Construction method of typical engineering project investment prediction model based on transformer substation

Abstract

The invention discloses a construction method based on a typical engineering project investment prediction model of a transformer substation. The construction method comprises the following specific steps: S1, collecting a representative typical power transformation engineering scheme and historical price data thereof; s2, preprocessing the collected data, including noise reduction, normalization processing and invalid data removal; s3, substituting the processed historical price data one by one by adopting a control variable method, performing mathematical modeling by utilizing a linear regression model, and quantifying the influence of each variable on the total cost; s4, monitoring independent variable information flow and dynamically adjusting the weight based on a dynamic asymmetric information regulation and control and self-adaptive offset mechanism; s5, verifying and correcting the linear regression model, and optimizing a prediction formula in combination with regional characteristics; and S6, according to dynamic changes of market and regional economy, adjusting the model in real time, and generating final cost prediction data. The problems that the regional adaptability is poor, and the cost scheme is not updated in time are effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of machine learning, and in particular to a method for constructing an investment prediction model based on a typical substation engineering project. Background Art

[0002] First, with the rapid development of modern power systems, substation projects, as a key link in power transmission and distribution, occupy an important position in power engineering management in terms of construction and investment forecasting. The investment forecasting of substation projects not only involves a large amount of capital investment, but also affects the stability and operating efficiency of the entire power system. In actual projects, there are many factors that affect the cost of substations, including material prices, labor costs, equipment costs, construction environment, etc. The fluctuations of these factors pose a huge challenge to the prediction of project costs.

[0003] Secondly, the existing substation project cost prediction methods are usually based on fixed historical data and standard processes, and estimate future project costs by establishing mathematical models. Most of these traditional methods use simple prediction models such as linear regression and cost-plus. However, these models often assume that market conditions are relatively stable and the relationship between various influencing factors is linear. This assumption has great limitations in practical applications. Especially in the face of dynamic changes in the market and the particularity of the regional economy, traditional models often cannot accurately reflect the real market situation, resulting in a large deviation between the prediction results and the actual costs.

[0004] Third, the existing technology still has the problem of imperfect data processing in cost forecasting. Usually, historical price data will be interfered by many external factors, such as economic fluctuations, policy changes, etc. These factors may cause more noise or outliers in the data. If these data are not fully preprocessed during the model construction process, the accuracy and reliability of the final prediction results will be seriously affected. In addition, the model update mechanism commonly used in the existing technology is relatively simple, and it is usually unable to respond to the dynamic changes of the market in real time, resulting in the model being difficult to continue to provide accurate prediction data after a period of application.

[0005] In summary, the existing substation project investment forecasting technology has the following defects: First, the market dynamics are not taken into consideration during the model construction process, and there is a lack of effective dynamic regulation mechanism, resulting in low prediction accuracy; second, it fails to fully combine regional economic characteristics, resulting in poor regional adaptability of the prediction model; third, there are deficiencies in the data processing link, and the existence of noise data and outliers weakens the reliability of the model; fourth, the model update mechanism is imperfect, and it is difficult to maintain accurate prediction capabilities when market conditions change. Therefore, how to provide a construction method based on a typical substation project investment forecasting model is a problem that technicians in this field urgently need to solve. Summary of the invention

[0006] One object of the present invention is to propose a method for constructing an investment forecasting model based on a typical substation engineering project. The present invention fully combines dynamic asymmetric information regulation and adaptive offset mechanism, and describes in detail the process of achieving accurate cost forecasting through real-time data monitoring, dynamic weight adjustment and model optimization. It has the advantages of strong regional adaptability, high prediction accuracy and excellent dynamic response capability.

[0007] According to an embodiment of the present invention, a method for constructing a typical substation project investment forecasting model comprises the following steps:

[0008] S1. Collect representative typical substation engineering schemes and historical price data in typical projects;

[0009] S2. Preprocess the collected data, including noise reduction, normalization, and removal of invalid data;

[0010] S3. Use the control variable method to substitute the processed historical price data one by one, replace the single independent variable, use the linear regression model to mathematically model the project cost, quantify the impact of each variable on the total cost, and calculate the impact coefficient;

[0011] S4. Based on dynamic asymmetric information regulation and adaptive offset mechanism, monitor the information flow of independent variables, identify influential independent variables and dynamically adjust the weights;

[0012] S5. Verify and correct the linear regression model, substitute historical real engineering data into the model, correct the data based on regional characteristics, and optimize and improve the prediction formula;

[0013] S6. According to the dynamic changes in market prices and regional economy, combined with dynamic data update mechanism and adaptive control mechanism, the model is adjusted in real time to reflect the latest market information and regional characteristics, and generate the final cost forecast data.

