Engineering intelligent management method and system

Abstract

The invention discloses an intelligent engineering management method and system, and relates to the technical field of engineering acceptance, and the method comprises the steps: calculating the occurrence frequency Fca of different types of repair reasons based on the classification statistics of the repair reasons in an engineering original data set Draw collected in real time, and generating a high-frequency rework factor frequency feature vector Vcau; the method can quickly lock the dominant factors with high occurrence of repair, accurately identify the weak links in the construction process, construct a comprehensive abnormal score calculation model, and calculate the comprehensive abnormal score San, thereby carrying out the quantitative evaluation of the overall construction state of the current project, obtaining an early warning signal Ealert in time, and outputting a high-frequency rework factor analysis report. According to the method, an accurate optimization decision basis is provided, early recognition and intervention of construction quality hidden dangers can be achieved on the premise that extra hardware investment is not increased, the situation that small-area repair problems are accumulated and evolved into large-scale quality defects is prevented, and the controllability of the construction process and the final delivery quality are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering acceptance, and in particular to an engineering intelligent management method and system. Background Art

[0002] In the modern construction industry, with the continuous acceleration of urbanization and the increasing requirements of residents for the quality of living environment, the field of construction engineering has been widely developed and applied. Among them, construction engineering not only covers the main structure construction and mechanical and electrical installation, but also includes the comprehensive improvement of the building appearance, interior space comfort and functionality, that is, engineering. As an important part of construction engineering, engineering mainly involves the beautification of indoor and outdoor spaces, layout optimization and functional improvement. Its quality is directly related to the overall user experience and commercial value of the building.

[0003] In the existing engineering management practice, although progress management software, quality acceptance record system and other tools have been initially applied, there is a general lack of systematic and quantitative monitoring methods for small-scale repair incidents during the construction process, especially those with a small area and strong concealment. These small-scale repair phenomena are often regarded as normal fluctuations and ignored, and fail to attract timely attention, resulting in the accumulation of problems and the evolution of large-scale quality defects, increasing the difficulty and cost of large-scale repairs in the later stage.

[0004] At the same time, the in-depth analysis of the causes of the repair phenomenon is also relatively weak. The existing management system focuses on result-oriented records, such as simply filing the number or category of repairs, but lacks in-depth exploration of specific factors that lead to high incidence of repairs, such as operational errors of specific construction personnel, quality fluctuations of specific material batches, and specific process defects. This management model lacks process data association and causal chain deduction, resulting in difficulty in accurately identifying the root cause even if a repair problem is discovered, and thus making it impossible to formulate targeted prevention and optimization measures. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an engineering intelligent management method and system, which solves the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an engineering intelligent management method, comprising the following steps: S1, real-time collection of completion data and repair data of each work surface during the construction process to form the original data set Draw; S2. Filter and process the original engineering data set Draw, extract small-area repair events within a fixed repair area, count the small-area repair events and the total amount of construction tasks per unit time, calculate the small-area repair rate Rsmall, and obtain the small-area repair rate feature vector Vsmall; S3, based on the classification statistics of the rework reasons in the original engineering data set Draw, calculate the occurrence frequency Fca of different types of rework reasons, and generate the frequency feature vector Vcau of high-frequency rework factors; S4. Integrate the acquired small area rework rate feature vector Vsmall and high frequency rework factor frequency feature vector Vcau, build a comprehensive anomaly score calculation model, and calculate the comprehensive anomaly score San under the current engineering construction status; S5. Compare the obtained comprehensive abnormality score San with the preset engineering status warning threshold Than to obtain the warning signal Ealert and the high-frequency rework factor analysis report.

[0007] Preferably, said S1 includes S11 and S12; S11. During the construction process, the construction completion data and repair data on each work surface in the decoration project are collected in real time through the construction record system and electronic recording equipment to form a preliminary collection data set Dco; The initial collection data set Dco includes the construction task number TWID, the construction completion timestamp TWT, the rework event number RID, the rework area RA, the rework reason code RC and the rework occurrence timestamp RT; S12, performing data preprocessing on the obtained preliminary collected data set Dco, wherein the data preprocessing includes format standardization, missing value filling and time sequence unification to form the engineering original data set Draw.

[0008] Preferably, said S2 includes S21; S21, filtering and processing the original engineering data set Draw, by filtering out the repair area RA less than the preset small area determination threshold ThRa, and then extracting the small area repair events within the small area determination threshold ThRa, forming a small area repair event subset Dsmall={RID(i), RA(i), RC(i), RT(i)|RA(i)<ThRa}; Among them, i represents the i-th rework event, RID(i), RA(i), RC(i), and RT(i) respectively represent the rework event number RID, rework area RA, rework reason code RC, and rework occurrence timestamp RT of the i-th rework event.

