A water conservancy construction progress prediction method based on big data

Abstract

The application relates to the technical field of progress prediction, in particular to a water conservancy construction progress prediction method based on big data, which comprises the following steps: constructing a time sequence curve based on field operation start and end time, attendance and equipment operation data synchronization, comparing progress and plan time distribution, extracting time sequence offset characteristics, corresponding completion proportion and target of subitems, judging node synchronization and offset, adjusting scheduling plan sequence, and outputting progress prediction control interval. The application realizes dynamic monitoring of operation progress and plan target through time sequence fusion of multi-source heterogeneous data, forms quantitative expression of progress state in combination with node offset identification, relies on a trend comparison mechanism to link prediction curves and field data changes, automatically generates signals for scheduling adjustment, optimizes task sequencing and node synchronization by using structured change information, enhances interval prediction capability by using segmented statistics, improves the pertinence and real-time response level of scheduling, and promotes the initiative and adaptability of construction progress prediction.

Description

Technical Field

[0001] This invention relates to the field of progress prediction technology, and in particular to a method for predicting the progress of water conservancy construction based on big data. Background Technology

[0002] Schedule forecasting primarily involves the methods and means of systematically managing and predicting the progress of operations in fields such as engineering, manufacturing, and logistics. It is widely used in various industry scenarios, including construction engineering, manufacturing, and information technology development. Traditional water conservancy construction schedule forecasting methods involve using static data such as historical project information, construction calendars, and personnel and equipment allocation plans, combined with the experience and judgment of project management personnel, to make preliminary estimates of construction progress using methods such as the critical path method, program review and approval technique, or linear schedule method.

[0003] Existing technologies for managing the progress of water conservancy construction rely on static data as the basis for progress judgment. The work logs and plan contents have problems such as long information update cycles and weak data timeliness. Progress status is usually added after the completion of nodes, which makes it impossible to capture fluctuations in on-site task execution and resource scheduling in real time. In actual construction, it is difficult to respond in a timely manner to the accumulation of progress deviations. The lack of linkage between static planning and on-site operations leads to phenomena such as delayed progress identification, slow response in scheduling, and delays in construction progress adjustment, which affect the continuity and response efficiency of the overall project management. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method for predicting the progress of water conservancy construction based on big data.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for predicting the progress of water conservancy construction based on big data, comprising the following steps:

[0006] S1: Based on the start and end times of on-site operations, attendance and equipment operation data, the progress data is synchronized according to a unified time base, and a time series curve is established according to the scheduling plan nodes. The time distribution between on-site progress and the planned target is analyzed to obtain the time series offset characteristics.

[0007] S2: Based on the time-series offset characteristics, compare the daily construction completion volume and the work report of the work team, check the completion ratio of each task against the planned tasks, determine the synchronization status of the current progress with the scheduling plan, and obtain the node offset identifier.

[0008] S3: Based on the node offset identifier, determine the progress change trend, analyze the historical prediction curve and the on-site multi-source data trajectory, fit the progress prediction curve and the change trend, and match the offset cause with the preset abnormal pattern library to obtain the task adjustment driving signal.

[0009] S4: Based on the task adjustment drive signal, adjust the scheduling plan order, reorder the sub-tasks, record the adjustment progress nodes in real time, compare the plan with the original order, determine the postponement nodes, and obtain the task order change information.

[0010] S5: Based on the task sequence change information, filter the task completion nodes, analyze the changes in the timing of the completion time and the adjusted planned nodes, and combine the progress time difference segmentation to segment and statistically analyze the progress fluctuation range to obtain the progress prediction and control interval.

[0011] The present invention is improved in that the time sequence offset features include project duration continuity, time sequence consistency and progress distribution type; the node offset identifier includes node code, progress offset category and association identifier; the task adjustment driving signal includes adjustment method, signal level and scheduling link; the task sequence change information includes adjustment node sequence, adjustment execution status and change impact range; and the progress prediction control interval includes interval number, control range and interval control status.

[0012] The present invention is improved in that the step of obtaining the time-series offset feature is specifically as follows:

[0013] S111: Based on the start and end times of on-site operations, attendance data, and equipment operating status, each data item is merged and time-aligned according to a unified time base, time series information is aggregated, and the continuous time sequence structure of each data type is sorted out to obtain a time series progress data group.

[0014] S112: Based on the time-series progress data set, compare the time series with the key progress nodes in the scheduling plan, classify and organize the data records of the associated segments of each node, and connect the node operation process with the node sequence to obtain the node time-series curve set.

[0015] S113: Based on the node time series curve set, determine the start and end intervals of the planned time and on-site progress of each node, and analyze the offset between nodes by combining the continuous operation status of each node and the node connection sequence to obtain the time series offset characteristics.

[0016] The present invention is improved in that the step of obtaining the node offset identifier is specifically as follows:

[0017] S211: Based on the time-series offset characteristics, combined with the on-site construction completion volume and the work report data of the work team, and according to the scheduling plan task objectives, the relationship between the work completion ratio and the planned task objectives is compared item by item, and the data is classified and organized through the corresponding relationship to obtain the work completion ratio analysis results.

[0018] S212: Based on the analysis results of the work completion ratio, determine the progress synchronization. By comparing the actual progress with the time nodes of the scheduling plan, analyze the lag and advance of the progress, evaluate the work status of each progress node on site, identify the deviation phenomenon, and obtain the progress synchronization judgment result.

[0019] S213: Based on the progress synchronization judgment result, according to the time interval and the connection between tasks, identify the time offset between each node, classify and label the offset type and node status, and combine the differences between on-site progress and planned tasks to obtain the node offset label.

[0020] The present invention is improved in that the step of obtaining the task adjustment drive signal is specifically as follows:

[0021] S311: Based on the node offset identifier, analyze the node code and progress offset category, match the scheduling plan with the on-site progress record, calculate the time sequence difference between the plan and the actual progress, identify the offset pattern, and obtain the node time sequence offset trend sequence.

[0022] S312: Based on the node time-series offset trend sequence, compare the offset changes with the actual progress, analyze the impact on progress based on equipment status and personnel attendance data, and calculate the matching situation between the actual and predicted progress curves using the following formula:

[0023] ;

[0024] The trend difference score was obtained, where, Refers to trend difference score, The predicted progress value refers to the predicted progress of the i-th node. This refers to the actual progress value, which is the actual progress of the i-th node. The fluctuation in equipment status refers to the fluctuation in the status of the j-th equipment. This refers to the intensity of staff attendance, specifically the intensity of attendance for the j-th staff member. This indicates the number of nodes participating in the matching. This indicates the total number of synchronous fluctuation data items;

[0025] S313: Based on the trend difference score, identify the abnormal nodes that have occurred, analyze their correlation with the current stage deviation phenomenon, and summarize the main related abnormal factors of the progress deviation by comparing equipment downtime and personnel attendance factors, and obtain the task adjustment driving signal.

