CN116070925A - Optimization method and system for power grid engineering scheme - Google Patents
Optimization method and system for power grid engineering scheme Download PDFInfo
- Publication number
- CN116070925A CN116070925A CN202111233493.XA CN202111233493A CN116070925A CN 116070925 A CN116070925 A CN 116070925A CN 202111233493 A CN202111233493 A CN 202111233493A CN 116070925 A CN116070925 A CN 116070925A
- Authority
- CN
- China
- Prior art keywords
- pso
- power grid
- model
- cost
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000005457 optimization Methods 0.000 title description 8
- 238000003062 neural network model Methods 0.000 claims abstract description 69
- 238000013210 evaluation model Methods 0.000 claims abstract description 29
- 238000004519 manufacturing process Methods 0.000 claims abstract description 12
- 238000012549 training Methods 0.000 claims description 46
- 230000003068 static effect Effects 0.000 claims description 37
- 238000010276 construction Methods 0.000 claims description 22
- 238000012360 testing method Methods 0.000 claims description 20
- 238000004422 calculation algorithm Methods 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000010219 correlation analysis Methods 0.000 claims description 8
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 claims description 7
- 229910052799 carbon Inorganic materials 0.000 claims description 7
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 claims description 6
- 238000007418 data mining Methods 0.000 claims description 6
- 229910002092 carbon dioxide Inorganic materials 0.000 claims description 3
- 239000001569 carbon dioxide Substances 0.000 claims description 3
- 230000009467 reduction Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 2
- 239000002245 particle Substances 0.000 description 21
- 238000013461 design Methods 0.000 description 9
- 238000013528 artificial neural network Methods 0.000 description 8
- 230000008901 benefit Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 238000004088 simulation Methods 0.000 description 6
- 238000007726 management method Methods 0.000 description 5
- 238000012706 support-vector machine Methods 0.000 description 5
- 230000003044 adaptive effect Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000012216 screening Methods 0.000 description 4
- 238000003860 storage Methods 0.000 description 3
- 230000001133 acceleration Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 229910000831 Steel Inorganic materials 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000002301 combined effect Effects 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000007728 cost analysis Methods 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000013439 planning Methods 0.000 description 1
- 238000004451 qualitative analysis Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Educational Administration (AREA)
- Entrepreneurship & Innovation (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Water Supply & Treatment (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Data Mining & Analysis (AREA)
- Public Health (AREA)
- Game Theory and Decision Science (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Primary Health Care (AREA)
- Mathematical Physics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a preferable method and a system of a power grid engineering scheme, comprising the following steps: calculating an influence index value of each power grid project scheme based on the acquired information of all the power grid project schemes; calculating the cost of each power grid project scheme based on the information of each power grid project scheme and a pre-established key influence index for influencing the project cost in combination with a pre-established cost assessment model; determining an optimal power grid engineering scheme based on the impact index value of each power grid engineering scheme and the manufacturing cost of the power grid engineering scheme; the cost evaluation model is constructed based on weights obtained by calculating a pre-trained PSO-SVM model and a pre-trained PSO-BP neural network model according to a prediction error standard deviation. The invention adopts two modes of image index value and cost to determine the optimal proposal according to the weight, and solves the problem of poor implementation of proposal caused by only considering the cost in the prior art.
Description
Technical Field
The invention relates to the field of big data analysis and power grid engineering, in particular to a method and a system for optimizing a power grid engineering scheme.
Background
The power grid enterprises serve as the support of national economy and play a fundamental role in economic construction. Currently, the market reform of the power industry is continuously deepening, and the great changes of the market supervision mode and the enterprise profit mode lead to greater pressure for power grid enterprises. How to fully utilize the advantages of big data information resources in the power grid industry in a gradually marketized competitive environment and develop corresponding big data analysis methods and tools, so that the fund use efficiency is optimized, the profit and the overall social benefit of the power grid enterprise are maximized, and the method and the device have become an important subject to be researched in the current stage of the power grid enterprise.
For a long time, the investment management of the power grid enterprises on the power grid construction projects still has a great defect, the projects can be researched and initially set up, the due functions of the projects are difficult to be exerted, the project investment can not be effectively controlled on the project source, a great amount of funds are idle, and the use efficiency of the funds is greatly reduced. At present, a perfect theory and method are lacked in the aspect of statistical analysis of company investment difference, and research on factors causing the difference of the staged investment conditions of power grid projects is still in a stage of qualitative analysis. Meanwhile, on the basis that the large data information management platform realizes information intercommunication and data sharing, the research of cost analysis and trend prediction is carried out by the power grid enterprises, and the method has an important role in fully utilizing the advantages of the large data to carry out effective analysis work and improving the data utilization rate. The power grid project cost prediction scheme is designed, the power grid project cost prediction precision is improved, the power grid project cost level is accurately predicted, management and control of power grid enterprises on project cost are facilitated, the fund investment degree is mastered, and the fund utilization efficiency is improved.
Before a power grid enterprise prepares to conduct project projects of a certain type, a plurality of project schemes are designed according to actual conditions of different areas, and then a preferable project scheme is selected from the project schemes.