[0014] Optionally, S3 includes the following steps:

[0015] S31. When calculating typical project costs, the control variable method is used to insert the collected historical price data into the model one by one, and only one independent variable is replaced each time, keeping the other independent variables unchanged, so as to calculate the impact of a single independent variable on the total project cost:

[0016] Y1=β0+β1X1;

[0017] Y2=β0+β2X2;

[0018] Y3=β0+β3X3;

[0019] …

[0020] Yn =β0+β n X n ;

[0021] S32. Record the cost trend of typical substation projects along with the change of each independent variable, and generate a complete cost change data set X n ;

[0022] S33, the cost change data set of the typical project X n Substitute into the linear regression model and construct the following regression equation:

[0023] Y=β0+β1X1+β2X2+…+β n X n +μ;

[0024] Among them, Y is the total cost of the project, X1, X2, ..., X n are the independent variables, β0 is the intercept term, β1,β2,...,β n is the regression coefficient of each variable, μ is the error term;

[0025] S34. Calculate the regression coefficients β1, β2, ..., β n , the regression coefficient will quantify the impact of each independent variable on the total project cost;

[0026] S35. Use historical real engineering data of the same standard process to verify the constructed regression model and test the model's goodness of fit R 2 Values, and significance of regression coefficients;

[0027] S36. During the verification process, if it is found that the model does not meet the expected statistical standards, adjust the model through data correction, recalculate the regression coefficient, and continuously optimize the regression formula;

[0028] S37. Store the finalized regression model and parameters into the database.

[0029] Optionally, S4 includes the following steps:

[0030] S41. Constructing enhanced dynamic asymmetric information matrix E-DAIM:

[0031] S411. Standardize historical data related to typical substation projects;

[0032] S412, calculate each variable X within time τ i For the independent variable X j Basic information transfer rate: To reflect the dynamic asymmetric information flow between independent variables, where Δt is the time delay parameter used to capture the lag effect between variables in the market;

[0033] S413, based on all calculated information transmission rates Taking into account the time decay effect and market volatility factors, an enhanced dynamic asymmetric information matrix E-DAIM is constructed. Each element EDAI in E-DAIM ij (t) is:

[0034]

[0035] Among them, EDAI ij (t) represents the independent variable X i At time t, j The information transmission intensity reflects the asymmetry between the two under market conditions. λ is the time decay coefficient, which controls the gradual weakening of the impact of historical information on the current situation. κ and ω are the amplitude and frequency adjustment coefficients of information transmission, which are used to introduce the impact of market cyclical fluctuations on information transmission.

[0036] S42. Based on the constructed E-DAIM matrix, monitor the information flow between the variables and calculate the information transmission intensity EDAI ij (t), identify the influential independent variables;

[0037] S43. To quantify the interaction effects between key independent variables, calculate the multivariate interaction control term C ij (t):

[0038]

[0039] Among them, C ij (t) is used to describe the compound interaction effect between multiple independent variables at different time points, η is the interaction regulation coefficient, which regulates the interaction between different variables, and γ is the nonlinear regulation index, which controls the intensity of the interaction effect. is the cyclical adjustment parameter, which is used to capture the effect of market cyclical changes on the interaction between independent variables;

[0040] S44. Integrate information flow and historical data, combine the monitored key independent variable information flow data and historical market data, and dynamically calculate the weights W of each variable in different time periods. i (t):

[0041]

[0042] Among them, W i (t) is the independent variable X at time t i The dynamic weight reflects the actual influence under current market conditions, α iis the initial adjustment coefficient, θ is the time-decreasing factor, and β is the logarithmic adjustment coefficient, which is used to smooth the impact of market fluctuations on the weight of the independent variable;

[0043] S45, the dynamically adjusted weight W i (t) Applied to the cost forecasting model to reflect the adjustment of the influence of market conditions on independent variables in real time;

[0044] S46. When the influence of the independent variable is detected to be too large, the composite nonlinear offset factor δ is calculated. i (t):

[0045]

[0046] Among them, δ i (t) is used to weaken the independent variable X i The excessive influence of the independent variable is prevented from causing inaccuracy of the model prediction results due to excessive influence of the independent variable. γ is the adjustment coefficient, which controls the sensitivity of the offset factor. R i and R j Refers to the independent variable X i and X j The weights are recursively adjusted at time t, ξ and φ are periodic fluctuation adjustment parameters;

