Hoisting equipment distribution visualization and monitoring method

Abstract

The invention discloses a hoisting equipment distribution visualization and monitoring method, and relates to the technical field of industrial hoisting safety, and the method comprises the following steps: generating a synchronously associated multi-source fusion data set; constructing a three-dimensional basic model of the hoisting operation area; dividing the three-dimensional dynamic model into a base layer, an equipment layer, a key component layer and a data layer by adopting a self-adaptive hierarchical rendering algorithm, adjusting the rendering precision of each layer according to a priority coefficient, and generating an equipment distribution visual interface; identifying an abnormal state of the equipment and positioning a fault position, and triggering a corresponding grading early warning signal according to an abnormal grade; and based on the abnormity type and the fault position corresponding to the early warning signal, a targeted disposal suggestion scheme is generated. The real-time monitoring, fault early warning and three-dimensional dynamic visualization of the hoisting equipment are realized by fusing space-time alignment, multi-source data synchronization, a digital twinning technology and a self-adaptive rendering algorithm, and the equipment safety and the operation efficiency are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of industrial hoisting safety technology, specifically a method for visualizing and monitoring the distribution of hoisting equipment. Background Technology

[0002] In industrial hoisting operations, the distribution monitoring and status management of hoisting equipment directly affect operational safety and efficiency. Current hoisting equipment management in the industry suffers from the following pain points: Existing monitoring systems often rely on single sensors or localized data collection, lacking effective integration of multi-source data such as equipment operating parameters, environmental data, and operational instructions. Inconsistent timestamps and missing data are frequent problems, preventing the formation of a complete equipment operation profile and impacting the accuracy of subsequent analysis. Traditional monitoring systems often use two-dimensional interfaces to display equipment locations, making it difficult to reconstruct the three-dimensional spatial relationships of the work area. Furthermore, model updates are lagging, failing to match the actual operating status of the equipment in real time. When equipment moves or the working environment changes, the visual interface easily becomes disconnected from the physical scene, making it difficult for managers to promptly grasp the dynamic distribution of equipment. Existing systems often rely on real-time parameter thresholds for anomaly detection, lacking the ability to predict equipment operating trends. Warnings are often triggered only after a fault occurs, missing opportunities for early intervention. Simultaneously, low fault location accuracy makes it difficult to quickly pinpoint faulty components, leading to low response efficiency. Existing technologies primarily focus on real-time fault monitoring, lacking a comprehensive health assessment system for the entire equipment lifecycle. This makes it impossible to predict equipment aging or potential failure risks in advance, easily leading to unplanned downtime, increasing maintenance costs and operational safety hazards.

[0003] With the development of industrial intelligence, the requirements for the precision, real-time and intelligent management of hoisting equipment are increasing. There is an urgent need for an integrated monitoring method that can realize multi-source data fusion, dynamic scene visualization, trend prediction and early warning, closed-loop control and health assessment to solve the shortcomings of existing technologies. Summary of the Invention

[0004] The purpose of this invention is to solve the problems mentioned above by proposing a method for visualizing and monitoring the distribution of hoisting equipment.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A method for visualizing and monitoring the distribution of hoisting equipment, the specific implementation process of which is as follows:

[0007] S1. Collect equipment operating parameters, spatial location data, environmental perception data and operation instruction data through sensor network, and use spatiotemporal alignment algorithm to perform timestamp matching and missing value filling on the collected multi-source data to generate a synchronously associated multi-source fusion dataset;

[0008] S2. Based on digital twin technology, and combined with the actual terrain, building layout and equipment parameters of the hoisting operation area, a three-dimensional basic model of the hoisting operation area is constructed. The multi-source fusion dataset is input into the three-dimensional basic model, the deviation value between the model parameters and the actual data is calculated, and the model parameters are dynamically calibrated based on the deviation value to generate a three-dimensional dynamic model consistent with the physical scene.