Disclosure of Invention
In order to solve the problem that in the prior art, only engineering cost is considered to determine power grid engineering, and the power grid engineering has poor feasibility, the invention provides a preferable method of a power grid engineering scheme, which comprises the following steps:
calculating an influence index value of each power grid project scheme based on the acquired information of all the power grid project schemes;
calculating the cost of each power grid project scheme based on the information of each power grid project scheme and a pre-established key influence index for influencing the project cost in combination with a pre-established cost assessment model;
determining an optimal power grid engineering scheme based on the impact index value of each power grid engineering scheme and the manufacturing cost of the power grid engineering scheme;
the cost evaluation model is constructed based on weights obtained by calculating a pre-trained PSO-SVM model and a pre-trained PSO-BP neural network model according to a prediction error standard deviation.
Preferably, the construction of the cost evaluation model includes:
acquiring the key influence indexes and historical data of engineering static investment;
constructing a sample set by using the historical data of the key influence indexes and the historical data of the engineering static investment;
dividing the sample set into a training set and a testing set;
respectively training a PSO-SVM model and a PSO-BP neural network model based on the training set to obtain a trained PSO-SVM model and a trained PSO-BP neural network model;
calculating weights of the trained PSO-SVM model and the PSO-BP neural network model based on the test set;
constructing the cost evaluation model based on the PSO-SVM model and the PSO-BP neural network model in combination with the calculated weight;
the key influence indexes are determined by calculating gray correlation degrees between each influence factor sequence and static investment of the overhead line based on a data mining algorithm of gray correlation analysis.
Preferably, the training set is used for training the PSO-SVM model and the PSO-BP neural network model respectively to obtain a trained PSO-SVM model and a trained PSO-BP neural network model, and the training set comprises the following steps:
and respectively taking the historical data of the key influence indexes in the training set as the input of a PSO-SVM model and a PSO-BP neural network model, respectively taking the historical data of engineering static investment in the training set as the output of the PSO-SVM model and the PSO-BP neural network model, and training the PSO-SVM model and the PSO-BP neural network model to obtain a trained PSO-SVM model and a trained PSO-BP neural network model.
Preferably, the calculating weights of the trained PSO-SVM model and PSO-BP neural network model based on the test set includes:
respectively taking historical data of key influence indexes in the test set as input of a PSO-SVM model and a PSO-BP neural network model, respectively outputting a first engineering static investment predicted value and a second engineering static investment predicted value, and comparing the first engineering static investment predicted value and the second engineering static investment predicted value with the static investment of the test set engineering to determine standard deviation of fitting errors of the PSO-SVM model and the PSO-BP neural network model;
and obtaining the weights of the PSO-SVM model and the PSO-BP neural network model by using the standard deviation of the fitting errors of the PSO-SVM model and the PSO-BP neural network model.
Preferably, the cost evaluation model is represented by the following formula:
f c ＝ω 1 f 1 +ω 2 f 2
wherein f 1 ，f 2 Unbiased predictive value f of PSO-SVM model and PSO-BP neural network model c Is the predicted value of the cost evaluation model omega 1 And omega 2 The weight coefficients of the PSO-SVM model and the PSO-BP neural network model are respectively.
Preferably, the weight coefficient omega of the PSO-SVM model 1 Calculated as follows:
wherein Var (e 2 ) For the prediction error variable of the PSO-BP neural network model, cov (e 1 ,e 2 ) For Var (e 1 ) A prediction error variable amount for the PSO-SVM model;
weight coefficient omega of PSO-SVM model 2 Calculated as follows:
preferably, the key impact indicator is determined by:
carrying out standardized treatment on engineering static investment and influence factor data thereof by adopting a normalization method;
calculating gray correlation degree between each influence factor subjected to standardized treatment and engineering static investment by adopting a gray correlation analysis method;
and selecting the influence factors with gray correlation degree larger than the set threshold as key influence indexes for predicting the construction cost of the overhead line.
Preferably, the calculating the impact index value of each power grid project scheme based on the obtained information of all power grid project schemes includes:
calculating the loss rate based on the ratio between the loss difference generated before and after the power grid engineering scheme is implemented and the loss electric quantity before the power grid engineering scheme is implemented;
calculating the carbon emission reduction amount based on the carbon dioxide emission difference value before and after the implementation of the power grid engineering scheme;
wherein the impact index value includes a loss rate and a carbon emission amount.
Preferably, the determining the optimal power grid project scheme based on the impact index value of each power grid project scheme and the cost of the power grid project scheme includes:
based on the image index value of each power grid project scheme and the manufacturing cost of the power grid project scheme, summing according to a preset weight value, and calculating the total value of each power grid project scheme;
and taking the power grid engineering scheme with the minimum total value as an optimal power grid engineering scheme.