[0047] S47. Real-time monitoring of market dynamics and weight changes of each variable, using sensitivity analysis to assess the marginal impact of each variable on the total cost Y(t), and calculating the dynamic offset feedback value F i (t):

[0048]

[0049] Among them, F i (t) is the adaptive dynamic offset feedback value, which reflects the actual impact of market changes on the independent variable. is the total cost Y(t) versus the independent variable X k The sensitivity coefficient of (t) indicates the direct impact of the change of the independent variable on the cost. and ψ are the time and periodicity adjustment factors;

[0050] S47, the calculated dynamic offset feedback value F i (t) Applied to the cost prediction model, the weights of key independent variables are dynamically adjusted, and the model is continuously optimized through an adaptive optimization algorithm.

[0051] Optionally, S5 includes the following steps:

[0052] S51. Continuously collect real-time data from the market and pre-process the collected real-time data, including data cleaning, outlier processing and standardization;

[0053] S52. Calculate the independent variable X based on real-time data i (t) is the rate of change compared to the previous time point t-1, and the formula is:

[0054]

[0055] Where ΔX i (t) is the independent variable X at time t i The rate of change of

[0056] S53. According to the rate of change of the independent variable, the regression coefficient β in the regression model is i (t) is dynamically adjusted, and the formula is:

[0057] β i (t) = β i (t-1)·(1+α·ΔX i (t));

[0058] Among them, α is the adjustment coefficient, which controls the adjustment range of the regression coefficient; β i (t) is the regression coefficient at time t;

[0059] S54, Substitute the adjusted regression coefficient: Substitute the dynamically adjusted regression coefficient β i (Substitute t0 into the linear regression model and update the predicted value of the total project cost Y(t) in real time:

[0060]

[0061] Among them, F i (t) is the dynamic offset feedback value calculated according to claim 4, and μ(t) is the error term;

[0062] S55, by calculating the real-time updated total project cost Y(t), output the latest forecast result, and use it for the formulation and adjustment of the project budget;

[0063] S56. Regularly compare the model's prediction results with actual engineering data to test the model's goodness of fit R 2 The significance of the regression coefficient β is adjusted according to the verification results. i (t) and dynamic offset feedback value F i (t), to correct for biases in the model;

[0064] S57. Store the updated model parameters, regression coefficients and dynamic adjustment results in the database, combine with historical data analysis, continuously optimize the model structure and parameter configuration, and adapt to future market changes.

[0065] Optionally, the S6 includes the following steps:

[0066] S61. Analyze the specific needs of the substation project, identify the scenarios where investment forecasting is required, and apply the model to specific scenarios based on different application scenarios and project requirements;

[0067] S62. Collect relevant regional market economic data and characteristic parameters based on the economic conditions and market characteristics of the project area;

[0068] S63. Using regional market economic data as independent variable Z j Import the model:

[0069]

[0070] Among them, Y adjusted Y(t) is the adjusted total project cost forecast value, Y(t) is the unadjusted total project cost forecast value, and Z j (t) is the regional characteristic parameter, γ j is the regional adjustment coefficient, which is used to reflect the impact of regional characteristics on the total construction cost;

[0071] S64, using a model that is dynamically adjusted and updated in real time, substitute the independent variable data X i (t) and regional characteristic parameter Z j (t), calculate the estimated investment cost of the substation project, the formula is:

[0072]

[0073] Among them, Y final (t) is the final predicted total cost;

[0074] S65. Analyze the calculation results, identify the key influencing factors, and make corresponding adjustment suggestions based on the forecast data;

[0075] S66. Output the prediction results and compare them with the actual needs of the project to evaluate the prediction accuracy of the model and verify the effectiveness of the model;

[0076] S67. Based on the evaluation of the forecast results, make further adjustment suggestions;

[0077] S68. The final prediction results of the model are used to support investment decisions for substation engineering projects.

[0078] The beneficial effects of the present invention are:

[0079] (1) The present invention proposes a method for constructing an investment forecasting model based on a typical substation project. By collecting and processing historical price data of typical substation projects and using preprocessing techniques such as noise reduction, normalization, and removal of invalid data, the noise and outliers in the data are eliminated, ensuring the accuracy and consistency of the model input data, thereby significantly improving the accuracy of cost forecasting, especially when faced with data complexity and a changing market environment.

[0080] (2) The present invention uses the control variable method and linear regression model to mathematically model the project cost, and then substitutes the processed historical price data one by one to quantify the impact of each variable on the total cost, and calculates the impact coefficient. This process can accurately identify and quantify the key factors that affect the total cost of the project, making the cost forecast more refined, facilitating managers to make targeted adjustments and controls on cost factors, and effectively improving the scientificity and accuracy of project cost management.

