Power grid provincial management industry equity investment risk assessment method

Abstract

The invention relates to the technical field of risk assessment, and discloses a power grid provincial management industry stock right investment risk assessment method, which comprises the following steps: a system identifies stock right investment risk types, determines main influence factors of different risks, constructs a GDP growth prediction model based on ARIMA, forms market risk quantitative assessment standard formulation, constructs a load prediction model based on SAO optimization LSTM, and provides a power grid provincial management industry stock right investment risk assessment system. The method comprises the following steps: establishing an operation risk level evaluation model based on Markov and LS-SVM, forming operation risk quantitative evaluation standard formulation, constructing an operation risk level evaluation model based on Markov and LS-SVM, realizing operation risk quantitative evaluation standard formulation, applying a neural network model, realizing scientific setting of market risk, operation risk and operation risk score weight, and obtaining a final risk evaluation result. According to the method, the actual current situation of the power grid equity investment is combined, the main risk type of the power grid equity investment is systematically evaluated through a series of advanced intelligent prediction methods, and therefore scientific and quantitative support is provided for investment decision making.

Description

Technical Field

[0001] The present invention relates to the technical field of risk assessment, and more specifically, to a method for assessing the risk of equity investment in a provincial power grid industry. Background Art

[0002] With the continuous development of the power industry and the gradual opening of the market, power grid equity investment has become an important investment area in the capital market. However, as an infrastructure related to social and economic development, the risk of investment decisions in power grids is relatively high, involving multiple aspects such as market, operation, and technology. Traditional risk assessment methods often rely on empirical judgments and lack scientific quantitative analysis, making it difficult to fully and accurately reflect the impact of various risks on investment results. Therefore, how to accurately assess the various risks in power grid equity investment and formulate reasonable risk management strategies has become a major challenge in current power grid investment decisions.

[0003] Existing grid investment risk assessment methods mainly focus on the assessment of a single risk type. For example, market risk usually relies on qualitative analysis of macroeconomic indicators (such as GDP, policy changes, etc.); operational risk is often predicted through historical analysis of grid load data, but most of these methods rely on simple statistical analysis and lack high-precision prediction models; and for operational risk, traditional fault prediction models or experience-based judgments are often used, lacking real-time and dynamic predictions of the grid operating status.

[0004] In order to meet this challenge, in recent years, with the rapid development of big data and artificial intelligence technology, more and more intelligent prediction and analysis methods have been applied to power grid investment risk assessment. For example, the ARIMA model is widely used in market risk prediction, the LSTM network has achieved remarkable results in time series load forecasting, and the Markov model and support vector machine (SVM) have also shown good adaptability in operational risk assessment. However, these methods still have problems such as poor model independence, inability to process multi-dimensional data, and difficulty in integrating different types of risk information. Therefore, a new comprehensive technology is urgently needed to conduct a comprehensive and accurate assessment of market risk, operational risk, and operation risk through multi-model fusion and deep learning, combined with the characteristics of power grid equity investment.

[0005] To this end, we propose a risk assessment method for equity investment in provincial-level power grid industries. Summary of the invention

[0006] The purpose of the present invention is to provide a method for evaluating the risk of equity investment in power grid provincial management industries to overcome the technical problems existing in the prior art.

[0007] In order to achieve the above technical objectives and the above technical effects, the present invention provides the following technical solutions:

[0008] A method for evaluating the risk of equity investment in a provincial power grid industry, the method comprising the following steps:

[0009] Step 1: Based on the actual status of power grid equity investment, systematically identify the types of equity investment risks and clarify the main influencing factors of different risks;

[0010] Step 2: Use the improved anti-entropy-grey correlation analysis method to determine the correlation between investment risk and each key influencing factor, and select the most important factors in different risk types as the basis for risk level determination;

[0011] Step 3: In view of market risk, we select GDP growth rate as the main discriminant factor, build a GDP growth forecasting model based on ARIMA, and formulate quantitative assessment standards for market risk;

[0012] Step 4: In view of operational risks, load growth is selected as the main discriminant factor, and a load forecasting model based on SAO optimized LSTM is constructed to form a quantitative assessment standard for operational risks;

[0013] Step 5: Aiming at the operation risk, a Markov and LS-SVM-based operation risk level assessment model is constructed to achieve the formulation of the quantitative assessment standard for operation risk;

[0014] Step six: Apply the neural network model to scientifically set the scoring weights for market risk, operational risk, and operation risk to obtain the final risk assessment results.

[0015] Preferably, in a method for assessing equity investment risks in a provincial power grid industry, the equity investment risk types in step 1 include market risk, operational risk and operation risk;

[0016] Factors influencing market risks include changes in electricity demand, GDP, changes in policies and regulations, fluctuations in energy prices, the improvement of electricity marketization, the emergence of external competitors, changes in the international market and adjustments in monetary policy;

[0017] The factors affecting operational risks include the failure rate of power grid equipment, overload of transmission lines, aging of equipment, aging degree of power grid facilities, construction of monitoring and early warning systems, safety management of power grids, enforcement of operating procedures, external attacks, network security issues and errors in power grid dispatching;

[0018] Factors affecting operational risks include load growth rate, equipment aging, maintenance, extreme weather caused by climate change, application of new technologies and personnel management.