The invention provides a preferable system of a power grid engineering scheme based on the same inventive concept, which comprises the following components:
the index value calculation module is used for calculating an influence index value of each power grid project scheme based on the acquired information of all the power grid project schemes;
the cost calculation module is used for calculating the cost of each power grid project scheme based on the information of each power grid project scheme and a pre-established key influence index for influencing the project cost in combination with a pre-established cost evaluation model;
the scheme optimizing module is used for determining an optimal power grid engineering scheme based on the influence index value of each power grid engineering scheme and the manufacturing cost of the power grid engineering scheme;
the cost evaluation model is constructed based on weights obtained by calculating a pre-trained PSO-SVM model and a pre-trained PSO-BP neural network model according to a prediction error standard deviation.
Compared with the prior art, the invention has the beneficial effects that:
1. the invention provides a preferable method of a power grid engineering scheme, which comprises the following steps: calculating an influence index value of each power grid project scheme based on the acquired information of all the power grid project schemes; calculating the cost of each power grid project scheme based on the information of each power grid project scheme and a pre-established key influence index for influencing the project cost in combination with a pre-established cost assessment model; determining an optimal power grid engineering scheme based on the impact index value of each power grid engineering scheme and the manufacturing cost of the power grid engineering scheme; the cost evaluation model is constructed based on weights obtained by calculating a pre-trained PSO-SVM model and a pre-trained PSO-BP neural network model according to a prediction error standard deviation. The invention adopts two modes of image index value and cost to determine the optimal proposal according to the weight, and solves the problem of poor implementation of proposal caused by only considering the cost in the prior art.
2. The invention can realize the comparison work of different engineering design schemes under the same scene, effectively reduce the repeated verification of different engineering quantities and shorten the project cost prediction comparison and optimization period in the process of the power grid project construction. Meanwhile, the invention can accurately predict the investment amount of the project, arrange a fund raising plan and ensure the rationality and the fund utilization rate of the number of the invested funds of the project.
3. In the technical scheme provided by the invention, the construction cost of the power grid project scheme is predicted by adopting the construction cost evaluation model formed by combining the PSO-SVM and the PSO-BP neural network, so that the fitting precision is higher, the nonlinear fitting capacity and the mapping capacity are stronger, the prediction stability is stronger, the prediction of the construction cost can be better realized, the prediction risk brought by a single model can be avoided, and the prediction result is more reliable.
4. According to the technical scheme provided by the invention, the prediction result of the engineering cost and the image index value are used as the basis, the optimization of different design schemes is carried out, and the comparison of the power grid engineering projects is rapidly realized. And meanwhile, project investment funds and funds raising plans are determined by an optimal power grid engineering scheme, project investment is guaranteed to be reasonable, funds waste and long-time idling are avoided to the greatest extent, meanwhile, the loss rate and carbon emission are considered, and the resource is utilized to the greatest extent while environmental protection and energy saving are realized.
Drawings
FIG. 1 is a schematic diagram of a preferred method of a grid engineering solution of the present invention;
FIG. 2 is a flowchart of the PSO-SVM algorithm of the present invention;
FIG. 3 is a flowchart of the PSO-BP neural network algorithm of the present invention;
FIG. 4 is a graph of the gray correlation of overhead line engineering impact factors of the present invention;
FIG. 5 is a graph of the PSO-SVM predictive fit of overhead line engineering cost according to the present invention;
fig. 6 is a schematic diagram of a preferred system of a grid engineering solution according to the invention.
Detailed Description
In order to make the solution of the embodiment of the present invention better understood by those skilled in the art, the embodiment of the present invention is further described in detail below with reference to the accompanying drawings and embodiments.
Example 1:
the invention provides a preferable method of a power grid engineering scheme, as shown in fig. 1, comprising the following steps:
step 1: calculating an influence index value of each power grid project scheme based on the acquired information of all the power grid project schemes;
step 2: calculating the cost of each power grid project scheme based on the information of each power grid project scheme and a pre-established key influence index for influencing the project cost in combination with a pre-established cost assessment model;
step 3: determining an optimal power grid engineering scheme based on the impact index value of each power grid engineering scheme and the manufacturing cost of the power grid engineering scheme;
the cost evaluation model is constructed based on weights obtained by calculating a pre-trained PSO-SVM model and a pre-trained PSO-BP neural network model according to a prediction error standard deviation.
Wherein for step 1: the detailed description of calculating the impact index value of each power grid project scheme based on the acquired information of all power grid project schemes is as follows:
calculating the loss rate based on the ratio between the loss difference generated before and after the power grid engineering scheme is implemented and the loss electric quantity before the power grid engineering scheme is implemented;
calculating the carbon emission reduction amount based on the carbon dioxide emission difference value before and after the implementation of the power grid engineering scheme;
wherein the impact index value includes a loss rate and a carbon emission amount.
