CN103440370A - Transmission and transformation project construction cost assessment method and device - Google Patents

Abstract

The invention provides a method and device for evaluating the cost of a power transmission and transformation project. The method includes: receiving input historical sample data of a power transmission and transformation project; initializing the number of iterations, inertia weight, learning factor, particle velocity, and particle The population scale of the swarm establishes the chaotic particle swarm model; optimizes the described chaotic particle swarm model parameters according to the chaotic particle swarm optimization algorithm; determines the iteration of the chaotic particle swarm model according to the described historical sample data and the optimized chaotic particle swarm model Optimum values of times, inertia weights, and learning factors; according to the determined iterations, inertia weights, and optimal values of learning factors, respectively determine the penalty coefficient, insensitivity coefficient, and kernel function parameter establishment of the least squares support vector machine model The least squares support vector machine model; receives the input actual sample data of the power transmission and transformation project; generates the cost evaluation result of the power transmission and transformation project according to the actual sample data of the power transmission and transformation project and the established least squares support vector machine model.

Description

Method and device for cost evaluation of power transmission and transformation project

technical field

The invention relates to the technical field of power transmission and transformation engineering, in particular to a method and device for cost evaluation of power transmission and transformation projects.

Background technique

The construction of power transmission and transformation projects generally has the characteristics of huge project investment, many fields involved, and complex influencing factors. Therefore, it has always been a difficult problem to control the construction cost of power transmission and transformation projects. The cost control of power transmission and transformation projects is based on the established cost target of the calculated and determined project cost and investment cost, strictly calculates, adjusts and supervises all costs in the cost formation process, reveals deviations, corrects them in time, and ensures the cost target realization. Therefore, in order to improve the utilization efficiency of resources, optimizing resource allocation must start from the goal of control—evaluating the project cost.

With the development of science and technology, many cost estimation methods have appeared in the field of domestic engineering cost, including: estimation index method, estimated budget method, exponential smoothing method, specific weight method, fuzzy mathematical calculation method, gray correlation degree calculation method, neural network Model method, empirical estimation method, etc. These methods can solve the rapid estimation of engineering cost in some specific historical periods and during the progress of engineering projects, but their common disadvantage is that they fix the most active factors in the competition and are difficult to adapt to the requirements of the market economic system. Ignoring the time value of funds, lack of dynamics, resulting in the separation of technology and economy, resulting in too large errors in estimated cost, it is still difficult to meet the actual needs of engineering construction in the development of a market economy.

In the 1990s, the theoretical research on machine learning of small sample data gradually matured, forming a relatively complete theoretical system - statistical learning theory. On this basis, in 1995, Vapnik proposed a new machine learning method - support Vector machine technology. The emergence of support vector machine technology provides an effective theoretical analysis basis for the learning of small sample data. It has been successfully applied in many fields and has become a new research hotspot in small sample learning. However, in actual research, it is found that it is still difficult to obtain stable and good learning results by relying solely on support vector machine technology for small-sample data learning. Improve the small sample data learning method of computer technology, design a scientific and reasonable small sample data intelligent learning improvement algorithm, and apply the improved algorithm to the rapid estimation of power transmission and transformation project cost, and meet the needs of the power transmission and transformation project construction process. The need for cost control and implementation of bidding activities.

At present, there are few studies on engineering cost estimation methods. In addition to the commonly used fixed budget estimation and list pricing methods, fuzzy mathematics, gray relational degree, artificial neural network, support vector machine and other methods are mainly used to analyze the engineering cost estimation methods. Were studied.

(1) Cost estimation based on fuzzy mathematics

Professor Wang Zhenxian first proposed that the construction cost itself is an inaccurate number with fuzzy ideas in the country, and combined with the actual application of fuzzy mathematics methods to engineering practice, he proposed a new method for quickly estimating the construction cost. Using the degree of membership to reflect the relationship between engineering projects, a group of typical engineering projects that are closest to the estimated project are selected as similar projects, and then the estimated cost of the estimated project is calculated from the actual value of this group of similar projects. Tang Xiaoyang et al. proposed to establish a random-fuzzy mathematical feature statistical method based on the theory of probability and fuzzy mathematics, and use the concept of fuzzy mathematical closeness to estimate the cost of sub-projects, and the superposition of sub-project costs constitutes the possible value of the overall project cost. Subsequently, Song Hongbin, Jiang Dehua, Li Cheng, Bi Xing and others made further research on the application of fuzzy mathematics in cost estimation.

