CN-121032165-B - Engineering cost control method and system based on big data

CN121032165BCN 121032165 BCN121032165 BCN 121032165BCN-121032165-B

Abstract

The invention relates to the technical field of engineering cost control and discloses an engineering cost control method and system based on big data. The method comprises the steps of collecting a full-period cost data stream of an engineering project, generating a standardized cost data set through data cleaning, constructing a dynamic cost feature library, extracting multidimensional features such as time sequence fluctuation, resource allocation dispersion, supplier association and the like, inputting the feature library into a cost anomaly detection model constructed by a pre-training deep learning network, outputting a cost deviation index, triggering a correction instruction if the cost deviation index exceeds a threshold value, matching historical cases to generate an optimization strategy set containing material replacement, construction period adjustment and supplier replacement, virtually deducting the optimization strategy through an engineering digital twin system, and screening a strategy with a predicted cost curve closest to a target value as a final implementation scheme. The method relies on big data and deep learning technology to realize the full-period dynamic management and control of engineering cost, and improves the accuracy and feasibility of cost control.

Inventors

CHENG YIHONG
ZHANG LEI
ZHANG BORUI

Assignees

福建农业职业技术学院

Dates

Publication Date: 20260508
Application Date: 20251103

Claims (7)

1. The engineering cost control method based on big data is characterized by comprising the following steps: Collecting a full-period cost data stream of an engineering project, wherein the full-period cost data stream comprises a design phase budget parameter, a construction phase material consumption parameter and an operation and maintenance phase loss parameter, and generating a standardized cost data set after abnormal values are removed by a data cleaning engine; Constructing a dynamic cost feature library, and extracting multidimensional features from the standardized cost data set, wherein the extracted features comprise time sequence fluctuation features, resource allocation discrete features and provider association features; Establishing a cost anomaly detection model, inputting a dynamic cost feature library into a pre-trained deep learning network, outputting a cost deviation index, and triggering a cost correction instruction when the deviation index exceeds a preset threshold; generating a cost optimization strategy set, matching similar scenes in a historical engineering case library according to the cost correction instruction, and calling a strategy generator to output an optimization strategy comprising a material replacement scheme, a construction period adjustment scheme and a supplier replacement scheme; executing a strategy verification flow, inputting an optimization strategy into an engineering digital twin system for virtual deduction, outputting a predicted cost curve after strategy execution, and screening a strategy with the predicted cost curve closest to a target value as a final execution scheme; the specific steps of constructing the dynamic cost feature library include: Separating the resource consumption rate characteristic and the fund flow period characteristic from the time sequence data block by adopting a characteristic decomposition algorithm, establishing a supplier association map, extracting the supply overlapping degree and price fluctuation synergistic characteristic among all material suppliers, and fusing the time sequence characteristic, the resource characteristic and the supplier characteristic into a dynamic cost characteristic vector set according to weight proportion; the specific steps of establishing the cost anomaly detection model comprise: The method comprises the steps of configuring structural parameters of a deep neural network, loading historical engineering abnormal data as training samples, adjusting network weight parameters through a back propagation algorithm, setting a cost deviation index calculation unit at an output layer, wherein the index is formed by the product of Euclidean distance of actual cost and budget cost and characteristic fluctuation amplitude; the specific steps for generating the cost optimization strategy set comprise: The method comprises the steps of establishing a three-dimensional search index in a historical engineering case library, wherein index dimensions comprise cost deviation amplitude, engineering types and abnormal feature combinations, adopting a neighbor search algorithm to locate the first N historical cases with highest similarity with the current abnormal scene, analyzing correction measures and effect data thereof in the historical cases to generate a strategy option list with effect evaluation weights, and carrying out cross combination on single strategies of materials, construction periods and provider dimensions through a strategy combiner to generate a composite optimization strategy.
2. The big data based engineering cost control method according to claim 1, wherein the specific steps of executing the policy verification procedure include: importing a BIM model and a resource scheduling plan of a current project in an engineering digital twin system; decomposing the composite optimization strategy into an executable material purchasing instruction, a construction progress instruction and a supplier switching instruction; Running multiple rounds of Monte Carlo simulation, and calculating cost prediction distribution under different market fluctuation situations; and selecting a strategy with the smallest variance and the closest median of the cost distribution to the budget target as a verification passing scheme.
3. The big data based engineering cost control method of claim 2, further comprising a real-time feedback adjustment step of: deploying an Internet of things data acquisition terminal in a strategy execution stage, and capturing actual material consumption data and construction progress deviation in real time; Dynamically comparing the real-time data with the predicted cost curve to generate an execution deviation coefficient; When the deviation coefficient exceeds the fault tolerance threshold, triggering a strategy re-optimization flow and re-executing digital twin verification.
4. The big data based engineering cost control method according to claim 3, wherein the specific steps of the trigger policy re-optimization procedure include: Analyzing the main source dimension of the execution deviation, and identifying dominant factors in material price mutation, construction efficiency reduction or supplier performance abnormality; according to the dominant factor type, adding search conditions in the historical case library, and narrowing the screening range of similar cases; extracting an emergency strategy in the newly added case and evaluating the adaptation degree of the emergency strategy with the current engineering scene; An incremental optimization strategy package is generated that includes risk hedging measures.
5. The big data based engineering cost control method according to claim 4, further comprising a cost knowledge graph updating step of: Packaging the cost data stream, the abnormal event and the execution strategy of the whole period of the project into a case data packet; extracting feature vectors and strategy effect indexes in the case data packet, and converting the feature vectors and strategy effect indexes into newly added nodes and side relations of the knowledge graph; updating the node vector representation through a graph embedding algorithm, and optimizing the semantic matching precision of the subsequent case retrieval.
6. The big data based engineering cost control method according to claim 5, further comprising a multi-project co-optimization step: establishing a cost association analysis channel crossing engineering projects in the knowledge graph; Identifying cost linkage effects caused by sharing suppliers among different projects or the same construction process; When a new item triggers a cost anomaly, historical policies of associated items are synchronously retrieved and collaborative optimization suggestions are generated.
7. A big data based engineering cost control system comprising a memory, a processor and a computer program stored in the memory and running on the processor, characterized in that the processor, when executing the computer program, realizes the steps of the big data based engineering cost control method according to any of the preceding claims 1 to 6.