[0081] (3) The present invention introduces a dynamic asymmetric information regulation and adaptive offset mechanism. By monitoring the information flow between independent variables, dynamically identifying and adjusting the weights of key independent variables, the model can self-regulate under real-time market conditions to maintain the accuracy and stability of the prediction results. This mechanism enables the cost prediction model to have a strong dynamic response capability, effectively respond to market fluctuations, ensure a high degree of consistency between the prediction results and the actual market situation, and improve the reliability of the project budget and the efficiency of economic management.

[0082] (4) In the process of model construction and optimization, the present invention verifies the model by substituting historical real engineering data into the model, and corrects and optimizes the data in combination with regional characteristics, thereby ensuring the applicability and accuracy of the model in a specific region. In particular, when regional economic characteristics are significant, the present invention can adjust the model according to the actual conditions of different regions, generate cost forecast data that is highly consistent with regional market characteristics, and significantly enhance the regional adaptability and forecast accuracy of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0084] Figure 1 It is a flow chart of the construction method of the typical substation engineering project investment forecasting model proposed by the present invention;

[0085] Figure 2 This is a data graph based on the typical substation project investment forecasting model proposed by the present invention;

[0086] Figure 3This is a schematic diagram of the mathematical modeling process of the linear regression model using the control variable method proposed in the present invention. DETAILED DESCRIPTION

[0087] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0088] refer to Figure 1-3 ,The construction method of the investment prediction model based on the typical substation project includes the following steps:

[0089] S1. Collect representative typical substation engineering schemes and historical price data in typical projects;

[0090] S2. Preprocess the collected data, including noise reduction, normalization, and removal of invalid data;

[0091] S3. Use the control variable method to substitute the processed historical price data one by one, replace the single independent variable, use the linear regression model to mathematically model the project cost, quantify the impact of each variable on the total cost, and calculate the impact coefficient;

[0092] S4. Based on dynamic asymmetric information regulation and adaptive offset mechanism, monitor the information flow of independent variables, identify influential independent variables and dynamically adjust the weights;

[0093] S5. Verify and correct the linear regression model, substitute historical real engineering data into the model, correct the data based on regional characteristics, and optimize and improve the prediction formula;

[0094] S6. According to the dynamic changes in market prices and regional economy, combined with dynamic data update mechanism and adaptive control mechanism, the model is adjusted in real time to reflect the latest market information and regional characteristics, and generate the final cost forecast data.

[0095] In this implementation, S3 includes the following steps:

[0096] S31. When calculating typical project costs, the control variable method is used to insert the collected historical price data into the model one by one, and only one independent variable is replaced each time, keeping the other independent variables unchanged, so as to calculate the impact of a single independent variable on the total project cost:

[0097] Y1=β0+β1X1;

[0098] Y2=β0+β2X2;

[0099] Y3=β0+β3X3;

[0100] …

[0101] Yn =β0+β n X n ;

[0102] S32. Record the cost trend of typical substation projects along with the change of each independent variable, and generate a complete cost change data set X n ;

[0103] S33, the cost change data set of the typical project X n Substitute into the linear regression model and construct the following regression equation:

[0104] Y=β0+β1X1+β2X2+…+β n X n +μ;

[0105] Among them, Y is the total cost of the project, X1, X2, ..., X n are the independent variables, β0 is the intercept term, β1,β2,...,β n is the regression coefficient of each variable, μ is the error term;

[0106] S34. Calculate the regression coefficients β1, β2, ..., β n , the regression coefficient will quantify the impact of each independent variable on the total project cost;

[0107] S35. Use historical real engineering data of the same standard process to verify the constructed regression model and test the model's goodness of fit R 2 Values, and significance of regression coefficients;

[0108] S36. During the verification process, if it is found that the model does not meet the expected statistical standards, adjust the model through data correction, recalculate the regression coefficient, and continuously optimize the regression formula;

[0109] S37. Store the finalized regression model and parameters into the database.

