A Big Data Management System for Factory Construction Risk Perception
By combining a cloud platform with the construction site, and utilizing structured rules and visual perception devices, construction behavior can be monitored in real time. This solves the problems of high deployment costs and inaccurate assessments in traditional factory construction risk management systems, and achieves more efficient construction risk management.
Patent Information
- Application Number
- CN202611127698.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-28
- Publication Date
- 2026-11-10
AI Technical Summary
Existing factory construction safety risk management systems rely on physical quantity sensors, which are costly to deploy, easily damaged, and difficult to quantify and assess the degree of compliance of construction entities with regulations, leading to difficulties in construction management.
The architecture combines a cloud platform with the construction site, using structured rule features and visual perception devices to monitor construction behavior in real time, assess construction deviations, and conduct risk analysis, thereby reducing reliance on physical quantity sensors.
It achieves more accurate construction risk assessment, reduces deployment and maintenance costs, improves the accuracy of risk assessment and its management guidance significance, and solves the problem of difficulty in implementing traditional systems.
Smart Images

Figure CN123070684A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of construction risk management technology, specifically a big data management system for factory construction risk perception. Background Technology
[0002] Existing factory construction safety risk management systems primarily revolve around physical quantity monitoring. This involves deploying sensors for stress, displacement, and settlement on structures such as tower cranes, foundation pits, and scaffolding to collect physical parameters in real time and trigger alarms when limits are exceeded. However, this approach faces significant challenges in practical engineering applications. The deployment cost of physical quantity sensors is extremely high; for a medium-sized factory project, the sensor procurement and installation costs alone can reach hundreds of thousands of yuan. Furthermore, the harsh construction site environment, with dust, vibration, and humidity leading to high sensor damage rates, severe drift, and significant data noise, results in many current factory construction projects still relying on traditional construction methods for management. Moreover, for enterprise managers, the core demand is not merely to obtain a bunch of physical parameter alarms, but to ensure strict adherence to regulations during construction. They desire a means to quantify and assess the degree of compliance by each construction entity, using this as a basis for risk assessment and management decisions.
[0003] In order to solve the above problems, this invention provides a big data management system for factory construction risk perception. Summary of the Invention
[0004] To address the problems of the above solutions, this invention provides a big data management system for factory construction risk perception.
[0005] The objective of this invention can be achieved through the following technical solutions: A big data management system for factory construction risk perception, including a cloud platform and a construction terminal; Furthermore, the cloud platform establishes communication connections with each construction site.
[0006] The cloud platform includes a cloud rule base, a construction configuration module, and a risk analysis module. The cloud-based rule base is used to store structured construction rules.
[0007] Furthermore, the structured rule features are triple rules, with the format being process, action sequence, and judgment condition. Each action is marked with the action weight coefficient corresponding to action missing deviation, timing misalignment deviation, action error deviation, and action execution qualified, respectively.
[0008] Furthermore, the stored structured construction rules are optimized and analyzed in real time, and the safety feedback data of each area when the structured rule features are met and when the structured rule features are not met are statistically analyzed in real time. The safety feedback data is extracted according to the preset feedback feature format to obtain the safety feedback feature set of each process. The rule risk consistency rate of the structured rule features of the action is calculated based on the safety feedback feature set. No rule adjustment is made when the rule risk consistency rate is greater than the threshold X1; When the rule risk consistency rate is not greater than the threshold X1, the structured rule features of the action are adjusted.
[0009] The construction configuration module obtains the construction project information corresponding to each construction end, matches structured rule features from the cloud rule base based on the construction project information, and transmits the structured rule features to the construction rule base of the construction end.
[0010] The risk analysis module is used to perform construction risk analysis, to statistically analyze the process risk value of each process in each factory construction project in real time, to perform coupled risk analysis based on the process risk value of each process, to obtain project risk results, and to send the project risk results to the construction end.
[0011] Furthermore, coupled risk analysis is performed based on the process risk values of each process, including: Calculate the average of the process risk values of all current parallel construction processes, and mark it as the multi-process foundation superposition risk value; obtain the dynamic coupling coefficient λ; The overall risk value of the project is calculated as λ × the cumulative risk value of multiple processes. The project risk outcome is determined based on the project's comprehensive coupled risk value.