[0009] Preferably, said S2 includes S22; S22. Extract the total number of construction tasks Ntask based on the original engineering data set Draw, extract the total number of small area repair events Nsmall based on the small area repair event subset Dsmall, divide them according to a fixed unit time, mark them as time periods t, and count the small area repair rate Rsmall(t) in time period t. After integrating all time periods, obtain the small area repair rate feature vector Vsmall={Rsmall(1), Rsmall(2), ..., Rsmall(t)|t∈M}, where M represents the total number of time periods divided according to a fixed unit time. The small area repair rate Rsmall(t) within time period t is obtained by the following calculation formula: ; Where Nsmall(t) represents the total number of small area repair events Nsmall within time period t, and Ntask(t) represents the total number of construction tasks Ntask within time period t.

[0010] Preferably, said S3 includes S31; S31, extracting the repair reason code RC based on the original engineering data set Draw, and classifying and counting the repair reason code RC to obtain the number of repair events Ncause for each type of repair reason code RC, and then integrating the number of repair events Ncause for all repair reason codes RC to obtain the event number set Nset; The number of repair events Ncause is obtained by the following calculation formula: ; Where Ncause(k) represents the number of repair events of repair reason category k, δ represents the conditional exponential function, and returns 1 when the condition of the conditional exponential function δ is met, otherwise it returns 0, N represents the total number of repair records, specifically the length of the original engineering data set Draw, and RC(i) represents the repair reason code of the i-th repair event.

[0011] Preferably, said S3 includes S32; S32, based on the obtained event quantity set Nset, calculate the occurrence frequency of each type of rework reason code RC, obtain the rework reason frequency vector Fca, and generate a high-frequency rework factor feature vector Vcau={Fca(1), Fca(2), ..., Fca(k)|k∈K} by integrating the rework reason frequency vector Fca of each type of rework reason code RC, where K represents the total number of types of rework reason codes RC; The repair reason frequency vector Fca is obtained by the following calculation formula: ; Where Fca(k) represents the repair reason frequency vector of repair reason category k.

[0012] Preferably, the S4 includes S41; S41, performing maximum value normalization processing on the obtained small area rework rate feature vector Vsmall, obtaining the normalized small area rework rate feature vector VsmallNorm after the maximum value normalization processing, performing square normalization processing on the obtained high-frequency rework factor frequency feature vector Vcau, obtaining the normalized high-frequency rework factor frequency feature vector VcauNorm after the square normalization processing, amplifying the difference of high-frequency factors, integrating the normalized small area rework rate feature vector VsmallNorm and the normalized high-frequency rework factor frequency feature vector VcauNorm, constructing a comprehensive anomaly score calculation model, and calculating the comprehensive anomaly score San under the current engineering construction status; The normalized small area repair rate feature vector VsmallNorm is obtained by the following maximum value normalization processing formula: ; Wherein, VsmallNorm(t) represents the normalized small area repair rate feature vector in time period t, Vsmall(t) represents the small area repair rate feature vector in time period t, and max(Vsmall) represents the maximum small area repair rate feature vector in all time periods; The normalized high-frequency rework factor frequency characteristic vector VcauNorm is obtained by the following square normalization processing formula: ; Where VcauNorm(k) represents the normalized high-frequency rework factor frequency feature vector of rework reason category k, and Fca(k) represents the rework reason frequency vector of rework reason category k; The comprehensive anomaly score San is obtained by the following calculation formula: ; Wherein, s1 and s2 represent the weight coefficients of the normalized small area rework rate feature vector VsmallNorm and the normalized high frequency rework factor frequency feature vector VcauNorm, respectively, and s1+s2=1. The specific value is set by the user, and K represents the total number of types of rework reason codes RC.

[0013] Preferably, the S5 includes S51; S51, comparing the obtained comprehensive abnormality score San with the preset engineering status warning threshold Than, judging whether there is an abnormality in the current engineering construction status according to the comparison result, and generating an early warning signal Ealert according to the comparison result; The warning signal Ealert is obtained through the following judgment generation method: When the comprehensive abnormality score San ≥ the engineering status warning threshold Than, the current engineering construction status is judged to be abnormal, and the warning signal Ealert=1 is generated; When the comprehensive abnormality score San ≥ the engineering status warning threshold Than, it is determined that there is no abnormality in the current engineering construction status, and the warning signal Ealert=0 is generated.

[0014] Preferably, the S5 includes S52; S52. According to the obtained warning signal Ealert, when the warning signal Ealert=1 occurs, the dominant rework factor Fdom in the current engineering construction cycle is generated based on the high-frequency rework factor characteristic vector Vcau, and all the rework records that meet the generation of the dominant rework factor Fdom are integrated. Specifically, the rework number RID is extracted to obtain the records in the original engineering data set Draw, and a high-frequency rework factor analysis report RcauReport is generated, including the rework reason category k, the number of reworks and the dominant rework factor Fdom. Then, according to the high-frequency rework factor analysis report RcauReport, an optimization suggestion is generated by extracting the specification requirements for the repair reason category k in the preset engineering construction specification for the active rework factor category; The dominant rework factor Fdom is obtained by the following calculation formula: ; In the formula, argmax k (Fca(k)) represents the repair reason frequency vector of the repair reason category k for which the maximum value is found.