[0110] In this implementation, S4 includes the following steps:

[0111] S41. Constructing enhanced dynamic asymmetric information matrix E-DAIM:

[0112] S411. Standardize historical data related to typical substation projects;

[0113] S412, calculate each variable X within time τ i For the independent variable X j Basic information transfer rate: To reflect the dynamic asymmetric information flow between independent variables, where Δt is the time delay parameter used to capture the lag effect between variables in the market;

[0114] S413, based on all calculated information transmission rates Taking into account the time decay effect and market volatility factors, an enhanced dynamic asymmetric information matrix E-DAIM is constructed. Each element EDAI in E-DAIM ij (t) is:

[0115]

[0116] Among them, EDAI ij (t) represents the independent variable X i At time t, j The information transmission intensity reflects the asymmetry between the two under market conditions. λ is the time decay coefficient, which controls the gradual weakening of the impact of historical information on the current situation. κ and ω are the amplitude and frequency adjustment coefficients of information transmission, which are used to introduce the impact of market cyclical fluctuations on information transmission.

[0117] S42. Based on the constructed E-DAIM matrix, monitor the information flow between the variables and calculate the information transmission intensity EDAI ij (t), identify the influential independent variables;

[0118] S43. To quantify the interaction effects between key independent variables, calculate the multivariate interaction control term C ij (t):

[0119]

[0120] Among them, C ij (t) is used to describe the compound interaction effect between multiple independent variables at different time points, η is the interaction regulation coefficient, which regulates the interaction between different variables, and γ is the nonlinear regulation index, which controls the intensity of the interaction effect. is the cyclical adjustment parameter, which is used to capture the effect of market cyclical changes on the interaction between independent variables;

[0121] S44. Integrate information flow and historical data, combine the monitored key independent variable information flow data and historical market data, and dynamically calculate the weights W of each variable in different time periods. i (t):

[0122]

[0123] Among them, W i (t) is the independent variable X at time t i The dynamic weight reflects the actual influence under current market conditions, αi is the initial adjustment coefficient, θ is the time-decreasing factor, and β is the logarithmic adjustment coefficient, which is used to smooth the impact of market fluctuations on the weight of the independent variable;

[0124] S45, the dynamically adjusted weight W i (t) Applied to the cost forecasting model to reflect the adjustment of the influence of market conditions on independent variables in real time;

[0125] S46. When the influence of the independent variable is detected to be too large, the composite nonlinear offset factor δ is calculated. i (t):

[0126]

[0127] Among them, δ i (t) is used to weaken the independent variable X i The excessive influence of the independent variable is prevented from causing inaccuracy of the model prediction results due to excessive influence of the independent variable. γ is the adjustment coefficient, which controls the sensitivity of the offset factor. R i and R j Refers to the independent variable X i and X j The weights are recursively adjusted at time t, ξ and φ are periodic fluctuation adjustment parameters;

[0128] S47. Real-time monitoring of market dynamics and weight changes of each variable, using sensitivity analysis to assess the marginal impact of each variable on the total cost Y(t), and calculating the dynamic offset feedback value F i (t):

[0129]

[0130] Among them, F i (t) is the adaptive dynamic offset feedback value, which reflects the actual impact of market changes on the independent variable. is the total cost Y(t) versus the independent variable X k The sensitivity coefficient of (t) indicates the direct impact of the change of the independent variable on the cost. and ψ are the time and periodicity adjustment factors;

[0131] S47, the calculated dynamic offset feedback value F i (t) Applied to the cost prediction model, the weights of key independent variables are dynamically adjusted, and the model is continuously optimized through an adaptive optimization algorithm.

[0132] In this implementation, S5 includes the following steps:

[0133] S51. Continuously collect real-time data from the market and pre-process the collected real-time data, including data cleaning, outlier processing and standardization;

[0134] S52. Calculate the independent variable X based on real-time data i (t) is the rate of change compared to the previous time point t-1, and the formula is:

[0135]

[0136] Where ΔX i (t) is the independent variable X at time t i The rate of change of

[0137] S53. According to the rate of change of the independent variable, the regression coefficient β in the regression model is i (t) is dynamically adjusted, and the formula is:

[0138] β i (t) = β i (t-1)·(1+α·ΔX i (t));

[0139] Among them, α is the adjustment coefficient, which controls the adjustment range of the regression coefficient; β i (t) is the regression coefficient at time t;

[0140] S54, Substitute the adjusted regression coefficient: Substitute the dynamically adjusted regression coefficient β i (t) is substituted into the linear regression model to update the predicted value of the total project cost Y(t) in real time:

[0141]

[0142] Among them, F i (t) is the dynamic offset feedback value calculated according to claim 4, and μ(t) is the error term;

[0143] S55, by calculating the real-time updated total project cost Y(t), output the latest forecast result, and use it for the formulation and adjustment of the project budget;

[0144] S56. Regularly compare the model's prediction results with actual engineering data to test the model's goodness of fit R 2 The significance of the regression coefficient β is adjusted according to the verification results. i (t) and dynamic offset feedback value F i (t), to correct for biases in the model;

[0145] S57. Store the updated model parameters, regression coefficients and dynamic adjustment results in the database, combine with historical data analysis, continuously optimize the model structure and parameter configuration, and adapt to future market changes.