[0026] The present invention is improved in that the step of obtaining the task sequence change information is specifically as follows:

[0027] S411: Based on the task adjustment driving signal, analyze the sub-tasks in the scheduling plan, optimize the original task list order according to task priority and resource allocation, and compare the start and end times and positions of each task before and after optimization to obtain the priority adjustment sorting sequence.

[0028] S412: Based on the priority adjustment sorting sequence, calculate the time change, task arrangement change and task distribution density between the original task sequence and the original task sequence, obtain the time offset measurement result, determine the time change of the task in the plan adjustment, and obtain the task time disturbance quantity group.

[0029] S413: Based on the task timing disturbance group, analyze the structural changes and state changes of each task node, determine the impact of task node reordering, optimize task execution state and resource association, and obtain task order change information.

[0030] The present invention is improved in that the step of obtaining the progress prediction and control interval is specifically as follows:

[0031] S511: Based on the task sequence change information, filter the task completion nodes, compare the actual completion time of each node with the adjusted planned node, and analyze the progress of each node through time sequence comparison to obtain a set of task completion nodes;

[0032] S512: Based on the set of task completion nodes, determine the time difference between the completion time of each node and the planned node, analyze the progress status of each node, classify and organize the offset of each time interval, and obtain the time difference segment set.

[0033] S513: Based on the time difference segment set, analyze the progress fluctuation of each segment, classify and label each fluctuating segment in combination with node offset information, determine the progress control requirements of each segment, and obtain the progress prediction control range.

[0034] The present invention is improved in that the unified time reference refers to using a standardized timeline as a reference for various data collection and progress node registration; the work team reporting refers to the daily on-site operation logs submitted by each construction team, showing the completed work volume, personnel attendance, machinery usage, and material consumption; and the progress change trend refers to judging the dynamic change direction of progress within a continuous time interval, such as increase, decrease, stabilization, acceleration, or deceleration.

[0035] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0036] In this invention, dynamic monitoring of work progress and planned objectives is achieved through time-series fusion of multi-source heterogeneous data. A quantitative expression of progress status is formed by combining node offset identifiers. Based on the trend comparison mechanism, the prediction curve is linked with changes in on-site data to automatically generate signals for scheduling adjustments. Structured change information is used to optimize task sorting and node synchronization. Segmented statistics are used to enhance interval prediction capabilities, improve the targeting and real-time response level of scheduling, and promote the initiative and adaptability of construction progress prediction. Attached Figure Description

[0037] Figure 1 This is a flowchart of the main steps of the present invention;

[0038] Figure 2 This is a flowchart illustrating the acquisition of temporal offset features in this invention;

[0039] Figure 3 This is a flowchart illustrating the process of obtaining the node offset identifier in this invention;

[0040] Figure 4 This is a flowchart illustrating the acquisition of the task adjustment drive signal in this invention.

[0041] Figure 5 This is a flowchart illustrating the process of obtaining task sequence change information in this invention.

[0042] Figure 6 This is a flowchart of the process for obtaining the progress prediction and control interval in this invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0044] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0045] All user-related information involved in this invention (including but not limited to biometric information, identity verification information, behavioral data, device information, and other data that can be used for identity verification and personalized services) is collected and processed with the user's full knowledge and voluntary consent. The collection, storage, and use of all information strictly comply with applicable national and regional laws and regulations, and meet relevant data protection standards and policy requirements. The use of data is limited to purposes necessary for providing the technical services of this invention, and reasonable technical and management measures will be taken to ensure the security and confidentiality of users' personal information in terms of information protection and privacy.

[0046] For examples, please refer to Figure 1 This invention provides a technical solution: a method for predicting the progress of water conservancy construction based on big data, comprising the following steps:

[0047] S1: Based on the start and end times of on-site operations, attendance data, and equipment operating status, multiple progress data are synchronized according to a unified time base. A time series curve is constructed according to the schedule plan progress nodes. By comparing the time distribution between on-site progress and the planned target, the time series offset characteristics are obtained.

[0048] S2: Based on the time-series offset characteristics, compare the daily construction completion volume, check the work report status of the work team, divide the work completion ratio into items according to the planned task objectives, judge the synchronization between the actual progress and the scheduling plan, focus on the offset difference of each progress point, and obtain the node offset identifier.

[0049] S3: Based on the node offset identifier, determine the progress change trend, analyze the change process of historical progress prediction curve and on-site multi-source data, compare the fitting degree of the prediction curve and the actual curve with the trend correspondence, and combine the offset phenomenon of the current stage to summarize the reasons and obtain the task adjustment driving signal.

[0050] S4: Based on the task adjustment driving signal, adjust the scheduling plan under the feedback, reorder all sub-tasks in sequence, adjust the priority parameters, record the corresponding progress nodes in real time, compare the planned order with the original task order in time sequence, determine the postponement nodes, and obtain the task order change information.

[0051] S5: Based on task sequence change information, filter task completion nodes, analyze the time sequence changes between completion time and adjusted planned nodes, combine construction progress time difference segmentation, perform segmented statistics on progress fluctuation range, and obtain progress prediction and control interval.

[0052] The time-series offset characteristics include project duration continuity, time sequence consistency, and progress distribution type; the node offset identifier includes node code, schedule offset category, and association identifier; the task adjustment driving signal includes adjustment method, signal level, and scheduling link; the task sequence change information includes adjusted node sequence, adjusted execution status, and change impact range; and the schedule prediction control interval includes interval number, control range, and interval control status.

[0053] In S1, a unified time base refers to using a standardized timeline (such as using Beijing time, project schedule calendar, etc.) as a reference for various data collection and progress node registration, so that all work, progress, equipment status and other information can be aligned on the same time axis; multiple progress data refers to various progress information from different sources, mainly including on-site work start and end times, personnel attendance records, mechanical equipment operating status, on-site logs, etc., all structured data that affect project progress; scheduling plan progress nodes refer to clearly defined project nodes such as key milestones, task start and end points, and phased acceptance points set in the project management system or scheduling plan table, used to compare with actual progress; time series curve refers to the dynamic progress change curve formed by arranging multiple progress data and scheduling plan nodes in chronological order, used to intuitively reflect the temporal relationship between the plan and actual progress; time distribution refers to the distribution characteristics of various progress data throughout the construction period, that is, the time segments of the actual on-site task start, end and duration, which are compared with the plan requirements to quantify the deviation.

