply equipment acquired includes wind power equipment, photovoltaic equipment, energy storage equipment and electric vehicles. The specific data acquired includes but is not limited to the voltage, current, frequency, active / reactive power, power factor and power generation data of wind power equipment; the DC side voltage and current, AC side power, power generation, light intensity and component temperature data of photovoltaic equipment; the battery state of charge (SOC, remaining power percentage), health state (SOH, battery life), charge and discharge cycle number, charge and discharge power and efficiency data of energy storage equipment; and the charging power, charging capacity, charging start and end time and remaining charging time data of electric vehicles;

[0097] It should be noted that the above real-time data is obtained using smart meters.

[0098] Preprocess the acquired real-time data and then upload it to the blockchain node;

[0099] The preprocessing process includes:

[0100] Data cleaning, used to remove duplicate data, fill missing values, and correct erroneous data;

[0101] Data standardization, which is used to convert data collected by different devices or sensors into a unified format or unit;

[0102] Each block also contains the Merkle root hash of the meter data and the smart contract execution result function. The pre-processed data is divided into blocks, a Merkle tree is generated, and the root hash is calculated. The Merkle root hash of the meter data is stored in each block to ensure that the data cannot be tampered with. The data packet is then digitally signed using the device's private key to verify the legitimacy of the data source. Furthermore, the smart contract automatically executes pre-set rules (such as voltage range verification and timestamp validity), and data that fails verification is directly rejected from being uploaded to the blockchain node.

[0103] The data receiver uses the Byzantine fault-tolerant consensus to verify the data of the blockchain node;

[0104] The data receiver can be a node of a power grid company, and then the Byzantine fault-tolerant consensus algorithm is run for verification.

[0105] If the verification fails, the data will be acquired again, pre-processed, uploaded to the blockchain node, and verified;

[0106] If the verification is successful, the data of the blockchain node is encrypted and transmitted to the power evaluation model through the transmission protocol;

[0107] The encryption process is to encrypt the data transmission channel through the TLS protocol to ensure that the data is not stolen or tampered with during transmission.

[0108] The power evaluation model performs a credibility evaluation on the data to obtain a credibility score;

[0109] Converting the credibility metric scores into credibility metric levels;

[0110] Execute corresponding disposal strategies according to different trust measurement levels.

[0111] Example 2

[0112] As an embodiment, the power evaluation model includes data quality dimension, system operation dimension, user behavior dimension and environmental factor dimension;

[0113] The data quality dimensions include the acquisition accuracy deviation rate index, the transmission packet loss rate index, the time series continuity index and the outlier ratio index;

[0114] The system operation dimensions include load fluctuation coefficient index, equipment reliability index index, network loss rate index and voltage qualification rate index;

[0115] The user behavior dimensions include a power consumption pattern similarity index, a demand deviation index, and an abnormal power consumption probability index;

[0116] The environmental factor dimensions include a temperature sensitivity coefficient index, a humidity impact index, and a seasonal adjustment factor index.

[0117] In addition, the acquisition accuracy deviation rate index is calculated using a hypothesis test (Z test), and the specific calculation formula is:

[0118]

[0119] In the formula is the sample mean, where the sample refers to a certain part of the electricity data collected by the smart meter. For example, the smart meter collects electricity data 100 times in total, and 10 of them are selected as samples; μ is the overall mean; σ is the sample standard deviation; n is the sample size; the Z value is used to judge whether the collected data deviates significantly from the expectation, thereby evaluating the accuracy.

[0120] The transmission packet loss rate indicator is calculated using Poisson distribution fitting, and the specific calculation formula is:

[0121]

[0122] Where λ is the average number of packet losses and k is the actual number of packet losses.

[0123] The time series continuity index is calculated using autocorrelation analysis, and the specific calculation formula is:

[0124]

[0125] Where ρ is the autocorrelation coefficient, which measures the continuity of electricity data in the time series and has a value range of [-1, 1]. If it is close to 1, it means that the data is highly correlated in time and has good continuity; if it is close to 0 or a negative value, it indicates that the data may be interrupted or irregular. is the sample mean, which refers to the mean of the time series, that is, n is the total length of the time series, k is the lag order (k = 1 indicates the correlation between adjacent moments), X t is the observed value of the time series at time t.