[0146] In this implementation, S6 includes the following steps:

[0147] S61. Analyze the specific needs of the substation project, identify the scenarios where investment forecasting is required, and apply the model to specific scenarios based on different application scenarios and project requirements;

[0148] S62. Collect relevant regional market economic data and characteristic parameters based on the economic conditions and market characteristics of the project area;

[0149] S63. Using regional market economic data as independent variable Z j Import the model:

[0150]

[0151] Among them, Y adjusted Y(t) is the adjusted total project cost forecast value, Y(t) is the unadjusted total project cost forecast value, and Z j (t) is the regional characteristic parameter, γ j is the regional adjustment coefficient, which is used to reflect the impact of regional characteristics on the total construction cost;

[0152] S64, using a model that is dynamically adjusted and updated in real time, substitute the independent variable data X i (t) and regional characteristic parameter Z j (t), calculate the estimated investment cost of the substation project, the formula is:

[0153]

[0154] Among them, Y final (t) is the final predicted total cost;

[0155] S65. Analyze the calculation results, identify the key influencing factors, and make corresponding adjustment suggestions based on the forecast data;

[0156] S66. Output the prediction results and compare them with the actual needs of the project to evaluate the prediction accuracy of the model and verify the effectiveness of the model;

[0157] S67. Based on the evaluation of the forecast results, make further adjustment suggestions;

[0158] S68. The final prediction results of the model are used to support investment decisions for substation engineering projects.

[0159] Embodiment 1:

[0160] In a large-scale substation construction project in a certain province, the project management team faced complex cost control challenges. Due to the frequent fluctuations in market material prices and the regional economic characteristics, traditional cost forecasting methods are difficult to accurately reflect the actual situation, resulting in a large deviation between budget preparation and actual expenditure. In recent years, due to the large fluctuations in construction material prices and labor costs in a certain province, project managers urgently need a more accurate and real-time investment forecasting method to effectively control project costs and ensure the smooth implementation of the project.

[0161] To meet this challenge, the project management team decided to apply the method of constructing a typical substation project investment forecasting model in this invention. This method collects and analyzes historical data of substation projects, combines dynamic asymmetric information regulation and adaptive offset mechanism, and aims to provide a more accurate cost forecasting model, thereby achieving more scientific budget preparation and cost control.

[0162] The project team first collected historical data on several typical substation projects in a province, covering the main cost components such as construction costs, installation costs, equipment purchase costs, and human resources and equipment costs from 20 to 24 years. These data were derived from the State Grid information price, local material prices, and other public price data, ensuring the accuracy and completeness of the data.

[0163] After the data collection was completed, the team conducted comprehensive pre-processing and standardized the data using advanced noise reduction, normalization and outlier removal techniques. This process eliminated random noise and abnormal fluctuations in historical data, ensuring the consistency and accuracy of the data in subsequent modeling.

[0164] Next, the team used the control variable method and linear regression model to conduct an in-depth analysis of the project costs. By inputting processed historical data one by one, the model can quantify the impact of each independent variable (such as labor costs, material costs, equipment costs, etc.) on the total cost and calculate the impact coefficient of each variable. This process not only helps the project team identify the key influencing factors in the cost, but also provides a scientific basis for managers to make decisions in cost control.

[0165] In order to further enhance the adaptability and dynamic response capabilities of the model, the team introduced a dynamic asymmetric information regulation and adaptive offset mechanism. By real-time monitoring and analyzing the information flow between independent variables, the model can automatically identify the key variables that have a greater impact on the project cost and dynamically adjust their weights according to market changes. This mechanism effectively solves the impact of market price fluctuations on budget preparation, enabling the model to accurately reflect the actual market situation and maintain the accuracy and stability of cost forecasts.

[0166] In the later stages of the project, the team input real substation project data into the model, combined with regional economic characteristics, to verify and revise the forecast model. Through this process, the model was further optimized to better reflect the actual market conditions in a certain region, and the cost forecast data generated was extremely accurate.

[0167] Data Proof and Beneficial Effects

[0168] Throughout the project implementation, the team conducted a detailed comparative analysis of the use of the method of the present invention and the traditional method, covering multiple stages and different types of expenses. These data further demonstrated the superiority of the method of the present invention in terms of accuracy, time efficiency and cost control.