[0054] In S2, the work report of each work team refers to the daily on-site operation logs submitted by each construction team, including the actual completed work volume, personnel attendance, machinery usage, and material consumption, which serve as important raw data for actual progress. The planned task target refers to the specific work targets and outputs that need to be completed for each time period and each task unit in the scheduling plan, such as how many cubic meters of concrete should be poured on a certain day. The item correspondence refers to the comparison between the actual work completion ratio and the planned task by sub-item (such as by project sub-item, process, work team, etc.) to ensure that each item is linked to the planned target. The synchronization status refers to the analysis of whether the actual progress and the planned progress are on the same rhythm at each item (time point / process), identifying whether they are lagging, ahead, or synchronized. The deviation difference refers to the progress deviation between the actual and the planned at a certain node or interval found in the synchronization analysis, which is quantified as ahead, behind, or consistent.

[0055] In S3, the progress change trend refers to a comprehensive judgment on the dynamic change direction of the actual progress within a continuous time interval, such as increase, decrease, stabilization, acceleration, or deceleration, distinguishing between normal progress and abnormal fluctuations; the change process refers to the change trajectory of historical and current multi-source progress data in the time series, including stage-specific abrupt changes and periodic fluctuations, which are reflected as curve trends; the degree of fit refers to the comparison and analysis of the predicted progress curve and the actual data curve using mathematical statistics or machine learning methods, and the degree of fit between the two is used to judge the accuracy of the prediction; the trend correspondence refers to the comparison of the consistency between the actual and predicted curves in terms of trends (growth, decline, plateau, etc.), used to discover the degree of conformity or deviation between the prediction and the actual operating rules; the deviation phenomenon in the current stage refers to the deviation between the actual progress and the prediction (or plan) at each node within the current scheduling cycle, and summarizing the specific reasons for its occurrence (such as equipment failure, weather impact, etc.).

[0056] In S4, the feedback-based scheduling plan refers to the latest scheduling plan after dynamic correction and optimization based on the prediction and monitoring results of the previous stage; sub-tasks refer to the specific work units broken down in the entire project, such as concrete pouring, rebar tying, and pipeline laying, each of which is scheduled independently; priority parameters refer to parameters used to distinguish the urgency or importance of each task, such as prioritizing urgent tasks and using sorting algorithms to adjust the execution order; schedule nodes refer to the specific start, end, intermediate checks, and other key progress time points of each task, used for project management; time sequence comparison refers to comparing the latest adjusted plan nodes with the original nodes one by one on the timeline to identify which tasks have been advanced, postponed, or remained unchanged; postponed nodes refer to tasks or schedule nodes that have been delayed due to plan adjustments.

[0057] In S5, the task completion node refers to the actual time point at which the task is completed on the construction site, which can be confirmed through various channels such as on-site work reports, equipment data, and supervisor acceptance; the adjusted planned node refers to the planned progress node after adjustment and optimization in the S4 stage, which is the benchmark for comparison with the actual task completion node; time difference segmentation refers to dividing the time interval between the actual completion node and the planned node in the time series, which facilitates the statistics of progress fluctuations and the tracing of causes; the progress fluctuation range refers to the specific intervals of progress being ahead of schedule, delayed, or normal by analyzing the time difference segmentation, and clarifying the magnitude and distribution of progress status changes at different stages.

[0058] Please see Figure 2 The specific steps for obtaining the temporal offset features are as follows:

[0059] S111: Based on the start and end times of on-site operations, attendance data, and equipment operating status, each data item is merged and time-aligned according to a unified time base, time series information is aggregated, and the continuous time sequence structure of each data type is sorted out to obtain a time series progress data group.

[0060] Using the start and end times of on-site operations, personnel attendance data, and equipment operating status as the raw data sources, the time field is first extracted from each of these three types of data and uniformly converted into the standard Beijing time format. Simultaneously, garbled characters, missing items, and format errors are removed from the data. For example, the format "2025 / 5 / 12 8:30" in a certain equipment startup record is converted to "2025-05-12, 08:30:00". Next, based on the job number or section number as the aggregation basis, multiple job records under the same number are sorted by time and adjacent data segments are merged. For example, a section of rebar tying task records with two segments from 08:00 to 12:00 and from 14:00 to 18:00 on May 10th should be merged into a continuous time period from 08:00 to 18:00 on May 10th. Finally, based on the employee number in the attendance data, multiple clock-in records of the same employee within a day are merged into one group. For example, if someone starts work at 08:02, 1 If someone leaves their post at 2:01, returns at 13:00, and finishes get off work at 18:03, their effective attendance for the day is calculated to be 9 hours. Then, the equipment operation status records are organized, merging multiple start-stop records for the same equipment number to generate daily operation time intervals, and identifying breakpoint areas. If there are records with more than 4 hours of no startup, it is considered an operational interruption. Subsequently, the three types of data are unified into a time format and mapped onto a continuous time axis in hours. Data is sliced ​​according to the actual records, and each hour is labeled with whether there is work, whether equipment is running, and whether personnel are present. This is summarized to form three types of time series segments for each day. For example, from 08:00 to 18:00 on May 10, 2025, the rebar binding operation segment is continuous, personnel attendance is continuous, and the equipment operation segment is interrupted for 1 hour from 11:30 to 12:30. The time periods are labeled and organized by day into structured data, forming a time-series progress data group.

[0061] S112: Based on the time-series progress data group, compare the key progress nodes in the time series and the scheduling plan, classify and organize the data records of the associated segments of each node, and connect the node operation process with the node sequence to obtain the node time-series curve set;

[0062] The start and end times of key progress nodes are extracted one by one from the scheduling plan, and each node is assigned a number and its planned duration in days. Then, in the generated time-series progress data group, attendance, work, and equipment data records for the corresponding segments are extracted according to time matching. For example, if the scheduling plan sets the "underground water pipeline laying" task to May 11th to May 20th, 2025, with the corresponding number G01, then all data entries within the time-series data group from May 11th to May 20th are filtered and categorized by data type. All equipment start / stop records within this segment are grouped into "Equipment Data - G01," and work records within this segment tagged with "underground water pipeline laying" are grouped into "Work Data - G01." The attendance records of team members tagged with "underground water pipeline laying" are grouped into "Work Data - G01." The records of the construction team performing the task are categorized into "Attendance Data - G01". The same operation is then performed on the next node G02 after node G01, generating multiple sets of node data segments with corresponding numbers. Next, according to the node order set in the scheduling plan, each node is sorted from early to late according to the planned time, and the consistency of its corresponding data segment order is verified. That is, it is confirmed whether G01 is followed by G02, and it is determined whether there is any overlap, intersecting or gap time between the two data segments. For example, if the end time of G01 is 18:00 on May 20, and the start time of G02 is 08:00 on May 21, then the two are marked as normal sequential continuation, and the node order set from G01 to Gn is constructed in sequence. The data segment sequence associated with each node is numbered, organized and stored in the node time series curve set.