Wherein for step 2: calculating the cost of each power grid project scheme based on the information of each power grid project scheme and a predetermined key influence index for influencing the project cost in combination with a pre-constructed cost assessment model, wherein the description of the cost of each power grid project scheme is as follows:
based on a data mining algorithm of gray correlation analysis, referring to historical cost data of power grid engineering, calculating gray correlation degree between each influence factor sequence and static investment of the overhead line, thereby determining key influence indexes of the overhead line on the construction cost as input variables of an overhead line construction cost prediction model;
based on the input variables of the overhead line engineering cost prediction model, respectively serving as PSO-SVM and PSO-BP intelligent prediction model prediction input variables, training and simulating sample data determined in a selected area, and predicting various costs of various projects;
based on the prediction results of the two intelligent models on the various costs of various projects, the weights of the two models are determined through error analysis of training values, and the optimal combination prediction is adopted to make intelligent prediction on the change trend of the overhead line project cost in the actual project. The preferred combination herein refers to a cost assessment model.
Example 2
The prediction accuracy of the construction cost is described below by a specific case for the cost evaluation model in the present invention, and the cost evaluation model is referred to as combined prediction in this embodiment:
through selecting 104 pieces of sample data of overhead line engineering in a region with developed economy from overhead line engineering data of 2015-2017 in a certain region, and respectively training and simulating a PSO-SVM, a PSO-BP and a preferable cost evaluation model, the prediction results are as follows:
(1) Sensitive factor screening
And determining the engineering type as overhead line engineering. And (3) calculating gray correlation degree between each influence factor sequence and static investment of the overhead line by adopting a data mining algorithm based on gray correlation analysis, and identifying and screening key influence indexes of the construction cost of the overhead line. In order to ensure that the screened influence factors have high correlation with static investment, 13 influence factors with gray correlation degree larger than 0.85 are selected by taking 0.85 as a threshold according to the difference of gray correlation degree of the influence factors and combining expert experience, and as shown in figure 4, the 13 key influence factors are respectively voltage level X1, basic concrete quantity X2, basic steel quantity X3, pole tower type X5, tower material quantity X6, tower base number X8, grounding number X9, topography X10, line length X12, wire price X15, wire sectional area X16, permanent occupation cost of tower base X17 economic crops, agricultural and sideline products and other compensation cost X18, which are used as key indexes for predicting the construction cost of the overhead line, namely the input variables of the construction cost prediction model of the overhead line.
(2)PSO-SVM
The basic idea of the PSO-SVM algorithm is to construct a particle swarm support vector machine model by utilizing the combination of a support vector machine and a particle swarm, and the penalty coefficient and the kernel function obtained through optimization are used as the support vector machine to serve as final model parameters. The algorithm comprises the following steps:
(1): the algorithm control parameters are initialized. The method mainly comprises the steps of enabling the population scale of particles to be P, inertia weight, iteration times t, acceleration weight coefficient C1 of the particles themselves, global acceleration weight coefficient C2, flight speed Vmax of the particles, and random numbers rand1 and rand2 which are mutually independent and uniformly distributed;
(2): initializing group particles, mainly random positions and speeds;
(3): calculating an adaptation value for each particle;
(4): comparing the adaptive value of each particle with the adaptive value of the historical best position, and taking the adaptive value as the individual best value if the adaptive value of each particle is good;
(5): comparing the individual optimum value of each particle with the adapted value of the global optimum position P, if so, taking it as the global optimum;
(6): calculating inertia weight, and updating particle speed and position;
(7): judging whether a termination condition is met, if yes, finishing searching, outputting optimal parameters of a support vector machine, otherwise, jumping to the step 3, and continuing searching;
(8): substituting the obtained optimal parameters into a support vector machine prediction model.
For a clearer description of the algorithm process, a learning algorithm flow chart is presented herein, as shown in fig. 2.
According to the steps, the specific data implementation mode is as follows:
1) And initializing data. And taking the key influence factors of the construction cost of the overhead line project obtained by screening as PSO-SVM prediction input variables, taking static investment of the overhead line project as output variables, and carrying out training simulation on 94 sample data in the economic developed area of Jiangsu province in 2015-2017.
2) Training and simulation. And establishing an intelligent prediction model based on the PSO-SVM according to the overhead line engineering historical data, training and learning the historical data, verifying a model training result by taking static investment cost of 10 engineering, and comparing the demonstration data and the prediction data as shown in the following table. And solving a relative fitting error according to a formula by utilizing the fitting output value and the actual value obtained by the PSO-SVM model, wherein the relative error formula is as follows:
wherein Y represents an actual value,
representing the fit value.
TABLE 1 PSO-SVM prediction results for overhead line engineering
| Engineering |
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
| Practical cost | 7755 | 1302 | 590 | 4326 | 1005 | 351 | 3046 | 3114 | 10668.8 | 6907.561 |
| Predicting cost | 6994 | 1301 | 590 | 4369 | 1011 | 241 | 3046 | 3093 | 4928 | 5361 |
| Fitting error | 0.098 | 0.001 | 0.009 | 0.010 | 0.006 | 0.313 | 0.008 | 0.007 | 0.191 | 0.224 |
In order to more intuitively embody the prediction effect of the PSO-SVM prediction model on the overhead line engineering cost, the predicted value and the actual value are drawn as shown in figure 5.