(2) Cost estimation based on gray relational degree

Qian Yongfeng proposed for the first time to use the generating function method of the gray system to make up for the lack of uncertainty in the adjustment coefficient in the fuzzy mathematical cost estimation model. Zhang Xiekui and Qian Yongfeng used the gray system theory to estimate construction costs, but only roughly considered the characteristics of the project and did not consider the weight of sub-projects. The accuracy of the estimated cost was low and it was difficult to popularize and apply. However, Xun Zhiyuan and Yu Caihua decomposed the estimated engineering project and similar engineering projects, took the sub-project as the starting point for calculation, combined the characteristics of the sub-project and the cost to calculate the correlation degree, and made up for the shortcomings in the models established by Zhang Xiekui and Qian Yongfeng , which improves the accuracy of the estimation results. Later, Zhang Chuanyou, Huang Baozhen, Liao Qixiang, etc. respectively combined the fuzzy closeness degree and the gray relational degree, improved the previous estimation model, and further improved the accuracy of estimation.

(3) Cost estimation based on artificial neural network

Shao Liangbin and Gao Shulin introduced the engineering cost artificial neural network estimation model and artificial intelligence estimation system software through the research on the basic principles of neural network, and combined with the example of shaft engineering in the mine project construction to analyze. Subsequently, more experts and scholars applied neural networks to study the cost estimation of different construction projects. Qiang Maoshan et al. conducted research on the cost estimation of hydropower projects, and opened up the application of neural networks in hydropower projects. Shen Jinshan, Zhao Xin, Yang Yi, Fu Hong and others respectively used BP neural network to establish a project cost estimation model. Yin Tao et al. studied the application of neural network in the cost estimation method of power transmission project. In recent years, there have been studies on the combination of neural network, clustering technology, genetic algorithm and other theories to improve the learning ability of the network. Deng Huanbin, Li Chiyu, etc. combined fuzzy mathematics and BP neural network to design a rapid estimation model of project cost. Wang Ying et al. combined soft computing methods, clustering techniques and fuzzy neural network theory to design a power line project cost prediction model. Xiong Yan designed a construction project cost estimation model using genetic algorithm combined with BP neural network theory.

(4) Cost estimation based on support vector machine

Wei Juntao studied the application of support vector machine in the cost estimation method of electric power transmission project. Jiang Lina used rough set and support vector machine methods to form an intelligent forecasting system, and solved the problem of low efficiency in construction cost forecasting. Hao Kuansheng et al proposed a construction project cost prediction method based on fuzzy least squares support vector machine. Wu Xiaojuan used the system prediction method of support vector machine, combined with the trend and characteristics of thermal power project cost, to establish a thermal power plant project cost prediction model. Wang Jinxiang, Xie Ying and others used support vector machines to evaluate the cost of highway projects. Peng Guangjin proposed an improved algorithm for intelligent learning of small sample data based on parameter optimization regression support vector machine, and applied the algorithm to rapid estimation of project cost.

The disadvantage of fuzzy mathematics is that the description of complex problems of project cost estimation is too simple, so the estimation results are naturally rough. The disadvantage of the gray relational theory is that it overestimates the cost similarity of different projects, and the calculation error is large, which makes it difficult to meet the accuracy requirements of the current project cost estimation within 10%. The disadvantage of neural network is that learning requires a large training sample size to ensure the robustness and convergence of the algorithm. The disadvantage of support vector machine is that the convergence speed is slow and the running time is long. In addition, the parameters are very sensitive to external changes and depend on experience.

Looking at the cost estimation methods mentioned above, most of them are still at a superficial level, either the application of the algorithm in a single aspect, or the lack of systematization, or only applicable to engineering fields with large historical engineering data. The cost estimation method of engineering data is basically not discussed in depth.