[0063] S113: Based on the node time series curve set, determine the start and end intervals of the planned time and on-site progress of each node, and combine the continuous operation status of each node and the node connection sequence to analyze the offset between nodes and obtain the time series offset characteristics.

[0064] Extract the planned start and end times from the time-series curves of each node and compare them with the start and end times of the actual data. Compare these times sequentially by node number to identify the offset in days or hours. For example, if node G01 is planned for May 11th to May 20th, but actual work started on May 13th and ended on May 21st, the offset start time is +2 days, and the offset end time is +1 day. Next, analyze whether the work process within that node is continuous, i.e., determine if there are any work interruptions at the hourly level. If work data is found to exist on a certain day, such as May 15th, from 08:00 to 10:00 and from 13:00 to 15:00, but from 10:00 to 13:00 is empty, then it is recorded as 3 hours of work on that day. If an interruption occurs, the end and start time differences between the current node and the next node are compared to determine if the nodes are tightly connected. For example, if G01 ends at 18:00 on May 21 and G02 starts at 08:00 on May 22, then the connection is tight. If there is a gap of more than 2 days, it is marked as an abnormal node connection. The offset data is then quantitatively evaluated, with ±1 day set as the allowable error range. If the offset exceeds this range, it is considered a significant offset. The offset amount, number of job interruptions, connection status, and other information are summarized to form the offset characteristics of the current node. For example, if node G01 records an offset start + 2 days, end + 1 day, 2 job interruptions, and normal connection, it constitutes a complete set of time-series offset characteristics.

[0065] Please see Figure 3 The specific steps for obtaining the node offset identifier are as follows:

[0066] S211: Based on the time series offset characteristics, combined with the on-site construction completion volume and the work report data of the work team, and according to the scheduling plan task objectives, the relationship between the work completion ratio and the planned task objectives is compared item by item, and the data is classified and organized through the corresponding relationship to obtain the work completion ratio analysis results.

[0067] Combining on-site construction completion data and work report data from work teams, the offset between the actual start and end times and the planned time for each construction task is extracted. The start and end dates on-site are recorded by task number. Then, based on the work report forms, the actual completed work volume, participating work teams, and number of man-days for each construction sub-item are extracted. The task number and time period corresponding to each record are verified to accurately correspond to the planned node. Subsequently, the scheduling plan task objectives are broken down to obtain the planned completed work volume for each construction day. If the scheduling plan sets a concrete pouring task from June 1st to June 5th, 2025, with a planned total of 500 cubic meters, it is broken down into 100 cubic meters per day based on a 5-day average task objective. This planned volume is compared with the actual completed volume of the task in the work report data. If the on-site work is completed on June 2nd... If the volume reaches 85 cubic meters, the completion rate for that day is 85%. Then, the completion rate of each task item is compared with the planned rate for the corresponding day to determine whether each work record is ahead of schedule, behind schedule, or in sync. Specifically, a deviation threshold of ±5% is used. If the completion rate for the day is between 95% and 105%, it is marked as sync; below 95%, it is marked as behind schedule; and above 105%, it is ahead of schedule. The above judgment results are recorded in the subsequent fields of each work record. Then, all records are classified and organized according to the work number to form the work completion trend of each construction task in different time periods. For example, if the rebar binding task numbered J203 is below 90% from June 1 to June 3, and exceeds 110% from June 4 to June 5, it is recorded as a type of early lag followed by late catch-up, forming the work completion rate analysis result for this task.

[0068] S212: Based on the analysis results of the work completion ratio, determine the progress synchronization. By comparing the actual progress with the time nodes of the scheduling plan, analyze the lag and advance of the progress, evaluate the work status of each progress node on site, identify deviation phenomena, and obtain the progress synchronization judgment result.

[0069] Extract the daily completion percentage of each task within the planned time and the actual operation time on the on-site timeline, and compare it with the key time nodes set in the scheduling plan. For example, if task J203 is planned from June 1st to June 5th, 2025, but the actual operation is from June 2nd to June 6th, then it is lagging by one day. Next, compare the daily completion percentage of task J203, calculate the difference between the daily completion percentage and the planned percentage, and then judge the continuous trend of the daily difference. If the completion percentage is lower than the planned value by more than 10% for two consecutive days, then the task is marked as lagging at the corresponding node. Conversely, if the completion percentage is higher than the planned completion percentage by more than 10% for two consecutive days, then it is in an advanced operation state. Subsequently, analysis... The difference between the cumulative completed amount and the planned cumulative completed amount at each node is used to determine whether the node is a lagging node if the completed amount is less than 95% of the total planned amount by the end of the planned period, a leading node if it exceeds 105%, and a synchronous node otherwise. The construction status records for that node are then checked for interruptions. For example, if there is no operational data in the equipment records on a certain day, and the attendance record shows only one person with no valid work records, that day is marked as an interrupted operation. The actual work status of each node is then categorized into three types: continuous, intermittent, and idle. Combined with the aforementioned leading and lagging classifications, each progress node is assigned a status label of "synchronous," "leading," "lagging," or "abnormal," thus forming a progress synchronization judgment result.

[0070] S213: Based on the progress synchronization judgment results, and according to the time interval and the connection between tasks, identify the time offset between each node, classify and label the offset type and node status, and combine the differences between on-site progress and planned tasks to obtain the node offset label;

[0071] By comparing the planned timeline with the on-site recorded timeline based on the start and end times of the construction tasks, the first step is to identify the number of days between the actual start and the planned start of each task node. Then, the second step is to identify the number of days between the actual end and the planned end. Subtracting these two offset values ​​yields the offset trend value. If the start offset is +1 day and the end offset is +3 days, it indicates a persistent lag, and an offset threshold range of ±1 day is set. Nodes exceeding this range need to be categorized and flagged. Next, the time continuity between nodes is checked. For example, if node N1 ends at 18:00 on June 10th and node N2 starts at 08:00 on June 12th, there is an over-lagging timeline. If there is a 1-hour gap, the two nodes are considered discontinuous and marked as "broken chain". If the difference between consecutive nodes does not exceed 8 hours, they are marked as "continuous". Based on the above offset types and the progress synchronization results, the offset reasons are classified by type. For example, if there is a work interruption of more than 3 days in the lagging node, it is marked as "lagging due to on-site factors". If it is completed ahead of schedule but the attendance data is dense and the equipment utilization rate increases, it is marked as "early due to concentrated resources". Combining the original node number, corresponding status and offset reason, the node offset identifier is output. For example, the identifier for number T302 is: lagging type, broken chain status, on-site impact type, forming a node offset identifier.