[0019] Preferably, in a method for evaluating the risk of equity investment in provincial power grid industries, the application of the improved anti-entropy-grey correlation analysis method in step 2 specifically includes the following steps:

[0020] The entropy value is calculated for each key influencing factor to reflect the uncertainty of its information;

[0021] The entropy value is converted into the anti-entropy value through the anti-entropy formula, which reflects the certainty and difference of the factor information. The larger the anti-entropy value, the greater the weight given.

[0022] The grey correlation analysis method is used to calculate the correlation coefficient between each factor and the reference sequence, and then the correlation between each key influencing factor and the risk level is obtained;

[0023] Combining the anti-entropy value and grey correlation degree, the factors most relevant to the risk level are selected as the basis for determining the risk level.

[0024] Preferably, in a method for assessing the risk of equity investment in a provincial power grid industry, the GDP growth forecasting model in step 3 is adjusted by three main parameters p, d, and q, wherein p represents the number of autoregressive terms, d represents the number of differences, and q represents the number of sliding average terms. The formulation of the market risk quantitative assessment standard includes the following steps:

[0025] Use historical GDP data to stabilize the series to ensure that the data meets the requirements of the ARIMA model;

[0026] Determine the most appropriate parameters (p, d, q), using the autocorrelation function (ACF) and partial autocorrelation function (PACF) plots to assist in the selection;

[0027] Use the selected ARIMA model for training to fit the trend of GDP growth;

[0028] Based on the model's forecast results, we develop quantitative assessment standards for market risks, and determine the different levels of market risks based on future trends in GDP growth.

[0029] Preferably, in a method for assessing the risk of equity investment in a provincial power grid industry, the formulation of the quantitative assessment standard for operational risk in step 4 includes the following steps:

[0030] Collect historical load data and pre-process it to ensure data is clean and normalized;

[0031] Use the SAO algorithm to optimize the structure and hyperparameters of the LSTM model and find the optimal configuration;

[0032] By training the optimized LSTM model, we can predict future load growth and obtain the load change trend in different time periods.

[0033] Based on the forecast results, formulate quantitative assessment standards for operational risks.

[0034] Preferably, in a method for assessing the risk of equity investment in a provincial power grid industry, the formulation of the quantitative assessment standard for operating risk in step five includes the following steps:

[0035] Markov chain model to evaluate the operation status of transmission lines;

[0036] The PCA algorithm extracts key influencing factors;

[0037] Least squares support vector machine (LS-SVM) is used to predict risk level;

[0038] Preferably, in a method for risk assessment of equity investment in a provincial power grid industry, the calculation of the risk assessment result in step six includes the following steps:

[0039] By collecting and integrating relevant data on market risk, operational risk and operation risk, the input features of the neural network model are constructed. These features include the predicted results of each risk and the relevant influencing factors.

[0040] Design the architecture of the neural network, select the appropriate number of layers and neurons, and use the back-propagation algorithm for training;

[0041] The model is trained through training data, and the weight of each risk score is adjusted to accurately reflect the contribution of each risk to the overall risk of power grid investment;

[0042] Using the trained neural network model, new risk data is input to calculate the final scores of market risk, operational risk and operation risk, and the comprehensive risk assessment results are calculated based on these scores.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] The present invention combines the actual status of power grid equity investment and systematically evaluates the main risk types of power grid equity investment through a series of advanced intelligent prediction methods, thereby providing scientific and quantitative support for investment decisions;

[0045] The present invention proposes a systematic, multi-dimensional and automated risk assessment method for power grid equity investment through the comprehensive application of improved anti-entropy-grey correlation analysis method, ARIMA model, SAO optimized LSTM model, Markov and LS-SVM model combined operation risk assessment method, and neural network model. This method can effectively identify and quantify different types of investment risks, provide a scientific and accurate investment decision support tool, and provide all-round risk management and control for power grid equity investment. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the description of the specific implementation methods will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 It is the overall flow chart of the present invention;

[0048] Figure 2 Schematic diagram of the LSTM neural network in the present invention;

[0049] Figure 3 It is a schematic diagram of the transmission line operation risk level in the present invention. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in 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. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] like Figure 1-3 The present embodiment is a method for evaluating the risk of equity investment in power grid provincial-level industries, and the method comprises the following steps:

[0052] Step 1: Based on the actual status of power grid equity investment, systematically identify the types of equity investment risks and further clarify the main influencing factors of different risks.

[0053] Market risks are closely related to changes in the macro-economy, policy environment and energy market. The market risks of power grid equity investment are affected by many factors, including changes in electricity demand, GDP, changes in policies and regulations, and energy price fluctuations. The improvement of the degree of marketization of electricity may change the flexibility of electricity prices and affect investment returns. The emergence of external competitors, changes in the international market and adjustments in monetary policy also have a significant impact on the risks of the power grid investment market.

[0054] Operational risk involves failures and emergencies during the actual operation of the power grid. Factors such as the failure rate of power grid equipment, overload of transmission lines, and aging of equipment directly determine the magnitude of operational risk. The aging degree of power grid facilities and the construction of monitoring and early warning systems also have an important impact on the occurrence of operational risks. The safety management of the power grid, the enforcement of operating procedures, and external attacks or network security issues may lead to emergencies, thereby increasing operational risks. Errors in power grid dispatching will also affect operational safety.

[0055] Operational risk focuses on the future profitability of the power grid, and load forecasting is the most important indicator for assessing this risk. The rate of load growth directly affects whether the power grid can operate efficiently and obtain sufficient revenue. Too fast a growth in load demand may cause the power grid to overload, thereby increasing the risk of equipment failure; while too slow a growth in load may affect the effective use of power grid assets and reduce profit potential. In addition, factors such as equipment aging, maintenance, extreme weather caused by climate change, the application of new technologies, and personnel management will also have an impact on the operational risk of the power grid. These factors together determine the operational efficiency and profitability of the power grid.

[0056] Step 2: Use the improved anti-entropy-grey correlation analysis method to determine the correlation between investment risk and each key influencing factor. Select the most important factors in different risk types as the basis for risk level determination.

[0057] The reasons for introducing anti-entropy are as follows: Enhance objectivity: Traditional grey correlation analysis adopts equal weighting, that is, it considers all key influencing factors to be equally important, which may lead to the neglect of key influencing factors' information. Anti-entropy method quantifies the information of key influencing factors and automatically adjusts the weights of key influencing factors, so that the analysis results can better reflect the actual situation; Reveal internal associations: Anti-entropy can highlight those key influencing factors with large variations and rich information, so as to give them more attention in correlation analysis, which is helpful to discover more refined correlation patterns; Reduce bias: When the correlation between key influencing factors is complex and scattered, anti-entropy-grey correlation analysis can effectively avoid the analysis bias caused by local correlation and improve the overall accuracy of the analysis.

[0058] The basic steps of the improved anti-entropy-grey correlation analysis method are as follows: Data preprocessing: Collect risk and its related key influencing factor data, perform necessary data cleaning, standardization or normalization processing to ensure the comparability between key influencing factors. Calculate the anti-entropy of key influencing factors: Calculate the entropy value of each key influencing factor to reflect the uncertainty of its information. Through the anti-entropy formula conversion, the higher the anti-entropy value, the more certain the information provided by the key influencing factor, the greater the difference, and the higher the weight should be given. Weight determination: Weight the key influencing factors based on the anti-entropy value. The weight reflects the relative importance of each key influencing factor in the analysis, overcoming the subjectivity of equal treatment. Construct reference sequence and comparison sequence: Usually, risk is selected as the reference sequence, and other key influencing factors are selected as the comparison sequence to construct the grey system model. Calculate the correlation coefficient and correlation degree: Apply the weighted difference sequence to calculate the correlation coefficient between each comparison sequence and the reference sequence, taking into account the weight determined by the anti-entropy. Calculate the grey correlation degree of each comparison sequence to obtain the correlation matrix.

[0059] In order to avoid the problem of weight imbalance caused by the excessive sensitivity of key influencing factors in the entropy weight method, the anti-entropy value method is used to objectively weight the key influencing factors. The specific steps are as follows:

[0060] (1) Construct the evaluation matrix of key influencing factors. Select n key influencing factors, k regions, and m years of data to establish the corresponding initial matrix, x ij is the value of the jth key influencing factor of the ith object, and the key influencing factor matrix X is evaluated. m for

[0061]

[0062] (2) Calculate the anti-entropy of each key influencing factor. Define the anti-entropy h′ as

[0063]

[0064] Among them, p j is the entropy value of the jth key influencing factor, 0≤p j ≤1, and

[0065] The anti-entropy h′ of each key influencing factor is

[0066]

[0067] Where: c ij is the entropy value of the jth key influencing factor under the i-th object,

[0068] (3) Normalization. The anti-entropy of each key influencing factor is further normalized to obtain:

[0069]

[0070] Where: d j is the objective weight of the jth key influencing factor.

[0071] In order to scientifically and rationally evaluate the impact of each key influencing factor on risk, the key influencing factors are combined and weighted as follows:

[0072]

[0073] Where: w j is the combined weight of the key influencing factor j; c j d j They are respectively the objective weight (objective weight calculated according to the entropy value) and subjective weight of the jth key influencing factor.

[0074] In power grid fault analysis and prediction, the combined weights of key influencing factors are crucial for accurately assessing the correlation between these factors and faults. The combined weight is the result of a reasonable integration of the subjective weight and objective weight of each key influencing factor, which can more comprehensively reflect the importance of each factor. The subsequent calculation of the correlation depends on these combined weights, reflecting the mutual relationship and interaction between the key influencing factors.

[0075] The fuzzy mathematics comprehensive evaluation method uses the ascending semi-trapezoidal distribution function as the membership function to fit and score each key influencing factor. In order to further quantify the correlation between the risk and its key influencing factors, the improved anti-entropy-grey correlation analysis method is used to determine the correlation between the two. Traditional grey correlation analysis establishes a correlation matrix by obtaining a differential sequence, and generally adopts equal weighting. This equal weighting method easily conceals the characteristic relationship between each key influencing factor and ignores the rich information contained in the key influencing factors themselves. Since the key influencing factors come from multiple aspects and levels, when the correlation relationship is relatively discrete, a trend of local correlation is likely to appear, resulting in analysis deviations. Therefore, anti-entropy is introduced on the basis of the entropy method to overcome the existing subjective problems. The specific steps are as follows. The calculation formula for the correlation between sequences in the traditional grey correlation analysis method is:

[0076]

[0077] Where: γ j is the correlation between power grid fault and key influencing factor j; rj is the correlation coefficient between the power grid fault and the key influencing factor j determined based on the combined weight, and the fixed weight value is 1 / n.

[0078] Define the residual degree of the jth key influencing factor as

[0079] T j =1-e j ;

[0080] Where: T j is the residual degree of the jth key influencing factor; e j is the ratio of the anti-entropy of the jth key influencing factor to the maximum entropy, that is, the relative entropy. The specific expression is

[0081]

[0082] Where: E j 、E max are the anti-entropy and maximum entropy of the jth key influencing factor respectively. j The larger the value is, the more obvious the characteristics of the key influencing factor are and the higher its importance is.

[0083] The proportional weight of other key influencing factors r is set to b(r), and the specific calculation formula is:

[0084]

[0085] Among them, T r is the residual degree of a key influencing factor r; replace 1 / n in the formula with b(r) to obtain the correlation degree γ between the risk and the key influencing factor j j The final expression is

[0086]

[0087] Step 3: In response to market risks, select GDP growth rate as the main discriminant factor, build a GDP growth forecast model based on ARIMA, and form a quantitative assessment standard for market risks.

[0088] When building a GDP growth forecast model based on ARIMA, the goal is to predict future GDP growth trends by analyzing historical GDP data, and to form a quantitative assessment standard for market risk based on this forecast. ARIMA (Autoregressive Integrated Moving Average) is a classic model for time series analysis, which is mainly used to forecast and model time series data, especially when there is no obvious seasonal fluctuation.

[0089] (1) Basic structure of the ARIMA model

[0090] The core of the ARIMA model consists of three parts:

[0091] Autoregressive (AR): AR models predict future values ​​by relying on past values, i.e., a linear relationship between current and past values.

[0092] Difference (I): By performing a difference operation on the time series, a non-stationary time series can be made stationary. Difference can eliminate the trend component.

[0093] Moving Average (MA): The MA model uses past error terms to predict current values.

[0094] The ARIMA model is usually expressed as ARIMA(p,d,q), where:

[0095] p is the order of the autoregressive term;

[0096] d is the number of differences;

[0097] q is the order of the sliding average term.

[0098] (2) The process of building an ARIMA model for GDP growth forecasting

[0099] Step 1: Data Preparation

[0100] Collect historical GDP data and calculate the GDP growth rate. The GDP growth rate is calculated using the following formula:

[0101]

[0102] Among them, GDP t represents the GDP in year t, GDP t-1 It is the GDP of the previous year.

[0103] Step 2: Stationarity test

[0104] The GDP growth rate series is tested for stationarity. The commonly used method is the ADF test (Augmented Dickey-Fuller Test) to determine whether the time series has a unit root, that is, whether it is a stationary series. If the series is not stationary, it needs to be differentiated to make it stationary.

[0105] Difference: Calculate the difference of time series data. The number of differences dd is usually 1 or 2 times until the series becomes stationary.

[0106] Step 3: Choose the order of the model

[0107] Use the autocorrelation function (ACF) and partial autocorrelation function (PACF) plots to determine the order pp and qq of the ARIMA model. Usually:

[0108] If the PACF graph is truncated at the pth order, it means that the order of the AR term is p;

[0109] If the ACF plot is truncated at the qth order, it means that the order of the MA term is q.

[0110] Step 4: Build an ARIMA model

[0111] After selecting appropriate p, d, q values, we use historical GDP growth data to fit the ARIMA model. Assuming we choose the ARIMA (1,1,1) model, the formula is:

[0112] Y t =α+φ 1 Y t-1 +θ 1 ∈ t-1 +∈ t

[0113] Y t is the GDP growth rate at time t;

[0114] α is a constant term;

[0115] φ 1 is the autoregressive coefficient;

[0116] θ 1 is the sliding mean coefficient;

[0117] ∈ t is the error term.

[0118] Step 5: Model Fitting and Prediction

[0119] By fitting the ARIMA model with historical data, we can get the model parameters (α, φ 1 ,θ 1 ), and use the model to predict GDP growth in the future. Assuming we need to predict the GDP growth rate in the next h period, we can use the following formula to predict:

[0120]

[0121] Evaluation criteria development:

[0122] High growth range (GDP growth rate greater than 6%): low market risk, assessed as Level 1 risk.

[0123] Medium-to-high growth range (GDP growth rate between 4% and 6%): medium market risk, assessed as level 2 risk.

[0124] Medium growth range (GDP growth rate between 2% and 4%): medium-to-high market risk, assessed as level 3 risk.

[0125] Stable growth range (GDP growth rate between 0% and 2%): higher market risk, assessed as level 4 risk.

[0126] Low growth range (GDP growth rate <0%): extremely high market risk, assessed as level 5 risk.

[0127] Step 4: In view of operational risks, load growth is selected as the main discriminant factor, and a load forecasting model based on SAO optimized LSTM is constructed to form a quantitative assessment standard for operational risks.

[0128] LSTM (Long Short-Term Memory Network) is a special recurrent neural network (RNN) suitable for processing and predicting time series data. Compared with traditional RNN, LSTM has stronger long-term memory ability and can effectively avoid the problem of gradient disappearance or explosion, so it is widely used in load forecasting.