(3) PSO-BP neural network
S1: initializing: firstly, determining input variables and output variables of a neural network model, and establishing a network topological structure, namely the number n of nodes of an input layer, the number h of nodes of an hidden layer and the number m of nodes of an output layer; initializing the position of particles
And speed->
Total number of particle swarms N, maximum iteration number T max Maximum and minimum values omega of inertia weight omega max And omega min Learning factor c 1 And c 2 Isoparametric parameters;
s2: training and simulation of sample data: inputting the input variable and the output variable data into the BP neural network model, and calculating the fitness of the particles to obtain the individual optimal and global optimal values of the particles. Comparing the individual with the pbest and the gbest, and recording the current position of the best particle; the fitness value of each particle was evaluated. If the value is better than the individual optimum, the individual optimum P bd Setting the particle to be a current value and updating the individual optimum of the particle; if the individual optimum in the particle is better than the current global optimum, setting the individual optimum as the global optimum of the particle, and updating the global extremum; optimizing weight and threshold of BP neural network by PSO algorithm, substituting the optimized value as initial weight of BP neural network into BP network for training, regulating weight and threshold by BP network training, stopping iteration when the Mean Square Error (MSE) of network performance index is smaller than preset error requirement or maximum iteration number, outputting result, otherwise continuing iteration until algorithm converges, PSO-BThe algorithm flow of the P neural network is shown in figure 3;
s3: and (3) predicting the manufacturing cost: substituting the historical data of each sensitive factor into the PSO-BP neural network model to predict each expense of each type of engineering;
s4: analysis of results: and analyzing the change trend of various costs of various projects according to the prediction result.
According to the steps, the specific data implementation mode is as follows:
1) And initializing data. According to the 13 key factors which influence the construction cost of the overhead line project and are screened out, the key factors are used as input variables of a PSO-BP intelligent prediction model, the static investment of the overhead line project is used as output variables, and 94 sample data of the economic developed area of Jiangsu province in 2015-2017 are trained and simulated.
2) Training and simulation. And establishing an intelligent PSO-BP neural network prediction model according to the overhead line engineering historical samples, dividing 94 sample data, wherein 84 sample data are used for model training and learning, taking 10 engineering data as a test set, and performing prediction result test, wherein the results of the test set are shown in the following table.
TABLE 3 PSO-BP prediction for overhead line engineering
In order to more intuitively embody the effect of predicting the overhead line engineering cost by the PSO-BP model.
(4) Preferably combined prediction
The combined prediction method is based on the maximum information utilization, and integrates the information contained in various single models to perform optimal combination. Thus, in most cases, the objective of improving the prediction results can be achieved by combining predictions. When the prediction accuracy of each model predicted value can be known, a weighted average method should be used to give a larger weight to the more accurate predicted value. The variance-covariance method (MV method) can be effectively used for solving the problem of weight selection.
Let f 1 ，f 2 Is two unbiased predictions about f c Is the combined predictor of the weighted average. The prediction errors are respectively e 1 ,e 2 And e c Taking ω 1 And omega 2 Is the corresponding weight coefficient, and omega 1 +ω 2 =1, have
f c ＝ω 1 f 1 +ω 2 f 2 (1)
Requirement f c Is also unbiased and the errors and their variances are respectively
e c ＝ω 1 e 1 +ω 2 e 2 (2)
This is due to: let x be (0) (t) is the original sequence, and has:
e 1 ＝x (0) (t)-f 1 (t)，e 2 ＝x (0) (t)-f 2 (t)，e c ＝x (0) (t)-f c (t) (3)
then:
thus, there are:
Var(e c )＝ω 1 2 Var(e 1 )+ω 2 2 Var(e 2 )+2ω 1 ω 2 cov(e 1 ,e 2 ) (5)
with respect to omega 1 For Var (e) c ) Obtaining minimum value to obtain
And omega 2 ＝1-ω 1 . Record Var (e) 1 )＝σ 11 ，Var(e 2 )＝σ 22 ，cov(e 1 ,e 2 )＝σ 12
The combined prediction weight coefficients of the two prediction methods are respectively
Due to e 1 、e 2 Independent of each other, sigma 12 =0, have
As can be seen from the above, there are
1) Model training and error, carrying out data training on a PSO-SVM prediction model and a PSO-BP neural network model in the two previous sections, and obtaining the error result as follows:
table 5 model training relative error contrast table
2) Determining model weight, and calculating standard deviation sigma of two model fitting errors according to the data in the table 1 Sum sigma 2 0.0492 and 0.1106 respectively, and then calculate the weights ω of the two models from equation (9) and equation (10) 1 =0.69 and ω 2 =0.31, where the model predictive value calculation formula is:
f c ＝0.69f 1 +0.31f 2 (12)
wherein f c Representing the predicted value of the cost evaluation model, f 1 Representing the predicted value, f, of the PSO-BP neural network model 2 Representing the PSO-SVM model predictive value.
3) Model solving, the optimal cost evaluation model predicted value and the predicted error are obtained according to the above formula solving as shown in the following table 6.