[0072] Please see Figure 4 The specific steps for obtaining the task adjustment drive signal are as follows:

[0073] S311: Based on node offset identifiers, analyze node codes and schedule offset categories, match scheduling plans with on-site progress records, calculate the time sequence difference between the plan and the actual progress, identify offset patterns, and obtain node time sequence offset trend sequences.

[0074] Extract the coding information and corresponding schedule offset category of each node. By traversing the node offset identifier table line by line, read the node number, start and end time, offset direction, and offset days for each node. For example, the pipeline foundation construction node with node number D105 is marked as a delayed type with an offset of +3 days. Its planned time period is from July 1 to July 6, 2025, and the actual work period is from July 4 to July 9. Then, match the node code with the scheduling plan record. Compare the time of node D105 in the scheduling plan table with the corresponding node work time in the field data. Calculate the difference between the planned start time and the actual start time of this node, which is +3 days. The difference between the planned end time and the actual end time is also +3 days, and it is recorded as a continuous delayed type. Then, add the offset value of this node to the schedule difference sequence and perform time-series difference calculations along the node sequence. The process involves statistical analysis, followed by plotting the time sequence of offset values ​​for multiple nodes. By comparing the changes in offset direction between adjacent nodes, it is possible to identify whether offset propagation exists. For example, if D105 offsets by +3 days, and subsequent nodes D106 offset by +2 days and D107 offset by +4 days, the offset is considered to have positive continuity and is classified as a progressive offset trend. If the offset value of a node changes abruptly, such as D108 offsetting by -2 days while the nodes before and after it are all positive offsets, then this node is recorded as a trend abrupt change point. The average offset value of all nodes is further calculated, and a trend threshold of ±2 days is set based on this average value. Nodes exceeding this threshold are marked as trend abnormal nodes. Through the above processing, the offset value, change trend, offset propagation relationship, and abrupt change of each node are uniformly recorded in the trend sequence table, and the node time-series offset trend sequence is output.

[0075] S312: Based on the node time-series offset trend sequence, compare the offset changes with the actual progress, analyze the impact on progress based on equipment status and personnel attendance data, and calculate the matching between the actual and predicted progress curves using the following formula:

[0076] ;

[0077] The trend difference score was obtained, where, The trend difference score indicates the degree of match between predicted and actual progress, and is used to measure the difference between the prediction and actual progress. This refers to the predicted progress value, which is the predicted progress for the i-th node, expressed in time (e.g., days, hours). It represents the progress that should be completed at this node according to the project plan. This refers to the actual progress value, which is the actual progress of the i-th node, expressed in time (e.g., days, hours). It represents the actual progress completed at that node during the construction process. This refers to the fluctuation of equipment status, specifically the status fluctuation of the j-th piece of equipment. It is a dimensionless value that reflects the dynamic fluctuations of the equipment during construction, such as downtime or operating status. This refers to the intensity of personnel attendance, specifically the intensity of the j-th personnel attendance. It is a dimensionless value that reflects the actual attendance of construction workers, such as the combination of the number of workers present and their working hours. This indicates the number of nodes participating in the matching. This indicates the total number of synchronous fluctuation data items;

[0078] Trend discrepancy scoring is a quantitative indicator that reflects the degree of matching between schedule forecasts and actual execution, taking into account the impact of equipment and personnel status. Through scoring, the accuracy of construction schedule forecasts and the synchronicity of actual progress can be assessed, providing a basis for subsequent schedule adjustments and decision-making. If A smaller discrepancy between the predicted and actual progress indicates a smaller difference in progress, suggesting a more accurate progress forecast and better execution of the construction plan; if... A large discrepancy indicates a significant difference between the forecast and the actual progress, requiring adjustments to the plan or responses to influencing factors (such as equipment failure or personnel issues).

[0079] Based on the node time-series offset trend sequence, the offset changes are further compared with the actual progress. The impact on progress is analyzed based on equipment status and personnel attendance data. First, by comparing the time-series offset of each node with the changing trend of the actual progress, combined with equipment status fluctuations (such as downtime or changes in equipment operating efficiency) and personnel attendance information (such as absences and overtime), the analysis is conducted. The degree of matching between the actual and predicted progress curves is obtained by weighting various influencing factors. In the specific execution process, equipment operating status (such as equipment failure time and operating status) and personnel attendance data (such as number of attendees and working hours) are first collected from the field. A trend difference score is calculated using a formula. The following raw values ​​are assumed:

[0080] Node A: Forecast Progress Days, actual progress sky;

[0081] Node B: Forecast Progress Days, actual progress sky;

[0082] Node C: Forecast Progress Days, actual progress sky;

[0083] Equipment status fluctuation (hours):

[0084] Device A: Hour;

[0085] Device B: Hour;

[0086] Equipment C: Hour;

[0087] Staff attendance intensity (percentage):

[0088] Person A: ;

[0089] Personnel B: ;

[0090] Personnel C:

[0091] Using the minimum-maximum normalization method:

[0092] Equipment status fluctuations (hourly) normalization:

[0093] ;

[0094] ;

[0095] ;

[0096] Normalized staff attendance intensity (percentage):

[0097] ;

[0098] ;

[0099] ;

[0100] Calculate the trend difference score based on the normalized data. :

[0101] First item: Difference between forecast and actual progress (sum of absolute values):

[0102] ;

[0103] Weighted sum of equipment status and personnel attendance (square root):

[0104] ;

[0105] ;

[0106] Overall Score:

[0107] ;

[0108] Define the following ranges to classify schedule deviations:

[0109] This range indicates that the difference between the predicted progress and the actual progress is small, and the progress synchronization is good. Under normal circumstances, the predicted progress is very close to the actual progress, the deviation is within an acceptable range, and the project is progressing normally.

[0110] This range indicates a small to moderate deviation in progress. Although there is some deviation, it is still within a reasonable range. At this time, although there is a delay or advance in the progress, the impact is small, and the project can still proceed as planned. Slight adjustments to some aspects are required.

[0111] This interval indicates a significant deviation from the schedule, with a large discrepancy between the schedule and the plan. The impact of equipment and personnel factors cannot be ignored. In this case, the project needs to be reassessed and adjusted. Adjustment strategies include accelerating the execution of certain tasks or adjusting resource allocation.

[0112] This interval indicates a large schedule deviation, a significant gap between the schedule and the scheduling plan, requiring a large-scale adjustment to the entire project schedule. The project is in a delayed state, and there are significant problems with the operating efficiency of equipment or personnel. It is urgent to optimize the scheduling plan, increase resources, or adjust task priorities.