Description

Engineering cost control method and system based on big data Technical Field The invention relates to the technical field of engineering cost control, in particular to an engineering cost control method and system based on big data. Background The engineering project cost control runs through the whole period of design, construction and operation and maintenance, and is core content for guaranteeing project economic benefit and smooth promotion. Along with the expansion of the engineering project scale and the improvement of the technical complexity, the cost data of each stage has the characteristics of complicated types, scattered sources and dynamic changes, and the traditional cost control method gradually exposes a plurality of limitations. Traditional cost control is dependent on manual statistics and experience judgment, one-sided performance often exists in a data acquisition link, only material consumption data in a construction stage are focused, and integration of budget parameters in a design stage and loss data in an operation and maintenance stage is ignored, so that a full-period view angle is lacking in cost analysis. Meanwhile, the manual processing of the data is easy to miss and error, and abnormal values are difficult to effectively remove, so that the accuracy and normalization of the cost data are insufficient, and the reliability of subsequent cost analysis is directly affected. In the aspect of cost feature extraction, the prior art usually only focuses on cost value change of a single dimension, and cannot fully mine multi-dimensional features such as time sequence fluctuation rules, resource allocation rationality, supplier association relations and the like, so that cost feature characterization is not comprehensive enough, and inherent logic and influence factors of cost change cannot be accurately reflected. This makes it difficult for the subsequent anomaly detection model to obtain an effective data support, failing to identify potential cost-off risks in time. In the process of anomaly detection and strategy generation, the traditional method mostly adopts a simple statistical model or fixed threshold value judgment, has poor adaptability to complex and changeable engineering cost data, and often has the condition of anomaly detection lag or erroneous judgment, so that the cost hyperbranched problem is enlarged. In addition, the formulation of the cost optimization strategy lacks effective linkage with the historical engineering cases, the strategy is not strong in pertinence, and an effective verification mechanism is lacking, so that the blind execution of the strategy can cause secondary cost risks, and the difficulty of cost management and control is further aggravated. Disclosure of Invention The invention aims to provide a big data-based engineering cost control method and a big data-based engineering cost control system, so as to solve the problems in the background technology. In order to achieve the above object, the present invention provides a big data based engineering cost control method, the method comprising: Collecting a full-period cost data stream of an engineering project, wherein the full-period cost data stream comprises a design phase budget parameter, a construction phase material consumption parameter and an operation and maintenance phase loss parameter, and generating a standardized cost data set after abnormal values are removed by a data cleaning engine; Constructing a dynamic cost feature library, and extracting multidimensional features from the standardized cost data set, wherein the extracted features comprise time sequence fluctuation features, resource allocation discrete features and provider association features; Establishing a cost anomaly detection model, inputting a dynamic cost feature library into a pre-trained deep learning network, outputting a cost deviation index, and triggering a cost correction instruction when the deviation index exceeds a preset threshold; generating a cost optimization strategy set, matching similar scenes in a historical engineering case library according to the cost correction instruction, and calling a strategy generator to output an optimization strategy comprising a material replacement scheme, a construction period adjustment scheme and a supplier replacement scheme; Executing a strategy verification flow, inputting an optimization strategy into an engineering digital twin system for virtual deduction, outputting a predicted cost curve after strategy execution, and screening a strategy with the predicted cost curve closest to a target value as a final execution scheme. Preferably, the specific steps of constructing the dynamic cost feature library include: performing time dimension slicing on the standardized cost data set to divide time sequence data blocks arranged according to construction progress nodes; separating the resource consumption rate characteristic and the fund flow period characteristic f