[0012] Furthermore, based on the process risk value of each process, the construction technology capability level for constructing the corresponding factory building is assessed. Coupled risk prediction analysis is then performed based on the construction technology capability level to obtain the project risk prediction results.
[0013] The construction terminal includes a construction rule base, a monitoring module, and a construction analysis module; The construction rule base is used to store structured rule features.
[0014] The monitoring module performs real-time monitoring during factory construction, obtains construction monitoring data, and performs feature processing on the construction monitoring data to obtain monitoring feature data.
[0015] Furthermore, the construction monitoring data is collected through fixed video surveillance equipment and mobile video surveillance equipment at the construction site.
[0016] Furthermore, the construction monitoring data is processed to identify features according to the structured rule features corresponding to each process and action sequence.
[0017] The construction analysis module matches the structured rule features of the corresponding process to the monitoring feature data from the construction rule base, calibrates the monitoring feature data according to the structured rule features, obtains the process risk value of the corresponding process, and sends the process risk value to the cloud platform.
[0018] Further, the monitoring feature data is calibrated based on the structured rule features, including: Identify the structured rule features to identify each action i of the process, i=1, 2, ..., n, where n is the total number of actions specified for the process; The monitoring feature data is evaluated based on structured rule features to obtain the action status of each action. The action status includes action missing deviation, timing misalignment deviation, action error deviation, and action execution qualified. The corresponding action weight coefficient ω is then identified based on the action status. i ; Calculate the process risk value according to the formula for calculating process risk value.
[0019] Furthermore, the formula for calculating the process risk value is as follows: ; In the formula: GD is the process risk value.
[0020] Furthermore, the formula for calculating the process risk value is as follows: ; In the formula: GD is the process risk value; f(x) i ) is the correction function. x i This indicates the action status of the corresponding action; when the action status is an action error deviation, the compliance deviation degree is calculated, and the compliance correction coefficient δ is calculated based on the compliance deviation degree.
[0021] Furthermore, a compliance correction factor is calculated based on the degree of compliance deviation, including: Identify the probability of security risks associated with different levels of compliance deviation for each process and action shared on the cloud platform. Based on the corresponding risk representation probability of the action, identify the action weight coefficient when the action is in the action error deviation state, and obtain the safety risk probability corresponding to the action weight coefficient. The compliance correction coefficient is calculated based on the probability of security risk and the representative probability of security risk.
[0022] Furthermore, the cloud platform facilitates probability sharing of security risks, including: The cloud platform aggregates and statistically analyzes the safety data corresponding to each action in each region when the action is in an error deviation state. Based on the safety data, it calculates the representative probability of safety risk for each action at different compliance deviations. The representative probability of safety risk refers to the mathematical processing of multiple safety risk probabilities corresponding to the same compliance deviation to obtain a representative safety risk probability, which is marked as the representative probability of safety risk. The representative probability of safety risk for each process and action at different compliance deviations is shared with the construction end in the corresponding region.
[0023] Furthermore, the construction analysis module includes an independent early warning unit, which compares the process risk value of each process with the preset early warning standard. When the process risk value exceeds the early warning standard, a risk warning is issued.
[0024] Compared with the prior art, the beneficial effects of the present invention are: This invention eliminates the reliance on exceeding physical quantity limits as a risk trigger condition. Instead, it uses specified execution deviations as the core input, allowing risk assessments to directly reflect the true state of on-site management and making the assessment results more meaningful for management guidance. Secondly, this invention employs a cloud-edge hybrid architecture. The cloud platform uniformly manages the specification library and iteratively refines rules by region, continuously improving specification accuracy with data accumulation, thus enhancing risk assessment accuracy and resolving the discrepancies caused by relying solely on rules for risk assessment. Finally, this invention uses visual perception devices as the primary data acquisition method, extracting construction behavior data from video streams through algorithms such as target detection, pose estimation, and optical character recognition. This significantly reduces reliance on physical quantity sensors, lowers deployment costs, and the visual devices are easy to install and maintain, solving the problem of difficult implementation of existing systems. Attached Figure Description
[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0026] Figure 1 This is a block diagram illustrating the principle of the present invention. Detailed Implementation
[0027] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.