[0126] The outlier ratio indicator is analyzed using the box plot method, and the specific calculation formula is:

[0127] Outlier value range:

[0128] Q3+1.5IQR<x<Q1-1.5IQR

[0129] Obtain the range of non-outlier values ​​based on outlier values:

[0130] Q3+1.5IQR≥X≥Q1-1.5IQR;

[0131] Calculate the proportion of non-outliers:

[0132]

[0133] Calculate the proportion of outliers:

[0134] x η =1-X η ;

[0135] Among them, x represents an outlier, x η represents the proportion of outliers; X represents non-outliers, X η Indicates the proportion of outliers; Q1 and Q3 are the lower quartile and upper quartile of the data respectively; IQR = Q3-Q1 represents the range of the middle 50% of the data.

[0136] In the system operation dimension, the load fluctuation coefficient index is calculated using the coefficient of variation method. The specific calculation formula is:

[0137]

[0138] Where, σ is the standard deviation of the load data; μ is the mean of the load data.

[0139] The equipment reliability index is calculated using Weibull distribution, and the specific calculation formula is:

[0140]

[0141] Where η is the scale parameter that determines the degree of “stretching” of the distribution; β is the shape parameter used to control the distribution shape; t is the equipment operation time;

[0142] The network loss rate index is analyzed using a regression model, and the specific calculation formula is:

[0143] y=β0+β1x1+ε;

[0144] Where β0 is the intercept term, which represents the baseline network loss rate when all independent variables are zero; β iis the regression coefficient, reflecting the marginal impact of each independent variable on the network loss rate; x i are influencing factors (independent variables), such as load current, voltage, line resistance, ambient temperature, and transmission distance in the power system; ε is a random error variable that is independent of each other and obeys the same normal distribution N(0,σ).

[0145] The voltage qualification rate index is evaluated using the process capability index, and the specific calculation formula is:

[0146]

[0147] Where μ is the process mean (voltage mean), which represents the voltage correction process; σ is the process (voltage) standard deviation, which is used to quantify process volatility; and (USL, LSL) are the upper and lower limits of voltage compliance.

[0148] In the user behavior dimension, the power consumption pattern similarity index is analyzed using cluster analysis (K-means), and the specific calculation formula is:

[0149]

[0150] Where x i represents the feature vector of the i-th user, μ k represents the center of the kth cluster, r ik is an indicator variable, if x i ∈k is 1, otherwise it is 0;

[0151] The demand deviation index is statistically analyzed using Mahalanobis distance, and the specific calculation formula is:

[0152]

[0153] Where (x-μ) represents the difference vector of two data points in each dimension, reflecting the direct difference; Σ -1 Represents the inverse matrix of the covariance matrix of the data set;

[0154] The abnormal power consumption probability index is calculated using the logistic regression method, and the specific calculation formula is:

[0155]

[0156] Where β represents the weight vector; β0 represents the bias term; X=[x1,x2,...,x n ] is the electricity consumption characteristic vector, such as x1 is the electricity consumption in the time period, x2 is the electricity consumption fluctuation rate, x3 is the historical deviation in the same period, and so on.

[0157] In the environmental factor dimension, the temperature sensitivity coefficient index is analyzed using the Pearson correlation coefficient, and the specific calculation formula is:

[0158]

[0159] The variable x i is the i-th temperature observation; variable y i is the i-th electricity consumption observation value, and are the means of variables x and y respectively, and the calculated result r ranges from -1 to 1, with positive values ​​indicating positive correlation and negative values ​​indicating negative correlation. The larger the absolute value, the stronger the correlation.

[0160] The humidity impact index adopts the partial least squares regression method, and the specific calculation formula is:

[0161] Y=XB T +E;

[0162] Where X represents the score matrix of the independent variable x (including humidity, temperature, wind speed and other environmental parameters), B represents the score matrix of the dependent variable y (such as electricity consumption, equipment power, etc.), and E represents the residual matrix;

[0163] The seasonal adjustment factor index is analyzed using the X-13ARIMA (seasonal adjustment model), and the specific calculation formula is:

[0164]

[0165] Where Y t Represents the observation value of the original time series at time t; T t Represents the “trend-cycle” component, reflecting long-term changes; S t represents the seasonal component, reflecting periodic fluctuations; k represents the window length; ω i Represents a symmetric weight coefficient to ensure smoothness; I i Represents irregular components, reflecting the impact of random noise or sudden events; ε t Represents the random error term in the ARIMA model, which obeys independent and identical distribution.