TABLE 6 overhead line engineering preference prediction results
As shown in the table above, the maximum relative error between the predicted value and the actual value of the overhead line project cost obtained by the combined prediction is 21%, the minimum relative error is 3.26%, and the average relative error of the overhead line project cost prediction is 10.3%. The method has the advantages that the fitting precision of the optimal combined prediction model is higher, the nonlinear fitting capacity and the mapping capacity are higher, the prediction of the overhead line engineering cost can be better realized, meanwhile, the combined prediction model considers the combined effect of the two models, the prediction stability is higher, the prediction risk brought by a single model can be avoided, and the prediction result is more reliable.
The specific embodiment can show that the combined prediction method for the power grid engineering cost constructed by the invention has higher prediction precision and stability than a single prediction method, and is beneficial to the management and control of the power grid engineering cost level. The specific power grid engineering combination prediction and optimization scheme is as follows:
(1) Historical cost data of the power grid engineering is collected, wherein the historical cost data comprises influence factor related data and cost related data.
(2) And screening key technical factors influencing the construction cost by adopting a data mining method of gray correlation analysis.
(3) And taking the screened key influencing factors as input variables, performing simulation training on the PSO-SVM and the PSO-BP prediction model to obtain a trained PSO-SVM prediction model and a trained PSO-BP neural network model, and determining weights of the two models through error analysis of training values to obtain a cost evaluation model.
(4) And (3) taking the scheme in the design stage of the power grid project as a benchmark, inputting the related data of the key influence factors of the current power grid project into a cost evaluation model for training for different design schemes, and calculating to obtain project cost predicted values corresponding to different power grid project design schemes.
(5) The project cost predicted values of different design schemes of the project are used as the reflection of the current scheme cost level, the comparison of technical and economic cost is rapidly carried out, and the design scheme with the optimal economic benefit is selected as the combined prediction and optimization result;
(6) And taking the optimized design scheme as an actual construction scheme, carrying out cost management by referring to a model prediction result, reasonably determining the investment level of the engineering project by referring to the prediction value, reasonably planning the engineering cost in each aspect, and realizing the balance of economic benefit and social benefit.
Example 3
Based on the same inventive concept, the invention also provides a preferable system of the power grid engineering scheme, as shown in fig. 6, comprising:
the index value calculation module is used for calculating an influence index value of each power grid project scheme based on the acquired information of all the power grid project schemes;
the cost calculation module is used for calculating the cost of each power grid project scheme based on the information of each power grid project scheme and a pre-established key influence index for influencing the project cost in combination with a pre-established cost evaluation model;
the scheme optimizing module is used for determining an optimal power grid engineering scheme based on the influence index value of each power grid engineering scheme and the manufacturing cost of the power grid engineering scheme;
the cost evaluation model is constructed based on weights obtained by calculating a pre-trained PSO-SVM model and a pre-trained PSO-BP neural network model according to a prediction error standard deviation.
The construction of the cost evaluation model comprises the following steps:
acquiring the key influence indexes and historical data of engineering static investment;
constructing a sample set by using the historical data of the key influence indexes and the historical data of the engineering static investment;
dividing the sample set into a training set and a testing set;
respectively training a PSO-SVM model and a PSO-BP neural network model based on the training set to obtain a trained PSO-SVM model and a trained PSO-BP neural network model;
calculating weights of the trained PSO-SVM model and the PSO-BP neural network model based on the test set;
constructing the cost evaluation model based on the PSO-SVM model and the PSO-BP neural network model in combination with the calculated weight;
the key influence indexes are determined by calculating gray correlation degrees between each influence factor sequence and static investment of the overhead line based on a data mining algorithm of gray correlation analysis.
Preferably, the training set is used for training the PSO-SVM model and the PSO-BP neural network model respectively to obtain a trained PSO-SVM model and a trained PSO-BP neural network model, and the training set comprises the following steps:
and respectively taking the historical data of the key influence indexes in the training set as the input of a PSO-SVM model and a PSO-BP neural network model, respectively taking the historical data of engineering static investment in the training set as the output of the PSO-SVM model and the PSO-BP neural network model, and training the PSO-SVM model and the PSO-BP neural network model to obtain a trained PSO-SVM model and a trained PSO-BP neural network model.
Preferably, the calculating weights of the trained PSO-SVM model and PSO-BP neural network model based on the test set includes:
respectively taking historical data of key influence indexes in the test set as input of a PSO-SVM model and a PSO-BP neural network model, respectively outputting a first engineering static investment predicted value and a second engineering static investment predicted value, and comparing the first engineering static investment predicted value and the second engineering static investment predicted value with the static investment of the test set engineering to determine standard deviation of fitting errors of the PSO-SVM model and the PSO-BP neural network model;
and obtaining the weights of the PSO-SVM model and the PSO-BP neural network model by using the standard deviation of the fitting errors of the PSO-SVM model and the PSO-BP neural network model.
