Engineering Cost Audit Method and System

Abstract

This application relates to the technical field of engineering project management, and particularly to a project cost auditing method and system, which includes collecting a number of historical engineering project documents, determining a baseline project, calculating the impact factors of all historical projects except the baseline project and the main material consumption ratios of each major material based on the historical engineering project documents; training a preset support vector machine model based on the impact factors of the historical projects and the main material consumption ratios of each major material to audit whether the project cost is qualified according to the impact factors of the project and the main material consumption ratios of each major material; calculating the impact factors of the target project and the main material consumption ratios of each major material; inputting the impact factors of the target project and the main material consumption ratios of each major material into the model, and outputting the project cost audit results of each major material of the target project. This application can achieve intelligent project cost auditing, reduce the dependence on manual experience, and thus complete efficient and accurate auditing.

Description

Technical Field

[0001] The present application relates to the technical field of engineering project management, and in particular to an engineering cost audit method and system. Background Art

[0002] Project cost refers to the estimated or actual construction costs of a project during the construction period. Accurate project cost review is not only related to the reasonable budget and cost control of the project, but also has important significance for ensuring project quality and reducing project risks. With the continuous expansion of project construction and the increase in project complexity, project cost accounting, as an important link to ensure the smooth implementation of the project, has received more and more attention for its scientificity and accuracy.

[0003] At present, the rapid development of the digital economy era has prompted the traditional construction engineering cost consulting business to move towards digital transformation. In the early consultation stage of construction projects, cost consulting companies often face the dilemma of insufficient design drawings or project information that cannot meet the depth requirements. In this context, it is required to be able to quickly review engineering projects in the absence of data. In the existing solutions, companies mostly rely on manual review methods, using historical project data accumulation and traditional experience to perform cost accounting to meet business needs.

[0004] However, the existing manual audit technology is overly dependent on individual experience, which makes it difficult to achieve efficient and relatively accurate audits in complex data environments. At the same time, with the rapid development of intelligent technology, how to achieve intelligent engineering cost audits, reduce reliance on manual experience, and thus complete efficient and accurate audits is a key problem that needs to be solved urgently. Summary of the invention

[0005] This application provides a method and system for project cost auditing, which can realize intelligent project cost auditing, reduce the reliance on manual experience, and thus complete efficient and accurate auditing. This application provides the following technical solutions:

[0006] In a first aspect, the present application provides a method for project cost auditing, the method comprising:

[0007] Collect several historical engineering project files, determine one of the historical projects as the benchmark project, and calculate the impact factors of all historical projects except the benchmark project and the main material consumption ratio of each main material based on the historical project files;

[0008] Based on the influencing factors of the historical projects and the main material consumption ratio of each main material, a preset support vector machine model is trained, and the support vector machine model is used to review whether the project cost is qualified according to the influencing factors of the project and the main material consumption ratio of each main material;

[0009] Obtain the target engineering project file, calculate the impact factor of the target project and the main material consumption ratio of each main material;

[0010] Input the impact factor of the target project and the main material consumption ratio of each main material into the trained support vector machine model, and the support vector machine model outputs the cost audit results of each main material of the target project.

[0011] In a specific implementable solution, calculating the impact factor of all historical projects except the benchmark project and the main material consumption ratio of each main material based on the historical engineering project file includes:

[0012] The number of floors and the above-ground building area of each historical project can be directly obtained from the construction drawings in the historical project file;

[0013] Obtain the building perimeter and building area of the standard floor of the project from the architectural drawings in the construction drawings, and use the calculation formula to obtain the perimeter coefficient equal to the building perimeter / building area;

[0014] Obtain the structural area and building area of the standard floor of the project from the structural drawings in the construction drawings, and use the calculation formula to obtain the structural area coefficient equal to the structural area / building area;

[0015] Indicating the main material In the th historical project, the total consumption, and its corresponding main material consumption The calculation formula of is that the main material consumption is equal to / the above-ground building area of the th historical project , and calculate the main material consumption of the corresponding main material in the benchmark project as ; The main material consumption of the main material is divided by the main material consumption of the corresponding material in the benchmark project to obtain the main material consumption ratio of the main material , and then the number of floors of the th historical project , perimeter coefficient and structural area coefficient are divided by the corresponding number of floors of the benchmark project , perimeter coefficient and structural area coefficient , to obtain the floor ratio , perimeter coefficient ratio and structural area ratio as the impact factor of the th historical project.​​

[0016] In a specific feasible implementation, training the preset support vector machine model based on the impact factor of the historical project and the main material consumption ratio of each main material includes:

[0017] Calculating the maximum reasonable deviation rate corresponding to the main material of each historical project;

[0018] Construct a two-dimensional plane, where the abscissa of the two-dimensional plane is the comprehensive impact factor and the ordinate is the main material consumption ratio under the influence of the comprehensive impact factor. Map the main material consumption ratio of each historical project under the influence of its comprehensive impact factor in the two-dimensional plane and represent it as a point;

[0019] Map the ratio of the maximum reasonable deviation rate corresponding to the main material of each historical project to the maximum reasonable deviation rate of the main material corresponding to the benchmark project in the two-dimensional plane and represent it as a line;

[0020] Design a loss function, and continuously adjust the parameters by the method of gradient descent until the loss function is minimized, obtain the optimal parameters that make all historical project data meet the qualified conditions, and complete the training work of the support vector machine model.