Contents of the invention

An embodiment of the present invention provides a method for evaluating the cost of a power transmission and transformation project, and the method includes:

Receive input historical sample data of power transmission and transformation projects;

Initialize the number of iterations of chaotic particle swarm, inertia weight, learning factor, particle speed, population size of particle swarm to establish chaotic particle swarm model;

Optimizing the described chaotic particle swarm model parameters according to the chaos particle swarm optimization algorithm;

Determine the number of iterations of the chaotic particle swarm model, the inertia weight, and the optimal value of the learning factor according to the historical sample data and the optimized chaotic particle swarm model;

Determine the penalty coefficient, insensitivity coefficient and kernel function parameters of the least squares support vector machine model to establish the least squares support vector machine model according to the determined number of iterations, the inertia weight, and the optimal value of the learning factor;

Receive the actual sample data of the input power transmission and transformation project;

According to the actual sample data of power transmission and transformation projects and the established least square support vector machine model, the cost evaluation results of power transmission and transformation projects are generated.

The present invention also provides a cost evaluation device for power transmission and transformation projects, which includes:

The data input module is used to receive historical sample data and actual sample data of the input power transmission and transformation project;

The chaotic particle swarm model initialization module is used to initialize the number of iterations of the chaotic particle swarm, the inertia weight, the learning factor, the particle velocity, and the population size of the particle swarm to establish the chaotic particle swarm model;

An optimization module, configured to optimize the parameters of the chaotic particle swarm model according to the optimization of the chaotic particle swarm optimization algorithm;

An optimal value determination module is used to determine the number of iterations of the chaotic particle swarm model, the inertia weight, and the optimal value of the learning factor according to the historical sample data and the optimized chaotic particle swarm model;

The least squares support vector machine module is used to determine the penalty coefficient, insensitivity coefficient and kernel function parameters of the least squares support vector machine model according to the determined iteration number, inertia weight, and optimal value of the learning factor to establish the least squares support vector machine model;

The evaluation module is used to generate the cost evaluation result of the power transmission and transformation project according to the actual sample data of the power transmission and transformation project and the established least square support vector machine model.

The invention can not only enable the investor to accurately estimate the cost of the new project in the feasibility study stage in the early stage of project construction, but also assist the budgetary reviewers in the preliminary design stage to conduct reasonable and rapid cost review, so as to provide the basis for investment decision-making , and can help the project construction unit to quickly determine the scope of the enterprise's quotation in the bidding activities, optimize the quotation strategy on the premise of ensuring the enterprise's benefits, and maximize the success rate of winning the bid.

In order to make the above and other objects, features and advantages of the present invention more comprehensible, preferred embodiments will be described in detail below together with the accompanying drawings.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained according to these drawings without any creative effort.

Fig. 1 provides a flow chart of a power transmission and transformation project cost evaluation method for an embodiment of the present invention;

Fig. 2 also provides a structural block diagram of a power transmission and transformation project cost evaluation device for the present invention;

Fig. 3 is the construction flow chart of chaotic particle swarm least squares support vector machine evaluation model in the embodiment of the present invention

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

As shown in Figure 1, an embodiment of the present invention provides a method for evaluating the cost of a power transmission and transformation project, and the method includes:

Step S101, receiving the input historical sample data and actual sample data of the power transmission and transformation project;

Step S102, initializing the number of iterations of the chaotic particle swarm, the inertia weight, the learning factor, the particle velocity, the population size of the particle swarm, and establishing the chaotic particle swarm model;

Step S103, optimizing the parameters of the chaotic particle swarm model according to the chaotic particle swarm optimization algorithm;

Step S104, according to the historical sample data and the optimized chaotic particle swarm model, determine the number of iterations of the chaotic particle swarm model, the inertia weight, and the optimal value of the learning factor;

Step S105, substituting the determined number of iterations, inertial weights, and optimal values of learning factors into the corresponding formulas to determine the penalty coefficient, insensitivity coefficient, and kernel function parameters of the least squares support vector machine model, and finally establish the least squares support vector machine model. vector machine model;

Step S106, generating cost evaluation results of the power transmission and transformation project according to the actual sample data of the power transmission and transformation project and the established least squares support vector machine model.

Preferably, in the embodiment of the present invention, the population size of the chaotic particle swarm is set according to the sample size of the historical sample data.

Preferably, in the embodiment of the present invention, when initializing the particle velocity, the corresponding coefficient is multiplied according to the number of iterations, the inertia weight, and the magnitude of the learning factor.

Preferably, in the embodiment of the present invention, principal component analysis is performed on the sample data to determine influencing factors.

In addition, as shown in Figure 2, the present invention also provides a cost evaluation device for power transmission and transformation projects, which includes:

The data input module 201 is used to receive the historical sample data and actual sample data of the input power transmission and transformation project;

The chaotic particle swarm model initialization module 202 is used to initialize the number of iterations of the chaotic particle swarm, inertia weights, learning factors, particle speeds, and the population size of the particle swarm to establish a chaotic particle swarm model;

An optimization module 203, configured to optimize the parameters of the chaotic particle swarm model according to the chaotic particle swarm optimization algorithm;

Optimum value determination module 204, for determining the number of iterations of the chaotic particle swarm model, the inertia weight, and the optimal value of the learning factor according to the historical sample data and the optimized chaotic particle swarm model;

The least squares support vector machine module 205 is used to determine the penalty coefficient, insensitivity coefficient and kernel function parameters of the least squares support vector machine model according to the determined number of iterations, the inertia weight, and the optimal value of the learning factor to establish the least squares multiplication support vector machine model;

The evaluation module 206 is configured to generate the cost evaluation result of the power transmission and transformation project according to the actual sample data of the power transmission and transformation project and the established least squares support vector machine model.

In the embodiment of the present invention, the chaotic particle swarm least squares support vector machine evaluation model proposes the background:

Support vector machines have shown many unique advantages in solving small-sample, nonlinear and high-dimensional machine learning problems. However, it is still difficult to achieve stable and good learning results by relying solely on support vector machine technology for small-sample data learning. For this reason, this technology applies a new type of improved algorithm--least squares support vector machine based on the original support vector machine theory. The biggest difference between the improved algorithm and the support vector machine algorithm is that the improved algorithm introduces the least squares Multiply the linear system into the support vector machine, replace the traditional support vector machine and use the quadratic programming method to solve the function estimation problem. This reduces the complexity of the model, simplifies the process of building the model, and improves the accuracy of the learning results.

With the deepening of the least squares support vector machine evaluation model in engineering applications, it also exposes some inevitable defects, the most prominent is the selection and optimization of model parameters. In the past, the selection of parameters generally relied on expert systems. Or set the initial value to search blindly, etc., which will inevitably affect the accuracy of the model in practical applications and cause a certain impact. Its deficiencies are manifested as follows:

① The penalty coefficient C determines the complexity of the model and the degree of punishment for fitting deviations greater than ε according to the characteristics of the sample data. If the C value is too large (>100) or too small (<10), the generalization performance of the system will be deteriorated due to over-learning or under-learning.

②The insensitivity coefficient ε indicates the system’s expectation of the error of the estimated function on the sample data. The larger the value of ε, the smaller the number of support vectors and the sparser the expression of the solution, but too large ε can also reduce the accuracy of regression estimation.

③The kernel function parameter σ precisely defines the structure of the high-dimensional characteristic space φ(x), thus controlling the complexity of the final solution. If the value of σ is too large or too small, the generalization performance of the system will deteriorate.

How to choose reasonable parameters has become a problem in the application process of support vector machine algorithm, and it is also the focus of current application research. However, the commonly used cross-validation trial calculation method is not only time-consuming, but also the search purpose is unclear, which leads to waste of resources, time-consuming and labor-intensive, and cannot effectively optimize the parameters. Therefore, it is necessary to find a new method that can reasonably and efficiently optimize the parameters of the least squares support vector machine model, making the estimation model flexible and intelligent, and more in line with the needs of actual power transmission and transformation engineering modeling.

The particle swarm optimization algorithm is simple to implement, but it has shortcomings such as weak local search ability, easy to fall into local optimum, and slow convergence speed in the later stage of evolution. Since chaotic motion has the characteristics of ergodicity, randomness, and sensitivity to initial conditions, this technology introduces the idea of chaos into the basic particle swarm optimization algorithm, improves the diversity of the population and the ergodicity of particle search, and improves the efficiency of particle swarm optimization. The ability of the algorithm to get rid of local extreme points improves the convergence speed and accuracy of the basic particle swarm optimization algorithm. Based on this, this technology considers using the chaotic particle swarm optimization algorithm to optimize the parameter settings in the model. The basic idea of the chaotic particle swarm optimization algorithm: 1) Use the chaotic sequence to initialize the position and velocity of the particles, which does not change the random nature of the particle swarm optimization algorithm when it is initialized, and at the same time can make good use of the chaotic characteristics to improve the population's Diversity and ergodicity of particle search, on the basis of generating a large number of initial groups, select the initial group. 2) Generate a new chaotic sequence based on the optimal position searched by the current entire particle swarm, and replace the position of a particle in the current particle swarm with the optimal position particle in the chaotic sequence. The search algorithm of chaotic sequence is introduced to generate many local optimal neighborhood points in the iteration, so as to help the inert particles escape from the local minimum point, so as to quickly search for the optimal solution.

Chaotic particle swarm least squares support vector machine evaluation model construction in the embodiment of the present invention, as shown in Figure 3 is the flow chart of the construction of the chaotic particle swarm least squares support vector machine evaluation model in the embodiment of the present invention:

①Initialize and set the size M of particle swarm, the maximum allowable number of iterations L, the inertia weight W, the learning factor D, and initialize the speed of each particle. It should be noted that due to the simultaneous optimization of L, W and D, the values of the three parameters are generally not in the same order of magnitude, and the corresponding coefficients should be multiplied when initializing the particle velocity.

② Chaos initialization particle position. Randomly generate a 3-dimensional vector with each component value between 0 and 1, and get N vectors as the initial population, and then carry each component to the value range of the L, W, and D parameters respectively, and finally calculate the particle The fitness value of the group is selected, and M solutions with better performance are selected from the N initial groups as the initial solutions, and N initial velocities are randomly generated.

③ If the particle fitness is better than the individual extremum, set the fitness value of the particle swarm to the new position.

④ Particle fitness is better than the global extremum, and the global extremum is set as the new position.

⑤ Update the speed and position of the particles.

⑥ Perform chaos optimization on the optimal position. The global extremum is mapped to the domain of the Logistic equation, and then the Logistic equation is used to iterate to generate the chaotic variable sequence, and then the produced chaotic variable sequence is returned to the original solution space through inverse mapping. In the original solution space, the fitness value of each feasible solution experienced by the chaotic variable is calculated, and the feasible solution with the best performance is obtained.

⑦ Replace the position of any particle in the current population with the best feasible solution.

⑧If the maximum number of iterations is met, stop searching, and the global optimal position is the parameter vector (L, W, D); otherwise, return to the third step.

⑨ Construct the sample mean square error e _RMSE for the parameters C, ε and σ that need to be optimized as the fitness function of the least squares support vector machine, and use it as the objective function of the particle swarm optimization algorithm after chaos optimization. When the least squares support vector When the sample root mean square error of the machine is the smallest, the corresponding C, ε and σ are the optimal parameters. Finally, the least squares support vector machine evaluation model for chaotic particle swarm optimization is established.

Below in conjunction with concrete embodiment the present invention is described in further detail:

This project establishes an evaluation model for the cost stage of the 500kv substation project of Jibei Electric Power Company. An evaluation model is established for eight items including production engineering construction costs, site-related single project construction costs, other costs, and static investment. Here, the static investment cost is taken as an example to establish an evaluation model. The input samples are x ₁ : rated voltage of medium voltage side, x ₂ : rated voltage of low voltage side, x _{3: substation type, x 4 :} earthquake intensity, x ₅ : heating area or not, x ₆ : station area, x ₇ : Area of the main control building, x ₈ : support amount, x ₉ : basic amount, x ₁₀ : excavation amount, x ₁₁ : main transformer with and without load voltage regulation, x ₁₂ : number of main transformers, x ₁₃ : current capacity, x ₁₄ : Number of wires on the high-voltage side, x ₁₅ : Number of wires on the medium-voltage side, x ₁₆ : Number of wires on the low-voltage side, x ₁₇ : Wiring type on the high-voltage side, x ₁₈ : Wiring type on the medium-voltage side, x ₁₉ : Wiring type on the low-voltage side, x ₂₀ : Number of reactors, x ₂₁ : Number of capacitors, x ₂₂ : Number of disconnectors, x ₂₃ : Number of transformers, x 24 : Number of lightning arresters, x ₂₅ : Number of switchgears, x ₂₆ : Type of power distribution device on the high voltage side _, x ₂₇ : Type of power distribution device on the medium voltage side, x ₂₈ : Type of power distribution device on the low voltage side, x ₂₉ : Power cable, x ₃₀ : Control cable, these 30 influencing factors form the new index of 9 principal components. Since the historical sample size of the substation project is 29, the population size of the particle swarm is set to 29, the maximum number of iterations L is 1000, and the two parameters W and D are encoded in binary, where the search range of W is set to [0, 100], and the search range of D is set to [0.1, 100]. The initial velocity of the particles is 2. Kernel functions commonly used in least squares support vector machines include radial basis function, polynomial function, linear function, etc. Research shows that radial basis function has strong generalization ability, so this project chooses radial basis function to obtain the best evaluation model. At the same time, according to the input historical samples, to find the optimal parameters C, ε and σ. The optimal parameters obtained through calculation are: C=23, ε=520, σ=1.24, and the corresponding evaluation results are shown in Table 1:

Table 1 Comparison of prediction results of different parameters

和y(i)分别为预测值和实际值，上述表1中的参数是在适应度函数e_RMSE最小的情况下求得的。in, in

and y(i) are the predicted value and the actual value respectively, and the parameters in the above table 1 are obtained when the fitness function e _RMSE is the smallest.

It can be seen from Table 1 that when C=23, ε=520, and σ=1.24, the error between the static investment cost of the substation project obtained by the evaluation model and the real static investment cost of the substation project is the smallest. Therefore, this project uses parameters C=23, ε=520, σ=1.24 to establish a power transmission and transformation project cost evaluation model based on chaotic particle swarm least squares support vector machine. Kernel functions commonly used in least squares support vector machines include radial basis function, polynomial function, linear function, etc. Research shows that radial basis function has strong generalization ability, so this embodiment chooses radial basis function in order to obtain the most good evaluation model.

The following is a sample verification of cost evaluation using this scheme:

1. Sample verification in the budget estimation stage of 500kv substation project

(1) Static investment

The static investment evaluation index group of the 500kv substation project budgeting stage consists of 9 indicators, combined with 19 actual engineering samples and 10 typical case samples, a total of 29 sample data, based on the principle of support vector machine, using Matlab software programming, with 27 The sample is used as a learning sample, and the other two are used as a calculation sample. The obtained evaluation model verification results are as follows:

Table 2 Static investment evaluation results of 500kv substation project budget stage

sample Actual value (10,000 yuan) Estimated value (10,000 yuan) deviation Anci 500kV Substation Project 34780 34430 1.01% Qinhuangdao Changli 500kV Substation 39810 37120 6.76%

(2) Electrical part installation engineering fee

The 500kv substation project estimate stage electrical part of the installation project cost evaluation index group consists of 7 indicators, combined with 19 actual engineering samples and 10 typical case samples, a total of 29 sample data, based on the principle of support vector machine, using Matlab software programming, Taking 27 samples as learning samples and the other 2 as measurement samples, the verification results of the evaluation model are as follows:

Table 3 500kv substation project estimation stage electrical part installation project cost evaluation results

sample Actual value (10,000 yuan) Estimated value (10,000 yuan) deviation Anci 500kV Substation Project 2706.946 2614.242 3.42% Qinhuangdao Changli 500kV Substation 2938.705 2697.542 8.21%

(3) Purchase fee for electrical equipment

The 500kv substation project estimate stage electrical part of the equipment purchase evaluation index group consists of 7 indicators, combined with 19 actual engineering samples and 10 typical case samples, a total of 29 sample data, based on the principle of support vector machine, using Matlab software programming, Taking 27 samples as learning samples and the other 2 as measurement samples, the verification results of the evaluation model are as follows:

Table 4 Evaluation results of the purchase cost of electrical equipment in the budget estimation stage of 500kv substation project

sample Actual value (10,000 yuan) Estimated value (10,000 yuan) deviation Anci 500kV Substation Project 20950.211 21835.432 4.23% Qinhuangdao Changli 500kV Substation 24786.166 23097.748 6.81%

(4) Construction costs of major production projects

The 500kv substation project budget estimate stage is composed of 4 indicators for the main production engineering construction cost evaluation index group, combined with 19 actual engineering samples and 10 typical case samples, a total of 29 sample data, based on the principle of support vector machine, using Matlab software programming , taking 27 samples as learning samples and the other 2 as measuring samples, the obtained evaluation model verification results are as follows:

Table 5 500kv substation project budget estimation stage main production engineering construction cost assessment results

sample Actual value (10,000 yuan) Estimated value (10,000 yuan) deviation Anci 500kV Substation Project 3066.131 2778.681 9.38% Qinhuangdao Changli 500kV Substation 2347.506 2135.341 9.04%

(5) Auxiliary production engineering equipment purchase fee

The 500kv substation project estimate stage auxiliary production engineering equipment purchase evaluation index group consists of 7 indicators, combined with 19 actual engineering samples and 10 typical case samples, a total of 29 sample data, based on the principle of support vector machine, using Matlab software programming , taking 27 samples as learning samples and the other 2 as measuring samples, the obtained evaluation model verification results are as follows:

Table 6 Evaluation results of the purchase cost of auxiliary production engineering equipment in the budget estimation stage of 500kv substation project

sample Actual value (10,000 yuan) Estimated value (10,000 yuan) deviation Anci 500kV Substation Project 90.136 96.686 7.26% Qinhuangdao Changli 500kV Substation 105.041 112.556 7.15%

(6) Auxiliary production engineering construction costs

The 500kv substation project estimate stage auxiliary production engineering construction cost evaluation index group consists of 4 indicators, combined with 19 actual engineering samples and 10 typical case samples, a total of 29 sample data, based on the principle of support vector machine, using Matlab software programming , taking 27 samples as learning samples and the other 2 as measuring samples, the obtained evaluation model verification results are as follows:

Table 7 500kv substation project budget estimate stage auxiliary production engineering construction cost assessment results

sample Actual value (10,000 yuan) Estimated value (10,000 yuan) deviation Anci 500kV Substation Project 563.120 595.184 5.69% Qinhuangdao Changli 500kV Substation 845.683 785.563 7.11%

(7) Construction costs for individual projects related to the station site

The 500kv substation project budget estimate stage and site-related single project construction cost evaluation index group consists of 4 indicators, combined with 19 actual engineering samples and 10 typical case samples, a total of 29 sample data, based on the principle of support vector machine, using Matlab software programming, 27 samples are used as learning samples, and the other 2 are used as measurement samples. The obtained evaluation model verification results are as follows:

Table 8 500kv substation project estimation stage and station site related single project construction cost evaluation results

sample Actual value (10,000 yuan) Estimated value (10,000 yuan) deviation Anci 500kV Substation Project 694.832 656.205 5.55% Qinhuangdao Changli 500kV Substation 841.952 785.098 6.75%

(8) Other expenses

The 500kv substation project budget estimation stage and other cost evaluation index groups are composed of 2 indicators, combined with 19 actual engineering samples and 10 typical case samples, a total of 29 sample data, based on the principle of support vector machine, using Matlab software programming, to 27 One sample is used as a learning sample, and the other two are used as a calculation sample. The obtained evaluation model verification results are as follows:

Table 9 Evaluation results of other expenses in the budget estimation stage of 500kv substation project

sample Actual value (10,000 yuan) Estimated value (10,000 yuan) deviation Anci 500kV Substation Project 3381.612 3628.81 7.31% Qinhuangdao Changli 500kV Substation 6811.763 6330.626 7.06%

Determination of safety range of new cost index for substation project

Combined with the above research content, the safety range of the cost indicators corresponding to the two learning samples in the budget estimation stage of the 500kv substation project is obtained as follows:

Table 10 500kv substation project budget estimate stage indicator safety interval

Beneficial effects brought by the technical solution of the present invention

(1) Starting from artificial intelligence technology, expand and improve related concepts and algorithms around particle swarm optimization algorithm, chaotic optimization algorithm, nonlinear kernel principal component analysis and support vector machine technology, which is beneficial to the intelligent learning algorithm of small sample data improvement of.

(2) In the field of engineering cost, this technology intends to propose a systematic engineering cost based on the small sample data intelligent learning improvement algorithm, combined with historical engineering cost data, through data preprocessing, data clustering, and data classification learning. Quick cost estimation method. This technology will not only enable investors to accurately estimate the cost of new projects in the preliminary feasibility study stage of project construction, but also assist budgetary reviewers in the preliminary design stage to conduct reasonable and rapid cost reviews, so as to provide a basis for investment decisions In addition, it can help the project construction unit to quickly determine the quotation range of the enterprise in the bidding activities, optimize the quotation strategy on the premise of ensuring the enterprise's benefits, and maximize the success rate of winning the bid.

In the present invention, specific examples have been applied to explain the principles and implementation methods of the present invention, and the descriptions of the above examples are only used to help understand the method of the present invention and its core idea; meanwhile, for those of ordinary skill in the art, according to this The idea of the invention will have changes in the specific implementation and scope of application. To sum up, the contents of this specification should not be construed as limiting the present invention.

Claims (8)

1. A power transmission and transformation project cost evaluation method, characterized in that, the method comprises: Receive input historical sample data of power transmission and transformation projects; Initialize the number of iterations of chaotic particle swarm, inertia weight, learning factor, particle speed, population size of particle swarm to establish chaotic particle swarm model; According to the chaotic particle swarm optimization algorithm, the parameters of the chaotic particle swarm model are optimized; Determine the number of iterations of the chaotic particle swarm model, the inertia weight, and the optimal value of the learning factor according to the historical sample data and the optimized chaotic particle swarm model; Determine the penalty coefficient, insensitivity coefficient and kernel function parameters of the least squares support vector machine model to establish the least squares support vector machine model according to the determined number of iterations, the inertia weight, and the optimal value of the learning factor; Receive the actual sample data of the input power transmission and transformation project; According to the actual sample data of power transmission and transformation projects and the established least square support vector machine model, the cost evaluation results of power transmission and transformation projects are generated. 2. The power transmission and transformation project cost evaluation method as claimed in claim 1, wherein the population size of the described initializing chaotic particle swarm comprises: setting the population size of the chaotic particle swarm according to the sample size of the historical sample data . 3. The power transmission and transformation project cost evaluation method as claimed in claim 1, characterized in that, the particle velocity of the described initialization chaotic particle swarm comprises: when initializing the particle velocity, according to the magnitude of the number of iterations, inertia weight, and learning factor Multiply by the corresponding coefficient. 4. The cost evaluation method of power transmission and transformation project according to claim 1, characterized in that, said method further comprises: performing principal component analysis on said sample data to determine influencing factors. 5. A cost evaluation device for power transmission and transformation projects, characterized in that the method includes: The data input module is used to receive historical sample data and actual sample data of the input power transmission and transformation project; The chaotic particle swarm model initialization module is used to initialize the number of iterations of the chaotic particle swarm, the inertia weight, the learning factor, the particle velocity, and the population size of the particle swarm to establish the chaotic particle swarm model; An optimization module, configured to optimize the parameters of the chaotic particle swarm model according to the optimization of the chaotic particle swarm optimization algorithm; An optimal value determination module is used to determine the number of iterations of the chaotic particle swarm model, the inertia weight, and the optimal value of the learning factor according to the historical sample data and the optimized chaotic particle swarm model; The least squares support vector machine module is used to determine the penalty coefficient, insensitivity coefficient and kernel function parameters of the least squares support vector machine model according to the determined iteration number, inertia weight, and optimal value of the learning factor to establish the least squares support vector machine model; The evaluation module is used to generate the cost evaluation result of the power transmission and transformation project according to the actual sample data of the power transmission and transformation project and the established least square support vector machine model. 6. The power transmission and transformation project cost evaluation device as claimed in claim 5, wherein the population size of the described initializing chaotic particle swarm comprises: setting the population size of the chaotic particle swarm according to the sample capacity of the described historical sample data . 7. The power transmission and transformation project cost evaluation device as claimed in claim 5, wherein the particle velocity of the described initializing chaotic particle swarm comprises: when initializing the particle velocity, according to the magnitude of the number of iterations, inertia weight, and learning factor Multiply by the corresponding coefficient. 8 . The cost evaluation device for power transmission and transformation projects according to claim 5 , further comprising: a principal component analysis module for performing principal component analysis on the sample data to determine influencing factors.

Images