Preferably, the cost evaluation model is represented by the following formula:
f c ＝ω 1 f 1 +ω 2 f 2
wherein f 1 ，f 2 Unbiased predictive value f of PSO-SVM model and PSO-BP neural network model c Is the predicted value of the cost evaluation model omega 1 And omega 2 The weight coefficients of the PSO-SVM model and the PSO-BP neural network model are respectively.
Preferably, the weight coefficient omega of the PSO-SVM model 1 Calculated as follows:
wherein Var (e 2 ) For the prediction error variable of the PSO-BP neural network model, cov (e 1 ,e 2 ) For Var (e 1 ) A prediction error variable amount for the PSO-SVM model;
weight coefficient omega of PSO-SVM model 2 Calculated as follows:
the data extraction and simulation subunit is used for establishing an intelligent prediction model of the PSO-SVM or PSO-BP neural network for the overhead line engineering historical samples, dividing 94 sample data, wherein 84 sample data are used for model training and learning, taking 10 engineering data as a test set, and testing a prediction result.
Preferably, the combination prediction module: and carrying out data training on the PSO-SVM prediction model and the PSO-BP neural network model to obtain an error result, calculating weights of the two models according to standard deviation of fitting errors of the two models, and carrying out optimization of different design schemes on prediction of the overhead line engineering cost according to a model prediction value calculation formula so as to quickly realize economic cost comparison in the power grid engineering project construction process.
It will be apparent that the described embodiments are some, but not all, embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The foregoing is illustrative of the present invention and is not to be construed as limiting thereof, but rather as providing for the use of additional embodiments and advantages of all such modifications, equivalents, improvements and similar to the present invention are intended to be included within the scope of the present invention as defined by the appended claims.
Claims (10)
1. A preferred method of grid engineering project comprising:
calculating an influence index value of each power grid project scheme based on the acquired information of all the power grid project schemes;
calculating the cost of each power grid project scheme based on the information of each power grid project scheme and a pre-established key influence index for influencing the project cost in combination with a pre-established cost assessment model;
determining an optimal power grid engineering scheme based on the impact index value of each power grid engineering scheme and the manufacturing cost of the power grid engineering scheme;
the cost evaluation model is constructed based on weights obtained by calculating a pre-trained PSO-SVM model and a pre-trained PSO-BP neural network model according to a prediction error standard deviation.
2. The method of claim 1, wherein the construction of the cost assessment model comprises:
acquiring the key influence indexes and historical data of engineering static investment;
constructing a sample set by using the historical data of the key influence indexes and the historical data of the engineering static investment;
dividing the sample set into a training set and a testing set;
respectively training a PSO-SVM model and a PSO-BP neural network model based on the training set to obtain a trained PSO-SVM model and a trained PSO-BP neural network model;
calculating weights of the trained PSO-SVM model and the PSO-BP neural network model based on the test set;
constructing the cost evaluation model based on the PSO-SVM model and the PSO-BP neural network model in combination with the calculated weight;
the key influence indexes are determined by calculating gray correlation degrees between each influence factor sequence and static investment of the overhead line based on a data mining algorithm of gray correlation analysis.
3. The method of claim 2, wherein training the PSO-SVM model and the PSO-BP neural network model based on the training set to obtain a trained PSO-SVM model and a trained PSO-BP neural network model, respectively, comprises:
and respectively taking the historical data of the key influence indexes in the training set as the input of a PSO-SVM model and a PSO-BP neural network model, respectively taking the historical data of engineering static investment in the training set as the output of the PSO-SVM model and the PSO-BP neural network model, and training the PSO-SVM model and the PSO-BP neural network model to obtain a trained PSO-SVM model and a trained PSO-BP neural network model.
4. The method of claim 2, wherein the calculating weights of the trained PSO-SVM model and PSO-BP neural network model based on the test set comprises:
respectively taking historical data of key influence indexes in the test set as input of a PSO-SVM model and a PSO-BP neural network model, respectively outputting a first engineering static investment predicted value and a second engineering static investment predicted value, and comparing the first engineering static investment predicted value and the second engineering static investment predicted value with the static investment of the test set engineering to determine standard deviation of fitting errors of the PSO-SVM model and the PSO-BP neural network model;
and obtaining the weights of the PSO-SVM model and the PSO-BP neural network model by using the standard deviation of the fitting errors of the PSO-SVM model and the PSO-BP neural network model.
5. The method of claim 3, wherein the cost assessment model is of the formula:
f c ＝ω 1 f 1 +ω 2 f 2
wherein f 1 ，f 2 Unbiased predictive value f of PSO-SVM model and PSO-BP neural network model c Is the predicted value of the cost evaluation model omega 1 And omega 2 The weight coefficients of the PSO-SVM model and the PSO-BP neural network model are respectively.
6. The method of claim 5, wherein the weight coefficient ω of the PSO-SVM model 1 Calculated as follows:
wherein Var (e 2 ) For the prediction error variable of the PSO-BP neural network model, cov (e 1 ,e 2 ) For Var (e 1 ) A prediction error variable amount for the PSO-SVM model;
weight coefficient omega of PSO-SVM model 2 Calculated as follows:
7. the method of claim 2, wherein the key impact indicator is determined by:
carrying out standardized treatment on engineering static investment and influence factor data thereof by adopting a normalization method;
calculating gray correlation degree between each influence factor subjected to standardized treatment and engineering static investment by adopting a gray correlation analysis method;
and selecting the influence factors with gray correlation degree larger than the set threshold as key influence indexes for predicting the construction cost of the overhead line.