[0169] Table 1 Data table

[0170]

[0171]

[0172] Through detailed data analysis and table comparison, it can be clearly seen that the actual application effect of the present invention in substation engineering projects is significantly better than that of traditional methods. In particular, in terms of budget accuracy, time efficiency and cost control, the present invention demonstrates its powerful advantages, providing project managers with more reliable and scientific tools, effectively reducing engineering investment risks and ensuring the smooth implementation of the project.

[0173] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for constructing an investment forecasting model based on a typical substation project, characterized in that: The method comprises the following steps: S1. Collect representative typical substation engineering schemes and historical price data in typical projects; S2, preprocessing the collected data, including noise reduction, normalization, and removal of invalid data; S3. Use the control variable method to substitute the processed historical price data one by one, replace the single independent variable, use the linear regression model to mathematically model the project cost, quantify the impact of each variable on the total cost, and calculate the impact coefficient; S4. Based on dynamic asymmetric information regulation and adaptive offset mechanism, monitor the information flow of independent variables, identify influential independent variables and dynamically adjust the weights; S5. Verify and correct the linear regression model, substitute historical real engineering data into the model, correct the data based on regional characteristics, and optimize and improve the prediction formula; S6. According to the dynamic changes in market prices and regional economy, combined with dynamic data update mechanism and adaptive control mechanism, the model is adjusted in real time to reflect the latest market information and regional characteristics, and generate the final cost forecast data.

2. The method for constructing a typical substation project investment forecasting model according to claim 1, characterized in that: The S3 comprises the following steps: S31. When calculating typical project costs, the control variable method is used to insert the collected historical price data into the model one by one, and only one independent variable is replaced each time, keeping the other independent variables unchanged, so as to calculate the impact of a single independent variable on the total project cost: Y1=β0+β1X1; Y2=β0+β2X2; Y3=β0+β3X3; … Y n =β0+β n X n ; S32. Record the cost trend of typical substation projects along with the change of each independent variable, and generate a complete cost change data set X n ; S33, the cost change data set of the typical project X n Substitute into the linear regression model and construct the following regression equation: Y=β0+β1X1+β2X2+…+β n X n +m; Among them, Y is the total cost of the project, X1, X2, ..., X n are the independent variables, β0 is the intercept term, β1,β2,...,β n is the regression coefficient of each variable, μ is the error term; S34. Calculate the regression coefficients β1, β2, ..., β n , the regression coefficient will quantify the impact of each independent variable on the total project cost; S35. Use historical real engineering data of the same standard process to verify the constructed regression model and test the model's goodness of fit R 2 Values, and significance of regression coefficients; S36. During the verification process, if it is found that the model does not meet the expected statistical standards, adjust the model through data correction, recalculate the regression coefficient, and continuously optimize the regression formula; S37. Store the finalized regression model and parameters into the database.

3. The method for constructing a typical substation project investment forecasting model according to claim 1, characterized in that: The S4 comprises the following steps: S41. Constructing enhanced dynamic asymmetric information matrix E-DAIM: S411. Standardize historical data related to typical substation projects; S412, calculate each variable X within time τ i For the independent variable X j Basic information transfer rate: To reflect the dynamic asymmetric information flow between independent variables, where Δt is the time delay parameter used to capture the lag effect between variables in the market; S413, based on all calculated information transmission rates Taking into account the time decay effect and market volatility factors, an enhanced dynamic asymmetric information matrix E-DAIM is constructed. Each element EDAI in E-DAIM ij (t) is: Among them, EDAI ij (t) represents the independent variable X i At time t, j The information transmission intensity reflects the asymmetry between the two under market conditions. λ is the time decay coefficient, which controls the gradual weakening of the impact of historical information on the current situation. κ and ω are the amplitude and frequency adjustment coefficients of information transmission, which are used to introduce the impact of market cyclical fluctuations on information transmission. S42. Based on the constructed E-DAIM matrix, monitor the information flow between the variables and calculate the information transmission intensity EDAI ij (t), identify the influential independent variables; S43. To quantify the interaction effects between key independent variables, calculate the multivariate interaction control term C ij (t0: Among them, C ij (t) is used to describe the compound interaction effect between multiple independent variables at different time points, η is the interaction regulation coefficient, which regulates the interaction between different variables, and γ is the nonlinear regulation index, which controls the intensity of the interaction effect. is the cyclical adjustment parameter, which is used to capture the effect of market cyclical changes on the interaction between independent variables; S44. Integrate information flow and historical data, combine the monitored key independent variable information flow data and historical market data, and dynamically calculate the weights W of each variable in different time periods. i (t): Among them, W i (t) is the independent variable X at time t i The dynamic weight reflects the actual influence under current market conditions, α i is the initial adjustment coefficient, θ is the time-decreasing factor, and β is the logarithmic adjustment coefficient, which is used to smooth the impact of market fluctuations on the weight of the independent variable; S45, the dynamically adjusted weight W i (t) Applied to the cost forecasting model to reflect the adjustment of the influence of market conditions on independent variables in real time; S46. When the influence of the independent variable is detected to be too large, the composite nonlinear offset factor δ is calculated. i (t): Among them, δ i (t) is used to weaken the independent variable X i The excessive influence of the independent variable is prevented from causing inaccuracy of the model prediction results due to excessive influence of the independent variable. γ is the adjustment coefficient, which controls the sensitivity of the offset factor. R i and R j Refers to the independent variable X i and X j The weights are recursively adjusted at time t, ξ and φ are periodic fluctuation adjustment parameters; S47. Real-time monitoring of market dynamics and weight changes of each variable, using sensitivity analysis to assess the marginal impact of each variable on the total cost Y(t), and calculating the dynamic offset feedback value F i (t): Among them, F i (t) is the adaptive dynamic offset feedback value, which reflects the actual impact of market changes on the independent variable. is the total cost Y(t) versus the independent variable X k The sensitivity coefficient of (t) indicates the direct impact of the change of the independent variable on the cost. and ψ are the time and periodicity adjustment factors; S47, the calculated dynamic offset feedback value F i (t) Applied to the cost prediction model, the weights of key independent variables are dynamically adjusted, and the model is continuously optimized through an adaptive optimization algorithm.