[0021] In a specific feasible implementation, the calculating the maximum reasonable deviation rate corresponding to the main material of each historical project includes:

[0022] The maximum reasonable deviation rate corresponding to the main material in the th historical project is calculated as follows:

[0023] ;

[0024] Wherein, is the reasonable allowable deviation rate of the project cost of the th historical project set artificially based on project cost experience, is the ratio of the combined price of the material part in the th historical project to the total project cost of this project, represents the number of types of main materials in the th historical project, represents the total amount of the main material in the th historical project, represents the corresponding unit price of the main material in the th historical project.

[0025] In a specific feasible implementation, mapping the main material consumption ratio of each historical project under the influence of its comprehensive influence factor in a two-dimensional plane and presenting it as a point includes:

[0026] Assign weights to the three influence factors using the weighted Euclidean norm, and then calculate the square root of the sum of the squares of the three influence factors to obtain the comprehensive influence factor , and the formula is as follows:

[0027] ;

[0028] Among them, , , are the weights assigned to the three influence factors respectively;

[0029] Based on the comprehensive influence factor, use the denominator attenuation model to calculate the main material consumption ratio of the main material in the th historical project under the influence of the comprehensive influence factor as follows:

[0030] ;

[0031] Among them, is a correction coefficient greater than 0;

[0032] The coordinates of the point on the two-dimensional plane where the main material consumption ratio of the main material in the th historical project is mapped under the influence of its comprehensive influence factor are .

[0033] In a specific feasible implementation, the design of the loss function, continuously adjusting the parameters by the method of gradient descent until the loss function is minimized, includes:

[0034] Design the loss function as shown below:

[0035] ;

[0036] Among them, represents the main material consumption ratio of the main material in the th historical project under the influence of the comprehensive influence factor, represents the maximum reasonable deviation rate of the main material in the th historical project, represents the maximum reasonable deviation rate of the main material in the reference version project, is the regularization coefficient, is the economic weight, reflecting the proportion of the main materials in the total project cost in the nth historical project, and the calculation method is as follows:

[0037] ;

[0038] Among them, represents the number of types of main materials in the nth historical project, represents the total consumption of the main materials in the nth historical project, represents the corresponding unit price of the main materials in the nth historical project, is the ratio of the combined price of the material part in the nth historical project to the total project cost of this project, represents the total cost of all main materials in the nth historical project;

[0039] By means of gradient descent, continuously adjust the parameter , until the loss function is minimized. After each adjustment, check that the points mapped by the main materials of each historical project should be below the horizontal decision line mapped by the corresponding maximum reasonable deviation rate. If there are still points exceeding it, continue to adjust the parameters until a set of optimal parameters is found so that all historical project data meet the qualified conditions.

[0040] In a specific implementable solution, after calculating the impact factor of the target project and the main material consumption ratio of each main material, it further includes:

[0041] Calculate the value range of the main material consumption ratio of each main material based on the number of floors, perimeter coefficient, and structural area coefficient of the project corresponding to each main material. The calculation formula is as follows:

[0042] ;

[0043] ;

[0044] Among them, is the sum of the impact factors of the target project corresponding to the main material , is the proportionality coefficient of the main material , is the deviation tolerance value of the main material ;

[0045] After calculating the value range of the main material consumption ratio of the main material, it is determined whether the main material consumption ratio of the main material calculated based on the target engineering project document before is within the value range. If it is within the value range, it indicates that the calculated main material consumption ratio of the main material is correct. If it is not within the value range, it indicates that there is a problem with the calculated main material consumption ratio of the main material, and the verification of the main material consumption ratio fails. At this time, it is necessary to recalculate.

[0046] In a second aspect, the present application provides a project cost auditing system, adopting the following technical solutions:

[0047] A project cost auditing system includes:

[0048] A historical data collection module, configured to collect a number of historical engineering project documents, determine one of the historical projects as a reference project, and calculate the impact factors of all historical projects except the reference project and the main material consumption ratio of each main material based on the historical engineering project documents:

[0049] A support vector machine model training module, configured to train a preset support vector machine model based on the impact factors of historical projects and the main material consumption ratio of each main material, and the support vector machine model is used to audit whether the project cost is qualified according to the impact factors of the project and the main material consumption ratio of each main material;

[0050] A target data acquisition module, configured to acquire a target engineering project document and calculate the impact factors of the target project and the main material consumption ratio of each main material;

[0051] A support vector machine model auditing module, configured to input the impact factors of the target project and the main material consumption ratio of each main material into the trained support vector machine model, and the support vector machine model outputs the project cost auditing results of each main material of the target project.

[0052] In a third aspect, the present application provides an electronic device, the device includes a processor and a memory; a program is stored in the memory, and the program is loaded and executed by the processor to implement a project cost auditing method as described in the first aspect.

[0053] In a fourth aspect, the present application provides a computer-readable storage medium, a program is stored in the storage medium, and when the program is executed by a processor, it is used to implement a project cost auditing method as described in the first aspect.

[0054] In summary, the beneficial effects of the present application at least include:

[0055] (1) By introducing normalization processing, impact factors, and a calculation method for reasonable deviation rates, the accuracy and scientificity of project cost auditing have been significantly improved. Specifically, the normalization method standardizes the main material consumption ratios of different projects, enabling projects of different scales and complexities to be compared and evaluated under the same standard. This eliminates the data asymmetry problem caused by project scale differences in traditional auditing methods. By selecting the benchmark project with the most complete data and comparing the data of the target project with it, the comparability and consistency of the data are ensured. The introduction of impact factors takes into account the influence of project characteristics on material consumption, making the auditing process more accurate and avoiding misjudgments caused by overly rough general rules or templates. At the same time, through the calculation method of reasonable deviation rates, combined with material unit prices, consumption quantities, and the proportion of total project costs, a reasonable deviation range for different material items is dynamically set to ensure the accuracy and flexibility of the audit. This technical solution has greatly improved the accuracy of project cost auditing, making the audit results more scientific and operable.