[0015] An engineering intelligent management system, comprising an engineering data acquisition module, an engineering vector generation module, a classification statistics module, a comprehensive evaluation module and a decision-making module; The engineering data collection module collects the completion data and repair data of each work surface in real time during the engineering construction process to form the engineering original data set Draw; The engineering vector generation module filters and processes the original engineering data set Draw, extracts small-area repair events within a fixed repair area, counts small-area repair events and the total amount of construction tasks per unit time, calculates the small-area repair rate Rsmall, and obtains the small-area repair rate feature vector Vsmall; The classification statistics module performs classification statistics based on the rework reasons in the original engineering data set Draw, calculates the occurrence frequency Fca of different types of rework reasons, and generates the frequency feature vector Vcau of high-frequency rework factors; The comprehensive evaluation module integrates the acquired small area repair rate feature vector Vsmall and high frequency rework factor frequency feature vector Vcau, builds a comprehensive anomaly score calculation model, and calculates the comprehensive anomaly score San under the current engineering construction status; The decision module compares the obtained comprehensive anomaly score San with the preset engineering status warning threshold Than to obtain the warning signal Ealert and high-frequency rework factor analysis report.

[0016] The present invention provides an engineering intelligent management method and system, which has the following beneficial effects: (1) Based on the real-time collected engineering original data set Draw, the classification statistics of the reasons for rework are calculated, the occurrence frequency Fca of different types of rework reasons is calculated, and the frequency characteristic vector Vcau of the high-frequency rework factor is generated, so that the dominant factors of high incidence of rework can be quickly locked, the weak links in the construction process can be accurately identified, and a comprehensive anomaly score calculation model is constructed to calculate the comprehensive anomaly score San, so as to quantitatively evaluate the overall construction status of the current project, obtain the early warning signal Ealert in time, and output the high-frequency rework factor analysis report, providing construction management personnel with accurate optimization decision-making basis. Without increasing additional hardware investment, it can realize early identification and intervention of construction quality hazards, prevent small-area rework problems from accumulating into large-scale quality defects, and greatly improve the controllability of the construction process and the final delivery quality.

[0017] (2) Through the systematic extraction and classification statistics of the rework reason codes RC in the original engineering data set Draw, a comprehensive quantitative analysis of the distribution characteristics of rework problems during the construction process can be achieved. Furthermore, based on the event number set Nset, the occurrence frequency of each type of rework reason code RC is calculated, the rework reason frequency vector Fca is obtained, and the high-frequency rework factor feature vector Vcau is generated by integration, so that the main rework factors affecting the quality of engineering construction can be accurately identified. This not only avoids the fuzzy attribution of rework problems caused by the lack of systematic data classification in traditional construction management, but also can dynamically grasp the changing trend of rework reasons during the construction process and timely discover potential systemic problems in the construction process.

[0018] (3) By comparing the comprehensive anomaly score San with the preset engineering status warning threshold Than, the warning signal Ealert is generated in real time, realizing the rapid identification and accurate alarm of construction status anomalies. Furthermore, when the warning signal Ealert=1 is generated, based on the high-frequency rework factor feature vector Vcau, the dominant rework factor Fdom in the current engineering construction cycle is dynamically extracted, and the relevant rework records are integrated to generate the high-frequency rework factor analysis report RcauReport. Based on RcauReport, targeted construction optimization suggestions are output to realize dynamic anomaly modeling based on the comprehensive analysis of multi-source rework characteristics. When the initial abnormality of the construction status occurs, the dominant rework factor can be quickly traced back and targeted optimization measures can be generated, effectively shortening the abnormality identification and management response time, and improving the timeliness and accuracy of the closed-loop processing of quality problems on the construction site. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic diagram of the steps of an engineering intelligent management method of the present invention; Figure 2 This is a schematic diagram of a block diagram of an engineering intelligent management system of the present invention; Figure 3 This is a statistical bar chart of the classification of rework reasons based on the original engineering data set Draw. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0021] Example 1 The present invention provides an engineering intelligent management method, please refer to Figure 1 , including the following steps: S1, real-time collection of completion data and repair data of each work surface during the construction process to form the original data set Draw; S2. Filter and process the original engineering data set Draw, extract small-area repair events within a fixed repair area, count the small-area repair events and the total amount of construction tasks per unit time, calculate the small-area repair rate Rsmall, and obtain the small-area repair rate feature vector Vsmall; S3, based on the classification statistics of the rework reasons in the original engineering data set Draw, calculate the occurrence frequency Fca of different types of rework reasons, and generate the frequency feature vector Vcau of high-frequency rework factors; S4. Integrate the acquired small area rework rate feature vector Vsmall and high frequency rework factor frequency feature vector Vcau, build a comprehensive anomaly score calculation model, and calculate the comprehensive anomaly score San under the current engineering construction status; S5. Compare the obtained comprehensive abnormality score San with the preset engineering status warning threshold Than to obtain the warning signal Ealert and the high-frequency rework factor analysis report.

[0022] In this embodiment, based on the real-time collected engineering original data set Draw, the construction completion data and repair data of each work surface can be fully recorded, providing a reliable data basis for subsequent quality analysis; by screening and processing the engineering original data set Draw, extracting small-area repair events within a fixed repair area, and calculating the small-area repair rate Rsmall and obtaining the small-area repair rate feature vector Vsmall, it is possible to achieve timely monitoring and quantitative analysis of micro-quality fluctuations in the construction process; further, based on the classification statistics of the repair reasons in the engineering original data set Draw, the occurrence frequency Fca of different types of repair reasons is calculated and the high-frequency rework factor frequency feature vector Vcau is generated, so that the dominant factors of high-incidence repairs can be quickly locked, and the high-frequency rework factor frequency feature vector Vcau in the construction process can be accurately identified. Weak links; By integrating the small-area rework rate feature vector Vsmall and the high-frequency rework factor frequency feature vector Vcau, a comprehensive anomaly score calculation model is constructed, and the comprehensive anomaly score San is calculated, so as to quantitatively evaluate the overall construction status of the current project; finally, the comprehensive anomaly score San is compared with the engineering status warning threshold Than, the warning signal Ealert is obtained in time, and the high-frequency rework factor analysis report is output, which provides construction management personnel with accurate optimization decision-making basis, and can achieve early identification and intervention of construction quality hazards without increasing additional hardware investment, prevent the accumulation of small-area rework problems from evolving into large-scale quality defects, and greatly improve the controllability of the construction process and the final delivery quality, which has significant practical value and promotion significance.

[0023] Example 2 Please refer to Figure 1 , specifically: the S1 includes S11 and S12; S11. During the construction process, the construction completion data and repair data on each work surface in the decoration project are collected in real time through the construction record system and electronic recording equipment to form a preliminary collection data set Dco; The initial collection data set Dco includes the construction task number TWID, the construction completion timestamp TWT, the rework event number RID, the rework area RA, the rework reason code RC and the rework occurrence timestamp RT; S12, performing data preprocessing on the obtained preliminary collected data set Dco, wherein the data preprocessing includes format standardization, missing value filling and time sequence unification to form the engineering original data set Draw.

[0024] Said S2 includes S21; S21, filtering and processing the original engineering data set Draw, by filtering out the repair area RA less than the preset small area determination threshold ThRa, and then extracting the small area repair events within the small area determination threshold ThRa, forming a small area repair event subset Dsmall={RID(i), RA(i), RC(i), RT(i)|RA(i)<ThRa}; Among them, i represents the i-th rework event, RID(i), RA(i), RC(i), and RT(i) respectively represent the rework event number RID, rework area RA, rework reason code RC, and rework occurrence timestamp RT of the i-th rework event.

[0025] The S2 includes S22; S22. Extract the total number of construction tasks Ntask based on the original engineering data set Draw, extract the total number of small area repair events Nsmall based on the small area repair event subset Dsmall, divide them according to a fixed unit time, mark them as time periods t, and count the small area repair rate Rsmall(t) in time period t. After integrating all time periods, obtain the small area repair rate feature vector Vsmall={Rsmall(1), Rsmall(2), ..., Rsmall(t)|t∈M}, where M represents the total number of time periods divided according to a fixed unit time. The small area repair rate Rsmall(t) within time period t is obtained by the following calculation formula: ; Where Nsmall(t) represents the total number of small area repair events in time period t, Ntask(t) represents the total number of construction tasks in time period t, Ntask; A specific example of the small area repair rate feature vector Vsmall is shown below: Example prerequisites: Set the small area judgment threshold ThRa=1.0; sampling period unit: divided by day; the example collection time period includes 2 days, that is, M=2; Relevant data in the original engineering data set Draw: Rework event number RID: 101; Rework area RA: 0.5; Rework occurrence timestamp RT: Day 1; Repair event number RID: 102; repair area RA: 1.2; repair occurrence timestamp RT: day 1; Rework event number RID: 103; Rework area RA: 0.8; Rework occurrence timestamp RT: Day 1; Repair event number RID: 104; repair area RA: 0.7; repair occurrence timestamp RT: the second day; Repair event number RID: 105; repair area RA: 0.6; repair occurrence timestamp RT: the second day; Repair event number RID: 106; repair area RA: 1.5; repair occurrence timestamp RT: the second day; The total number of construction tasks Ntask in time period t=1: 50; The total number of construction tasks Ntask in time period t=1: 48; Filter the small area repair event subset Dsmall: Small area repair event subset Dsmall = {RID(i), RA(i), RC(i), RT(i) | RA(i) < ThRa = 1.0} Eligible repair events on Day 1: Repair event number RID: 101 and Repair event number RID: 103; Eligible repair events on the second day: Repair event number RID: 104 and repair event number RID: 105; Calculate the small area repair rate Rsmall (1) = 0.04 in the time period t = 1; Calculate the small area repair rate Rsmall (2) ≈ 0.0417 in the time period t = 2; The small area repair rate feature vector Vsmall={Rsmall(1)=0.04, Rsmall(2)≈0.0417}.

[0026] In this embodiment, the construction task number TWID, construction completion timestamp TWT, repair event number RID, repair area RA, repair reason code RC and repair occurrence timestamp RT are collected in real time through the construction record system and electronic recording equipment to form a preliminary collected data set Dco, and the original engineering data set Draw is constructed through format standardization, missing value filling and time sequence unification preprocessing, which ensures the consistency and comparability of the data from the source and provides a solid data foundation for subsequent intelligent analysis. At the same time, for small-area repair events during the construction process, by screening and processing the original engineering data set Draw, small-area repair events with a repair area RA less than the preset small area judgment threshold ThRa are extracted to form a small-area repair event subset Dsmall, and based on the total number of construction tasks Ntask and the total number of small-area repair events Nsmall in a unit time period t, the small-area repair rate Rsmall(t) is dynamically calculated, and finally the small-area repair rate feature vector Vsmall is obtained. This can not only ensure that the data association between repair events and construction tasks is highly accurate and timely, but also realize the fine-grained tracking and analysis of the dynamic evolution trend of the quality of the construction process in different time periods, effectively avoiding the hidden accumulation of quality problems caused by weak data foundation or delayed identification of repair events, and further improving the accuracy and response efficiency of engineering process management.

[0027] Example 3 Please refer to Figure 1 and Figure 3 , specifically: the S3 includes S31; S31, extracting the repair reason code RC based on the original engineering data set Draw, and classifying and counting the repair reason code RC to obtain the number of repair events Ncause for each type of repair reason code RC, and then integrating the number of repair events Ncause for all repair reason codes RC to obtain the event number set Nset; The number of repair events Ncause is obtained by the following calculation formula: ; Where Ncause(k) represents the number of repair events of repair reason category k, δ represents the conditional exponential function, and returns 1 when the condition of the conditional exponential function δ is met, otherwise it returns 0, N represents the total number of repair records, specifically the length of the original engineering data set Draw, and RC(i) represents the repair reason code of the i-th repair event.

[0028] The S3 includes S32; S32, based on the obtained event quantity set Nset, calculate the occurrence frequency of each type of rework reason code RC, obtain the rework reason frequency vector Fca, and generate a high-frequency rework factor feature vector Vcau={Fca(1), Fca(2), ..., Fca(k)|k∈K} by integrating the rework reason frequency vector Fca of each type of rework reason code RC, where K represents the total number of types of rework reason codes RC; The repair reason frequency vector Fca is obtained by the following calculation formula: ; Where Fca(k) represents the repair reason frequency vector of repair reason category k.

[0029] In this embodiment, by systematically extracting and classifying and counting the repair reason codes RC in the original data set Draw of the project, a comprehensive quantitative analysis of the distribution characteristics of repair problems during the construction process can be achieved. Specifically, by extracting the repair reason codes RC and performing classification and statistics, the number of repair events Ncause corresponding to each type of repair reason is accurately obtained, and the number of repair events of all repair reason codes RC is integrated to form an event number set Nset, which can not only clearly grasp the occurrence frequency and distribution trend of various types of repair problems, but also provide solid data support for the subsequent identification of high-frequency rework factors. Further, based on the event number set Nset, the frequency of occurrence of each type of repair reason code RC is calculated, the repair reason frequency vector Fca is obtained, and the high-frequency rework factor feature vector Vcau is generated by integration, so that the main repair factors affecting the quality of engineering construction can be accurately identified. Through the above-mentioned detailed steps, not only the fuzzy attribution of repair problems caused by the lack of systematic data classification in traditional construction management is avoided, but also the trend of changes in repair reasons can be dynamically grasped during the construction process, and potential systemic problems in the construction process can be discovered in time, effectively supporting the refined management of construction optimization and process control.

[0030] Example 4 Please refer to Figure 1 Specifically: S4 includes S41; S41, performing maximum value normalization processing on the obtained small area rework rate feature vector Vsmall, obtaining the normalized small area rework rate feature vector VsmallNorm after the maximum value normalization processing, performing square normalization processing on the obtained high-frequency rework factor frequency feature vector Vcau, obtaining the normalized high-frequency rework factor frequency feature vector VcauNorm after the square normalization processing, amplifying the difference of high-frequency factors, integrating the normalized small area rework rate feature vector VsmallNorm and the normalized high-frequency rework factor frequency feature vector VcauNorm, constructing a comprehensive anomaly score calculation model, and calculating the comprehensive anomaly score San under the current engineering construction status; The normalized small area repair rate feature vector VsmallNorm is obtained by the following maximum value normalization processing formula: ; Wherein, VsmallNorm(t) represents the normalized small area repair rate feature vector in time period t, Vsmall(t) represents the small area repair rate feature vector in time period t, and max(Vsmall) represents the maximum small area repair rate feature vector in all time periods; The normalized high-frequency rework factor frequency characteristic vector VcauNorm is obtained by the following square normalization processing formula: ; Where VcauNorm(k) represents the normalized high-frequency rework factor frequency feature vector of rework reason category k, and Fca(k) represents the rework reason frequency vector of rework reason category k; The comprehensive anomaly score San is obtained by the following calculation formula: ; Wherein, s1 and s2 represent the weight coefficients of the normalized small area rework rate feature vector VsmallNorm and the normalized high frequency rework factor frequency feature vector VcauNorm, respectively, and s1+s2=1. The specific value is set by the user, and K represents the total number of types of rework reason codes RC.

[0031] The S5 includes S51; S51, comparing the obtained comprehensive abnormality score San with the preset engineering status warning threshold Than, judging whether there is an abnormality in the current engineering construction status according to the comparison result, and generating an early warning signal Ealert according to the comparison result; The warning signal Ealert is obtained through the following judgment generation method: When the comprehensive abnormality score San ≥ the engineering status warning threshold Than, the current engineering construction status is judged to be abnormal, and the warning signal Ealert=1 is generated; When the comprehensive abnormality score San ≥ the engineering status warning threshold Than, it is determined that there is no abnormality in the current engineering construction status, and the warning signal Ealert=0 is generated.

[0032] The S5 includes S52; S52. According to the obtained warning signal Ealert, when the warning signal Ealert=1 occurs, the dominant rework factor Fdom in the current engineering construction cycle is generated based on the high-frequency rework factor characteristic vector Vcau, and all the rework records that meet the generation of the dominant rework factor Fdom are integrated. Specifically, the rework number RID is extracted to obtain the records in the original engineering data set Draw, and a high-frequency rework factor analysis report RcauReport is generated, including the rework reason category k, the number of reworks and the dominant rework factor Fdom. Then, according to the high-frequency rework factor analysis report RcauReport, an optimization suggestion is generated by extracting the specification requirements for the repair reason category k in the preset engineering construction specification for the active rework factor category; The dominant rework factor Fdom is obtained by the following calculation formula: ; In the formula, argmax k (Fca(k)) represents the repair reason frequency vector of the repair reason category k for which the maximum value is found.

[0033] In this embodiment, the normalized small area rework rate feature vector Vsmall is normalized to the maximum value to obtain the normalized small area rework rate feature vector VsmallNorm, and the high frequency rework factor frequency feature vector Vcau is square normalized to obtain the normalized high frequency rework factor frequency feature vector VcauNorm, so as to effectively amplify the local abnormal differences in the construction process during feature integration and enhance the sensitivity of the overall abnormal trend. Subsequently, a comprehensive abnormal score calculation model is constructed based on the normalized feature vector, and the early warning signal Ealert is generated in real time by comparing the comprehensive abnormal score San with the preset engineering status early warning threshold Than, thereby realizing the rapid identification and accurate alarm of the abnormal construction status. Further, when the early warning signal Ealert=1 is generated, based on the high frequency rework factor feature vector Vcau, the dominant rework factor Fdom in the current engineering construction cycle is dynamically extracted, and the high frequency rework factor analysis report RcauReport is generated by integrating the relevant rework records, and the targeted construction optimization suggestions are output based on RcauReport. Through the above-mentioned detailed steps, the present invention can not only realize dynamic anomaly modeling based on comprehensive analysis of multi-source rework characteristics, but also can quickly trace the dominant rework factors and generate targeted optimization measures when initial anomalies in the construction status occur, effectively shortening the anomaly identification and management response time, and improving the timeliness and accuracy of closed-loop processing of quality problems on the construction site.