[0166] Example 3

[0167] As an embodiment, the power evaluation model performs a credibility metric evaluation on the data to obtain a credibility metric score, and the specific steps are as follows:

[0168] The weight ratio of the four dimensions in the power evaluation model is defined as 4:3:2:1, that is, the weight ratio of the data quality dimension, system operation dimension, user behavior dimension and environmental factor dimension is 4:3:2:1;

[0169] The indicator weight allocation algorithm is used to allocate weights to the indicators of each dimension, that is, the indicator weights allocated to each indicator;

[0170] In addition, since the power evaluation model uses a strategy of simultaneously calculating multi-dimensional data, it combines real-time monitoring data, equipment operation status, temperature and humidity and other environmental factors, uses a weighted regression model and weighted least squares (WLS) to correct the weight distribution results, which can improve the credibility of the results.

[0171] The indicator weight allocation algorithm is specifically as follows:

[0172] Define weight distribution to quantify the impact of each device on the final power calculation results in different environments. Use a weighted regression model to define indicator weights:

[0173]

[0174] Among them, ω i is the weight of the corresponding indicator, S i is the historical credibility score, which can be calculated by the data consistency, abnormal frequency and equipment health status (such as failure rate) within a sliding time window (such as 30 days), and the value range is [0,1]. i is the environmental stability factor, which is used to quantify the impact of the environment on the data collected by the device, such as the temperature fluctuation rate or the humidity coefficient of variation; γ is the attenuation coefficient, which is used to control the suppression strength of environmental fluctuations on the weight (the default value is γ = 0.1). It is adaptively adjusted through Bayesian optimization to avoid overfitting;

[0175] The weighted least squares method is used to correct the weight distribution results. The goal of the weighted least squares model is set to minimize the weighted residual sum of squares, specifically:

[0176]

[0177] where y i is the original calculation result; x ik are independent variables, including equipment capacity (x i1 ), environmental parameters (temperature x i2 , humidity x i3 etc.); β is the regression coefficient, which represents the joint impact of various factors on the amount of electricity;

[0178] Where Y={y i}={y1,y2,...,y i} is the original calculation result (such as power load value);

[0179] is the independent variable matrix, each row corresponds to an observation sample (such as data at a certain moment), and each column corresponds to a feature (including the intercept term); β is the regression coefficient, which represents the influence weight of each independent variable on the dependent variable.

[0180] It should be noted that one indicator weight combination calculated according to the indicator weight allocation algorithm is:

[0181] Acquisition accuracy deviation rate index (35%), transmission packet loss rate index (30%), time series continuity index (20%), and abnormal value ratio index (15%);

[0182] Load fluctuation coefficient index (30%), equipment reliability index (25%), network loss rate index (25%), voltage qualification rate index (20%);

[0183] Power consumption pattern similarity index (40%), demand deviation index (35%), abnormal power consumption probability index (25%);

[0184] Temperature sensitivity coefficient index (45%), humidity impact index (30%), and seasonal adjustment factor index (25%).

[0185] Calculate the value of the indicator in each dimension and normalize it. Multiply the normalized value by the corresponding indicator weight and add them together to obtain the dimension value.

[0186] It's important to note that the normalization process includes minimum-maximum normalization, which maps the indicator's value to the [0, 1] range, or Z-score normalization, which normalizes based on the mean and standard deviation. The data quality dimension includes the collection accuracy deviation rate, transmission packet loss rate, and outlier ratio; the system operation dimension includes the network loss rate; and the user behavior dimension includes the demand deviation index. Z-score normalization is used for the temporal continuity of the data quality dimension, the load fluctuation coefficient and equipment reliability index of the system operation dimension, the power consumption pattern similarity and abnormal power consumption probability of the user behavior dimension, and the temperature sensitivity coefficient, humidity impact, and seasonal adjustment factor of the environmental factor dimension.

[0187] Furthermore, min-max normalization includes the following steps:

[0188] Define the minimum value x for each indicator based on historical data or theoretical range min and the maximum value x max , and then calculate the normalized value, the formula is as follows

[0189]

[0190] Where x is the original indicator value.

[0191] Z-score normalization involves the following steps, as needed:

[0192] Based on historical data, calculate the mean μ and standard deviation σ of each indicator, and then calculate the normalized value. The formula is as follows

[0193]

[0194] In addition, the comprehensive score range is adjusted to [0,100] after normalization, which directly corresponds to the credibility percentage and is used for subsequent level division (Level A, B, C, D).

[0195] Multiply the value of each dimension by the corresponding dimension weight and add them together to obtain the credibility measurement score.

[0196] According to the above weight distribution combination, the score of the data quality dimension can be calculated as

[0197] S 数据质量 =(0.35×x 采集精度 )+(0.3×x 传输丢包 )+(0.2×x 时序连续性 )+(0.15×x 异常值占比 )

[0198] Similarly, the scores for the system operation, user behavior, and environmental factors dimensions are calculated separately.

[0199] The comprehensive score is generated as follows:

[0200] The scores of the four dimensions are weighted and summed according to the total weight (40%, 30%, 20%, 10%) to obtain the final comprehensive score:

[0201] S=0.4×S 数据质量 +0.3×S 系统运行 +0.2×S 用户行为 +0.1×S 环境因素

[0202] Example 4

[0203] As an embodiment, the credibility measurement levels include level A, level B, level C and level D, wherein the scoring range of level A is (credibility score ≥ 90%); the scoring range of level B is (75% ≤ credibility score ≤ 90%); the scoring range of level C is (60% ≤ credibility score ≤ 75%); and the scoring range of level D is (credibility score < 60%).

[0204] According to needs, the treatment strategy corresponding to the level A is that no correction is required; the treatment strategy corresponding to the level B is that local noise exists and correction is required; the treatment strategy corresponding to the level C is that re-examination is required; and the treatment strategy corresponding to the level D is direct elimination.

[0205] Example 5

[0206] like Figure 2The distributed resource power trust measurement and evaluation system based on blockchain technology includes a data acquisition module, a blockchain node, a data reception and evaluation module, and a hierarchical disposal module.

[0207] The data acquisition module is used to acquire real-time data of the distributed power supply equipment and pre-process the real-time data;

[0208] The blockchain node is used to receive and store pre-processed real-time data;

[0209] The data reception and evaluation module is used to verify and receive data from blockchain nodes, and to perform a credibility evaluation on the data;

[0210] The hierarchical handling module is used to convert the credibility metric scores into credibility metric levels and execute corresponding handling strategies according to different credibility metric levels.

[0211] Example 6

[0212] As attached Figure 3 An electronic device as shown is characterized by comprising:

[0213] Processor, memory, communication interface;

[0214] The memory is used to store executable instructions of the processor;

[0215] The processor is configured to execute the above-mentioned distributed resource power credibility measurement evaluation method based on blockchain technology by executing the executable instructions.

[0216] A readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the distributed resource power trust measurement evaluation method based on blockchain technology is implemented.

[0217] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A distributed resource power trust evaluation method based on blockchain technology is characterized by: Specifically include: Obtain real-time data of distributed power supply equipment; Preprocess the acquired real-time data and then upload it to the blockchain node; The data receiver uses the Byzantine fault-tolerant consensus to verify the data of the blockchain node; If the verification fails, the data will be acquired again, pre-processed, uploaded to the blockchain node, and verified; If the verification is successful, the data of the blockchain node is encrypted and transmitted to the power evaluation model through the transmission protocol; The power evaluation model performs a credibility evaluation on the data to obtain a credibility score; Converting the credibility metric scores into credibility metric levels; Execute corresponding disposal strategies according to different trust measurement levels.

2. The distributed resource power credibility evaluation method based on blockchain technology according to claim 1 is characterized by: The power evaluation model includes data quality dimensions, system operation dimensions, user behavior dimensions, and environmental factor dimensions; The data quality dimensions include the acquisition accuracy deviation rate index, the transmission packet loss rate index, the time series continuity index and the outlier ratio index; The system operation dimensions include load fluctuation coefficient index, equipment reliability index index, network loss rate index and voltage qualification rate index; The user behavior dimensions include a power consumption pattern similarity index, a demand deviation index, and an abnormal power consumption probability index; The environmental factor dimensions include a temperature sensitivity coefficient index, a humidity impact index, and a seasonal adjustment factor index.

3. The distributed resource power credibility evaluation method based on blockchain technology according to claim 2 is characterized by: The calculation formula of the acquisition accuracy deviation rate index is: In the formula is the sample mean; μ is the population mean; σ is the sample standard deviation; n is the sample size; The calculation formula of the transmission packet loss rate indicator is: Where λ is the average packet loss rate, and k is the actual number of packet losses; The calculation formula of the temporal continuity index is: Where ρ is the autocorrelation coefficient, is the sample mean, n is the total length of the time series, k is the lag order, X t is the observed value of the time series at time t; The specific steps for calculating the outlier ratio index are as follows: Outlier value range: Q3+1.5IQR<x<Q1-1.5IQR Obtain the range of non-outlier values ​​based on outlier values: Q3+1.5IQR≥X≥Q1-1.5IQR; Calculate the proportion of non-outliers: Calculate the proportion of outliers: x η =1-X η ; Among them, x represents an outlier, x η represents the proportion of outliers; X represents non-outliers, X η Indicates the proportion of non-outliers; Q1 and Q3 are the lower quartile and upper quartile of the data respectively; IQR = Q3-Q1 represents the range of the middle 50% of the data.

4. The distributed resource power credibility evaluation method based on blockchain technology according to claim 2 is characterized by: The calculation formula of the load fluctuation coefficient index is: Where, σ is the standard deviation of load data; μ is the mean of load data; The calculation formula of the equipment reliability index is: Where η is the scale parameter; β is the shape parameter, and t is the equipment operation time; The calculation formula of the network loss rate index is: y=β0+β1x1+β2x2+...+β n x n +e; Where β0 is the intercept term; β i is the regression coefficient, reflecting the marginal impact of each independent variable on the network loss rate; x i is the influencing factor (independent variable); ε is a random error variable that is independent of each other and obeys the same normal distribution N(0,σ); The calculation formula of the voltage qualification rate index is: Where μ is the process mean; σ is the process standard deviation; (USL, LSL) are the upper and lower limits of voltage compliance.

5. The distributed resource power credibility evaluation method based on blockchain technology according to claim 2 is characterized by: The calculation formula of the power consumption pattern similarity index is: Where x i represents the feature vector of the i-th user, μ k represents the center of the kth cluster, r ik is an indicator variable, if x i ∈k is 1, otherwise it is 0; The calculation formula of the demand deviation index indicator is: Where (x-μ) represents the difference vector of two data points in each dimension; Σ -1 Represents the inverse matrix of the covariance matrix of the data set; The calculation formula of the abnormal power consumption probability index is: Where β represents the weight vector; β0 represents the bias term, and X is the electricity consumption feature vector.

6. The distributed resource power credibility evaluation method based on blockchain technology according to claim 2 is characterized by: The calculation formula of the temperature sensitivity coefficient index is: The variable x i is the i-th temperature observation; variable y i is the i-th electricity consumption observation value, and are the means of variables x and y respectively; The calculation formula of the humidity impact index is: Y=XB T +E; Where X represents the score matrix of the independent variable x, B represents the score matrix of the dependent variable y, and E represents the residual matrix; The calculation formula of the seasonal adjustment factor index is: Where, T t Represents the "trend-cycle" component; S t represents the seasonal component; k represents the window length; ω i Represents the symmetric weight coefficient; I i Represents irregular components; ε t Represents the random error term in the ARIMA model.

7. The distributed resource power credibility evaluation method based on blockchain technology according to claim 1 is characterized by: The power evaluation model performs a credibility evaluation on the data to obtain a credibility score, and the specific steps are as follows: The weight ratio of the four dimensions in the power evaluation model is defined as 4:3:2:1; The indicator weight allocation algorithm is used to allocate weights to the indicators of each dimension; Calculate the value of the indicator in each dimension and normalize it. Multiply the normalized value by the corresponding indicator weight and add them together to obtain the dimension value. Multiply the value of each dimension by the corresponding dimension weight and add them together to obtain the credibility measurement score.

8. The distributed resource power credibility evaluation method based on blockchain technology according to claim 7 is characterized by: The indicator weight allocation algorithm is specifically as follows: The weighted regression model is used to define the indicator weights: Among them, ω i is the weight of the corresponding indicator, S i Score historical credibility; E i is the environmental stability factor; γ is the attenuation coefficient; The weighted least squares method is used to correct the weight distribution results: where y i is the original calculation result; x ik is the independent variable; β is the regression coefficient.

9. The distributed resource power credibility evaluation method based on blockchain technology according to claim 1 is characterized by: The credibility measurement levels include level A, level B, level C and level D, where the scoring range of level A is (credibility score ≥ 90%); the scoring range of level B is (75% ≤ credibility score ≤ 90%); the scoring range of level C is (60% ≤ credibility score ≤ 75%); and the scoring range of level D is (credibility score < 60%).

10. The distributed resource power credibility evaluation method based on blockchain technology according to claim 9 is characterized by: The treatment strategy corresponding to the level A is that no correction is required; the treatment strategy corresponding to the level B is that local noise exists and correction is required; the treatment strategy corresponding to the level C is that re-examination is required; and the treatment strategy corresponding to the level D is direct elimination.

Images