[0056] (2) By adopting a support vector machine model, this application has realized the intelligence and automation of the project cost auditing process. Traditional project cost auditing relies on manual experience and rules, while the technical solution of this application uses a data-driven method to train the model with historical project data, thus realizing intelligent auditing discrimination. During the training process, the model is continuously optimized through the denominator decay model and the gradient descent algorithm, and can adaptively adjust the discrimination boundary of the model to identify reasonable and unreasonable material consumption quantities. Specifically, the denominator decay model adjusts the influence of different samples in the feature space, avoiding the model's over-sensitivity to single abnormal data and enhancing the model's adaptability to diverse project characteristics. At the same time, the gradient descent algorithm optimizes the model parameters, enabling the trained model to have strong generalization ability and be able to handle projects of different types and scales. During the auditing process, by inputting the data of the target project, the model can automatically output the audit results, including marking abnormal material items, replacing the traditional manual auditing process, and realizing the full automation of project cost auditing, greatly improving the efficiency and accuracy.

[0057] By first collecting a number of historical engineering data, including the number of floors, building area, perimeter coefficient, structural area coefficient of each project, and the actual consumption of main building materials, and calculating the main material consumption ratio of each material, select the project with the most complete and representative data as the reference project, and use the main material consumption ratios of each main material of this project as the standard to normalize the main material consumption of the remaining projects, obtaining the standardized consumption ratio and influencing factors of each project. Subsequently, use the historical engineering data to train a preset support vector machine model so that it learns how to review whether the project cost is qualified based on the influencing factors of the project and the main material consumption ratio of each material. Finally, use the trained support vector machine model to review whether the cost of each material in the target project is qualified, realizing intelligent project cost review, reducing the dependence on manual experience, and thus completing efficient and accurate review.

[0058] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly and implement it according to the content of the specification, the following takes the preferred embodiments of this application and describes them in detail in conjunction with the drawings as follows. Brief Description of the Drawings

[0059] Figure 1 It is a schematic flowchart of the project cost review method in the embodiment of this application.

[0060] Figure 2 It is a schematic flowchart of training a preset support vector machine model based on the main material consumption ratio and influencing factors of historical projects in the embodiment of this application.

[0061] Figure 3 It is a data comparison chart of the number of floors, perimeter coefficient, and structural area coefficient in the historical project data and the main material consumption ratio of cast-in-place concrete walls and columns.

[0062] Figure 4 It is a schematic overall flowchart of the project cost review method in the embodiment of this application.

[0063] Figure 5 It is a structural block diagram of the project cost review system in the embodiment of this application.

[0064] Figure 6 It is a block diagram of the electronic device for project cost review in the embodiment of this application. Detailed Description of the Embodiment

[0065] The following combines the drawings and embodiments to further describe the specific implementation manners of this application in detail. The following embodiments are used to illustrate this application but not to limit the scope of this application.

[0066] Optionally, this application takes the engineering cost auditing method provided in each embodiment as an example for illustration in an electronic device. The electronic device is a terminal or a server. The terminal can be a mobile phone, a computer, a tablet computer, etc. This embodiment does not limit the type of the electronic device.

[0067] Referring to Figure 1 , which is a schematic flowchart of the engineering cost auditing method provided in an embodiment of this application. The method at least includes the following steps:

[0068] Step S101: Collect several historical engineering project files, determine one of the historical projects as the reference project, and calculate the influence factors of all historical projects except the reference project and the main material consumption ratio of each main material based on the historical engineering project files.

[0069] Specifically, first collect historical engineering project files. The historical engineering project files include construction drawings and result files, etc. The acquisition channels of the historical engineering project files include but are not limited to public engineering demonstration platforms and enterprise historical data. Subsequently, based on the data integrity principle, determine one of the historical projects as the reference project. The data integrity principle is that compared with other engineering project files, the project information in the selected engineering project files is complete, the drawing annotations are complete, the drawn graphics are accurate, the dimension annotations are correct, etc., that is, the project with the most complete data among the engineering project files is used as the reference project.

[0070] After determining the reference project, the number of floors and the above-ground building area of each historical project can be directly obtained from the construction drawings in the historical project files. Then, the building perimeter and building area of the standard floor of the project can be obtained from the architectural drawings in the construction drawings, and the perimeter coefficient is calculated using the formula: perimeter coefficient = building perimeter / building area. Then, the structural area and building area of the standard floor of the project can be obtained from the structural drawings in the construction drawings, and the structural area coefficient is calculated using the formula: structural area coefficient = structural area / building area. Then, calculate the main material consumption of all main materials in a certain project among the n historical projects. Indicates the main material in the th historical project. The total consumption is calculated and summarized according to the calculation standards issued by the engineering cost industry. The corresponding main material consumption The calculation formula of is: main material consumption is equal to / the above-ground building area of the th historical project , and calculate the main material consumption of the corresponding main material in the reference project as ; Then divide the main material consumption of the main material by the main material consumption of this material in the benchmark project to obtain the main material consumption ratio of the main material . Then divide the number of floors , perimeter coefficient and structural area coefficient of the th historical project by the corresponding number of floors , perimeter coefficient and structural area coefficient of the benchmark project to obtain the floor ratio , perimeter coefficient ratio and structural area ratio as the influencing factors of the th historical project.

[0071] It should be noted that the calculated main material consumption ratio , floor ratio , perimeter coefficient ratio and structural area ratio are all normalized so that the values are all between 0 and 1. The influencing factors of different materials under each project are the same. In addition, since historical project items are all archived after passing the review, the cost review results of historical projects are all qualified.

[0072] Step S102: Train a preset support vector machine model based on the influencing factors of historical projects and the main material consumption ratio of each main material. The support vector machine model is used to review whether the project cost is qualified according to the influencing factors of the project and the main material consumption ratio of each main material.

[0073] In step S102, the preset support vector machine model in this application is an existing model structure. By training the input data of historical projects, the model learns how to map to a suitable decision space according to the input data, and then judges whether the project cost is qualified by comparing the obtained output with the preset standard. This preset model structure itself is the same as the existing SVM structure, but in specific applications, this application will perform specific designs on the input features and decision rules to make it applicable to the scenario of project cost review. Refer to Figure 2 , which is a schematic flow chart of training a preset support vector machine model based on the main material consumption ratio and influencing factors of historical projects in an embodiment of this application. The method at least includes the following steps:

[0074] Step S1021: Calculate the maximum reasonable deviation rate corresponding to the main materials of each historical project.

[0075] In step S1021, in the project cost audit, each major material has an allowable fluctuation range, which is defined by the maximum reasonable deviation rate. The maximum reasonable deviation rate represents the maximum deviation degree allowed for the main material consumption ratio under the current project conditions. It not only reflects the reasonable fluctuation of material use but also reflects the influence of the proportion of the material cost in the total cost of the project.

[0076] Specifically, the data preprocessing layer in the preset support vector machine model first calculates the maximum reasonable deviation rate corresponding to the main materials of all historical projects except the benchmark project based on the historical project files. Specifically, for the th historical project, the maximum reasonable deviation rate corresponding to the main materials is calculated as follows: The calculation method is as follows:

[0077] ;

[0078] Where is the reasonable allowable deviation rate of the project cost of the th historical project set artificially based on project cost experience. It should be noted that according to project cost experience, for the same type of historical projects, their corresponding are defaulted to be equal. is the ratio of the combined price of the material part in the th historical project to the total cost of the project. represents the number of types of main materials in the th historical project. represents the total consumption of the main material in the th historical project. represents the corresponding unit price of the main material in the th historical project. , and can all be obtained from the achievement files of the corresponding projects.

[0079] In step S1022, a two-dimensional plane is constructed. The abscissa of the two-dimensional plane is the comprehensive influence factor, and the ordinate is the main material consumption ratio under the influence of the comprehensive influence factor. The main material consumption ratio of each historical project under the influence of its comprehensive influence factor is mapped in the two-dimensional plane and presented as a point.

[0080] In step S1022, after the data preprocessing layer of the support vector machine model calculates the maximum reasonable deviation rate corresponding to the main materials of each historical project, a two-dimensional plane is constructed in the plane mapping layer. Among them, the abscissa of the two-dimensional plane is the comprehensive influence factor, and the ordinate is the main material consumption ratio under the influence of the comprehensive influence factor.

[0081] Specifically, first, the weighted Euclidean norm is used to assign weights to the three influence factors, and then the square root of the sum of the squares of the three influence factors is calculated to obtain the comprehensive influence factor. The formula is as follows:

[0082] ;

[0083] Among them, , , are the weights assigned to the three influence factors respectively. It should be noted that currently, , , are uncertain and need to be determined during subsequent model training. In practice, compared with the linear weighted summation in the prior art, in the prior art, the three factors are independently accumulated, and the amplification effect of the simultaneous high values of multiple factors on the overall influence cannot be reflected. The Euclidean norm sums and takes the square root after squaring each component, which can naturally amplify the comprehensive value when multiple factors are relatively large. For example, when all three influence factors are relatively high, grows faster than simple summation, and can better reflect the high-risk state under complex engineering conditions. In addition, in multi-dimensional space, the Euclidean norm corresponds to the distance from a point to the origin, which can intuitively reflect the distance of the engineering conditions from the optimal (low-risk) state in the influence factor space. At the same time, the weight only determines the relative importance of each factor in the sum of squares, and does not change the positive contribution characteristics of the factor itself, which is conducive to separate optimization during training.

[0084] Subsequently, based on the comprehensive influence factor, the denominator decay model is used to calculate the main material consumption ratio of the main material in the th historical project under the influence of the comprehensive influence factor as follows:

[0085] ;

[0086] Among them, is a correction coefficient greater than 0, which is used to reflect how much the consumption ratio should be reduced for each unit increase in the comprehensive influence factor. Similarly, it is also uncertain and needs to be determined during subsequent model training. In practice, compared with the linear superposition in the prior art, when the comprehensive influence factor is very large in linear superposition calculation, will increase linearly, which may lead to Far beyond the reasonable range, it is not conducive to the stability of the model. In reality, the more extreme the engineering conditions, the marginal impact of the influencing factor on the main material consumption ratio often decreases, and the linear superposition cannot reflect this saturation effect. If the denominator attenuation model is adopted, as increases, the denominator increases linearly, making gradually approach a certain lower limit, reflecting the saturation effect of the influencing factor on the consumption ratio. The more demanding the engineering conditions, the smaller the additional influence effect, which conforms to the law that the material consumption does not decrease infinitely in extreme projects in reality. In addition, will always fall within the interval and will not cause an explosive increase in the model output due to the excessive of the comprehensive influencing factor, which is helpful for the numerical stability of model training.

[0087] Finally, the point coordinates on the two-dimensional plane where the main material in the historical project is mapped according to the main material consumption ratio under the influence of its comprehensive influencing factor are . Since is not determined for the time being, the position of the point where the main material of each historical project is mapped according to the main material consumption ratio under the influence of its comprehensive influencing factor on the two-dimensional plane is currently uncertain.

[0088] Step S1023: Map the ratio of the maximum reasonable deviation rate corresponding to the main material of each historical project to the maximum reasonable deviation rate corresponding to the main material of the reference version project within the two-dimensional plane and show it as a line.

[0089] In step S1023, after the mapping of the points is completed, for the main material consumption ratio of the main material of each historical project under the influence of its comprehensive influencing factor, map the ratio of its corresponding maximum reasonable deviation rate to the maximum reasonable deviation rate corresponding to the main material of the reference version project into the two-dimensional plane and show it in the form of a horizontal decision line.

[0090] Specifically, for the maximum reasonable deviation rate corresponding to the main material in the historical project, the mapped horizontal decision line in the two-dimensional plane is as follows:

[0091] ;

[0092] The mapped horizontal line is parallel to the horizontal axis of the two-dimensional plane and is located at the ordinate .

[0093] Step S1024: Design a loss function and continuously adjust the parameters through gradient descent until the loss function is minimized, obtaining the optimal parameters that satisfy the qualified conditions for all historical engineering data, thus completing the training of the support vector machine model.

[0094] In step S1024, since historical engineering projects are archived after being audited and approved, the cost audit results of historical projects are all qualified, which means that the points mapped by the main materials of each historical project should be below the corresponding mapped horizontal decision line.

[0095] Design a loss function As follows:

[0096] ;

[0097] Where, represents the main material consumption ratio under the influence of the comprehensive influence factor of the -th historical project, represents the maximum reasonable deviation rate of the main material in the -th historical project, represents the maximum reasonable deviation rate of the main material in the reference version project, is the regularization coefficient, used to balance the over-standard penalty and parameter smoothing, preventing the parameters from being too large or overfitting. is the economic weight, reflecting the proportion of the main material in the total project cost of the -th historical project, and the calculation method is as follows:

[0098] ;

[0099] Where, represents the number of types of main materials in the -th historical project, represents the total consumption of the main material in the -th historical project, represents the corresponding unit price of the main material in the -th historical project, is the ratio of the combined price of the material part to the total project cost in the -th historical project, represents the -thThe total cost of all major materials in a historical project. In a project, the cost differences of different major materials are often significant. If all samples are treated equally (with the same loss weight), the model may prioritize optimizing even minor deviations in inexpensive materials, wasting training resources. After introducing the economic weight, the higher the cost of a material, the greater the proportion of the corresponding penalty for exceeding the standard in the total loss. The model will first learn how to correct high-cost materials more accurately.

[0100] It should be noted that and have the numerical basic element conditions for difference comparison, that is, they both contain the main materials main material consumption ratio / , and the following two derivation formulas can prove:

[0101] Formula 1: ;

[0102] Formula 2: ;

[0103] Among them, is the ratio of the combined price of the material part in the benchmark project to the total cost of the project, represents the total cost of all major materials in the benchmark project, represents the th historical project, represents the corresponding unit price of the main material in the benchmark project, represents the main material per square meter content in the benchmark project, represents the main material in the th historical project per square meter content. The above parameters can all be obtained from the result documents of the corresponding project.

[0104] In implementation, after the loss function is designed, the loss function is used to measure how many material items in the historical project have a main material consumption ratio exceeding their maximum reasonable deviation rate. Through the method of gradient descent, the parameters are continuously adjusted until the loss function is minimized. After each adjustment, it is checked that the points mapped by the main materials of each historical project should be below the horizontal decision line mapped by the corresponding maximum reasonable deviation rate. If there are still points exceeding it, the parameters are continued to be adjusted. As the number of iterations increases, the loss function finally stabilizes. At this time, a set of optimal parameters is found, making all historical project data meet the qualified conditions. After the parameters are determined, the training work of the support vector machine model is completed.

[0105] In addition, preferably, only one set of optimal parameters is obtained when determining the final optimal parameters of the model because a regularization term is considered in the design of the loss function. , which makes the entire loss function strictly convex in the parameter space. "Strictly convex" is a professional term in optimization theory, used to describe a function that has only one lowest point in the entire parameter space and has a shape like a "smooth bowl", thus ensuring the uniqueness of the global optimal solution. Under a strictly convex objective function, regardless of the initial point, as long as the optimization algorithm (such as gradient descent) converges, the same global minimum point - that is, a set of optimal . If there is no regularization, the simple "zero penalty" interval may correspond to multiple sets of parameters that can make all historical samples qualified. After adding regularization, among these multiple sets of qualified solutions, only the set of parameters with the smallest norm will minimize the overall loss, thus ensuring the uniqueness of the solution.

[0106] It should be noted that in the review system of this application, the core goal is to evaluate whether the cost of each material in the project is reasonable, and the rationality of the cost is mainly reflected in the balance between the material usage and its cost. The main material consumption ratio of the main materials reflects the intensity of material usage in the project, while the impact factor illustrates the adjustment effect of project conditions on the usage of main materials. The main material consumption ratio affected by the comprehensive impact factor accurately reflects the actual usage of main materials under specific project conditions. At the same time, the weight of the main material cost and the overall project cost is also integrated into the calculation formula of the maximum deviation rate. Therefore, when comparing the main material consumption ratio affected by the comprehensive impact factor with its corresponding maximum deviation rate, if the main material consumption ratio affected by the comprehensive impact factor is lower than the maximum deviation rate, it means that under the correction of the impact factor, the usage of the main material is within a reasonable cost control range, indicating that the cost review of the main material is qualified. On the contrary, if the main material consumption ratio affected by the comprehensive impact factor is equal to or greater than the maximum deviation rate, it means that the usage of the main material is excessive or the cost control is insufficient, which may lead to cost overruns, and thus it is judged as unqualified. The above method integrates the key elements of cost control, namely material usage and material cost, into the review decision in the form of quantitative indicators, and then realizes an effective judgment on the rationality of the cost of each main material.

[0107] Step S103: Obtain the target engineering project file, and calculate the impact factor of the target project and the main material consumption ratio of each main material.

[0108] In step S103, first obtain the target engineering project file, that is, the engineering project file to be reviewed, and calculate the impact factor of the target project and the main material consumption ratio of each main material in the same way as in step S101 above. Taking a certain main material as an example, the main material consumption ratio is , It is the consumption of the main materials in the target project of the main materials, and the consumption of the main materials in the reference project of the main materials. Correspondingly, the influence factors corresponding to the target project are the floor area ratio , the perimeter coefficient ratio and the structural area ratio .

[0109] In addition, preferably, after calculating the influence factors of the target project and the consumption ratio of the main materials of each main material, the consumption ratio of the main materials of each main material calculated is verified based on the influence factors of the target project.

[0110] Specifically, referring to Figure 3 , which is a data comparison chart of the floor number, perimeter coefficient, and structural area coefficient in the historical project data in the embodiment of the present application and the consumption ratio of the main materials of the cast-in-place concrete walls and columns. Through the analysis of the historical project data, the present application infers that there is a certain positive correlation between the consumption ratio of the main materials of each main material and the floor number, perimeter coefficient, and structural area coefficient of its corresponding project. Therefore, after calculating the consumption ratio of the main materials of each main material, the value range of the consumption ratio of the main materials of this material is calculated based on the floor number, perimeter coefficient, and structural area coefficient of the project corresponding to this main material. The calculation formula is as follows:

[0111] ;

[0112] ;

[0113] Among them, is the sum of the influence factors of the target project corresponding to the main material , is the proportionality coefficient of the main material , which is used to reflect the response intensity of the consumption ratio of the main materials of this material to the combined value of the influence factors and is obtained through training with historical project data. Specifically, collect the data of multiple historical engineering projects containing the material , calculate its corresponding value and the consumption ratio of the main materials, and estimate it through statistical fitting methods such as the least squares method or linear regression. is the deviation tolerance value of the main material , which represents the reasonable floating range of the consumption ratio of the main materials under the given influence factors and is also determined through training with historical project data. Specifically, for each historical project, calculate the residual between and the actual consumption ratio of the main materials, and statistically take its maximum value as The design of the above formula is easy to implement and interpret. It can automatically fit the response coefficient of the consumption ratio to the structural complexity and the reasonable deviation range according to the characteristics of each material through historical engineering data, avoiding subjective limitations, thereby improving the objectivity and accuracy of the review. At the same time, the independent modeling method for each material takes into account the personalized needs of usage differences, has good engineering adaptability and scalability, and is convenient for rapid deployment and continuous optimization in the actual cost review system.

[0114] For example, use the calculation of the slope of simple linear regression in the least squares method The formula is as follows:

[0115] ;

[0116] Among them, for a certain main material , assume there are historical engineering project samples. Denote the sum of the influencing factors of each project as , and the corresponding main material consumption ratio as , is the mean value of all , is the mean value of all main material consumption ratios .

[0117] After calculating the value range of the main material consumption ratio of the main material, judge whether the main material consumption ratio of the main material calculated based on the target engineering project file is within the value range. If it is within the value range, it means that the calculated main material consumption ratio of the main material is correct. If it is not within the value range, it means that there is a problem with the calculated main material consumption ratio of the main material, and the verification of the main material consumption ratio fails. At this time, it is necessary to recalculate.

[0118] Step S104: Input the influencing factors of the target project and the main material consumption ratio of each main material into the trained support vector machine model, and the support vector machine model outputs the cost review results of each main material of the target project.

[0119] In step S104, input the influencing factors of the target project and the main material consumption ratio of each main material into the trained support vector machine model. The support vector machine model outputs the review results of each main material of the target project. If the review result output by the model is unqualified, mark and remind the material item for the convenience of subsequent engineers to quickly locate and check and proofread.

[0120] To sum up, combined with Figure 4, this application first collects a number of historical engineering data, including the number of floors, building area, perimeter coefficient, structural area coefficient of each project, and the actual consumption of the main building materials, calculates the main material consumption ratio of each main material, selects the project with the most complete and representative data as the reference project, and uses the main material consumption ratios of this project as the standard to standardize the main material consumption of the remaining projects, obtaining the standardized consumption ratio and influencing factors of each project. Subsequently, the preset support vector machine model is trained using the historical engineering data to enable it to learn how to review whether the project cost is qualified based on the influencing factors of the project and the main material consumption ratio of each main material. Finally, the trained support vector machine model is used to review whether the cost of each main material in the target project is qualified, realizing intelligent project cost review, reducing the dependence on manual experience, and thus completing efficient and accurate review.

[0121] Figure 5 is the structural block diagram of a project cost review system provided by an embodiment of this application. The system at least includes the following modules:

[0122] Historical data collection module, used to collect a number of historical project files, determine one historical project as the reference project, and calculate the influencing factors of all historical projects except the reference project and the main material consumption ratio of each main material based on the historical project files;

[0123] Support vector machine model training module, used to train the preset support vector machine model based on the influencing factors of the historical project and the main material consumption ratio of each main material. The support vector machine model is used to review whether the project cost is qualified according to the influencing factors of the project and the main material consumption ratio of each main material;

[0124] Target data acquisition module, used to acquire the target project file, and calculate the influencing factors of the target project and the main material consumption ratio of each main material;

[0125] Support vector machine model review module, used to input the influencing factors of the target project and the main material consumption ratio of each main material into the trained support vector machine model, and the support vector machine model outputs the cost review results of each main material of the target project.

[0126] For relevant details, refer to the above method embodiment.

[0127] Figure 6 is the block diagram of an electronic device provided by an embodiment of this application. The device at least includes a processor 401 and a memory 402.

[0128] The processor 401 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 401 may be implemented in at least one of the following hardware forms: DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 401 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the wake state, also known as the CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 401 may be integrated with a GPU (Graphics Processing Unit), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 401 may further include an AI (Artificial Intelligence) processor, and the AI processor is used to process computing operations related to machine learning.

[0129] The memory 402 may include one or more computer-readable storage media, and the computer-readable storage media may be non-transitory. The memory 402 may further include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash storage devices. In some embodiments, the non-transitory computer-readable storage media in the memory 402 is used to store at least one instruction, and the at least one instruction is used to be executed by the processor 401 to implement the project cost audit method provided in the method embodiments of the present application.

[0130] In some embodiments, the electronic device may further optionally include: a peripheral device interface and at least one peripheral device. The processor 401, the memory 402, and the peripheral device interface may be connected through a bus or signal lines. Each peripheral device may be connected to the peripheral device interface through a bus, signal lines, or a circuit board. Schematically, the peripheral devices include, but are not limited to: a radio frequency circuit, a touch display screen, an audio circuit, and a power supply, etc.

[0131] Of course, the electronic device may also include fewer or more components, and this embodiment does not limit this.

[0132] Optionally, the present application also provides a computer-readable storage medium, and a program is stored in the computer-readable storage medium, and the program is loaded and executed by the processor to implement the project cost audit method in the above method embodiments.

[0133] Optionally, the present application further provides a computer product, which includes a computer-readable storage medium. A program is stored in the computer-readable storage medium and is loaded and executed by a processor to implement the project cost auditing method in the above method embodiments.

[0134] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0135] The above embodiments only represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A project cost audit method, characterized in that, The method includes: Collect a number of historical engineering project documents, determine one of the historical projects as the baseline project, and calculate the impact factors of all historical projects except the baseline project and the main material consumption ratio of each main material based on the historical engineering project documents; the main material consumption ratio of each main material is the ratio of the main material consumption of this main material to the main material consumption of the corresponding material in the baseline project, and the impact factor of the nth historical project is the number of floors of this project , perimeter coefficient , and structural area coefficient respectively to the number of floors , perimeter coefficient , and structural area coefficient corresponding to the baseline project; Training a preset support vector machine model based on the impact factors of the historical projects and the main material consumption ratios of each main material, where the support vector machine model is used to verify whether the project cost is qualified according to the impact factors of the project and the main material consumption ratios of each main material; Obtaining the target project engineering documents, and calculating the impact factors of the target project and the main material consumption ratios of each main material; After calculating the impact factors of the target project and the main material consumption ratios of each main material, it further includes: calculating the value range of the main material consumption ratio of each main material based on the number of floors, perimeter coefficient, and structural area coefficient of the project corresponding to each main material, and the calculation formula is as follows: ; ； Among them, is the main material and is the sum of the influence factors of the corresponding target project, is the main material and is the proportionality coefficient, is the main material and is the allowable deviation value, is the consumption of the main material in the target project while is the consumption of the main material in the reference project ; , , are respectively the number of floors, perimeter coefficient and structural area coefficient of the target project;​​ After calculating the value range of the main material consumption ratio of this main material, determine whether the main material consumption ratio of this main material calculated previously based on the target project engineering documents is within the value range. If it is within the value range, it means that the calculated main material consumption ratio of this main material is correct. If it is not within the value range, it means that the calculated main material consumption ratio of this main material has problems and the main material consumption ratio verification fails. At this time, it is necessary to recalculate; Inputting the impact factors of the target project and the main material consumption ratios of each main material into the trained support vector machine model, and the support vector machine model outputs the cost audit results of each main material of the target project.

2. The engineering cost audit method according to claim 1, wherein Calculating the impact factors of all historical projects except the baseline project and the main material consumption ratios of each main material based on the historical project engineering documents includes: The number of floors and the above-ground building area of each historical project can be directly obtained from the construction drawings in the historical project documents; Obtaining the building perimeter and building area of the standard floor of the project from the architectural drawings in the construction drawings, and using the calculation formula to obtain the perimeter coefficient equal to the building perimeter / building area; Obtaining the structural area and building area of the standard floor of the project from the structural drawings in the construction drawings, and using the calculation formula to obtain the structural area coefficient equal to the structural area / building area; Indicates the main material The total consumption in the th historical project, and the corresponding main material consumption The calculation formula for is that the main material consumption equals / the above-ground building area of the th historical project , and calculate the main material consumption of the corresponding main material in the reference version project ; The main material consumption of the main material is divided by the main material consumption of the reference version project for this material to obtain the main material consumption ratio of the main material . Then, divide the number of floors , perimeter coefficient and structural area coefficient of the th historical project by the corresponding number of floors , perimeter coefficient and structural area coefficient of the reference version project to obtain the floor ratio , perimeter coefficient ratio and structural area ratio as the influencing factors of the th historical project.

3. The engineering cost audit method according to claim 2, wherein Training a preset support vector machine model based on the impact factors of the historical projects and the main material consumption ratios of each main material includes: Calculating the maximum reasonable deviation rate corresponding to the main material of each historical project; calculating the maximum reasonable deviation rate corresponding to the main material of each historical project includes: No. The main materials in the historical project Corresponding to the maximum reasonable deviation rate The calculation method is as follows: ； Among them, is the reasonable allowable deviation rate of the project cost of the th historical project set artificially based on project cost experience, is the ratio of the combined price of the material part in the th historical project to the total project cost of this project, represents the number of types of main materials in the th historical project, represents the main material in the th historical project, represents the corresponding unit price of the main material in the th historical project; Constructing a two-dimensional plane, where the abscissa of the two-dimensional plane is the comprehensive impact factor and the ordinate is the main material consumption ratio under the influence of the comprehensive impact factor. Map the main material consumption ratio of each historical project under the influence of its comprehensive impact factor in the two-dimensional plane and show it as a point; Map the ratio of the maximum reasonable deviation rate corresponding to the main material of each historical project to the maximum reasonable deviation rate corresponding to the main material of the baseline project in the two-dimensional plane and show it as a line; Designing a loss function, and continuously adjusting the parameters by the method of gradient descent until the loss function is minimized, obtaining the optimal parameters that make all historical project data meet the qualified conditions, and completing the training work of the support vector machine model.

4. The engineering cost audit method according to claim 3, wherein Mapping the main material consumption ratio of each historical project under the influence of its comprehensive influence factor in a two-dimensional plane and presenting it as a point includes: The weighted Euclidean norm is used to assign weights to the three influencing factors, and then the square root of the sum of the squares of the three influencing factors is calculated to obtain the comprehensive influencing factor , and the formula is as follows: ； Among them, , , are the weights assigned to the three influencing factors respectively; Using the denominator decay model based on the comprehensive impact factor, calculate the main material consumption ratio of the main materials in the th historical project under the influence of the comprehensive impact factor as follows: as follows: ； Among them, is a correction coefficient greater than 0; The main materials in the first historical project, and the point coordinates on the two-dimensional plane where the main material consumption ratio affected by its comprehensive influence factor is mapped are .

5. The project cost audit method according to claim 4, characterized in that The described design of the loss function, continuously adjusting parameters through the gradient descent method until the loss function is minimized, includes: Design loss function As shown below: ； Among them, represents the main material consumption ratio under the influence of the comprehensive influence factor of the main materials in the th historical project, and represents the maximum reasonable deviation rate of the main materials in the th historical project, represents the maximum reasonable deviation rate of the main materials in the reference version project, is the regularization coefficient, is the economic weight, reflecting the proportion of the main materials in the th historical project in the total project cost, and the calculation method is as follows: in the total project cost, and the calculation method is as follows: ； Among them, represents the number of types of main materials in the th historical engineering project, represents the total consumption of the main material in the th historical project, represents the corresponding unit price of the main material in the th historical project, is the ratio of the combined price of the material part in the th historical project to the total cost of the project, represents the total cost of all main materials in the th historical project; By means of gradient descent, continuously adjust the parameters , until the loss function is minimized. After each adjustment, the points of the main material mapping of each historical project should be below the horizontal decision line of the corresponding maximum reasonable deviation rate mapping. If there are still points exceeding the line, continue to adjust the parameters until a set of optimal parameters is found so that all historical project data meet the qualified conditions.

6. A project cost audit system, characterized in that, Includes: Historical data collection module, which is used to collect a number of historical engineering project files, determine one of the historical projects as the baseline project, and calculate the impact factors of all historical projects except the baseline project and the main material consumption ratio of each main material based on the historical engineering project files; the main material consumption ratio of each main material is the ratio of the main material consumption of this main material to the main material consumption of the corresponding material in the baseline project, and the impact factor of the nth historical project is the number of floors of this project , perimeter coefficient and structural area coefficient respectively compared with the number of floors , perimeter coefficient and structural area coefficient of the baseline project; A support vector machine model training module, which is used to train a preset support vector machine model based on the influence factors of historical projects and the main material consumption ratio of each main material. The support vector machine model is used to review whether the project cost is qualified according to the influence factors of the project and the main material consumption ratio of each main material; A target data acquisition module, which is used to obtain the target project engineering document and calculate the influence factors of the target project and the main material consumption ratio of each main material; After calculating the influence factors of the target project and the main material consumption ratio of each main material, it further includes: Calculating the value range of the main material consumption ratio of each main material based on the number of floors, perimeter coefficient, and structural area coefficient of the project corresponding to each main material. The calculation formula is as follows: ; ； Among them, is the main material and is the sum of the impact factors of the corresponding target project is the proportionality coefficient of the main material and is the deviation tolerance of the main material is the main material in the target project is the consumption of the main material of the main material in the target project is the consumption of the main material of the main material in the reference project , , are respectively the number of floors, perimeter coefficient, and structural area coefficient of the target project; after calculating the value range of the consumption ratio of the main material, it is judged whether the consumption ratio of the main material calculated based on the target project document before is within the value range. If it is within the value range, it indicates that the calculated consumption ratio of the main material is correct. If it is not within the value range, it indicates that there is a problem with the calculated consumption ratio of the main material, and the verification of the consumption ratio of the main material fails. At this time, it is necessary to recalculate; A support vector machine model review module, which is used to input the influence factors of the target project and the main material consumption ratio of each main material into the trained support vector machine model, and the support vector machine model outputs the cost review results of each main material of the target project.

7. An electronic device, characterized in that, The device includes a processor and a memory; a program is stored in the memory, and the program is loaded and executed by the processor to implement a project cost review method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, A program is stored in the storage medium, and when the program is executed by a processor, it is used to implement a project cost review method according to any one of claims 1 to 5.

Images