8. The method of claim 1, wherein calculating an impact index value for each grid project plan based on the obtained information for all grid project plans comprises:
calculating the loss rate based on the ratio between the loss difference generated before and after the power grid engineering scheme is implemented and the loss electric quantity before the power grid engineering scheme is implemented;
calculating the carbon emission reduction amount based on the carbon dioxide emission difference value before and after the implementation of the power grid engineering scheme;
wherein the impact index value includes a loss rate and a carbon emission amount.
9. The method of claim 8, wherein said determining an optimal grid project plan based on the impact indicator value for each grid project plan and the cost of the grid project plan comprises:
based on the image index value of each power grid project scheme and the manufacturing cost of the power grid project scheme, summing according to a preset weight value, and calculating the total value of each power grid project scheme;
and taking the power grid engineering scheme with the minimum total value as an optimal power grid engineering scheme.
10. A preferred system for a power grid project, comprising:
the index value calculation module is used for calculating an influence index value of each power grid project scheme based on the acquired information of all the power grid project schemes;
the cost calculation module is used for calculating the cost of each power grid project scheme based on the information of each power grid project scheme and a pre-established key influence index for influencing the project cost in combination with a pre-established cost evaluation model;
the scheme optimizing module is used for determining an optimal power grid engineering scheme based on the influence index value of each power grid engineering scheme and the manufacturing cost of the power grid engineering scheme;
the cost evaluation model is constructed based on weights obtained by calculating a pre-trained PSO-SVM model and a pre-trained PSO-BP neural network model according to a prediction error standard deviation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111233493.XA CN116070925A (en) | 2021-10-22 | 2021-10-22 | Optimization method and system for power grid engineering scheme |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111233493.XA CN116070925A (en) | 2021-10-22 | 2021-10-22 | Optimization method and system for power grid engineering scheme |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116070925A true CN116070925A (en) | 2023-05-05 |
Family
ID=86177193
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111233493.XA Pending CN116070925A (en) | 2021-10-22 | 2021-10-22 | Optimization method and system for power grid engineering scheme |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116070925A (en) |
-
2021
- 2021-10-22 CN CN202111233493.XA patent/CN116070925A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Moonchai et al. | Short-term forecasting of renewable energy consumption: Augmentation of a modified grey model with a Kalman filter | |
| CN112686464A (en) | Short-term wind power prediction method and device | |
| CN106991212B (en) | Root strength prediction method based on GA _ PSO (genetic Algorithm-particle swarm optimization) GRNN (generalized regression neural network) algorithm | |
| CN111639783A (en) | Line loss prediction method and system based on LSTM neural network | |
| CN106485348A (en) | A kind of Forecasting Methodology of transaction data and device | |
| CN112884236B (en) | Short-term load prediction method and system based on VDM decomposition and LSTM improvement | |
| CN111695754A (en) | Electric power Internet of things information security risk assessment method and device | |
| CN113983646A (en) | Air conditioner interaction end energy consumption prediction method based on generation countermeasure network and air conditioner | |
| CN116187835A (en) | Data-driven-based method and system for estimating theoretical line loss interval of transformer area | |
| CN115470862A (en) | Dynamic self-adaptive load prediction model combination method | |
| CN114880806A (en) | New energy automobile sales prediction model parameter optimization method based on particle swarm optimization | |
| Pan et al. | Dynamic Task Allocation Algorithm for Hiring Workers that Learn. | |
| CN113762591A (en) | Short-term electric quantity prediction method and system based on GRU and multi-core SVM counterstudy | |
| CN112182835A (en) | Power system reliability assessment method and system considering wind power uncertainty and energy storage regulation | |
| CN114648178B (en) | Operation and maintenance strategy optimization method of electric energy metering device based on DDPG algorithm | |
| CN116070925A (en) | Optimization method and system for power grid engineering scheme | |
| CN112581311B (en) | Method and system for predicting long-term output fluctuation characteristics of aggregated multiple wind power plants | |
| CN114881344A (en) | Training method, device and medium for building energy consumption prediction model | |
| CN113987261A (en) | Video recommendation method and system based on dynamic trust perception | |
| CN114372558A (en) | Residential electricity load prediction method, medium and equipment based on multi-model fusion | |
| CN111291464A (en) | Dynamic equivalence method and device for power system | |
| Al-Ahmadi et al. | Spatial optimization of urban cellular automata model | |
| CN117273392B (en) | Furniture production decision method and device, electronic equipment and storage medium | |
| CN113705067B (en) | Microgrid optimization operation strategy generation method, system, equipment and storage medium | |
| Sousa et al. | Adaptive learning in multiagent systems: a forecasting methodology based on error analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |