Building engineering cost analysis system and method

Abstract

The invention discloses a construction engineering cost analysis system and method, and particularly relates to the technical field of construction cost behavior monitoring, and the method comprises the following steps: collecting the business information and field process information of an execution stage project, carrying out the field-level matching with a budget detail table, and constructing a data set; performing risk judgment according to the contract plan information and the actual execution information, and judging whether to start a risk identification process; executing multi-dimensional information cross-analysis to identify cost risk symptoms when the conditions are met; calculating dynamically updated first and second risk assessment factors based on the identification result, and inputting the first and second risk assessment factors into a risk assessment model to generate a comprehensive cost risk score; according to the method, accurate matching of the budget information and the execution data is realized, and the real-time sensing capability of the cost data is improved; through a dynamic risk identification and cross analysis mechanism, the early discovery and judgment capability of the cost abnormity is enhanced; and a two-factor evaluation and scoring model is constructed, so that quantitative expression and intelligent early warning of the cost risk are realized.

Description

Technical Field

[0001] This invention relates to the field of construction cost behavior monitoring technology, and more specifically, to a construction project cost analysis system and method. Background Technology

[0002] Construction project cost analysis, as a crucial component of the whole-process cost control system, aims to dynamically track resource usage, cost deviations, and contract performance during project execution, promptly identify potential risks, and assist project managers in achieving reasonable cost control. Currently, most common cost analysis systems in the industry are capable of periodic calculations and static comparisons based on project budgets and payment data, enabling the recording and analysis of some explicit cost deviations.

[0003] With the increasing scale and extended execution cycles of construction projects, the data dimensions involved in project operation have significantly increased. Heterogeneous information, including on-site material consumption, machinery operation behavior, and changes in payment schedules, all impact cost execution results. Against this backdrop, existing systems still face technical challenges in the fusion and modeling of multi-dimensional dynamic data, the proactive identification of risk indicators, and the structured expression of risk levels. On the one hand, most existing methods rely on fixed-period analysis or manual triggering mechanisms, lacking the ability to autonomously determine the timing of analysis based on changes in the project's execution status, potentially leading to resource waste or delayed problem detection. On the other hand, the manifestations of cost deviations are becoming increasingly complex, and traditional methods often struggle to promptly capture hidden cost risks arising from behavioral chains such as data distortion, process deviations, and abnormal contract performance schedules.

[0004] While some systems have initially constructed risk assessment models, they still primarily rely on static indicator settings and rule matching, lacking the ability to continuously model data evolution trends across different periods. This makes it impossible to quantify the persistence, clustering, and trend of risks. When faced with cross-cycle, multi-dimensional, and strongly coupled data changes, such models often struggle to output timely and decision-making-guided assessment results, limiting the shift in engineering cost control from passive response to proactive early warning. Therefore, this paper proposes a construction engineering cost analysis system and method to address these issues. Summary of the Invention

[0005] To achieve the above objectives, the present invention provides the following technical solution: A method for analyzing the cost of construction projects includes the following steps: Collect business information and on-site process information of construction projects in the execution phase, and then perform field-level data matching with the budget details table set at the start of the project to form a data set reflecting the status of contract execution and resource consumption. Based on the contract planning information and actual execution information in the dataset, risk assessment is conducted to determine whether the conditions for initiating the cost risk identification process have been met. Under the premise that the conditions for initiation are met, perform multi-dimensional information cross-analysis to determine whether there are any signs of cost risk in the project; Based on the identified symptom events, a first risk assessment factor and a second risk assessment factor are calculated. The first risk assessment factor is calculated based on the cumulative deviation ratio between each expense in the budget details and the actual expenditure in the current period. The second risk assessment factor is calculated based on the cumulative frequency of abnormal data, the extent of exceeding limits, and the duration of the trend. Both factors are dynamically updated with time windows as the dimension. The first and second risk assessment factors are input into the risk assessment model to generate a comprehensive cost risk score. If the comprehensive cost risk score exceeds the set risk threshold, the cost anomaly verification process will be automatically initiated, and corresponding risk warnings and handling suggestions will be generated based on the score result level.

[0006] In a preferred embodiment, the business information includes the total cost stipulated in the project contract, the phased performance plan, and the actual payment time record; the on-site process information includes the raw material requisition record and the usage time of machinery and equipment at the construction site.

[0007] In a preferred embodiment, risk assessment refers to: Time series data of the performance stage sequence, payment interval, raw material usage quantity and machinery usage time are extracted separately according to fixed cycles. The continuous offset between the planned value and the corresponding actual value is calculated for each item to form a difference data sequence. By setting offset amplitude threshold, behavior continuity judgment rules and payment rhythm fluctuation range, abnormal segments appearing in each sequence are judged. If any abnormal segment exists in any sequence, the cost risk identification process is triggered. Furthermore, the cycle in which the offset amplitude of any data item in the same cycle exceeds its budget threshold is defined as an abnormal cycle.

[0008] In a preferred embodiment, the multidimensional information cross-analysis operation refers to: Analyze whether each performance node is overdue, whether the construction progress plan is inconsistent with the on-site records, whether the total consumption of raw materials continuously exceeds the budget range, and whether the payment behavior deviates from the contract schedule. If any indicator meets the warning trigger condition, it is determined that the project has signs of cost risk.

[0009] In a preferred embodiment, the cost anomaly verification process refers to performing abnormal data backtracking, reconciling contract performance plans with payment records, and comparing material consumption records with actual inventory data.

[0010] In a preferred embodiment, the calculation of the first risk assessment factor includes the following steps: The relative offset rate of each expense category in multiple consecutive monitoring periods is extracted. For each expense category, the period segment in which the continuous deviation from the budget threshold is identified, and the cumulative sum of the offset rate increase in each period segment is calculated to obtain the cumulative value of the continuous offset of the expense category. For each expense category, the cumulative value of continuous offset is further calculated to obtain its abnormal persistence index. The abnormal persistence index is the product of the number of periods in which the single-period increase in the offset rate of the expense category exceeds the preset sensitivity threshold within the continuous offset interval and the cumulative value of continuous offset. This is used to quantify the combined effect of continuous offset intensity and abnormal trigger density. The expense category with the largest abnormality persistence index is selected from all expense categories. The cumulative value of the continuous offset of this expense category is multiplied by the abnormality persistence index after nonlinear compression. The resulting product is used as the first risk assessment factor for the current period to express the comprehensive risk performance between the continuous accumulation capacity of cost offset and the density of abnormal triggers.

[0011] In a preferred embodiment, nonlinear compression processing refers to: Obtain the original value of the anomaly persistence index and use it as the growth variable. Then, perform a curve transformation on the growth variable with a gradually decreasing increment, so that as the value of the growth variable increases, the rate of increase of the corresponding transformation result gradually slows down, and the contribution of a single new anomaly trigger to the overall risk value gradually weakens. The compressed result obtained after the transformation is used as the nonlinear compressed value of the anomaly persistence index.

[0012] In a preferred embodiment, the calculation of the second risk assessment factor includes the following steps: After arranging the abnormal periods in multiple continuous monitoring periods in chronological order, a sliding window of fixed time length is used to traverse them sequentially from front to back. If two or more abnormal periods appear in each sliding window, all abnormal periods in that window are classified as an associated risk node and removed from the sequence to be processed. After all windows have been traversed, the remaining unclassified single abnormal period is classified as an independent risk node, and each abnormal period participates in the attribution judgment only once. For each associated risk node, the number of abnormal cycles within that node is counted, and the sum of all abnormal offset amplitudes within that node is calculated. Then, the number of abnormal cycles and the total offset amplitude are multiplied together to obtain the risk value of that associated node. For each independent risk node, the single abnormal offset amplitude of that node is used as the risk value of that node. The risk values ​​of all nodes are arranged in chronological order of their appearance. The risk value of the earliest node is recorded directly. For each subsequent node, its risk value is multiplied by its order of appearance in the entire node sequence, so that nodes that appear later in time have a higher impact weight. The risk values ​​of all nodes after sequential amplification are accumulated, and the accumulated result is used as the second risk assessment factor for the current period to reflect the cumulative risk effect of abnormal behavior gradually increasing over time.

[0013] In a preferred embodiment, the steps for generating the comprehensive cost risk score include: The first and second risk assessment factors for the current monitoring period are normalized to map the original values ​​to a unified range and form standardized feature data. The first and second risk assessment factors, standardized for the current period, are used as inputs to form a two-dimensional time-series feature matrix according to their arrangement in multiple consecutive periods. This matrix is ​​then input into the trained convolutional neural network structure, i.e., the risk assessment model. The risk assessment model extracts features and matches risk levels for the risk change trends in multiple periods, and outputs the comprehensive cost risk score for the current period. If the comprehensive cost risk score exceeds the set risk threshold, the comprehensive cost risk score will be compared with the preset risk level range, the corresponding level standard will be matched, and the corresponding risk alarm information and management and handling suggestions will be automatically generated.

[0014] In a preferred embodiment, a construction project cost analysis system specifically includes: The data acquisition module is used to collect business information and on-site process information of construction projects in the execution phase, and to perform field-level data matching between the information and the budget details table set at the start of the project to form a data set reflecting the status of contract execution and resource consumption. The risk assessment module is used to assess risks based on contract planning information and actual execution information in the dataset, and to determine whether the conditions for initiating the cost risk identification process have been met. The symptom identification module is used to perform multi-dimensional information cross-analysis operations to determine whether there are any signs of cost risk in the project, provided that the activation conditions are met. The assessment and calculation module is used to calculate the first risk assessment factor and the second risk assessment factor based on the identified symptom events. The first risk assessment factor is calculated based on the cumulative offset ratio between each expense in the budget details and the actual expenditure in the current period. The second risk assessment factor is calculated based on the cumulative frequency of abnormal data, the extent of exceeding the limit, and the duration of the trend. Both factors are dynamically updated with time windows as the dimension. The risk scoring module is used to input the first risk assessment factor and the second risk assessment factor into the risk assessment model to generate a comprehensive cost risk score; The alarm response module is used to automatically enter the cost anomaly verification process when the comprehensive cost risk score exceeds the set risk threshold, and generate corresponding risk warnings and handling suggestions based on the score result level.

[0015] The technical effects and advantages of this invention are as follows: This invention collects business information and on-site process information during the execution phase and performs field-level data matching with the budget details set at project initiation. This forms a data set reflecting the contract execution and resource consumption status, achieving dynamic integration of budget control logic and execution data chain during the actual progress of construction projects. Traditional cost analysis methods are mostly based on static budgets and post-audit models, making it difficult to capture the frequently changing progress status, payment behavior, and on-site resource usage during construction, often resulting in delayed or unrealistic analysis conclusions. This invention, through field-level data matching, not only achieves accurate comparison between budget details and actual data but also supports horizontal (between different cost items), vertical (between different periods), and multi-source (between different systems) correlation identification at the data structure level, thus providing a comprehensive reconstruction of cost deviation trends at the source. The construction process of this data set is highly scalable, applicable to various construction scenarios and project types, and does not rely on a specific information platform. This improves the compatibility, universality, and engineering adaptability of cost analysis methods, providing a high-quality, structured input foundation for subsequent risk assessment and model evaluation.

[0016] This invention, by constructing a triggering condition and risk symptom judgment mechanism, transforms cost risk identification from static judgment to dynamic behavioral analysis, significantly improving the early detection capability and timeliness of handling cost anomalies. Traditional project management often employs periodic budget reviews or post-event debriefing based on financial settlement milestones. This approach lacks the ability to perceive real-time anomalies during execution, especially in dynamic changes such as project delays, material waste, and payment mismatches, easily overlooking early signs and allowing risks to evolve into systemic problems. This invention, however, sets clear risk judgment triggering conditions, such as behavioral indicators like deviations in performance sequence, fluctuations in payment rhythm, and exceeding resource usage limits, and periodically compares these with contract plan information. Based on meeting specific thresholds and behavioral continuity, it determines whether to trigger subsequent risk symptom analysis. Furthermore, the system, based on multi-dimensional information cross-analysis, performs logical verification and temporal consistency judgment on data such as construction progress deviations, abnormal on-site consumption, and payment delays, ensuring the discovery of potential cost risk signals across multiple independent data sources. This achieves proactive and systematic risk perception, reducing losses caused by judgment lag in traditional methods.

[0017] This invention generates a comprehensive cost risk score by calculating a first risk assessment factor and a second risk assessment factor and introducing them into a risk assessment model. This achieves the transformation of multi-dimensional dynamic indicators into a unified risk expression, constructing a quantifiable and predictable intelligent risk assessment mechanism. In the complex execution environment of construction projects, a single risk indicator often cannot fully reflect the overall picture of cost deviations. Therefore, this invention structurally integrates cost deviation signals from different dimensions. The first risk assessment factor focuses on the offset relationship between the budget and actual expenditures, reflecting the intensity of performance deviations in different cost categories. The second risk assessment factor focuses on the temporal aggregation characteristics of abnormal data, reflecting the frequency, magnitude, and trend evolution of risk events. Both assessment factors are updated on a time window scale, exhibiting good dynamic adaptability. Subsequently, the two factors are input into the risk assessment model, forming a temporal feature matrix based on their arrangement order over multiple periods. Nonlinear feature extraction is then performed through a trained neural network structure, ultimately outputting a comprehensive cost risk score. The scoring results can not only quantify the risk level, but also compare it with preset thresholds and level ranges, enabling automatic triggering of alarms, verification and handling suggestions, forming an intelligent full-process cost risk management system, which significantly improves project management efficiency and risk control level. Attached Figure Description

[0018] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings; Figure 1 This is a schematic diagram of a construction project cost analysis method according to the present invention.

[0019] Figure 2 This is a schematic diagram of a construction engineering cost analysis system according to the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Reference Figure 1 - Figure 2 The following examples were obtained: Example 1: A method for analyzing the cost of a construction project, comprising the following steps: The process involves collecting business information and on-site process information of construction projects in the execution phase, and then performing field-level data matching with the budget details set at the project initiation stage to form a data set reflecting the status of contract execution and resource consumption. The significance of this step is to ensure that the analyzed data not only covers the content agreed in the contract, but also accurately reflects the actual execution status of the construction site. By establishing a correspondence between project objectives and the current execution status through field-level data matching, a data foundation is provided for subsequent analysis, avoiding judgment errors caused by information gaps.

[0022] Based on the contract plan information and actual execution information in the dataset, risk assessment is conducted to determine whether the conditions for initiating the cost risk identification process have been met. The significance of this step is to build a dynamic triggering mechanism to avoid indiscriminate analysis of all projects or any point in time. Instead, the cost risk identification process is activated only when a critical deviation between the performance information and the actual execution data is detected, thereby improving resource utilization efficiency and ensuring the timeliness of risk discovery.

[0023] Under the premise that the initiation conditions are met, perform multi-dimensional information cross-analysis to determine whether there are signs of cost risk in the project. The significance of this step is that by comparing information from multiple dimensions (such as performance progress, payment behavior, on-site records, etc.), it can identify combined risk manifestations that are difficult to detect from a single dimension, enabling the system to detect its precursors before the risks are fully exposed, thereby strengthening the initiative and foresight of cost management.

[0024] Based on the identified warning events, a first risk assessment factor and a second risk assessment factor are calculated. The first risk assessment factor is calculated based on the cumulative deviation ratio between each expense in the budget details and the actual expenditure in the current period. The second risk assessment factor is calculated based on the cumulative frequency of abnormal data, the extent of exceeding limits, and the duration of the trend. Both factors are dynamically updated using a time window as the dimension. The significance of this step is that it quantifies the performance of cost risk into two traceable and analyzable factors, establishes assessment models from the perspectives of cost deviation and abnormal behavior characteristics, and continuously updates them using a time window approach. This makes the risk assessment dynamic, sensitive, and comparable, providing controllable input for subsequent scoring.

[0025] The first and second risk assessment factors are input into the risk assessment model to generate a comprehensive cost risk score. If the comprehensive cost risk score exceeds a set risk threshold, the process automatically enters the cost anomaly verification stage, and corresponding risk warnings and handling suggestions are generated based on the score level. The significance of this step lies in its ability to comprehensively analyze the impact of two types of key risk factors on the project cost execution status based on a standardized and systematic assessment mechanism. By using threshold judgment, potential risks are transformed into explicit response actions, achieving a closed loop from risk discovery and level identification to intelligent alarms and management command triggering, thereby improving the timeliness and accuracy of risk response.

[0026] In this implementation, the construction project cost analysis method first collects business information and on-site process information of construction projects in the execution phase. Then, it performs field-level data matching with the budget details table set at project initiation to form a data set reflecting the contract execution and resource consumption status. Specifically, a multi-dimensional data capture interface is established to collect business information and on-site process information of construction projects in the execution phase. Business information includes the total cost stipulated in the project contract, phased performance plans, and actual payment time records; on-site process information includes raw material requisition records and machinery usage time at the construction site. Specifically, the total cost is directly extracted from the contract system as a total amount field; the phased performance plan is marked with target completion time and amount allocation item by item in the form of a plan node list; and the actual payment time records are based on the payment timestamps in the financial system. Raw material requisition records are mainly statistically analyzed on a daily basis and aggregated by material category; machinery usage time is calculated from hourly data output by the equipment's intelligent sensors and stored in groups by daily equipment number.

[0027] The budget details set at project initiation are standardized in terms of field structure. Each expense category, timeline, material quantity, and usage period in the budget is broken down into standard field formats, including: budget project number, budget amount, budget start and end period, budget unit, and budget resource code, establishing a field-level mapping relationship. For example, a budget entry stating "Concrete C30 strength grade pouring quantity 300m³, budget unit price yuan / m³" will be broken down into "Resource type: Concrete", "Quantity: 300", "Unit: Cubic meter", "Budget unit price: 420", and "Budget amount: 126000".

[0028] The collected business information and on-site process information are compared one by one with the budget details table according to the field mapping standard to complete the field-level data matching operation, forming a dual data field set with real-time execution value and budget target value. During the field-level data matching process, the system automatically performs three-dimensional matching according to the budget project number, resource code and time period to ensure that the raw material requisition records can be accurately classified into the corresponding budget items. For example, "15 tons of cement requisitioned on October 15" is matched to "Budget item number A1042 Cement Quantity in October 20 tons", and its consumption ratio is recorded as 15 / 20, that is, 75%.

[0029] After field-level data matching is completed, a dataset reflecting the status of contract execution and resource consumption is constructed. This dataset includes fields such as budget amount, planning period, resource type, actual completion time, actual resource usage, and actual expenditure for each budget item. This dataset supports rolling updates in time windows and generates structured reports for subsequent factor calculation and risk assessment modules. Through this dataset, real-time extraction and analysis of data such as "Budget item B2025 originally planned payment date was October 31, 2024, actual payment date was November 10, 2024, and offset days were 10 days" and "Budget item C3030 raw material budget usage was 1000kg, current cumulative usage was 1200kg, and offset ratio was 20%" are possible.

[0030] In this implementation, risk assessment is performed based on the contract plan information and actual execution information in the dataset to determine whether the conditions for initiating the cost risk identification process have been met. Specifically, this includes the following steps: extracting the performance stage sequence, payment interval, raw material usage quantity, and machinery usage time from the dataset at fixed intervals to construct continuous time series data. The fixed interval can be set to seven days, ten days, or other time windows that match the project progress cycle. For example, using a seven-day monitoring cycle, the system automatically extracts "contract plan stage sequence T1-T5" and "planned payment node weekly interval of 30 days" from the dataset, while also extracting values ​​such as "actual cement consumption" and "excavator usage time" within each cycle to form corresponding planned and actual sequences.

[0031] The system calculates continuous offset values ​​between the extracted planned values ​​and their corresponding actual values, forming a difference data sequence. Taking cement as an example, if the planned cement consumption for the third cycle is 20 tons, while the actual consumption is 24 tons, the offset for that cycle is positive 4 tons; if the planned consumption for the fourth cycle is 20 tons, but the actual consumption is only 17 tons, the offset is negative 3 tons. The system arranges the offset values ​​for all cycles into an offset curve and calculates the relative offset magnitude for each cycle, which is the difference between the actual and planned values ​​divided by the planned value. For example, a 4-ton offset represents a 20% offset relative to a 20-ton plan.

[0032] Based on the established difference data sequence, preset offset amplitude thresholds, behavior continuity judgment rules, and payment rhythm fluctuation ranges are applied to determine abnormal segments. The offset amplitude threshold can be set to 15%, the behavior continuity rule is set to "two or more consecutive periods with the same offset direction and amplitude exceeding the threshold," and the payment rhythm fluctuation range is set to "no more than five days between the actual payment time and the planned payment time is considered normal." For example, if the raw material usage exceeds 17%, 20%, and 25% respectively in three consecutive periods, the system automatically marks the segment as an abnormal segment; if payment records show delays of eight and nine days in two periods respectively, the segment is also judged as abnormal.

[0033] A period in which the deviation of any data point exceeds its budget threshold is defined as an abnormal period, and this abnormal period serves as the basis for subsequent analysis. If, in a certain period, only cement usage deviates by 22%, while other indicators show no deviation, the system still classifies that period as an abnormal period. If any time series contains an abnormal segment, the system determines that the conditions for initiating the cost risk identification process have been met and transmits this initiation signal to the subsequent symptom identification module to activate the further risk analysis mechanism.

[0034] An abnormal segment refers to a continuous section in continuous time series data where a certain indicator (such as the order of contract fulfillment stages, payment intervals, raw material usage quantity, or machinery usage time) deviates beyond a set threshold, exhibits a consistent direction of deviation, or violates preset behavioral logic within two or more consecutive periods. For example, a project's budgeted excavator usage time is 100 hours per week. Actual records are: Period 1: 115 hours (+15%); Period 2: 120 hours (+20%); Period 3: 118 hours (+18%). If the system sets a 15% deviation threshold, then periods 1 to 3 meet the criteria of continuous exceeding the limit and consistent direction, constituting an abnormal segment. The core function of abnormal segments is to provide a logical basis for initiating the cost risk identification process. Once an abnormal segment exists in any monitoring indicator, it is considered that the conditions for initiating the cost risk identification process have been met.

[0035] In this implementation, under the premise that the initiation conditions are met, a multi-dimensional information cross-analysis operation is performed to determine whether there are signs of cost risk in the project. Specifically, this includes the following steps: analyzing whether each performance node is overdue, and identifying overdue situations by comparing the planned completion time of each node with the actual completion time in the contract. The system calculates the difference between the actual completion time and the planned completion time for each performance node. If this difference is greater than a preset threshold of five days, the node is considered overdue. For example, if node A was planned to be completed on June 1st, but actually completed on June 8th, then it is delayed by seven days, meeting the overdue condition. The system summarizes the overdue information of multiple nodes to form a performance lag indicator curve.

[0036] Based on the identification of overdue milestones, the system analyzes whether the construction schedule is consistent with the on-site records. Specifically, it compares the quantities of work for each stage in the contract plan with the actual quantities of work completed in the on-site process information. For example, if the planned quantity for the structural topping-out stage is 1,500 cubic meters of concrete, but the on-site records show that only 1,000 cubic meters have been completed, the deviation rate is 33%. If the deviation rate exceeds 20% for two consecutive periods, the system determines that there is a schedule consistency deviation for that stage and associates this information with the overdue milestones to form a joint symptom signal.

[0037] The system analyzes whether the total consumption of raw materials continuously exceeds the budgeted range. Taking steel reinforcement as an example, the budget stipulates that the use of steel reinforcement in a certain three-cycle period should not exceed 50 tons. The on-site requisition data are 18 tons, 20 tons, and 22 tons, totaling 60 tons, which exceeds the budgeted range by 20 percentage points. If the total consumption for three consecutive cycles exceeds the budgeted range, and the maximum deviation in a single cycle exceeds 15%, the "continuous exceedance" judgment is triggered, and the system marks the anomaly as a sign of resource overconsumption.

[0038] Based on schedule deviations and resource offsets, the system analyzes whether payment behavior deviates from the contract schedule and exhibits a delay trend. The system extracts payment schedule and payment time records from the contract, constructs a time series, and calculates the average difference between the actual payment time and the schedule over the most recent three periods. If this average exceeds seven days and there are two or more consecutive periods of delay, a delay trend in payment is identified. The system cross-checks the payment delay trend with three types of information: delayed performance nodes, inconsistent construction progress, and abnormal resource consumption. If any indicator meets the early warning trigger condition, the system determines that the project has signs of cost risk and sends a risk activation signal to the factor calculation module.

[0039] In this implementation, when the comprehensive cost risk score exceeds a set risk threshold, the system automatically enters the cost anomaly verification process. This process is used to deeply verify the source of the cost anomaly and its risk severity, specifically including the following steps: Performing anomaly data backtracking. The system automatically identifies the key period that caused the comprehensive cost risk score to exceed the limit and traces the original data of the two periods before and after it, including the execution status of the performance plan, resource consumption records, financial payment data, etc., generating a cross-period anomaly data chain. Taking the fifth period of a project as an example, the cost risk score of this period is 0.95, exceeding the threshold of 0.8. The system loads the data of this period and the two periods before and after it, a total of three periods, into the backtracking engine to form a complete execution trajectory.

[0040] The risk threshold is set based on statistical distribution analysis of comprehensive cost risk scores from a large number of historical construction projects. This is combined with backtesting of cost anomalies in actual projects to identify significant critical points between the score values ​​and the probability of actual risk events. The system prioritizes using clustering algorithms and percentile distribution methods to segment the score data, selecting locations with drastic changes in score concentration or a significant increase in event trigger frequency as initial threshold candidates. Then, by cross-comparing the score trajectories of known risk projects, a set of score intervals with discriminative, generalizable, and trigger-stable characteristics is ultimately determined as the risk threshold standard. This ensures that the threshold effectively identifies real risks while avoiding misjudgments and interference.

[0041] Based on the abnormal data chain, the system performs a reconciliation operation between the contract performance plan and payment records. The system compares each stage of performance milestones set in the contract with the payment time and amount recorded in the financial system, and generates a performance matching index based on the reconciliation results. If a payment record corresponding to a performance milestone is more than ten days late, or the payment amount is less than 90% of the contract amount for that stage, the system marks that stage as a period of performance-financial discrepancy. For example, if the planned payable amount for the sixth cycle is 500,000 yuan, but only 400,000 yuan is actually paid and it is five days late, the system identifies this as an abnormal reconciliation item.

[0042] After reconciliation, the system further compares material consumption records with actual inventory data. The system horizontally matches material requisition records in the on-site process information with the actual inventory quantities in the warehousing system and calculates the inventory-to-consumption matching rate for each key material. If this matching rate is below 80%, or if it decreases for two consecutive periods, it is identified as an anomaly in material flow. For example, if a project's concrete budget is 300 cubic meters per period, the inventory record shows 280 cubic meters, and on-site requisition is 350 cubic meters, the matching rate is only 80%, indicating a possible over-requisition / under-requisition or fraudulent consumption issue. Based on the results of anomaly data backtracking, reconciliation of performance plans and payment records, and comparison of material consumption records with actual inventory data, the system comprehensively generates an anomaly verification report. Anomalies are categorized into three types: structural deviation, behavioral delay, and data inconsistency, achieving automation, layered response, and visualized output of the cost anomaly verification process.

[0043] In this implementation, the first risk assessment factor is used to characterize the comprehensive risk performance between the continuous accumulation capacity of cost deviation and the density of abnormal triggers, and to extract the relative deviation rate of each expense category in multiple consecutive monitoring periods. Specifically, the system extracts the planned value and the actual expenditure value of each expense category in the budget details for each monitoring period, and calculates its relative deviation rate using the formula: (actual expenditure minus budget value) divided by the budget value. For example, the budget value for "template materials" in a certain project is 100,000 yuan in periods 1 to 5, while the actual expenditures are 105,000, 110,000, 108,000, 122,000, and 130,000 yuan respectively, corresponding to deviation rates of 5%, 10%, 8%, 22%, and 30%. For each expense category, the period segment in which continuous deviations from the budget threshold are identified, and the cumulative sum of the deviation rate increase in each period segment is calculated as the cumulative value of continuous deviation for that expense category. Taking an offset threshold of 8% as an example, the template material deviates from the threshold in the 2nd to 5th periods. The system identifies an abnormal segment consisting of four consecutive periods, with the corresponding offset rate increase being: from 10% to 30%, with adjacent increases of 5%, -2%, and 14%. The sum of the absolute values ​​is 21%, meaning that the cumulative value of the continuous offset of this expense category in this segment is 21%.

[0044] For each cost category, the cumulative value of continuous offset is further used to calculate its abnormal persistence index. The system counts the number of periods within the continuous offset interval where the single-period increase in offset rate exceeds a preset sensitivity threshold. Using 5% as the sensitivity threshold, if two periods in the template material segment have a single-period increase exceeding this value, the abnormal persistence index for that cost category is the number of periods (2) multiplied by the aforementioned cumulative continuous offset value of 21%, resulting in 42%. This quantifies the combined effect of persistent offset intensity and abnormal trigger density. The cost category with the largest abnormal persistence index is selected from all cost categories, and its cumulative continuous offset value is multiplied by the abnormal persistence index after nonlinear compression. The resulting product is used as the first risk assessment factor for the current period. For example, if the abnormal persistence index for "template materials" is 42%, the largest among all cost items, and its cumulative continuous offset value is 21%, with a compressed index of 26%, then the final output first risk assessment factor for this cost category is 5.46, used as the core indicator for cost risk in this period.

[0045] The setting of budget thresholds and sensitivity thresholds is based on statistical analysis of historical project data and the tolerance for deviations in project management standards, ensuring both general applicability and effective identification of abnormal behavior. Budget thresholds are typically used to determine whether the deviation between actual expenditures and the budget is within a reasonable range. They are set based on the normal fluctuation range of various cost categories in similar projects, such as the average deviation rate and standard deviation of structural materials and labor costs in historical data. This allows for the determination of a percentage deviation that covers the vast majority of normal deviations while excluding extreme anomalies, serving as the judgment threshold.

[0046] Sensitive thresholds are used to identify abnormal "increases" in the process of offset changes, emphasizing the ability to respond to trend amplification. The basis for setting them is to statistically model the single-cycle offset increase characteristics in the early stage of major cost deviation events in historical projects, extract key increase thresholds with predictive value, and dynamically classify them in combination with the sensitivity differences of project type, construction stage and cost category, so as to ensure that risks can be captured when they first appear, thereby effectively supporting the subsequent early warning and intervention process.

[0047] In this embodiment, nonlinear compression processing is used to map the original value of the anomaly persistence index to a compressed value with stronger risk expression convergence, so as to avoid the risk assessment factor being dominated by a single anomaly under extreme conditions. The original value of the anomaly persistence index corresponding to each cost category is obtained as the input growth variable for nonlinear compression processing. Taking template material cost as an example, in a certain monitoring period, its original value of the anomaly persistence index, calculated by previous steps, is 36.5. The system sets this value as the input basis for compression processing. A transformation curve with gradually decreasing increments is constructed based on the numerical range of the growth variable to express the compression trend of the growth variable. This embodiment uses the natural logarithm curve as the nonlinear transformation function. The specific transformation form is: the result of adding one to the growth variable is used as the input of the logarithmic function, that is, the form of "processing the growth variable plus one with the natural logarithm function" is adopted, that is, the compression effect is expressed by "natural logarithm (growth variable plus one)". For example, after the growth variable of 36.5 is transformed by this function, its corresponding compressed value is approximately 3.64.

[0048] The natural logarithm compression function is used as the standard compression channel to process the anomaly persistence index of all cost categories sequentially. The compression result is rounded to two decimal places to form the corresponding nonlinear compressed value. This compressed result is used to replace the original anomaly persistence index in the subsequent product calculation of the first risk assessment factor, thereby suppressing the contribution of continuously triggered anomalies at the numerical level. The system binds the processed nonlinear compressed value to the risk expression structure of the cost category in the current period and multiplies it with its continuous offset cumulative value to generate the first risk assessment factor. In the example of the template material above, the continuous offset cumulative value is 21.38, and its anomaly persistence index after nonlinear compression is 3.64. Therefore, the final output first risk assessment factor is 77.75, which is used to enter the next stage of the comprehensive cost risk scoring process.

[0049] In this embodiment, the calculation of the second risk assessment factor is mainly used to measure the clustering characteristics and risk evolution trend of multiple abnormal cycles over time. The calculation steps include the following four stages: First, multiple abnormal cycles continuously monitored by the system are initially sorted according to their actual occurrence time order, and then traversed using a sliding window mechanism with a fixed time length. In this embodiment, the sliding window length is set to sixty days. Starting from the earliest abnormal cycle, the system slides the window forward day by day, determining whether there are two or more abnormal cycles in each window. If so, all abnormal cycles within that window are classified as a single associated risk node and removed from the sequence of abnormal cycles to be processed. Taking a project as an example, five abnormal cycles were found within a two-month period. The system identified two associated risk nodes containing three and two abnormal cycles, respectively. After traversing all abnormal cycles, isolated abnormal cycles not included in any window are classified as independent risk nodes. Each abnormal cycle participates in the attribution judgment only once, avoiding duplicate risk inclusion. For each categorized associated risk node, internal indicators are calculated to count the number of abnormal cycles within that node. Simultaneously, the abnormal offset magnitude of each abnormal cycle within that node is extracted and summed. Taking the first associated risk node as an example, it contains three abnormal cycles with offsets of 5%, 8%, and 4% respectively, totaling 17%. The number of abnormal cycles for this node is three. The system multiplies this number by the total offset magnitude to obtain a risk value of 51 for this node. For each independent risk node, the system directly uses its corresponding single abnormal offset magnitude as the node's risk value, reflecting its basic contribution to the overall risk in an isolated state.

[0050] The risk values ​​of all associated and independent risk nodes are sorted according to their first appearance time and then weighted sequentially. After sorting, the earliest appearing node's risk value is recorded directly without weighting; for the second and subsequent nodes, the risk value is multiplied by their chronological sequence number to increase the impact of later-stage risks. For example, if the third node in the sequence has a risk value of 30 and a sequence number of 3, its weighted risk value will be 90. This approach effectively reflects the amplifying effect of later-stage risk clustering on the continuous impact on project costs. The system then sums all the sequentially weighted node risk values ​​sequentially, and the sum is the second risk assessment factor for the current period. This value is continuously updated within a dynamic time window and serves as an important input factor for generating subsequent comprehensive cost risk scores. It is used to fully express the clustering intensity, stage evolution, and trend diffusion characteristics of abnormal behavior over time, assisting in subsequent risk level identification and response strategy formulation.

[0051] In the implementation of a construction project cost analysis method, to achieve dynamic and accurate quantitative assessment of the cost risk status of a construction project during the execution phase, this embodiment designs and implements a comprehensive cost risk score generation step. The score result is used to determine whether the construction project currently has an abnormal cost risk exceeding the control threshold, and triggers the subsequent cost anomaly verification process accordingly. The score generation step specifically includes the following steps: To address the differences in the original numerical scale, dimensions, and variation range of risk indicators from different sources, and to avoid indicator imbalance in subsequent assessment modeling, this embodiment uses a maximum-minimum normalization method to normalize the first and second risk assessment factors of the current monitoring period. Specifically, the system calls the maximum and minimum values ​​of each risk factor within the historical thirty monitoring periods as boundary benchmark values, and performs the following mapping logic on the current value: Normalization result = (current period value - minimum value) ÷ (maximum value - minimum value); for example, if the first risk assessment factor of the current period is 0.45, the historical minimum value is 0.15, and the maximum value is 0.75, then the normalized result is 0.5. Normalization makes the two risk factors comparable and on the same order of magnitude, which facilitates subsequent convolution processing and model weight allocation.

[0052] To reflect the evolution trend and accumulation pattern of risk factors over time, this embodiment extracts the corresponding values ​​of the first and second risk assessment factors in chronological order for the first twenty consecutive monitoring periods based on the normalized first and second risk assessment factors, constructing a two-dimensional time-series feature matrix with twenty rows and two columns. Each row corresponds to a monitoring period, and each column represents the standardized value of the first and second risk assessment factors for that period.

[0053] After the model is built, in order to further improve the robustness of model training and prediction, the system will perform zero-mean unit variance standardization on the two-dimensional time series feature matrix. That is, by calculating the mean and standard deviation of each column, each element in the matrix is ​​subtracted from the mean of that column and then divided by the standard deviation of that column, so that the data is data-oriented and the variation scale is unified, ensuring that the model will not have gradient explosion or weight shift problems during processing.

[0054] After the feature matrix is ​​prepared, the system inputs it into the trained convolutional neural network structure for analysis, outputting the comprehensive cost risk score for the current period. The convolutional neural network structure is pre-constructed and trained using this method, with training data sourced from a large amount of historical construction project monitoring data, including standardized risk factor sequences within continuous periods and their corresponding expert-determined risk level labels. A supervised learning mechanism is used during training, with cross-entropy as the loss function, and parameters are updated using a backpropagation algorithm combined with stochastic gradient descent. The training objective is to minimize the difference between the predicted value and the actual risk level label. The network structure consists of two one-dimensional convolutional layers with eight kernels, a size of five, and a stride of one. Each layer is followed by a batch normalization layer and an activation function processing layer, finally connected to a fully connected output layer. The output is a single risk score value between zero and one, representing the relative intensity of the comprehensive cost risk for the current period. For example, if in the input two-dimensional time series feature matrix, it is found that both risk factors have risen continuously in the last three periods, and the value of the second risk assessment factor fluctuates significantly higher, then the network may output a risk score close to 0.85, indicating that there is a significant risk of abnormal cost in the current period.

[0055] To classify and identify potential risks revealed by comprehensive cost risk scoring and implement timely and effective measures, this embodiment designs and executes the following steps after completing the comprehensive cost risk scoring to automatically generate risk alarm information and management suggestions, ensuring that anomalies can be identified and responded to scientifically during the construction project execution phase. In this embodiment, the system first sets multiple risk level intervals based on the distribution patterns of comprehensive cost risk scores for historical construction projects, forming a scoring level mapping table. For example, the interval from 0 to 0.45 is defined as "low risk level," the interval from 0.45 to 0.65 is defined as "medium risk level," and the interval above 0.65 is defined as "high risk level," allowing users to dynamically adjust according to project characteristics. To ensure the objectivity and adaptability of the mapping standard, this level classification standard is automatically generated through cluster analysis and density segmentation of the scoring distribution of more than one hundred historical projects, while introducing percentile indicators to ensure that the high-risk interval focuses on the most problematic anomalies requiring intervention. Once the level interval standard is set, it will serve as the basic parameter for subsequent comparison of scoring results.

[0056] After completing the risk classification, the system compares the current period's comprehensive cost risk score with the set risk level ranges one by one, identifies its level position, and assigns the corresponding level label. For example, if the current period's score is 0.73, it is identified as "high-risk level" according to the level mapping table. This step not only completes the classification of the score results but also provides the instruction basis for the risk alarm information generated by the subsequent system. The system automatically records the level label in the risk file for that period and identifies it in the visualization platform using icons, colors, or risk status bars, forming a monitoring interface with prompting information.

[0057] Once the current period is identified as medium-risk or high-risk, the system immediately and automatically generates a risk alarm message containing multiple structured fields. This information includes, but is not limited to: the current period's risk score; the current risk level; the trend analysis conclusion (e.g., the score has increased compared to the previous five periods); the dominant cost category indicated by the high-risk factor; suggested investigation paths (e.g., focusing on verifying actual labor costs or raw material usage); and the contract nodes and payment plan numbers that may correspond to the risk. The generated alarm message will be simultaneously sent to the project manager, cost supervisor, financial auditor, and the engineering company's risk control department through various means such as platform message push, project management system interface transmission, and email distribution, ensuring that key positions are simultaneously aware of the risk status and can respond in a unified manner.

[0058] Based on the risk level, the system outputs management and handling suggestions and initiates corresponding response plans. After sending risk alarm information, the system will also automatically output management and handling suggestions corresponding to the current level. For medium-risk levels, suggestions include strengthening contract performance verification, temporarily freezing approval authority for overspending items, and supplementing construction process data; while for high-risk levels, the system will suggest immediately initiating a cost anomaly verification process, restricting key fund expenditures, and submitting an audit application, among other intensive measures. For example, if a construction project scores 0.75 in the twelfth monitoring cycle, and the system identifies it as a high-risk level, it will prompt: "The current risk score is at a high-risk level. Please prioritize checking the trajectory of labor cost expenditures and the deviation of payment nodes. It is recommended to initiate a cost anomaly verification process and suspend subsequent non-contractual fund disbursements."

[0059] Example 2: A construction project cost analysis system, specifically including: The data acquisition module is used to collect business information and on-site process information of construction projects in the execution phase, and to perform field-level data matching between the information and the budget details table set at the start of the project to form a data set reflecting the status of contract execution and resource consumption. The risk assessment module is used to assess risks based on contract planning information and actual execution information in the dataset, and to determine whether the conditions for initiating the cost risk identification process have been met. The symptom identification module is used to perform multi-dimensional information cross-analysis operations to determine whether there are any signs of cost risk in the project, provided that the activation conditions are met. The assessment and calculation module is used to calculate the first risk assessment factor and the second risk assessment factor based on the identified symptom events. The first risk assessment factor is calculated based on the cumulative offset ratio between each expense in the budget details and the actual expenditure in the current period. The second risk assessment factor is calculated based on the cumulative frequency of abnormal data, the extent of exceeding the limit, and the duration of the trend. Both factors are dynamically updated with time windows as the dimension. The risk scoring module is used to input the first risk assessment factor and the second risk assessment factor into the risk assessment model to generate a comprehensive cost risk score; The alarm response module is used to automatically enter the cost anomaly verification process when the comprehensive cost risk score exceeds the set risk threshold, and generate corresponding risk warnings and handling suggestions based on the score result level.

[0060] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0061] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0062] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0063] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0064] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for analyzing the cost of construction projects, characterized in that, Includes the following steps: Collect business information and on-site process information of construction projects in the execution phase, and then perform field-level data matching with the budget details table set at the start of the project to form a data set reflecting the status of contract execution and resource consumption. Based on the contract planning information and actual execution information in the dataset, risk assessment is conducted to determine whether the conditions for initiating the cost risk identification process have been met. Under the premise that the conditions for initiation are met, perform multi-dimensional information cross-analysis to determine whether there are any signs of cost risk in the project; Based on the identified symptom events, a first risk assessment factor and a second risk assessment factor are calculated. The first risk assessment factor is calculated based on the cumulative deviation ratio between each expense in the budget details and the actual expenditure in the current period. The second risk assessment factor is calculated based on the cumulative frequency of abnormal data, the extent of exceeding limits, and the duration of the trend. Both factors are dynamically updated with time windows as the dimension. The first and second risk assessment factors are input into the risk assessment model to generate a comprehensive cost risk score. If the comprehensive cost risk score exceeds the set risk threshold, the cost anomaly verification process will be automatically initiated, and corresponding risk warnings and handling suggestions will be generated based on the score result level.

2. The method for analyzing construction project costs according to claim 1, characterized in that, Business information includes the total cost stipulated in the project contract, the phased performance plan, and the actual payment time record; on-site process information includes the raw material requisition record and the usage time of machinery and equipment at the construction site.

3. The method for analyzing construction project costs according to claim 1, characterized in that, Risk assessment refers to: Time series data of the performance stage sequence, payment interval, raw material usage quantity and machinery usage time are extracted separately according to fixed cycles. The continuous offset between the planned value and the corresponding actual value is calculated for each item to form a difference data sequence. By setting offset amplitude threshold, behavior continuity judgment rules and payment rhythm fluctuation range, abnormal segments appearing in each sequence are judged. If any abnormal segment exists in any sequence, the cost risk identification process is triggered. Furthermore, the cycle in which the offset amplitude of any data item in the same cycle exceeds its budget threshold is defined as an abnormal cycle.

4. The method for analyzing construction project costs according to claim 1, characterized in that, Multidimensional information cross-analysis refers to: Analyze whether each performance node is overdue, whether the construction progress plan is inconsistent with the on-site records, whether the total consumption of raw materials continuously exceeds the budget range, and whether the payment behavior deviates from the contract schedule. If any indicator meets the warning trigger condition, it is determined that the project has signs of cost risk.

5. The method for analyzing construction project costs according to claim 3, characterized in that, The cost anomaly verification process refers to the process of backtracking abnormal data, reconciling contract performance plans with payment records, and comparing material consumption records with actual inventory data.

6. The method for analyzing construction project costs according to claim 5, characterized in that, The calculation of the first risk assessment factor includes the following steps: The relative offset rate of each expense category in multiple consecutive monitoring periods is extracted. For each expense category, the period segment in which the continuous deviation from the budget threshold is identified, and the cumulative sum of the offset rate increase in each period segment is calculated to obtain the cumulative value of the continuous offset of the expense category. For each expense category, the cumulative value of continuous offset is further calculated to obtain its abnormal persistence index. The abnormal persistence index is the product of the number of periods in which the single-period increase in the offset rate of the expense category exceeds the preset sensitivity threshold within the continuous offset interval and the cumulative value of continuous offset. This is used to quantify the combined effect of continuous offset intensity and abnormal trigger density. The expense category with the largest abnormality persistence index is selected from all expense categories. The cumulative value of the continuous offset of this expense category is multiplied by the abnormality persistence index after nonlinear compression. The resulting product is used as the first risk assessment factor for the current period to express the comprehensive risk performance between the continuous accumulation capacity of cost offset and the density of abnormal triggers.

7. The method for analyzing construction project costs according to claim 6, characterized in that, Nonlinear compression processing refers to: Obtain the original value of the anomaly persistence index and use it as the growth variable. Then, perform a curve transformation on the growth variable with a gradually decreasing increment, so that as the value of the growth variable increases, the rate of increase of the corresponding transformation result gradually slows down, and the contribution of a single new anomaly trigger to the overall risk value gradually weakens. The compressed result obtained after the transformation is used as the nonlinear compressed value of the anomaly persistence index.

8. The method for analyzing the cost of construction projects according to claim 7, characterized in that, The calculation of the second risk assessment factor includes the following steps: After arranging the abnormal periods in multiple continuous monitoring periods in chronological order, a sliding window of fixed time length is used to traverse them sequentially from front to back. If two or more abnormal periods appear in each sliding window, all abnormal periods in that window are classified as an associated risk node and removed from the sequence to be processed. After all windows have been traversed, the remaining unclassified single abnormal period is classified as an independent risk node, and each abnormal period participates in the attribution judgment only once. For each associated risk node, the number of abnormal cycles within that node is counted, and the sum of all abnormal offset amplitudes within that node is calculated. Then, the number of abnormal cycles and the total offset amplitude are multiplied together to obtain the risk value of that associated node. For each independent risk node, the single abnormal offset amplitude of that node is used as the risk value of that node. The risk values ​​of all nodes are arranged in chronological order of their appearance. The risk value of the earliest node is recorded directly. For each subsequent node, its risk value is multiplied by its order of appearance in the entire node sequence, so that nodes that appear later in time have a higher impact weight. The risk values ​​of all nodes after sequential amplification are accumulated, and the accumulated result is used as the second risk assessment factor for the current period to reflect the cumulative risk effect of abnormal behavior gradually increasing over time.

9. The method for analyzing the cost of construction projects according to claim 8, characterized in that, The steps for generating a comprehensive cost risk score include: The first and second risk assessment factors for the current monitoring period are normalized to map the original values ​​to a unified range and form standardized feature data. The first and second risk assessment factors, standardized for the current period, are used as inputs to form a two-dimensional time-series feature matrix according to their arrangement in multiple consecutive periods. This matrix is ​​then input into the trained convolutional neural network structure, i.e., the risk assessment model. The risk assessment model extracts features and matches risk levels for the risk change trends in multiple periods, and outputs the comprehensive cost risk score for the current period. If the comprehensive cost risk score exceeds the set risk threshold, the comprehensive cost risk score will be compared with the preset risk level range, the corresponding level standard will be matched, and the corresponding risk alarm information and management and handling suggestions will be automatically generated.

10. A construction project cost analysis system, based on the construction project cost analysis method according to any one of claims 1-9, characterized in that, Specifically, it includes: The data acquisition module is used to collect business information and on-site process information of construction projects in the execution phase, and to perform field-level data matching between the information and the budget details table set at the start of the project to form a data set reflecting the status of contract execution and resource consumption. The risk assessment module is used to assess risks based on contract planning information and actual execution information in the dataset, and to determine whether the conditions for initiating the cost risk identification process have been met. The symptom identification module is used to perform multi-dimensional information cross-analysis operations to determine whether there are any signs of cost risk in the project, provided that the activation conditions are met. The assessment and calculation module is used to calculate the first risk assessment factor and the second risk assessment factor based on the identified symptom events. The first risk assessment factor is calculated based on the cumulative offset ratio between each expense in the budget details and the actual expenditure in the current period. The second risk assessment factor is calculated based on the cumulative frequency of abnormal data, the extent of exceeding the limit, and the duration of the trend. Both factors are dynamically updated with time windows as the dimension. The risk scoring module is used to input the first risk assessment factor and the second risk assessment factor into the risk assessment model to generate a comprehensive cost risk score; The alarm response module is used to automatically enter the cost anomaly verification process when the comprehensive cost risk score exceeds the set risk threshold, and generate corresponding risk warnings and handling suggestions based on the score result level.

Images