[0129] SAO (Social-Aware Optimization) optimization is an optimization algorithm based on group intelligence, inspired by social group behavior (such as bird flocks foraging, fish schools swimming, etc.). In the LSTM model, the SAO optimization algorithm can be used to optimize the hyperparameters of LSTM (such as learning rate, hidden layer size, training cycle, etc.), thereby improving the prediction accuracy of the model.

[0130] Traditional neural networks have the problem of gradient vanishing due to the failure to consider the input at any previous time, which makes it difficult for the neural network to converge and the accuracy is low. Figure 2 The LSTM neural network shown is designed to effectively transmit and express information in long time series without causing useful information from a long time ago to be forgotten.

[0131] The input of the forget gate is the hidden state h of the previous moment t-1 and the current input x t , output f t Used to control which information in memory cells should be forgotten at a certain moment, f is the bias vector of the forget gate. The relationship between input and output is as follows:

[0132] f t =σ(W f [h t-1 ,x t ]+b f )

[0133] In the LSTM neural network, the input of the input gate is the hidden state h of the previous moment. t-1 and the current input x t , output i t Used to determine the information in the current input that should be added to the memory cell, b i is the bias vector of the input gate. The relationship between its input and output is:

[0134] i t =σ(W i [h t-1 ,x t ]+b i )

[0135] In addition, the input of the memory cell is the hidden state h of the previous moment t-1 and the current input x t , the output is the current cell C t Used to update memory cells, b c is the bias vector of the output gate.

[0136]

[0137]

[0138] Where:

[0139] ——External output information of the memory unit module at time t.

[0140] When using the SAO algorithm to optimize LSTM parameters, the algorithm is first initialized to determine the number of molecule clusters N and the dimension of the problem dim. Then, a batch of N rows and dim columns of molecules are randomly generated.

[0141]

[0142] Where:

[0143] Z——position information matrix of all molecules;

[0144] Z min , Z max ——the upper and lower limits of the molecular position;

[0145] Z rand ——Random numbers, used to randomly generate different molecular positions;

[0146] Z 1,1 ——The position of the first molecule in the first dimension.

[0147] The snow melting optimization algorithm used in this patent uses a dual population mechanism to divide the molecular population P into two sub-populations P a and P b The number of molecules in the population is N a and N b , the expression is:

[0148]

[0149] N b (t+1)=N b (t)-1

[0150] Low risk level (load growth > 8%): Grid demand is growing rapidly, power supply capacity is sufficient, investment returns are high, and operating risks are low.

[0151] Medium-low risk level (load growth of 6%-8%): Grid demand grows steadily and operations are stable, but there may be certain pressure to upgrade facilities.

[0152] Medium risk level (load growth of 4%-6%): The growth of grid demand has slowed down, and there is certain pressure on technology and equipment maintenance, with medium operational risks.

[0153] Medium-to-high risk level (load growth of 2%-4%): Grid demand grows slowly, equipment utilization decreases, and operational risks are high.

[0154] High risk level (load growth <2%): grid demand is stagnant or declining, equipment is idle, investment returns are limited, and operational risks are greatest.

[0155] Step 5: Build an operation risk level assessment model based on Markov and LS-SVM for operation risk, and formulate a quantitative assessment standard for operation risk.

[0156] (1) Select the factors affecting basic cost forecasting

[0157] When constructing a basic cost prediction model, the selected factors must be directly related to the core cost drivers of grid maintenance, daily operations and equipment replacement, including the asset size of transmission lines, the average operating life of transmission lines, the failure rate of transmission lines, the operating risk level of transmission lines, the amount of power transmitted, and the load rate.

[0158] (2) Determine the transmission line operation risk assessment level

[0159] The operational risk of a transmission line can be described by the probability of an adverse event occurring and the degree of harm to the line after the adverse event occurs. R represents the amount of risk that an adverse event may cause to the transmission line; p represents the probability that an adverse event may cause a failure in the transmission line; and h represents the degree of damage to the transmission line caused by an adverse event. Among them, p can also represent the probability of an adverse event occurring p 1 and the probability p of an adverse event leading to a transmission line failure 2 ; h can be expressed by the loss of line transmission capacity Y and the influence coefficient S, as follows:

[0160] R=f(p,h)=p·h

[0161] p=p 1 ·p 2

[0162] h=Y·S

[0163] Since there are many risk factors that affect the operation of transmission lines, the main consideration is the impact of external environmental factors on the lines. Different environmental factors appear at different times, and have different risks and impacts on the operation of transmission lines. Therefore, the risk period of external environmental factors and the risk weight caused by environmental factors are introduced into the expression of the degree of damage to the transmission line due to adverse events, that is, S i =a i ·ω i a is the influence coefficient of the risk that the line may encounter due to environmental influencing factor i; i represents the risk period of environmental factor i; ω i is the weight of environmental factor i. By sorting out the above formula, we can get:

[0164]

[0165] Because the loss of transmission capacity of the transmission line cannot be quantified, p 2 The value multiplied by Y is the transmission capacity loss, that is, y = p 2 Y, can be converted to:

[0166]

[0167] Combining the Regulations on Safety Risk Management of Power Grid Operation and the value range of R, the risk and risk level of transmission line operation are obtained, such as Figure 3 shown.

[0168] When conducting risk analysis, the complexity and diversity of influencing factors will increase the difficulty and complexity of the analysis. In order to avoid repeated calculation of related influencing factors, redundant influencing factors should be eliminated, key influencing factors should be extracted, and the difficulty of risk assessment should be reduced.

[0169] The first step is to use PCA to extract key influencing factors. Key influencing factors are representative and can describe most of the information in many influencing factors. By extracting key influencing factors, the risk assessment process of transmission line operation can be simplified.

[0170] Collect the influencing factors that may cause transmission line failure, standardize the influencing factors, and form a matrix Q=[Q 1 ,Q 2 ,...,Q n ] T ,or:

[0171]

[0172] According to the relationship between multiple influencing factors, a correlation matrix is ​​established: R = (r ij )m×m=Q′Q, and find the eigenvalue λ of R 1 ≥λ 2 ≥...≥λ m ≥0, and the eigenvector α 1 ,α 2 ,...,α m .

[0173] Determine the number of influencing factors. Calculate the cumulative impact rate of the first p influencing factors using the following formula: :

[0174]

[0175] use The total impact rate of the p impact factor on the transmission line operation failure can be described. Then use these p influencing factors to represent the information of m influencing factors (p<m), and record H=[H 1 ,H 2 ,...,H p ] T ,By calculating the cumulative impact rate of the first p influencing factors, the influencing factors with small contribution rate and redundant ones are eliminated.

[0176] Then calculate the comprehensive impact value of the influencing factors. Weight these p key influencing factors and calculate the comprehensive impact value W of the transmission line when multiple influencing factors act together; calculate the degree of influence and weight G of each influencing factor on the transmission line.

[0177] W=λH=[λ 1 λ 2 ... P ]H

[0178] =(λ 1 α 1 +λ 2 α 2 +...+λ p α p ) T ×[Q 1 Q 2 ...Q P ] T

[0179] From the above formula, we can get the weight of the impact factor on the comprehensive score:

[0180] G=[k 1 ,k 2 ,...,k p ]=(λ 1 α 1 +λ 2 α 2 +...+λ p α p ) T

[0181] The weight G of the impact factor is normalized. The larger the weight of the impact factor, the more representative the impact factor is. Therefore, when G>0.5, the corresponding impact factor is selected as the key impact factor.

[0182] Transmission line failure is a random event that can be described by the Markov chain model. For the power system, the Markov chain represents the operating state of the transmission line at t+1. t+1 Only the operating state O of the transmission line at time t tThe decision is independent of the running state at any time before t. By using this characteristic of Markov, we can reduce the dependence on the historical state, reduce the amount of calculation, and improve the operation efficiency. The Markov process is defined as:

[0183] O{T t+1 =t t+1 |T 0 =t 0 ,T 1 =t 1 ,...,T n =t n}=O{T t+1 =t t+1 |T n :

[0184] λ represents the probability of transmission line failure, μ represents the ability to repair the transmission line after failure, and combining the many factors that affect the failure of transmission lines, a probability model for changes in the operating status of transmission lines is established.

[0185]

[0186] The Markov process has the following equation in the limit state:

[0187] PM=P

[0188] P=[p N p 1 p 2 p 3 p 4 p 5 ]

[0189]

[0190] Use p N represents the probability of normal operation of the transmission line; p 1 、p 2 、p 3 、p 4 、p 5 、p 6 They represent the probability of line failure caused by 6 factors. The probability of transmission line failure caused by external factors is:

[0191]

[0192] There are many factors that affect the normal operation of transmission lines, and the main consideration is the impact of external factors on the operation of transmission lines. External factors include bird damage, lightning strikes, strong winds, external forces and many other factors. It is necessary to establish a probability model that includes the above many external factors:

[0193]

[0194] According to the research, P lim (l) = 1.4P max (l), where the severity of damage after a line failure is:

[0195]

[0196] The purpose of constructing a transmission line operation risk assessment model is to find the normal operation risk level of the transmission line corresponding to the key influencing factors. The least squares support vector machine algorithm is used to identify and classify transmission line faults.

[0197] Since the key influencing factors of line operation risk are nonlinear characteristics, functions are needed to map and transform the nonlinear key influencing factors of training data and convert the nonlinear influencing factors into linear equations.

[0198]

[0199] For ω,ξ i ,b,α i Find the partial derivatives and use KKT to remove ω and ξ i , we get the following linear equation:

[0200]

[0201] Ω=y k y h (φ(x k )) T φ(x h )=y k y h K(x k

[0202]

[0203] where k, h = 1, 2, ..., n; x k ,x h ,y h ,y k is the training sample vector; α is the i The vector formed; Y = [y 1 ,y 2 ,...,y n ], 1 v =[1,1,...,1]; K(x k ,x h ) is the kernel function; I is the unit vector.

[0204] Secondly, the classification equation is used to classify the line operation faults.

[0205]

[0206] Where: φ is the nonlinear mapping function vector, x i and i is the training sample data. In this way, the nonlinear key influencing factors are converted into linear data for solution. The transformed linear transmission line operation risk key influencing factors are input into the least squares support vector machine risk assessment model established above, and the operation risk faced by the transmission line can be obtained.

[0207] Since γ and σ 2 Directly affects the accuracy of the final transmission line fault classification. Through continuous iterative updates, until the optimal γ and σ are found 2 .

[0208] Step six, finally apply the neural network model to scientifically set the scoring weights of market risk, operational risk, and operation risk to obtain the final risk assessment results.

[0209] In the risk assessment process, the scoring results of market risk, operational risk and operation risk need to be weighted according to their contribution to the overall investment risk to obtain the final comprehensive risk assessment result. In order to achieve this goal, a neural network model can be used to scientifically set the weight of each risk score to ensure that the impact of each risk dimension on the final assessment result is in line with the actual situation.

[0210] Input layer: The input layer of the neural network receives scoring data from different risk dimensions, including market risk score, operational risk score, and operation risk score. The market risk score is 1-5, the operational risk score is 1-5, and the operation risk score is 1-5. The score of each dimension is used as the input feature of the neural network.

[0211] Hidden layer: The neural network processes the input data in a weighted manner through multiple hidden layers. By training the neural network, the weights of each input feature (market risk, operational risk, and operation risk) can be automatically adjusted so that the model can learn the importance of these risk scores to the final assessment results. The hidden layer processes the input data through a nonlinear activation function to extract useful feature information.

[0212] Output layer: The output layer of the neural network gives the final comprehensive risk score. This score represents the risk level of the overall power grid equity investment and is usually mapped to a fixed interval (such as 1-5, representing different risk levels).

[0213] Weight setting: The neural network automatically learns the weight of each risk dimension through the training process. By adjusting the weight parameters in the neural network, the model can automatically identify the impact of market risk, operational risk and operation risk on the final investment risk assessment. The scientific setting of weights can reflect the importance of each dimension in different investment environments. For example, if the impact of market risk is greater in a certain market environment, the neural network will automatically increase the weight of the market risk score and reduce the weight of other dimensions.

[0214] Training process: During the training process, the neural network is trained with historical data. The loss function is used to evaluate the gap between the predicted value and the true value, and the weights are updated through the back-propagation algorithm.

[0215] The first hidden layer is the result of cluster analysis. Therefore, combined with the cluster analysis results, the number of nodes in the first hidden layer is 4, which further improves the accuracy of prediction. The output vector of the first hidden layer is:

[0216] Q=(q 1 ,q 2 ,…,q m ) T ;

[0217] The output vector of the second hidden layer is:

[0218] M=(m 1 ,m 2 ,…,m m ) T ;

[0219] The output vector of the output layer is: the actual risk score corresponding to the moment to be predicted.

[0220] O=(o 1 ,o 2 ,…,o l ) T ;

[0221] Among them, the output vector is the weight of different risks.

[0222]

[0223] Assume that the activation function of each layer of nodes in the network is an S-type function, and the input of the first layer i node in the network is denoted as net i , the output is recorded as o i , the output of the kth node in the output layer is y k , then the input of the jth node in the middle layer is:

[0224]

[0225] o j=f(net j )

[0226]

[0227] Define the network error as the difference between the expected output and the actual output, then we have If the output layer has i neurons, the square error between the actual output and the expected output is defined as:

[0228]

[0229] Since the BP algorithm corrects the weights according to the negative gradient of the error E, the modification of the weights can be expressed as:

[0230] W m+1 =w m +Δw m =w m -λg m

[0231] Where m represents the number of iterations,

[0232] Among them, λ is the step size of learning.

[0233] The optimal path is optimized by combining the gradient descent method and the Gauss-Newton method, and the fault occurrence is output. At the beginning, λ takes a large number, which is equivalent to the gradient descent method with a small step size; as the optimal value is approached, λ decreases to zero, and S(X(k)) turns from the negative gradient direction to the λ direction of the Gauss-Newton method. Usually, when S(X(k))<f(X(k)), λ is reduced, otherwise λ is increased. Through optimal search optimization, the convergence speed can be increased by dozens or even hundreds of times.

[0234]

[0235] Let η k =1, then x k+1 =x k +S(x k )

[0236]

[0237] Because it is the output layer, is the actual output value, according to e k The definition of and square error can be obtained:

[0238]

[0239] According to e k The definition of can be obtained:

[0240]

[0241] According to the above formula We can get:

[0242]

[0243] According to the above formula We can get:

[0244]

[0245] Finally we get:

[0246]

[0247] Now let the learning error of the output layer be:

[0248] σ k =e k f′(net k )

[0249] have to:

[0250]

[0251] The weight modification of the hidden layer neural unit Δw kj :

[0252]

[0253] According to the above formula and j =f(net j ) can be obtained:

[0254]

[0255] Because we are looking for the change in the weight of the hidden layer, we should consider the effect of the previous layer on it, which is inherent:

[0256]

[0257] according to Knowable

[0258]

[0259] According to We can get:

[0260]

[0261] Bundle Bring-in Derived:

[0262]

[0263] Let the learning error of the hidden layer be:

[0264]

[0265] The goal of training is to minimize the error between the model prediction value and the actual risk assessment value. Through the back-propagation algorithm, the weight values ​​in the network are gradually adjusted so that the final output risk score is as close to the actual assessment value as possible. New investment data is predicted through the trained neural network model. Based on the scores of market risk, operational risk and operation risk, the neural network will give the final investment risk assessment results.