[0034] Example 5 An engineering intelligent management system, please refer to Figure 2,Specifically: including engineering data acquisition module, engineering vector generation module, classification and statistics module, comprehensive evaluation module and decision-making module; The engineering data collection module collects the completion data and repair data of each work surface in real time during the engineering construction process to form the engineering original data set Draw; The engineering vector generation module filters and processes the original engineering data set Draw, extracts small-area repair events within a fixed repair area, counts small-area repair events and the total amount of construction tasks per unit time, calculates the small-area repair rate Rsmall, and obtains the small-area repair rate feature vector Vsmall; The classification statistics module performs classification statistics based on the rework reasons in the original engineering data set Draw, calculates the occurrence frequency Fca of different types of rework reasons, and generates the frequency feature vector Vcau of high-frequency rework factors; The comprehensive evaluation module integrates the acquired small area repair rate feature vector Vsmall and high frequency rework factor frequency feature vector Vcau, builds a comprehensive anomaly score calculation model, and calculates the comprehensive anomaly score San under the current engineering construction status; The decision module compares the obtained comprehensive anomaly score San with the preset engineering status warning threshold Than to obtain the warning signal Ealert and high-frequency rework factor analysis report.

[0035] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An engineering intelligent management method, characterized in that: The following steps are involved: S1, real-time collection of completion data and repair data of each work surface during the construction process to form the original data set Draw; S2. Filter and process the original engineering data set Draw, extract small-area repair events within a fixed repair area, count the small-area repair events and the total amount of construction tasks per unit time, calculate the small-area repair rate Rsmall, and obtain the small-area repair rate feature vector Vsmall; S3, based on the classification statistics of the rework reasons in the original engineering data set Draw, calculate the occurrence frequency Fca of different types of rework reasons, and generate the frequency feature vector Vcau of high-frequency rework factors; S4. Integrate the acquired small area rework rate feature vector Vsmall and high frequency rework factor frequency feature vector Vcau, build a comprehensive anomaly score calculation model, and calculate the comprehensive anomaly score San under the current engineering construction status; S5. Compare the obtained comprehensive abnormality score San with the preset engineering status warning threshold Than to obtain the warning signal Ealert and the high-frequency rework factor analysis report.

2. The engineering intelligent management method according to claim 1, characterized in that: Said S1 includes S11 and S12; S11. During the construction process, the construction completion data and repair data on each work surface in the decoration project are collected in real time through the construction record system and electronic recording equipment to form a preliminary collection data set Dco; The initial collection data set Dco includes the construction task number TWID, the construction completion timestamp TWT, the rework event number RID, the rework area RA, the rework reason code RC and the rework occurrence timestamp RT; S12, performing data preprocessing on the obtained preliminary collected data set Dco, wherein the data preprocessing includes format standardization, missing value filling and time sequence unification to form the engineering original data set Draw.

3. The engineering intelligent management method according to claim 1, characterized in that: Said S2 includes S21; S21, filtering and processing the original engineering data set Draw, by filtering out the repair area RA less than the preset small area determination threshold ThRa, and then extracting the small area repair events within the small area determination threshold ThRa, forming a small area repair event subset Dsmall={RID(i), RA(i), RC(i), RT(i)|RA(i)<ThRa}; Among them, i represents the i-th rework event, RID(i), RA(i), RC(i), and RT(i) respectively represent the rework event number RID, rework area RA, rework reason code RC, and rework occurrence timestamp RT of the i-th rework event.

4. The engineering intelligent management method according to claim 3, characterized in that: The S2 includes S22; S22. Extract the total number of construction tasks Ntask based on the original engineering data set Draw, extract the total number of small area repair events Nsmall based on the small area repair event subset Dsmall, divide them according to a fixed unit time, mark them as time periods t, and count the small area repair rate Rsmall(t) in time period t. After integrating all time periods, obtain the small area repair rate feature vector Vsmall={Rsmall(1), Rsmall(2), ..., Rsmall(t)|t∈M}, where M represents the total number of time periods divided according to a fixed unit time. The small area repair rate Rsmall(t) within time period t is obtained by the following calculation formula: ； Where Nsmall(t) represents the total number of small area repair events Nsmall within time period t, and Ntask(t) represents the total number of construction tasks Ntask within time period t.

5. The engineering intelligent management method according to claim 1, characterized in that: The S3 includes S31; S31, extracting the repair reason code RC based on the original engineering data set Draw, and classifying and counting the repair reason code RC to obtain the number of repair events Ncause for each type of repair reason code RC, and then integrating the number of repair events Ncause for all repair reason codes RC to obtain the event number set Nset; The number of repair events Ncause is obtained by the following calculation formula: ； Where Ncause(k) represents the number of repair events of repair reason category k, δ represents the conditional exponential function, and returns 1 when the condition of the conditional exponential function δ is met, otherwise it returns 0, N represents the total number of repair records, specifically the length of the original engineering data set Draw, and RC(i) represents the repair reason code of the i-th repair event.

6. The engineering intelligent management method according to claim 5, characterized in that: The S3 includes S32; S32, based on the obtained event quantity set Nset, calculate the occurrence frequency of each type of rework reason code RC, obtain the rework reason frequency vector Fca, and generate a high-frequency rework factor feature vector Vcau={Fca(1), Fca(2), ..., Fca(k)|k∈K} by integrating the rework reason frequency vector Fca of each type of rework reason code RC, where K represents the total number of types of rework reason codes RC; The repair reason frequency vector Fca is obtained by the following calculation formula: ； Where Fca(k) represents the repair reason frequency vector of repair reason category k.

7. The engineering intelligent management method according to claim 1, characterized in that: The S4 includes S41; S41, performing maximum value normalization processing on the obtained small area rework rate feature vector Vsmall, obtaining the normalized small area rework rate feature vector VsmallNorm after the maximum value normalization processing, performing square normalization processing on the obtained high-frequency rework factor frequency feature vector Vcau, obtaining the normalized high-frequency rework factor frequency feature vector VcauNorm after the square normalization processing, amplifying the difference of high-frequency factors, integrating the normalized small area rework rate feature vector VsmallNorm and the normalized high-frequency rework factor frequency feature vector VcauNorm, constructing a comprehensive anomaly score calculation model, and calculating the comprehensive anomaly score San under the current engineering construction status; The normalized small area repair rate feature vector VsmallNorm is obtained by the following maximum value normalization processing formula: ； Wherein, VsmallNorm(t) represents the normalized small area repair rate feature vector in time period t, Vsmall(t) represents the small area repair rate feature vector in time period t, and max(Vsmall) represents the maximum small area repair rate feature vector in all time periods; The normalized high-frequency rework factor frequency characteristic vector VcauNorm is obtained by the following square normalization processing formula: ； Where VcauNorm(k) represents the normalized high-frequency rework factor frequency feature vector of rework reason category k, and Fca(k) represents the rework reason frequency vector of rework reason category k; Comprehensive formula to obtain: ； Wherein, s1 and s2 represent the weight coefficients of the normalized small area rework rate feature vector VsmallNorm and the normalized high frequency rework factor frequency feature vector VcauNorm, respectively, and s1+s2=1. The specific value is set by the user, and K represents the total number of types of rework reason codes RC.

8. The engineering intelligent management method according to claim 1, characterized in that: The S5 includes S51; S51, comparing the obtained comprehensive abnormality score San with the preset engineering status warning threshold Than, judging whether there is an abnormality in the current engineering construction status according to the comparison result, and generating an early warning signal Ealert according to the comparison result; The warning signal Ealert is obtained through the following judgment generation method: When the comprehensive abnormality score San ≥ the engineering status warning threshold Than, the current engineering construction status is judged to be abnormal, and the warning signal Ealert=1 is generated; When the comprehensive abnormality score San ≥ the engineering status warning threshold Than, it is determined that there is no abnormality in the current engineering construction status, and the warning signal Ealert=0 is generated.

9. The engineering intelligent management method according to claim 8, characterized in that: The S5 includes S52; S52. According to the obtained warning signal Ealert, when the warning signal Ealert=1 occurs, the dominant rework factor Fdom in the current engineering construction cycle is generated based on the high-frequency rework factor characteristic vector Vcau, and all the rework records that meet the generation of the dominant rework factor Fdom are integrated. Specifically, the rework number RID is extracted to obtain the records in the original engineering data set Draw, and a high-frequency rework factor analysis report RcauReport is generated, including the rework reason category k, the number of reworks and the dominant rework factor Fdom. Then, according to the high-frequency rework factor analysis report RcauReport, an optimization suggestion is generated by extracting the specification requirements for the repair reason category k in the preset engineering construction specification for the active rework factor category; The dominant rework factor Fdom is obtained by the following calculation formula: ； In the formula, argmax k (Fca(k)) represents the repair reason frequency vector of the repair reason category k for which the maximum value is found.

10. An engineering intelligent management system, applied to an engineering intelligent management method according to any one of claims 1 to 9, characterized in that: It includes engineering data acquisition module, engineering vector generation module, classification statistics module, comprehensive evaluation module and decision-making module; The engineering data collection module collects the completion data and repair data of each work surface in real time during the engineering construction process to form the engineering original data set Draw; The engineering vector generation module filters and processes the original engineering data set Draw, extracts small-area repair events within a fixed repair area, counts small-area repair events and the total amount of construction tasks per unit time, calculates the small-area repair rate Rsmall, and obtains the small-area repair rate feature vector Vsmall; The classification statistics module performs classification statistics based on the rework reasons in the original engineering data set Draw, calculates the occurrence frequency Fca of different types of rework reasons, and generates the frequency feature vector Vcau of high-frequency rework factors; The comprehensive evaluation module integrates the acquired small area repair rate feature vector Vsmall and high frequency rework factor frequency feature vector Vcau, builds a comprehensive anomaly score calculation model, and calculates the comprehensive anomaly score San under the current engineering construction status; The decision module compares the obtained comprehensive anomaly score San with the preset engineering status warning threshold Than to obtain the warning signal Ealert and high-frequency rework factor analysis report.

Images