[0009] S3. An adaptive layered rendering algorithm is adopted to divide the 3D dynamic model into a base layer (terrain, buildings), an equipment layer (main body of hoisting equipment), a key component layer (hook, outriggers, winch mechanism) and a data layer (operating parameters, status indicators). Combining the user's operating perspective, the importance level of the equipment and the data update frequency, the rendering priority coefficient of each layer is calculated, and the rendering accuracy of each layer is adjusted according to the priority coefficient to generate a visual interface of equipment distribution.

[0010] S4. Input the device operation data in the multi-source fusion dataset into the improved LSTM algorithm. Analyze the changing trend of device operation parameters through the algorithm, output parameter prediction results, call the preset safety rule base, incorporate the parameter prediction results and real-time operation data into the rule matching, identify abnormal device status and locate fault location, and trigger corresponding graded early warning signals according to the abnormality level.

[0011] S5. Based on the anomaly type and fault location corresponding to the early warning signal, generate targeted handling suggestions.

[0012] Preferably, the spatiotemporal alignment algorithm in S1 is specifically as follows:

[0013] The timestamps of the device's GPS positioning data and sensor data are matched, and linear interpolation is used to fill in the missing data gaps. The alignment formula is as follows:

[0014]

[0015] In the formula, For the target time Alignment data, , They are time points , The original collected data;

[0016] The collected equipment operating parameters include lifting capacity, lifting speed, slewing angle, hydraulic system pressure, and engine speed. Environmental sensing data includes wind speed, light intensity in the work area, and ground subsidence.

[0017] Preferably, S2 specifically includes:

[0018] The basic three-dimensional scene is constructed using BIM+GIS fusion technology, and the CAD model and technical parameters of the hoisting equipment are imported to restore the appearance, internal structure and moving joints of the equipment.

[0019] The edge computing node receives the multi-source fusion dataset output by S1 in real time, extracts the device position and attitude parameters from the data and compares them with the corresponding parameters of the 3D basic model to obtain the model update deviation value. When the deviation value is greater than 0.5%, the original model parameters are corrected in combination with the scene adaptation coefficient (value range 0.98-1.02).

[0020] The model update frequency is dynamically adjusted according to the equipment's operating status. The update frequency is 10Hz when the equipment is in operation and 1Hz when it is in standby mode. After each update, a calibrated 3D dynamic model is output.

[0021] Preferably, the adaptive layered rendering algorithm of S3 specifically includes:

[0022] The 3D dynamic model is divided into four layers according to scene elements: the base layer (terrain, buildings), the equipment layer (main body of hoisting equipment), the key component layer (hook, outriggers, winch mechanism), and the data layer (operating parameters, status indicators);

[0023] Define a rendering priority coefficient, which is obtained by weighted summation of device importance weight (core device 0.6, auxiliary device 0.3), data update frequency weight (high frequency 0.3, low frequency 0.1) and view distance weight (near distance 0.4, far distance 0.1), and the sum of the three weights is 1;

[0024] The rendering precision of each level is adjusted according to the priority coefficient. When the coefficient is not less than 0.4, high-precision rendering with no less than 500,000 triangles is used. When the coefficient is less than 0.4, lightweight rendering with no more than 100,000 triangles is used. The final output is a visualization interface of device distribution.

[0025] Preferably, the improved LSTM algorithm of S4 is used for device operation trend analysis. The algorithm structure includes an input layer, a forget gate, an input gate, a cell state layer, an output gate, and a fully connected layer. The forget gate uses the ELU activation function, which directly outputs the input value when the input value is greater than 0, and outputs the result of the natural constant raised to the power of the input value minus 1 when the input value is less than or equal to 0.

[0026] The device operation data in the multi-source fusion dataset output by S1 is input into the improved LSTM algorithm in time series order, and the algorithm predicts the changing trend of key device parameters in the next 5 minutes.

[0027] The preset safety rule base is built based on the equipment safety operation procedures and historical fault data. It includes parameter threshold rules, trend anomaly rules and associated fault rules. The parameter prediction results and real-time operation data are substituted into the rule base for matching, and the abnormal equipment status identification results and fault location information are output to trigger the corresponding graded early warning signals.