4. The method for constructing a typical substation project investment forecasting model according to claim 1, characterized in that: The S5 comprises the following steps: S51. Continuously collect real-time data from the market and pre-process the collected real-time data, including data cleaning, outlier processing and standardization; S52. Calculate the independent variable X based on real-time data i (t) is the rate of change compared to the previous time point t-1, and the formula is: Where ΔX i (t) is the independent variable X at time t i The rate of change of S53. According to the rate of change of the independent variable, the regression coefficient β in the regression model is i (t) is dynamically adjusted, and the formula is: b i (t)=βi(t-1)·(1+α·ΔX i (t)); Among them, α is the adjustment coefficient, which controls the adjustment range of the regression coefficient; β i (t) is the regression coefficient at time t; S54, Substitute the adjusted regression coefficient: Substitute the dynamically adjusted regression coefficient β i (t) is substituted into the linear regression model to update the predicted value of the total project cost Y(t) in real time: Among them, F i (t) is the dynamic offset feedback value calculated according to claim 4, and μ(t) is the error term; S55, by calculating the real-time updated total project cost Y(t), output the latest forecast result, and use it for the formulation and adjustment of the project budget; S56. Regularly compare the model's prediction results with actual engineering data to test the model's goodness of fit R 2 The significance of the regression coefficient β is adjusted according to the verification results. i (t) and dynamic offset feedback value F i (t), to correct for biases in the model; S57. Store the updated model parameters, regression coefficients and dynamic adjustment results in the database, combine with historical data analysis, continuously optimize the model structure and parameter configuration, and adapt to future market changes.

5. The method for constructing a typical substation project investment forecasting model according to claim 1, characterized in that: The S6 comprises the following steps: S61. Analyze the specific needs of the substation project, identify the scenarios where investment forecasting is required, and apply the model to specific scenarios based on different application scenarios and project requirements; S62. Collect relevant regional market economic data and characteristic parameters based on the economic conditions and market characteristics of the project area; S63. Using regional market economic data as independent variable Z j Import the model: Among them, Y adjusted Y(t) is the adjusted total project cost forecast value, Y(t) is the unadjusted total project cost forecast value, and Z j (t) is the regional characteristic parameter, γ j is the regional adjustment coefficient, which is used to reflect the impact of regional characteristics on the total construction cost; S64, using a model that is dynamically adjusted and updated in real time, substitute the independent variable data X i (t) and regional characteristic parameter Z j (t), calculate the estimated investment cost of the substation project, the formula is: Among them, Y final (t) is the final predicted total cost; S65. Analyze the calculation results, identify key influencing factors, and make corresponding adjustment suggestions based on the forecast data; S66. Output the prediction results and compare them with the actual needs of the project to evaluate the prediction accuracy of the model and verify the effectiveness of the model; S67. Based on the evaluation of the forecast results, make further adjustment suggestions; S68. The final prediction results of the model are used to support investment decisions for substation engineering projects.

Images