[0113] Based on the above interval division, In Within this range, it indicates a significant schedule deviation, with a large difference between the schedule and the plan. Equipment and personnel factors also have a certain impact on the schedule, suggesting that the project is either behind schedule or ahead of schedule. It is necessary to analyze and optimize the relevant aspects, and take measures such as accelerating construction or readjusting resource allocation to control the schedule deviation and avoid further impacting the overall project schedule.

[0114] S313: Based on trend difference scoring, identify abnormal nodes, analyze their correlation with the current stage deviation, and summarize the main related abnormal factors of progress deviation by comparing equipment downtime and personnel attendance factors to obtain task adjustment driving signals.

[0115] Extract the node numbers and offset values ​​marked as trend abrupt changes from the preceding segment, and record their positions in the trend sequence. Then, analyze the node offset data to confirm whether it meets the anomaly judgment criteria. For example, if a node's offset value is opposite to the offset direction of its preceding and following nodes and the difference exceeds 3 days, it is considered an abnormal node. For instance, if node T209 has an offset of -2 days, while T208 has +4 days and T210 has +3 days, then T209 is marked as abnormal. Next, extract the construction stage information of the node, such as the cable trench backfilling stage. Further extract the equipment operation records and personnel attendance records for the corresponding date from the field data, and check the changes in the total daily equipment start-up and shutdown time and the number of attendees during this stage. Compare the total equipment operating time of the day containing the abnormal node with the previous and following days. For example, if the equipment operated for 10 hours the previous day... If the number of employees on a given day is only 3 hours, it is recorded as an equipment malfunction. Then, it is determined whether there is a significant difference between the number of employees on that day and the average number on the previous and following days. For example, if the number of employees on the previous two days was 12, but only 5 were present on that day, it is determined that there is an abnormal fluctuation in attendance. If the change in either equipment or attendance exceeds a set threshold of 20%, it is recorded as a "resource anomaly" type of offset. If there is no significant change in either equipment or attendance, the material supply data and external environment records for that day are further checked. If the records show that there was a rainstorm and construction was suspended on that day, it is marked as a "weather impact" type of offset. All the summarized results are established according to the node number to establish a corresponding relationship, and the task adjustment driving signal corresponding to each node offset is output. The content includes: trigger node number, adjustment reason classification, resource impact identification field, whether to suggest changing the schedule, etc., to obtain the task adjustment driving signal.

[0116] Please see Figure 5 The specific steps for obtaining task order change information are as follows:

[0117] S411: Based on the task adjustment driving signal, analyze the sub-tasks in the scheduling plan, optimize the original task list order according to task priority and resource allocation, and compare the start and end times and positions of each task before and after optimization to obtain the priority adjustment sorting sequence.

[0118] Extract the task number, corresponding offset type, scheduling adjustment suggestion, and construction stage information. Then, obtain a list of all sub-tasks according to the scheduling plan in the construction project. Extract the planned start time, planned end time, associated resource type, and current status record for each task. Then, filter by tasks marked as "suggested adjustment" in the task adjustment drive signal, mark the tasks as tasks to be adjusted, and extract their original task order index. For example, if the task number T308 was originally ranked 12th, and the adjustment suggestion is "advance", then its original order is recorded as 12, and the target order is the earlier area. Then, perform priority scoring processing on all tasks to be adjusted. The priority score consists of three factors: task urgency level, resource availability score, and construction dependency index. The task urgency level is manually set by the project manager as an integer between 1 and 5, with smaller values ​​indicating higher priority. The resource availability score is determined by the current adjustment... The resource idle rate is calculated within the cycle. Resources with an idle rate greater than 80% are assigned a score of 3, those between 50% and 80% are assigned a score of 2, and those less than 50% are assigned a score of 1. The construction dependency index is calculated based on the number of prerequisite dependencies for the task. A dependency of 0 is assigned a score of 3, 1-2 dependencies are assigned a score of 2, and 3 or more dependencies are assigned a score of 1. The scores of the three items are added together to obtain the total score, which ranges from 3 to 11. Tasks are arranged from highest to lowest total score to generate a new task sorting list. Then, the start time and task position number before and after the adjustment are compared, and the difference between the original order and the adjusted order is recorded. For example, if T308 is adjusted from the 12th position to the 5th position, the start time is moved from July 20, 2025 to July 15, and the end time is moved from July 27 to July 22, the task sorting span is 7 positions, and the time adjustment is 5 days earlier. All the order change information of the adjusted tasks is recorded as a sorting comparison table, and the priority adjustment sorting sequence is output.

[0119] S412: Based on priority adjustment of the sorted sequence, calculate the time change, task arrangement change, and task distribution density compared to the original task sequence, using the following formula:

[0120] ;

[0121] Obtain the time-series offset measurement results, determine the time-series changes of the task during plan adjustments, and obtain the task time-series disturbance quantity set, where... Indicates task The temporal offset metric results are used to reflect the intensity of structural changes in the task after reordering. Indicates task The time parameter offset difference, i.e., the amount of change in the task's schedule before and after the adjustment. Indicates task The local task density within a given time period, i.e., the distribution of the number of relevant task nodes within that time period. Indicates task The difference in position in task sorting, that is, the magnitude of the change in the order of the task in the sequence before and after the adjustment;

[0122] The time offset metric reflects the intensity of dynamic disturbances experienced by each task at both the temporal and structural levels due to schedule adjustments. It is a quantitative result for measuring the scope of impact and key milestones of construction schedule changes. If a task's time arrangement changes significantly, or its ranking position changes noticeably, and the local task density is low, then the time offset metric result for that task will be... The larger the value, the more affected the task is by the adjustment; if the time and order changes are small, or if there are many tasks around it, the result value will be small, indicating that the task is relatively stable and less affected by the adjustment.

[0123] Task A, with task code A, was invoked. Its original start time was 09:00, and its adjusted start time was 09:45. The time offset was recorded. Meanwhile, the maximum time offset of all tasks is extracted to be 60 minutes. After processing using the maximum value normalization method, we have: Then, the number of concurrent tasks in the time segment where this task is located is counted as 4, and the maximum number of concurrent tasks is 6. The normalization method is the same as before, resulting in: Next, we analyze that the index of this task in the original sorting is 2, and the index after optimization is 4. Therefore, the change in its sorting structure is as follows: This parameter is a dimensionless discrete quantity and does not participate in the normalization process. Substituting the above data into the formula:

[0124] ;

[0125] Temporal offset measurement results of the task Based on the preset disturbance intensity range, they are divided into the following three categories:

[0126] when When the time is determined to be a low-disturbance section, it means that the task adjustment is small and the sorting structure and time arrangement remain stable;

[0127] when When this is the case, it is determined to be a medium-disturbance segment, indicating that the task has a certain degree of rearrangement and offset in time or structure;

[0128] when When a high-disturbance section is identified, it indicates drastic task adjustments, requiring a thorough assessment of its impact on scheduling rhythm and resource distribution.