[0028] Preferably, the specific rules of the security rule base include:

[0029] Set safety threshold ranges for each operating parameter, obtain the real-time change rate of the device operating parameters, and determine the trend as abnormal when the change rate exceeds the preset critical change rate and the duration exceeds 3 seconds;

[0030] A parameter correlation model is established based on Bayesian network. When the hydraulic system pressure exceeds the threshold and the lifting speed is less than 0.3 times the rated speed, it is determined to be a hydraulic system failure. The probability of failure is calculated by combining historical failure data.

[0031] Preferably, the tiered early warning mechanism specifically includes:

[0032] Level 1 warning: When S4 detects that a single parameter exceeds the threshold but does not trigger the associated fault rule, it generates a Level 1 warning signal, outputs the warning prompt through a green pop-up window in the visual interface, and records the abnormal parameter and the time of occurrence.

[0033] Level 2 warning: When S4 detects that multiple parameters exceed the threshold and trigger abnormal trend rules, it generates a level 2 warning signal and pushes the warning information through the sound and light alarm and system SMS.

[0034] Level 3 warning: When S4 identifies that the associated fault rule is triggered and the predicted fault probability is not less than 80%, a level 3 warning signal is generated, the high-risk operation circuit of the equipment is immediately cut off, and the emergency phone number of the management personnel is dialed, and the fault location result and handling flowchart are pushed.

[0035] Preferably, it also includes a device operation trajectory tracking function: by using GPS and inertial navigation fusion technology, the time series of device spatial position data in S1 is collected, the device movement trajectory and operation posture are recorded, and a trajectory dataset containing timestamps, three-dimensional coordinates, rotation angle and pitch angle is generated.

[0036] Preferably, S5 specifically includes:

[0037] Establish a historical case database, use the cosine similarity algorithm to obtain feature matching degree, retrieve cases from the case database with a feature similarity of not less than 85% with the current warning signal, and extract the handling measures from the cases;

[0038] The security rule base in S4 is called, and the extracted measures are optimized and adjusted in combination with the fault handling rules to generate three candidate handling solutions.

[0039] The analytic hierarchy process (AHP) was used to evaluate candidate solutions. Evaluation indicators included disposal time (weight 0.4), disposal cost (weight 0.3), and safety risk (weight 0.3). The comprehensive score of each solution was calculated, and the optimal disposal solution with the highest comprehensive score was output.

[0040] Preferably, it also includes a device health assessment function:

[0041] Collect equipment operation data from S1, fault records from S4, and equipment maintenance records to construct a health assessment model. The formula for the health assessment model is:

[0042]

[0043] In the formula, The running status score is calculated based on the degree to which parameters deviate from the threshold, with a maximum score of 100. The score is based on the frequency of failures (100 points for no failures, and 30 points deducted for each Level 3 failure). To maintain the target score (calculated based on the maintenance plan completion rate, with a maximum score of 100 points). For health status;

[0044] according to The value is used to classify the device health as excellent. ,good ,generally Poor There are four health levels. When the health status is poor, a maintenance reminder will be automatically pushed, and a maintenance plan including maintenance items and suggested time will be generated.

[0045] Compared with the prior art, the beneficial effects of the present invention are:

[0046] 1. Through multi-source data fusion and intelligent algorithm analysis, this method can monitor the operating status of hoisting equipment in real time and predict equipment parameter change trends based on LSTM, identifying potential faults in advance and triggering a tiered early warning mechanism in a timely manner. This early warning can significantly reduce the risk of equipment failure and ensure operational safety.

[0047] 2. The adoption of an adaptive layered rendering algorithm makes the 3D model of the hoisting operation area not only highly realistic, but also dynamically adjusts the rendering accuracy according to the importance of the equipment, the data update frequency and the user's perspective. This provides operators with a clear and intuitive equipment distribution map and improves the efficiency and accuracy of operation decisions.

[0048] 3. By leveraging historical case databases and health assessment functions, this method can optimize fault handling plans and provide personalized maintenance suggestions based on equipment operating data and historical fault records, thereby improving the accuracy and efficiency of equipment maintenance. At the same time, the health assessment function helps determine the current status of the equipment, facilitating advance maintenance and extending the equipment's service life. Attached Figure Description

[0049] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0050] Figure 1 This is a flowchart of the overall method of the present invention. Detailed Implementation

[0051] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0053] Please see Figure 1 As shown, a method for visualizing and monitoring the distribution of hoisting equipment is described, and the specific implementation process is as follows:

[0054] S1. Collect equipment operating parameters, spatial location data, environmental perception data and operation instruction data through sensor network, and use spatiotemporal alignment algorithm to perform timestamp matching and missing value filling on the collected multi-source data to generate a synchronously associated multi-source fusion dataset;

[0055] S2. Based on digital twin technology, and combined with the actual terrain, building layout and equipment parameters of the hoisting operation area, a three-dimensional basic model of the hoisting operation area is constructed. The multi-source fusion dataset is input into the three-dimensional basic model, the deviation value between the model parameters and the actual data is calculated, and the model parameters are dynamically calibrated based on the deviation value to generate a three-dimensional dynamic model consistent with the physical scene.

[0056] S3. An adaptive layered rendering algorithm is adopted to divide the 3D dynamic model into a base layer (terrain, buildings), an equipment layer (main body of hoisting equipment), a key component layer (hook, outriggers, winch mechanism) and a data layer (operating parameters, status indicators). Combining the user's operating perspective, the importance level of the equipment and the data update frequency, the rendering priority coefficient of each layer is calculated, and the rendering accuracy of each layer is adjusted according to the priority coefficient to generate a visual interface of equipment distribution.

[0057] S4. Input the device operation data in the multi-source fusion dataset into the improved LSTM algorithm. Analyze the changing trend of device operation parameters through the algorithm, output parameter prediction results, call the preset safety rule base, incorporate the parameter prediction results and real-time operation data into the rule matching, identify abnormal device status and locate fault location, and trigger corresponding graded early warning signals according to the abnormality level.

[0058] S5. Based on the anomaly type and fault location corresponding to the early warning signal, generate targeted handling suggestions.

[0059] The spatiotemporal alignment algorithm in S1 is specifically as follows:

[0060] The timestamps of the device's GPS positioning data and sensor data are matched, and linear interpolation is used to fill in the missing data gaps. The alignment formula is as follows:

[0061]

[0062] In the formula, For the target time Alignment data, , They are time points , The original collected data;

[0063] The collected equipment operating parameters include lifting capacity, lifting speed, slewing angle, hydraulic system pressure, and engine speed. Environmental sensing data includes wind speed, light intensity in the work area, and ground subsidence.

[0064] Using GPS positioning timestamps as a unified benchmark to achieve multi-source data anchoring solves the fragmentation problems of asynchronous time and mismatched dimensions in traditional multi-sensor data. By clarifying the collection range of core operating parameters and key environmental factors, a data dimension of equipment and environment is constructed. Compared with the solution of collecting parameters from a single device, it provides more comprehensive data support for subsequent fault diagnosis. The application of linear interpolation method controls the error of missing value filling within ±2%, ensuring data integrity.

[0065] S2 specifically includes:

[0066] The basic three-dimensional scene is constructed using BIM+GIS fusion technology, and the CAD model and technical parameters of the hoisting equipment are imported to restore the appearance, internal structure and moving joints of the equipment.

[0067] The edge computing node receives the multi-source fusion dataset output by S1 in real time, extracts the device position and attitude parameters from the data and compares them with the corresponding parameters of the 3D basic model to obtain the model update deviation value. When the deviation value is greater than 0.5%, the original model parameters are corrected in combination with the scene adaptation coefficient (value range 0.98-1.02).

[0068] The model update frequency is dynamically adjusted according to the equipment operating status. The update frequency is 10Hz when the equipment is in operation and 1Hz when it is in standby. After each update, a calibrated 3D dynamic model is output.

[0069] By integrating the detailed modeling capabilities of BIM with the macro-scene representation advantages of GIS, a three-dimensional integrated construction of equipment microstructures and regional macro-environment is achieved. Edge computing nodes are introduced to enable local data processing and real-time model calibration, solving the problem of cloud modeling latency. A pioneering deviation threshold triggering plus state adaptation update mechanism is used to ensure the consistency between the model and the physical scene (deviation ≤0.5%), while reducing resource consumption through differentiated update frequency. Compared with fixed frequency update schemes, resource consumption is reduced by more than 60%.

[0070] The adaptive layered rendering algorithm of S3 specifically includes:

[0071] The 3D dynamic model is divided into four layers according to scene elements: the base layer (terrain, buildings), the equipment layer (main body of hoisting equipment), the key component layer (hook, outriggers, winch mechanism), and the data layer (operating parameters, status indicators);

[0072] Define a rendering priority coefficient, which is obtained by weighted summation of device importance weight (core device 0.6, auxiliary device 0.3), data update frequency weight (high frequency 0.3, low frequency 0.1) and view distance weight (near distance 0.4, far distance 0.1), and the sum of the three weights is 1;

[0073] The rendering precision of each level is adjusted according to the priority coefficient. When the coefficient is not less than 0.4, high-precision rendering with no less than 500,000 triangles is used. When the coefficient is less than 0.4, lightweight rendering with no more than 100,000 triangles is used. The final output is a device distribution visualization interface.

[0074] Breaking through the resource waste bottleneck of traditional 3D rendering with uniform precision across the entire scene, a multi-dimensional dynamic priority evaluation system is established. By quantifying three core influencing factors—equipment importance, data timeliness, and user perspective—intelligent allocation of rendering resources is achieved. This ensures high-precision display of key components (≥500,000 triangles) while simplifying rendering of non-core areas, improving interface response speed by more than 30% and solving the industry visualization problems of unclear details and operational lag.

[0075] The improved LSTM algorithm of S4 is used for device operation trend analysis. The algorithm structure includes an input layer, a forget gate, an input gate, a cell state layer, an output gate, and a fully connected layer. The forget gate uses the ELU activation function, which directly outputs the input value when the input value is greater than 0, and outputs the result of the natural constant raised to the power of the input value minus 1 when the input value is less than or equal to 0.

[0076] The device operation data in the multi-source fusion dataset output by S1 is input into the improved LSTM algorithm in time series order, and the algorithm predicts the changing trend of key device parameters in the next 5 minutes.

[0077] The preset safety rule base is built based on the equipment safety operation procedures and historical fault data. It includes parameter threshold rules, trend anomaly rules and associated fault rules. The parameter prediction results and real-time operation data are substituted into the rule base for matching, and the abnormal equipment status identification results and fault location information are output to trigger the corresponding graded early warning signals.

[0078] By optimizing the LSTM forget gate performance through the ELU activation function, the gradient vanishing problem of traditional LSTM when the parameter fluctuations are small is solved, and the parameter prediction error in the next 5 minutes is controlled within ±3%. The predicted data is jointly matched with real-time data to realize the transformation from post-event alarm to pre-event prediction. Compared with the monitoring scheme that only relies on real-time thresholds, the lead time for anomaly identification is improved.

[0079] The specific rules of the security rule base include:

[0080] Set safety threshold ranges for each operating parameter, obtain the real-time change rate of the device operating parameters, and determine the trend as abnormal when the change rate exceeds the preset critical change rate and the duration exceeds 3 seconds;

[0081] A parameter correlation model is established based on Bayesian network. When the hydraulic system pressure exceeds the threshold and the lifting speed is less than 0.3 times the rated speed, it is determined to be a hydraulic system failure. The probability of failure is calculated by combining historical failure data.

[0082] By introducing a dual criterion of change rate and duration, false alarms triggered by a single threshold are avoided, improving the accuracy of trend anomaly identification. A parameter correlation model is constructed based on a Bayesian network, overcoming the limitations of traditional single-parameter independent judgment. This enables precise location of multi-parameter coupled faults and quantifies the probability of fault occurrence, providing data support for tiered early warning systems. Compared to manual experience-based judgment, the accuracy of fault location is improved.

[0083] The tiered early warning mechanism specifically includes:

[0084] Level 1 warning: When S4 detects that a single parameter exceeds the threshold but does not trigger the associated fault rule, it generates a Level 1 warning signal, outputs the warning prompt through a green pop-up window in the visual interface, and records the abnormal parameter and the time of occurrence.

[0085] Level 2 warning: When S4 detects that multiple parameters exceed the threshold and trigger abnormal trend rules, it generates a level 2 warning signal and pushes the warning information through the sound and light alarm and system SMS.

[0086] Level 3 warning: When S4 identifies that the associated fault rule is triggered and the predicted fault probability is not less than 80%, a level 3 warning signal is generated, the high-risk operation circuit of the equipment is immediately cut off, and the emergency phone number of the management personnel is dialed, and the fault location result and handling flowchart are pushed.

[0087] It also includes equipment operation trajectory tracking function: through GPS and inertial navigation fusion technology, it collects the time series of equipment spatial position data in S1, records the equipment movement trajectory and operation posture, and generates a trajectory dataset containing timestamps, three-dimensional coordinates, rotation angle and pitch angle.

[0088] S5 specifically includes:

[0089] Establish a historical case database, use the cosine similarity algorithm to obtain feature matching degree, retrieve cases from the case database with a feature similarity of not less than 85% with the current warning signal, and extract the handling measures from the cases;

[0090] The security rule base in S4 is called, and the extracted measures are optimized and adjusted in combination with the fault handling rules to generate three candidate handling solutions.

[0091] The analytic hierarchy process (AHP) was used to evaluate candidate solutions. Evaluation indicators included disposal time (weight 0.4), disposal cost (weight 0.3), and safety risk (weight 0.3). The comprehensive score of each solution was calculated, and the optimal disposal solution with the highest comprehensive score was output.

[0092] It also includes a device health assessment function:

[0093] Collect equipment operation data from S1, fault records from S4, and equipment maintenance records to construct a health assessment model. The formula for the health assessment model is:

[0094]

[0095] In the formula, The running status score is calculated based on the degree to which parameters deviate from the threshold, with a maximum score of 100. The score is based on the frequency of failures (100 points for no failures, and 30 points deducted for each Level 3 failure). To maintain the target score (calculated based on the maintenance plan completion rate, with a maximum score of 100 points). For health status;

[0096] according to The value is used to classify the device health as excellent. ,good ,generally Poor There are four health levels. When the health status is poor, a maintenance reminder will be automatically pushed, and a maintenance plan including maintenance items and suggested time will be generated.

[0097] A three-dimensional health assessment model for operation, failure, and maintenance is constructed. Through quantitative formulas, the health status of equipment is accurately represented, breaking through the passive mode of traditional post-failure maintenance. Based on the health level, proactive maintenance reminders are triggered, realizing the transformation from post-maintenance to predictive maintenance. Compared with existing technologies, unplanned downtime is reduced, equipment maintenance costs are lowered, and the value of the equipment throughout its entire life cycle is extended.

[0098] In summary, the advantages of this invention are:

[0099] By deploying a sensor network covering the device body, operating environment and control terminal, and combining it with a spatiotemporal alignment algorithm, timestamp matching and missing value filling of multi-source data are achieved, generating a synchronously associated multi-source fusion dataset. This solves the traditional data fragmentation problem, provides a high-quality data foundation for subsequent analysis, improves data synchronization accuracy to the millisecond level, and controls the missing value filling error within ±2%.

[0100] A 3D basic model is built based on digital twin technology. The model parameters are dynamically calibrated by combining multi-source fusion data to ensure that the 3D dynamic model is highly consistent with the physical scene (deviation value ≤0.5%). At the same time, an adaptive layered rendering algorithm is adopted to dynamically adjust the rendering accuracy according to the user's perspective and the importance of the device. This not only ensures the visualization details of core devices and key components (number of triangles ≥500,000), but also reduces the resource consumption of non-critical areas, and improves the response speed of the visualization interface by more than 30%.