[0028] like Figure 1 As shown, a big data management system for factory construction risk perception includes a cloud platform and a construction terminal; The cloud platform is deployed on a server cluster, while the construction end is deployed on edge computing servers or industrial control computers at the construction site.
[0029] The cloud platform communicates with each construction site.
[0030] For example, there are two communication channels between the cloud platform and the construction end: a rule synchronization channel and a data feedback channel. The rule synchronization channel is used to distribute the cloud rule library and updates to the construction end, and the data feedback channel is used to upload data from the construction end to the cloud platform. Rule synchronization adopts an incremental update mechanism, distributing only the changed parts, such as controlling the data volume of a single synchronization to within 10MB. Data feedback adopts a compressed upload mechanism, such as controlling the average daily data volume of a single project to within 50MB.
[0031] The cloud platform includes a cloud rule base, a construction configuration module, and a risk analysis module. The cloud-based rule base is used to store structured construction rules.
[0032] Structured construction rules are structured rule features extracted from construction rules.
[0033] For example, the structured rule feature is a triple rule, with the format being process, action sequence, and judgment condition. Each action is marked with a corresponding action weight coefficient, which has four values corresponding to action missing deviation, timing misalignment deviation, action error deviation, and action execution qualified values, such as 0.9, 0.3, 0.6, and 0, respectively. The weight coefficient is set according to the safety risk situation of the action under the background of action missing deviation, timing misalignment deviation, and action error deviation. If set by professionals, the action weight coefficient can also be set according to the historical maximum risk probability under the corresponding background. Action execution qualified is considered as 0.
[0034] For example, a centralized maintenance of the national construction specification database covers construction rule data such as national standard GB series, industry standard JGJ series, local standard DB series, project-level special plans, and safety technical briefing texts; The pre-trained BERT-BiLSTM-CRF model is used to extract information from the construction rule data, including process name information (such as foundation pouring, rebar tying, formwork installation), operation action information (such as accepting rebar, installing formwork, checking supports, pouring concrete), quantitative parameter information (such as bolt torque value ≥40N·m, concrete curing time ≥14 days, formwork verticality deviation ≤5mm), and rule constraint information. The rule constraint information is extracted from the text through a relation extraction algorithm to identify the preconditions and temporal constraints between actions (such as must be completed before pouring, cross-operation is strictly prohibited, and double confirmation is required).
[0035] Using a temporal relation extraction algorithm, temporal dependencies between actions are established based on time conjunctions (such as "first...then...finally...", "before...", "can only be performed after...") in the construction rule data. For example, for the foundation pouring process, the extraction result is: inspecting steel bars → installing formwork → inspecting supports → pouring concrete. There is a strict sequential dependency between these four actions; the subsequent actions cannot be performed if the preceding action is not completed.
[0036] The extraction results are transformed into structured triple rules, for example: Foundation pouring (process): Reinforcing steel acceptance → Formwork installation → Support inspection → Concrete pouring (action sequence), formwork verticality deviation ≤ 5mm and reinforcing steel protective layer thickness deviation ≤ 3mm (judgment criteria).
[0037] In one embodiment, version control is implemented on the cloud-based rule base. Each rule has a version number, source tag (national standard / industry standard / local standard / project), regional tag (province / city), and effective time, and supports retrieval by project type and regional tag.