[0266] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0267] It should be noted that relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that an article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprises a..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0268] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest range consistent with the principles and novel features disclosed herein.

[0269] It should be understood that the present application is not limited to what has been described above and shown in the drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for evaluating the risk of equity investment in provincial power grid industries, characterized in that: The method comprises the following steps: Step 1: Based on the actual status of power grid equity investment, systematically identify the types of equity investment risks and clarify the main influencing factors of different risks; Step 2: Use the improved anti-entropy-grey correlation analysis method to determine the correlation between investment risk and each key influencing factor, and select the most important factors in different risk types as the basis for risk level determination; Step 3: In view of market risk, we select GDP growth rate as the main discriminant factor, build a GDP growth forecasting model based on ARIMA, and formulate quantitative assessment standards for market risk; Step 4: In view of operational risks, load growth is selected as the main discriminant factor, and a load forecasting model based on SAO optimized LSTM is constructed to form a quantitative assessment standard for operational risks; Step 5: Aiming at the operation risk, a Markov and LS-SVM-based operation risk level assessment model is constructed to achieve the formulation of the quantitative assessment standard for operation risk; Step six: Apply the neural network model to scientifically set the scoring weights for market risk, operational risk, and operation risk to obtain the final risk assessment results.

2. A method for evaluating the risk of equity investment in power grid provincially managed industries according to claim 1, characterized in that: The types of equity investment risks in step 1 include market risk, operational risk and operation risk; Factors influencing market risks include changes in electricity demand, GDP, changes in policies and regulations, fluctuations in energy prices, the improvement of electricity marketization, the emergence of external competitors, changes in the international market and adjustments in monetary policy; The factors affecting operational risks include the failure rate of power grid equipment, overload of transmission lines, aging of equipment, aging degree of power grid facilities, construction of monitoring and early warning systems, safety management of power grids, enforcement of operating procedures, external attacks, network security issues and errors in power grid dispatching; Factors affecting operational risks include load growth rate, equipment aging, maintenance, extreme weather caused by climate change, application of new technologies and personnel management.

3. A method for evaluating the risk of equity investment in power grid provincially managed industries according to claim 1, characterized in that: The application of the improved anti-entropy-grey correlation analysis method in step 2 specifically includes the following steps: a. Calculate the entropy value for each key influencing factor to reflect the uncertainty of its information; b. The entropy value is converted into the anti-entropy value through the anti-entropy formula, which reflects the certainty and difference of the factor information. The larger the anti-entropy value, the greater the weight given; c. Use grey correlation analysis to calculate the correlation coefficient between each factor and the reference sequence, and then derive the correlation between each key influencing factor and the risk level; d. Combine the anti-entropy value and grey correlation degree to select the factors most relevant to the risk level as the basis for determining the risk level.

4. A method for evaluating the risk of equity investment in power grid provincially managed industries according to claim 1, characterized in that: In step 3, the GDP growth forecast model is adjusted by three main parameters p, d, and q, where p represents the number of autoregressive terms, d represents the number of differences, and q represents the number of sliding average terms. The formulation of the market risk quantitative assessment standard includes the following steps: a. Stabilize the series using historical GDP data to ensure that the data meets the requirements of the ARIMA model; b. Determine the most appropriate parameters (p, d, q), using the autocorrelation function (ACF) and partial autocorrelation function (PACF) plots to assist in the selection; c. Use the selected ARIMA model for training to fit the trend of GDP growth; d. Based on the forecast results of the model, formulate quantitative assessment standards for market risks, and judge the different levels of market risks according to the changing trends of future GDP growth.

5. A method for evaluating the risk of equity investment in power grid provincially managed industries according to claim 1, characterized in that: The formulation of the quantitative assessment criteria for operational risk in step 4 includes the following steps: a. Collect historical load data and pre-process it to ensure data cleanliness and standardization; b. Use the SAO algorithm to optimize the structure and hyperparameters of the LSTM model and find the optimal configuration; c. By training the optimized LSTM model, we can predict the future load growth and obtain the load change trend in different time periods; d. Based on the forecast results, formulate quantitative assessment standards for operational risks.

6. A method for evaluating the risk of equity investment in power grid provincially managed industries according to claim 1, characterized in that: The formulation of the operational risk quantitative assessment standard in step 5 includes the following steps: a. Markov chain model to evaluate the operation status of transmission lines; b. PCA algorithm extracts key influencing factors; c. Least squares support vector machine (LS-SVM) for risk level prediction; 7. A method for evaluating the risk of equity investment in power grid provincially managed industries according to claim 1, characterized in that: The calculation of the risk assessment results in step 6 includes the following steps: a. By collecting and integrating relevant data on market risk, operational risk and operation risk, the input features of the neural network model are constructed. These features include the prediction results of each risk and the relevant influencing factors. b. Design the architecture of the neural network, select the appropriate number of layers and neurons, and use the back-propagation algorithm for training; c. Use training data to learn the model and adjust the weight of each risk score so that it can accurately reflect the contribution of various risks to the overall risk of power grid investment; d. Using the trained neural network model, input new risk data, calculate the final scores of market risk, operational risk and operation risk, and calculate the comprehensive risk assessment results based on these scores.

Images