[0101] By using an improved LSTM algorithm to predict the trend of key equipment parameters within the next 5 minutes (prediction error ±3%), and combining it with a preset safety rule base (including parameter thresholds, trend anomalies, and associated fault rules), the system achieves a shift from post-event warning to pre-event prediction. The graded warning mechanism triggers differentiated handling based on the level of anomaly (e.g., immediately cutting off high-risk circuits in a level 3 warning), improving fault location accuracy to the component level and shortening the warning response time to less than 1 second, effectively reducing the failure rate and accident risk.

[0102] The solution generation method combines case-based reasoning with rule-based reasoning. It generates the optimal solution based on a historical case library (similarity ≥ 85%) and a security rule library, avoiding reliance on human experience. At the same time, the solution results are fed back to the data collection stage, forming a closed loop of collection, monitoring, diagnosis, solution and feedback. The efficiency of handling similar problems is improved by 40%, and the standardization of management processes is significantly improved.

[0103] By integrating equipment operation data, fault records, and maintenance information through a health assessment model, the health status of equipment is quantified and divided into four levels. When the health status is poor, maintenance reminders are automatically pushed, potential fault risks are predicted in advance, unplanned downtime is reduced by more than 25%, equipment maintenance costs are reduced by 15% to 20%, and the service life of equipment is extended.

[0104] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for hoisting equipment distribution visualization and monitoring, characterized in that, The implementation process is as follows: S1, through the sensor network acquisition equipment operation parameter, spatial position data, environmental perception data and operation instruction data, using time space alignment algorithm to collect multi-source data time stamp matching and missing value filling, generate synchronous correlation multi-source fusion data set; S2, based on digital twin technology, combined with the actual terrain of hoisting operation area, building layout and equipment parameters, construct hoisting operation area three-dimensional basic model; S3, using adaptive hierarchical rendering algorithm, the three-dimensional dynamic model is divided into basic layer, equipment layer, key component layer and data layer, combined with user operation view angle, equipment importance level and data update frequency, calculate the rendering priority coefficient of each level, according to the priority coefficient adjusts the rendering precision of each level, generates equipment distribution visualization interface; S4, the equipment operation data in the multi-source fusion data set is input into the improved LSTM algorithm, the output parameter prediction result is output, the equipment abnormal state is identified and the fault position is located, according to the abnormal level, the corresponding hierarchical warning signal is triggered; S5, based on the abnormal type and fault position corresponding to the warning signal, generate targeted disposal scheme.

2. The method of claim 1, wherein, The space-time alignment algorithm in S1 is as follows: The time stamp of the equipment GPS positioning data is matched with the time stamp of the sensor data, and the linear interpolation method is used to fill the data missing gap, and the alignment formula is: In the formula, alignment data of the target time point , , original collection data of the time point , , respectively The collected equipment operation parameters include lifting weight, lifting speed, rotation angle, hydraulic system pressure and engine speed, the environmental perception data includes wind speed, operation area illumination intensity and ground settlement.

3. The method of claim 1, wherein, S2 specifically includes: BIM+GIS fusion technology is used to construct the basic three-dimensional scene, the CAD model and technical parameters of hoisting equipment are imported, and the appearance, internal structure and joint motion of the equipment are restored; Through the edge computing node, the multi-source fusion data set output by S1 is received in real time, the equipment position, attitude parameter and corresponding parameter of three-dimensional basic model in the data are extracted and compared, the model update deviation value is obtained, when the deviation value is greater than 0.5%, the original model parameter is modified combined with scene adaptation coefficient, the model update frequency is dynamically adjusted according to the equipment running state, and the calibrated three-dimensional dynamic model is output after each update.

4. The method of claim 1, wherein, The adaptive hierarchical rendering algorithm of S3 specifically includes: The three-dimensional dynamic model is divided into four levels of basic layer, equipment layer, key component layer and data layer according to scene elements; Define rendering priority coefficient, which is obtained by weighted sum of equipment importance weight, data update frequency weight and view angle distance weight, adjust the rendering precision of each level according to the priority coefficient, and finally output the equipment distribution visualization interface.

5. The hoisting equipment distribution visualization and monitoring method of claim 1, characterized in that: The improved LSTM algorithm of S4 is used for equipment operation trend analysis, and the algorithm structure includes input layer, forgetting gate, input gate, cell state layer, output gate and full connection layer, wherein the forgetting gate adopts ELU activation function; The equipment operation data in the multi-source fusion data set output by S1 is input into the improved LSTM algorithm according to time sequence, and the change trend of the key parameters of the equipment is predicted; The preset safety rule base is constructed based on equipment safety operation procedures and historical failure data, and includes parameter threshold rules, trend anomaly rules and associated failure rules. The parameter prediction result and real-time operation data are substituted into the rule base for matching, and an equipment abnormal state recognition result and failure positioning information are output, triggering a corresponding hierarchical early warning signal.

6. The method of hoist equipment distribution visualization and monitoring according to claim 5, characterized in that, The specific rules of the safety rule base include: A safety threshold interval of each operation parameter is set, and a real-time change rate of the equipment operation parameter is obtained. When the change rate exceeds a preset critical change rate and the duration exceeds 3 seconds, it is determined that a trend anomaly occurs. A parameter association model is established based on a Bayesian network. When the hydraulic system pressure exceeds the threshold value and the lifting speed is less than 0.3 times the rated speed at the same time, it is determined that a hydraulic system failure occurs, and a failure probability is calculated based on historical failure data.

7. The method of claim 1, wherein, The hierarchical early warning mechanism specifically includes: First-level warning: When S4 identifies that a single parameter exceeds the threshold value but does not trigger the associated failure rule, a first-level warning signal is generated, and a warning prompt is output through a visual interface green pop-up window, and the abnormal parameter and the occurrence time are recorded; Second-level warning: When S4 identifies that multiple parameters exceed the threshold value and triggers the trend anomaly rule, a second-level warning signal is generated, and warning information is pushed through a sound and light alarm and a system SMS; Third-level warning: When S4 identifies that the associated failure rule is triggered and the predicted failure probability is not less than 80%, a third-level warning signal is generated, the device high-risk operation loop is immediately cut off, the management personnel emergency phone is dialed, and the failure positioning result and the disposal process chart are pushed.

8. The method of claim 1, wherein, It also includes a device operation trajectory tracing function: through GPS and inertial navigation fusion technology, time series of device spatial position data in S1 are collected, device movement trajectory and operation posture are recorded, and a trajectory data set containing time stamp, three-dimensional coordinates, rotation angle and pitch angle is generated.

9. The method of claim 1, wherein, S5 specifically includes: A historical disposal case library is established, a cosine similarity algorithm is used to obtain a feature matching degree, a case with a similarity degree not less than 85% to the current warning signal feature is retrieved from the case library, and disposal measures in the case are extracted; The safety rule base in S4 is called, and the extracted measures are optimized and adjusted in combination with the failure disposal rules to generate a candidate disposal scheme; An analytic hierarchy process is used to evaluate the candidate scheme, the evaluation indexes include disposal time, disposal cost and safety risk, a comprehensive score of each scheme is calculated, and an optimal disposal recommendation scheme with the highest comprehensive score is output.

10. The method of claim 1, wherein, It also includes a device health degree evaluation function: Device operation data in S1, failure records in S4 and device maintenance records are collected, and a health degree evaluation model is constructed, wherein the health degree evaluation model formula is: wherein is an operational status score, is a failure frequency score, is a maintenance compliance score, is a health score; According to The device health degree is divided into four levels of excellent, good, general and poor, and when the health degree is poor, a maintenance reminder is automatically pushed, and a maintenance plan including maintenance items and recommended time is generated.

Images