[0129] The calculation result of task A is ,satisfy The interval conditions are thus classified as medium disturbance segment. This result indicates that the task has significant changes in both structural ordering and time adjustment due to scheduling plan optimization, reaching the quantitative triggering standard for task disturbance classification. As a member task in the task time-series disturbance group obtained in the current step, its disturbance intensity will be used in subsequent steps to determine whether the task status has changed or whether resource conflicts have been triggered.

[0130] S413: Based on the task timing disturbance group, analyze the structural changes and state changes of each task node, determine the impact of task nodes after sorting adjustment, optimize task execution state and resource association, and obtain task order change information.

[0131] The difference between the change in start and end times of each task after the reordering adjustment and the original planned time is calculated to obtain the initial and final disturbance values ​​for each task. This disturbance value is used as the core parameter to determine the degree of impact on the task. Next, the dependency structure between tasks is analyzed, that is, the predecessor and successor task numbers of each task are read to form a task structure network diagram. It is determined whether there are situations where predecessor tasks are postponed after the reordering adjustment but subsequent tasks are not adjusted synchronously. If such order reversal exists, it is marked as a structural breakpoint. For example, if task T403 originally depended on T402 and T402 was completed on July 18th, and after the adjustment, T403 is moved to July 16th, then T403 is marked as a "reverse-order task". Then, the task status changes are analyzed, checking whether the construction status field of each task has changed after the adjustment, such as task T3. If the original status is "Not Started", and the adjusted start date is earlier than the current date, and there are already construction records, then the status is changed to "In Progress". Then, the resource configuration fields of each task are compared and analyzed to extract the types of equipment and personnel resources required for the adjusted task. The idle equipment table and empty personnel table in the current scheduling cycle are compared to determine whether there are resource conflicts. For example, if task T405 needs to use 2 pump trucks on July 20, but both pump trucks are occupied by T406 on that day, then it is marked as a resource conflict task, and the start date of T405 is postponed to July 21 for reconfirmation of status. After completing the above analysis, the structural position, start and end time, status field, and resource field before and after the task adjustment are compared to output the sequence change number, disturbance days, resource adaptation status, and changed structure identifier of each task, generating task sequence change information.

[0132] Please see Figure 6 The specific steps for obtaining the progress prediction control range are as follows:

[0133] S511: Based on task sequence change information, filter task completion nodes, compare the actual completion time of each node with the adjusted planned node, and analyze the progress of each node through time sequence comparison to obtain a set of task completion nodes;

[0134] Extract each task adjustment record, read the start and end times of the change, task number, reason for change, and original planned time for each task node. Filter the actual completed data in the construction progress record one by one. Match the completed task nodes with their task numbers and timestamps to the corresponding planned node numbers and adjusted time periods. Compare the actual completion time with the adjusted planned time item by item. For example, if the adjusted planned completion time for task M205 is August 18, 2025, but the on-site work report shows an actual completion time of August 20, the offset is calculated to be +2 days, and it is recorded as a delayed completion node. If the adjusted planned time for node M206 is August 25, and the actual completion time is August 24... If the actual completion time is before the adjusted planned time, the offset is -1 day, and it is recorded as an early completion node. Further, extract all completed tasks and nodes whose actual completion time is before or after the adjusted planned time. Generate a triplet of task number, planned completion time, actual completion time, and time difference for each node. Sort the nodes by node number to create a node comparison list. Arrange all completed nodes in the adjusted order and check if there are any discrepancies between the actual completion order and the adjusted order. If task M207 is scheduled after M208, but its actual completion time is two days earlier than M208, it is recorded as an execution order anomaly. Integrate the fields such as actual completion time, adjusted time, offset, and execution order to output a set of job completion nodes.

[0135] S512: Based on the set of task completion nodes, determine the time difference between the completion time of each node and the planned node, analyze the progress status of each node, classify and organize the offset of each time interval, and obtain the time difference segment set.

[0136] Extract the planned time and actual completion time for each node sequentially by node number, calculate the absolute difference between the two, and record the direction of the difference as "ahead of schedule" or "behind schedule". Set the time difference judgment threshold to ±1 day. If the difference is within this range, it is marked as "synchronous". Classify and statistically analyze the time difference type corresponding to each node, and group consecutive nodes of the same type into the same time interval. For example, if nodes M201 to M204 are all in a delayed state, with +2, +1, +3, and +2 days respectively, they are classified into the delayed interval T1, with the interval spanning from the planned completion time of M201 to the actual completion time of M204. Once the actual completion time ends, similarly, if M205 to M207 are in an advanced state, with -1, -2, and -1 days respectively, they are classified as the advanced interval T2. Then, according to the time axis order, the difference segments are divided, and the start node, end node, start time, end time, and offset type of each segment are recorded. The average offset days of each segment are further calculated and a judgment label is added. If the absolute value of the average offset is ≥3 days, it is marked as "severe offset", 1~2 days is "moderate offset", and less than 1 day is "minor offset". After distinguishing the levels, the difference statistics of each time offset segment are output to form a time difference segment set.

[0137] S513: Based on the time difference segment set, analyze the progress fluctuation of each segment, classify and label each fluctuating segment in combination with node offset information, determine the progress control requirements of each segment, and obtain the progress prediction control range.

[0138] The start time, end time, offset type, and offset level of each segment are extracted. First, the number of nodes in each segment is counted; for example, segment T1 contains 6 lagging nodes, and segment T2 contains 4 leading nodes. The consistency percentage of node offset directions within each segment is calculated. If the consistency percentage is ≥80%, the segment's volatility type is considered "unidirectional stable"; otherwise, it is "multidirectional mixed." Next, the node offset value sequence within each segment is sorted, and the maximum and minimum offset values ​​are extracted. If the difference between the maximum and minimum values ​​exceeds 5 days, the segment is recorded as a "high volatility zone"; otherwise, it is a "low volatility zone." This is then combined with... The offset reasons recorded in the previous stage node offset identifier are classified and statistically analyzed for the node reason fields in each segment. If more than 80% of the node offset reason fields in a certain segment are "equipment problem", then the segment is marked as "equipment-affected fluctuation zone". Then, the control level is divided into each segment, and the control level is set into three categories: "strong intervention", "normal scheduling" and "no adjustment required", which correspond to the segments with offset levels of "severe offset", "moderate offset" and "minor offset" respectively. The number, start and end time, offset type, fluctuation status, cause category and control level of each segment are integrated and output to form the progress prediction control interval.