[0038] In one embodiment, due to actual safety differences, simply following specifications may not be accurately applicable to certain projects in some cases. This is because national standards are set according to general standards, and for some special situations, higher or lower safety rule requirements may be required, since the ultimate goal is to conduct construction risk assessment. However, assessments can also be based on the general rules in the above embodiments to determine whether the rules are met for construction. Therefore, to improve the accuracy of subsequent construction risk assessment, the structured rule features in the cloud-based rule base are continuously optimized and adjusted, including: Real-time statistics are collected on safety feedback data for each region when it meets and does not meet the structured rule characteristics. That is, the occurrence of safety accidents under the background of meeting and not meeting the structured rule characteristics respectively. Regardless of whether the assessment is risk-free or risky, if a safety accident occurs, there is a safety problem; if no safety accident occurs, there is no safety problem. The safety feedback data is processed according to a preset feedback feature format to obtain the safety feedback feature set corresponding to each process. The safety feedback feature set is composed of various safety feedback feature units. For example, the feedback feature format is set as process, action, risk assessment result (risk assessment result determined by the subsequent risk analysis module and independent early warning unit) and actual risk result (whether the actual risk has occurred). If risk is handled after an early warning, the risk assessment result is considered to be correct. The processes and actions corresponding to the same safety feedback feature set are the same. The rule risk consistency rate of the structured rule features (referring to the rule features for the action) of the action is calculated based on the safety feedback feature set. Consistency means that the assessment has risk and the actual result has risk. If the assessment has risk but there is no risk, or if the assessment has no risk but a safety risk actually occurs, it is considered inconsistent. The proportion of the number of safety feedback feature units that meet the consistency criteria is calculated to the total number and marked as the rule risk consistency rate. When the rule risk consistency rate is greater than the threshold X1, no rule adjustment will be made. For example, if it is 95%, adjustments will be made according to user needs. A minimum value can be set by the cloud platform.
[0039] When the rule risk consistency rate is not greater than the threshold X1, the structured rule features of the action are adjusted, which is generally done and verified by professionals.
[0040] For example, the structured rule features corresponding to the action can be extracted from the original structured rule features, and region-specific constraint parameters can be added. For example, if the original structured rule features stipulate that "the steel reinforcement acceptance must be completed before concrete pouring", the supplementary provision can be "the positioning verification of the embedded parts must also be completed"; if the original structured rule features stipulate the sequence of each action, the rule can be adjusted so that the sequence of a certain action can be adjusted within a preset range.
[0041] In one embodiment, a proportional coefficient can be preset for various scenarios, such as assessment having risk and actual result having risk; assessment having risk but actual no risk; and assessment having no risk but actual occurrence of safety risk. The coefficient is not always calculated as 1. For example, the proportional coefficient is set to be greater than 1 for the scenario where there is no risk in the assessment but actual occurrence of safety risk, while the other proportional coefficients remain 1. The risk consistency rate is calculated based on this rule.
[0042] The construction configuration module obtains construction project information from each construction end, such as the location, construction project document, construction briefing and other related information; matches structured rule features from the cloud rule base based on the construction project information, such as matching based on region and specific construction; and transmits the structured rule features to the construction rule base at the construction end.
[0043] In one embodiment, for ease of use on the construction end, the models and algorithms required on the construction end can be pre-configured on the cloud platform, and models and algorithms adapted to various situations can be configured. Subsequently, based on the analysis needs of the construction end, the construction configuration module configures the models and algorithms for the construction end according to the construction project information and the construction end information.
[0044] The risk analysis module is used to perform construction risk analysis, and to statistically analyze the process risk value of each process in each construction project of each factory in real time. Based on the process risk value of each process, coupled risk analysis is performed to obtain the project risk results. The project risk results are sent to the construction end, which then provides early warning prompts to the management personnel. Alternatively, an independent early warning unit can issue early warnings.
[0045] In one embodiment, coupled risk analysis is performed based on the process risk values of each process. An intelligent analysis model is established based on machine learning, deep learning algorithms, etc. The training set is set manually and trained. The training set includes input data and output data. The input data is project information and the process risk values of each process. The output data is the project risk results, such as whether there is a risk and the reason for the risk.
[0046] In one embodiment, the construction technical capability of the current construction can also be determined based on the existing process risk values of each process. For example, each construction technical capability level can be preset, and the construction technical capability level corresponding to the process risk value combination of different processes can be predetermined, or various process risk value combinations corresponding to each construction technical capability level can be determined. The construction technical capability level of each process can be determined according to its process risk value, and the process risk value of subsequent processes can be estimated according to the construction technical capability level, thereby realizing coupled risk prediction analysis and obtaining project risk prediction results.
[0047] For example, a dual-model fusion architecture of feature engineering + random forest coupling analysis model + time series LSTM risk prediction model is adopted to realize multi-process risk coupling quantitative analysis and intelligent prediction of future risk situation respectively. Input feature set: divided into risk features, construction quality features, and rule adaptation features; Process risk characteristics: risk value of each process, number of high-risk processes; Coupling and correlation characteristics: number of parallel processes, types of cross-operation process combinations, and frequency of process risk superposition; Construction quality characteristics: compliance rate of procedures, frequency of violations, rectification closure rate, and construction quality stability coefficient; Rule adaptation characteristics: regional structured rule satisfaction status, rule adaptation deviation value.