[0139] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A big data-based water conservancy construction progress prediction method, characterized in that, The method comprises the following steps: S1: synchronizing progress data according to a unified time reference based on field operation start and end time, attendance and equipment operation data, establishing a time sequence curve according to a scheduling plan node, analyzing the time distribution between field progress and plan targets, aligning each operation time period with the plan time period by mapping the field operation start and end time and the scheduling plan node to a unified time axis to obtain time distribution differences, and obtaining time sequence offset characteristics; S2: comparing daily construction completion amounts based on the time sequence offset characteristics, checking team work reporting, corresponding operation completion proportions according to plan task targets, judging the synchronization between actual progress and scheduling plans, focusing on offset differences of each progress point, and obtaining node offset identifiers; S3: judging progress change trends based on the node offset identifiers, analyzing change processes of historical progress prediction curves and field multi-source data, comparing the fitting degree of prediction curves and actual curves by trend corresponding relations, matching with pre-set offset characteristic templates in an abnormal mode library, identifying offset causes, and obtaining task adjustment driving signals; S4: adjusting scheduling plans under feedback, reordering all subtask items in turn, recording adjustment corresponding progress nodes in real time after adjusting priority parameters, comparing plan sequences with original task sequences in time sequence, determining postponement nodes, and obtaining task sequence change information; S5: filtering operation completion nodes based on the task sequence change information, analyzing time sequence changes of completion time and adjusted plan nodes, segmenting progress time differences, dividing progress fluctuations into multiple time difference intervals according to the time difference between operation completion time and adjusted plan nodes, and statistically analyzing the number of nodes in each interval to obtain progress prediction control intervals. The time sequence offset characteristics include construction period continuity, time sequence consistency and progress distribution types, the node offset identifiers include node codes, progress offset categories and association identifiers, the task adjustment driving signals include adjustment modes, signal levels and scheduling links, the task sequence change information includes adjustment node sequences, adjustment execution states and change influence ranges, and the progress prediction control intervals include interval serial numbers, control ranges and interval control states.

2. The big data-based water construction progress prediction method according to claim 1, characterized in that, The acquisition step of the time sequence offset characteristics is specifically as follows: S111: performing format merging and time alignment on each data item according to a unified time reference based on field operation start and end time, attendance data and equipment operation states, aggregating time sequence information, and analyzing continuous time sequence structures of each data type to obtain a time sequence progress data group; S112: comparing time sequences with key progress nodes in a scheduling plan based on the time sequence progress data group, classifying and arranging data records of each node associated section, connecting node operation processes and node sequences, and obtaining a node time sequence curve set; S113: judging the start and end intervals of each node plan time and field progress based on the node time sequence curve set, analyzing node offset conditions in combination with continuous operation states of each node and node connection sequences, and obtaining time sequence offset characteristics.

3. The big data based water construction progress prediction method according to claim 1, characterized in that, The obtaining step of the node offset identification is specifically: S211: Based on the timing offset feature, combined with the field construction completion quantity and team reporting data, according to the scheduling plan task target, the relationship between the operation completion ratio and the plan task target is compared item by item, and the data is classified and arranged through the corresponding relationship, and the operation completion ratio analysis result is obtained; S212: Based on the operation completion ratio analysis result, judge the progress synchronization, by comparing the actual progress and the time node of the scheduling plan, analyze the lag and advance of the progress, evaluate the operation state of each progress node in the field, identify the offset phenomenon, and obtain the progress synchronization judgment result; S213: Based on the progress synchronization judgment result, according to the time interval and operation connection situation, identify the time offset between nodes, classify and identify the offset type and node state, and combine the difference between field progress and plan task, obtain the node offset identification.

4. The big data based water construction progress prediction method according to claim 1, wherein, The obtaining step of the task adjustment driving signal is specifically: S311: Based on the node offset identification, analyze the node code and progress offset category, match according to the scheduling plan and field progress record, calculate the timing difference between the plan and the actual progress, identify the offset rule, and obtain the node timing offset trend sequence; S312: Based on the node timing offset trend sequence, compare the offset change and the actual progress, analyze the influence on the progress according to the equipment state and personnel attendance data, calculate the matching situation of the actual and predicted progress curve, and use the formula: ； a trend difference score is obtained, wherein, a trend difference score, a predicted progress value, which is a predicted progress of the ith node, an actual progress value, which is an actual progress of the ith node, a device state fluctuation, which is a state fluctuation of the jth device, a personnel attendance intensity, which is an intensity of the jth personnel attendance, n represents a number of nodes participating in matching, and m represents a total number of synchronous fluctuation data items; S313: Based on the trend difference score, judge the abnormal node, analyze the association with the current stage offset phenomenon, compare the equipment downtime and personnel attendance factors, summarize the main associated abnormal factors of progress deviation, and obtain the task adjustment driving signal.

5. The big data based water construction progress prediction method according to claim 1, wherein, The obtaining step of the task sequence change information is specifically: S411: Based on the task adjustment driving signal, analyze the subtasks in the scheduling plan, optimize the original task list sequence according to the task priority and resource allocation, and compare the start and end time and position of each task before and after optimization, obtain the priority adjustment sorting sequence; S412: Based on the priority adjustment sorting sequence, calculate the time change, task arrangement change and task distribution density between the original task sequence, obtain the timing offset measurement result, judge the timing change of the task in the plan adjustment, and obtain the task timing disturbance quantity group; S413: Based on the task timing disturbance quantity group, analyze the task node structure change and state change, judge the influence of the task node after the sorting adjustment, optimize the task execution state and resource association, and obtain the task sequence change information.

6. The big data based water construction progress prediction method according to claim 1, wherein, The obtaining step of the progress prediction control interval is specifically: S511: Based on the task sequence change information, screen the operation completion nodes, compare the actual completion time of each node with the adjusted plan node one by one, analyze through time sequence comparison, sort out the progress of each node, and obtain the operation completion node set; S512: Based on the job completion node set, the time difference between the completion time of each node and the planned node is judged, the progress state of each node is analyzed, the offset of each time interval is classified and arranged, and the time difference segment set is obtained; S513: Based on the time difference segment set, the progress fluctuation of each segment is analyzed, each fluctuation segment is classified and labeled combined with node offset information, the progress control requirement of each segment is judged, and the progress prediction control interval is obtained.

7. The big data based water construction progress prediction method according to claim 1, wherein, The unified time reference refers to the standardized timeline as the reference for data collection and progress node registration of various types, the team report refers to the daily submission of the construction team, the completed quantity, personnel attendance, mechanical labor and material consumption of the field operation log, and the progress change trend refers to the judgment of the dynamic change direction of the increase, decrease, stability, acceleration or deceleration of the progress in the continuous time interval.

Images