[0048] Output result: Random forest coupling model outputs: risk coupling weights for each process, overall project coupling risk value, risk coupling level (low / medium / high / major), and core risk coupling combination; The LSTM time series forecasting model outputs: project risk values for the next 12 hours / 24 hours, risk change trends (stable / rising / worsening), and predictions of high-risk hazards.
[0049] For example, during a certain construction period, the project's operational status was as follows: three processes were carried out simultaneously: formwork erection (risk value 0.52), rebar construction (risk value 0.48), and high-altitude operations (risk value 0.65). The risks of each process were all within a moderate and controllable range, and no traditional single-risk warning was triggered.
[0050] Feature input: Input the risk values of the three processes, the parallel coupling status, the current construction compliance rate (92%), and the rule satisfaction status into the trained random forest coupling model; Model calculation: The model retrieves the dynamic coupling coefficients of each process, calculates the amplification effect of multiple processes, corrects the independent risk deviation of a single process, and finally outputs the real-time comprehensive coupling risk value of the project, which is 0.79. Result determination: The model determines that the current risk level is high risk, and accurately identifies multi-process coupled complex hidden dangers that are missed by traditional assessment methods; Coupling and source tracing: The model automatically outputs that the core source of risk is the cross-coupled operation of "template erection + high-altitude operation", providing a targeted basis for precise on-site management.
[0051] In one embodiment, coupled risk analysis is performed based on the process risk values of each process, including: Calculate the average of the process risk values of all current parallel construction processes and mark it as the multi-process foundation superposition risk value; train historical accident samples through a random forest model, and output the dynamic coupling coefficient λ in real time according to the current process combination type, parallel quantity, and risk matching degree.
[0052] Coupling coefficient rules: Single-process operation: λ=1.0 (no scale-up effect); Crossover of ordinary processes: 1.0 < λ < 1.5 (slight overlap); Cross-coupling of high-risk processes (high-altitude work + formwork + installation, etc.): λ≥1.5 (strong amplification effect).
[0053] The overall risk value of the project is calculated as λ × the cumulative risk value of multiple processes. The project risk outcome is determined based on the comprehensive coupled risk value of the project, and an early warning is issued if the risk exceeds the threshold.
[0054] The construction terminal includes a construction rule base, a monitoring module, and a construction analysis module; The construction rule base is used to store the structured rule features of each process, so as to facilitate matching the corresponding structured rule features according to the current construction process.
[0055] The monitoring module is used to perform real-time monitoring during factory construction, obtain construction monitoring data, and process the construction monitoring data to obtain monitoring feature data.
[0056] In one embodiment, visual sensing devices are the primary monitoring method. For example, comprehensive video surveillance equipment can be deployed at the construction site to achieve full monitoring of the construction. Drones and other devices can also be used for supplementary monitoring.
[0057] In one embodiment, additional monitoring equipment such as infrared monitoring devices can be used for auxiliary monitoring to achieve more comprehensive monitoring of the construction process.
[0058] In one embodiment, construction monitoring data is processed to identify the construction monitoring data using current image, video and other recognition technologies, obtain monitoring feature data for subsequent comparison with structured rule features, and perform targeted identification by combining relevant process information, action sequences and other information.
[0059] For example, the target detection algorithm uses the YOLOv8 model to identify three types of targets in the construction site video stream: construction workers (distinguishing by trade: steelworkers, formwork workers, bricklayers, etc.), construction materials (steel bars, formwork, bricks, etc.), and machinery and equipment (tower cranes, elevators, vibrators, etc.). At the same time, it identifies the status of the targets (such as the status of wearing safety helmets, wearing reflective vests, and the status of equipment operation).
[0060] Pose estimation algorithm: The MediaPipe Pose model is used to extract the skeletal key point sequence of the construction workers from the video stream. 33 human skeletal key points (including key points of the top of the head, face, shoulders, elbows, hands, hips, knees, feet and torso) are used to reconstruct the operation action through the spatiotemporal changes of the key points.
[0061] Optical character recognition algorithm: The PaddleOCR model is used to read text information from on-site signs, inspection reports, and handover records, such as concrete strength grade C30 and steel bar specification HRB400Φ20.
[0062] The construction analysis module matches the structured rule features of the corresponding process to the monitoring feature data from the construction rule base, and calibrates the monitoring feature data according to the structured rule features. This mainly includes whether the action is executed, whether the execution order of the action is correct, and whether the execution result of the action meets the judgment conditions, so as to obtain the process risk value of the corresponding process; and sends the process risk value to the cloud platform.
[0063] In one embodiment, calibrating the monitoring feature data according to the structured rule features includes: Identify the structured rule features to identify each action i of the process, i=1, 2, ..., n, where n is the total number of actions specified for the process; Based on the structured rule features, each action corresponding to the monitoring feature data is evaluated to obtain the action status of each action. The action status includes action missing deviation, timing misalignment deviation, action error deviation, and action execution qualified. The corresponding action weight coefficient ω is identified based on the action status. i ; The process risk value is calculated using the formula for calculating process risk value. The formula is as follows: ; In the formula: GD is the process risk value.
[0064] In one embodiment, when correcting the operational state of an error deviation, the calculation of the process risk value includes: When the action status is an action error deviation, calculate the compliance deviation degree, and calculate the compliance correction coefficient based on the compliance deviation degree; The calculation formula is: ; In the formula: GD is the process risk value; f(x) i ) is the correction function. x i This indicates the action status of the corresponding action, which includes action missing deviation, timing misalignment deviation, action error deviation, and action execution qualified; δ is the compliance correction coefficient calculated based on the compliance deviation degree; when the action status is not an action error deviation, the compliance correction coefficient is 1.
[0065] In one embodiment, calculating the compliance deviation includes: The monitoring feature data corresponding to the corresponding action is compared with the structured rule features to calculate the compliance deviation, that is, the degree of deviation from the rule standard. For example, if the specified rebar spacing is 20cm, but the actual spacing is 25cm, the compliance deviation is (25-20)÷20=0.25. If it is not a numerical value, such as the difference in rebar specifications, the compliance deviation is calculated based on the difference in specifications between the rebar and the designed rebar. For example, the compliance deviation can be calculated based on the tensile strength of the rebar. Specifically, the degree of deviation from the standard rule can be calculated based on various methods, such as preset compliance deviation calculation methods or specific compliance deviations for various actions under different conditions when there is an error deviation.
[0066] In one embodiment, calculating the compliance correction factor based on the compliance deviation includes: The cloud platform aggregates safety data corresponding to action error deviations from various regions. Based on this safety data, it statistically analyzes the safety risk probability corresponding to different compliance deviations for each action in that region. It identifies the safety risk probability corresponding to the action weight coefficient when the action is at an error deviation. Based on the two safety risk probabilities, it calculates a deviation correction coefficient. This process is repeated to obtain a large number of deviation correction coefficients corresponding to compliance deviations. Mathematical processing is then performed to determine the deviation correction coefficient corresponding to different compliance deviations for each action, and these coefficients are marked as compliance correction coefficients. Various mathematical processing methods, such as mode, variance, mean, and fitting, are used to determine representative deviation correction coefficients. Subsequently, the construction end can match these coefficients.
[0067] In one embodiment, the construction analysis module includes an independent early warning unit. This unit compares the risk value of each process step with a preset early warning standard. When the risk value exceeds the standard, a risk warning is issued. The early warning standard is a corresponding threshold, set according to the construction project management needs. It is typically preset by the cloud platform and subsequently configured by relevant project management personnel based on requirements.
[0068] Enables direct early warning at construction sites.
[0069] The above formulas are all numerical calculations after removing dimensions. The formulas are obtained by software simulation based on a large amount of data and are closest to the real situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained by simulation based on a large amount of data.
[0070] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.
Claims
1. A big data management system for factory construction risk perception, characterized in that, This includes both cloud platforms and construction terminals; The cloud platform includes a cloud rule base, a construction configuration module, and a risk analysis module; the construction terminal includes a construction rule base, a monitoring module, and a construction analysis module. The cloud-based rule base is used to store structured construction rules; The construction configuration module obtains construction project information corresponding to each construction end, matches structured rule features from the cloud rule base based on the construction project information, and transmits the structured rule features to the construction rule base of the construction end. The risk analysis module is used to perform construction risk analysis, to statistically analyze the process risk value of each process in each factory construction project in real time, to perform coupled risk analysis based on the process risk value of each process, to obtain project risk results, and to send the project risk results to the construction end. The construction rule base is used to store structured rule features; The monitoring module performs real-time monitoring during factory construction, obtains construction monitoring data, and performs feature processing on the construction monitoring data to obtain monitoring feature data. The construction analysis module matches the structured rule features of the corresponding process to the monitoring feature data from the construction rule base, calibrates the monitoring feature data according to the structured rule features, obtains the process risk value of the corresponding process, and sends the process risk value to the cloud platform.
2. The factory construction risk perception big data management system according to claim 1, characterized in that, The structured rule features are triple rules, with the format being process, action sequence, and judgment condition. Each action is marked with the action weight coefficient corresponding to action missing deviation, timing misalignment deviation, action error deviation, and action execution qualified, respectively.
3. The factory construction risk perception big data management system according to claim 1, characterized in that, The stored structured construction rules are optimized and analyzed in real time. Safety feedback data for each area that meets and does not meet the structured rule features are statistically analyzed in real time. The safety feedback data is extracted according to the preset feedback feature format to obtain the safety feedback feature set of each process. The rule risk consistency rate of the structured rule features of the action is calculated based on the safety feedback feature set. No rule adjustment is made when the rule risk consistency rate is greater than the threshold X1; When the rule risk consistency rate is not greater than the threshold X1, the structured rule features of the action are adjusted.
4. The factory construction risk perception big data management system according to claim 1, characterized in that, Coupled risk analysis is performed based on the process risk values of each process, including: Calculate the average of the process risk values of all current parallel construction processes, and mark it as the multi-process foundation superposition risk value; obtain the dynamic coupling coefficient λ; The overall risk value of the project is calculated as λ × the cumulative risk value of multiple processes. The project risk outcome is determined based on the project's comprehensive coupled risk value.
5. The factory construction risk perception big data management system according to claim 1, characterized in that, The risk analysis module assesses the construction technology capability level of the corresponding factory building based on the process risk value of each process, and performs coupled risk prediction analysis based on the construction technology capability level to obtain the project risk prediction results.
6. The factory construction risk perception big data management system according to claim 1, characterized in that, The construction monitoring data is collected through fixed and mobile video surveillance equipment at the construction site.
7. The factory construction risk perception big data management system according to claim 1, characterized in that, The monitoring feature data is calibrated based on the structured rule features, including: Identify structured rule features to identify each action i in the process, i = 1, 2, ..., n, where n is the total number of actions included in the process; The monitoring feature data is evaluated based on structured rule features to obtain the action status of each action. The action status includes action missing deviation, timing misalignment deviation, action error deviation, and action execution qualified. The corresponding action weight coefficient ω is then identified based on the action status. i ; Calculate the process risk value according to the formula for calculating process risk value.
8. The factory construction risk perception big data management system according to claim 7, characterized in that, The formula for calculating the risk value of a process is: ; In the formula: GD is the process risk value; f(x) i ) is the correction function. x i This indicates the action status of the corresponding action; when the action status is an action error deviation, the compliance deviation degree is calculated, and the compliance correction coefficient δ is calculated based on the compliance deviation degree.
9. A big data management system for factory construction risk perception according to claim 8, characterized in that, The compliance correction factor is calculated based on the compliance deviation, including: Identify the probability of security risks associated with different levels of compliance deviation for each process and action shared on the cloud platform. Based on the corresponding risk representation probability of the action, identify the action weight coefficient when the action is in the action error deviation state, and obtain the safety risk probability corresponding to the action weight coefficient. The compliance correction coefficient is calculated based on the probability of security risk and the representative probability of security risk.
10. The factory construction risk perception big data management system according to claim 1, characterized in that, The construction analysis module includes an independent early warning unit, which compares the process risk value of each process with the preset early warning standard. When the process risk value exceeds the early warning standard, a risk warning is issued.
