Tower crane safety monitoring method based on Internet of Things

Abstract

The invention provides a tower crane safety monitoring method based on the Internet of Things, and relates to the technical field of artificial intelligence and machine learning, and the tower crane safety monitoring method comprises the following steps: based on load long-term data, performing prediction through an LSTM model to obtain a prediction result; performing real-time adjustment through a linear regression model based on load short-term data, environment data and the prediction result to obtain a load prediction value; and based on the load prediction value, comparing the load prediction value with a preset safety threshold, and if the prediction value is greater than the preset safety threshold, triggering early warning.

Description

Technical Field

[0001] This application belongs to the technical field of artificial intelligence and machine learning, and specifically relates to a safety monitoring method for tower cranes based on the Internet of Things. Background Art

[0002] In recent years, with the rapid development of the construction industry, tower cranes, as one of the core equipment at the construction site, have gradually introduced intelligent technologies such as sensor networks, the Internet of Things (IoT), and artificial intelligence (AI) to achieve real-time monitoring of parameters such as load, tilt angle, and wind speed. The application of these technologies has significantly improved the safety and operating efficiency of tower cranes.

[0003] Existing tower crane safety monitoring systems mainly include functions such as data collection, processing, and warning. For example, Patent CN201910809774.1 discloses a method for realizing tower crane operation monitoring and alarm by comparing operator information with data stored in the server by the main control computer; Patent CN201120455834.3 proposes a high-precision positioning system based on GPS modules and RFID tags for monitoring the distance between tower cranes and providing warning signals.

[0004] Although the above technologies have improved the safety of tower cranes to a certain extent, there are still some deficiencies:

[0005] The construction site environment is complex and changeable, and sensor data is easily interfered with, affecting the reliability of the system. Although deep learning algorithms such as LSTM can effectively process complex time series data, their computational complexity is high and requires high-performance hardware support, increasing costs and energy consumption. It is difficult to meet the requirements of high-precision prediction and rapid response simultaneously using a single algorithm, and it performs poorly in the case of scarce data or rapid environmental changes. Existing systems fail to dynamically adjust the model according to different working scenarios, resulting in poor adaptability. Summary of the Invention

[0006] This application provides a safety monitoring method for tower cranes based on the Internet of Things to solve one of the above technical problems.

[0007] The technical solution adopted by this application is as follows:

[0008] An embodiment of this application provides a safety monitoring method for tower cranes based on the Internet of Things, including:

[0009] Based on long-term load data, predict through the LSTM model to obtain a prediction result;

[0010] Based on short-term load data, environmental data, and the prediction result, perform real-time adjustment through a linear regression model to obtain a load prediction value;

[0011] Based on the predicted load value, compare it with a preset safety threshold. If the predicted value is greater than the preset safety threshold, an alarm is triggered.

[0012] According to an embodiment of the present application, the prediction is performed through an LSTM model based on long-term load data to obtain a prediction result, specifically:

[0013] The load sensor continuously records the load weight data of the tower crane, converts the original signal into a digital signal, stores it in the form of a time series, and forms a continuous historical data set;

[0014] The original data is processed by filtering, denoising, and standardization to eliminate outliers and environmental interference;

[0015] The LSTM model processes the input sequence layer by layer according to the time step. Based on the statistical law of long-term data, the LSTM model outputs the load prediction result within a certain future time window.

[0016] According to an embodiment of the present application, the real-time adjustment is performed through a linear regression model based on short-term load data, environmental data, and the prediction result to obtain a load predicted value, specifically:

[0017] Obtain the real-time load weight data in the short term from the load sensor;

[0018] Synchronously collect the real-time data of the environmental sensor;

[0019] Use the future load change trend predicted by the LSTM model as the benchmark input. The prediction result captures the complex patterns of long-term historical data;

[0020] The linear regression model takes the short-term load data, environmental data, and the LSTM prediction result as input features, performs weighted calculation through preset weight allocation, and comprehensively evaluates the contribution of each factor to the final predicted value;

[0021] Dynamically adjust the weight ratio according to the changes in the current environment, and the linear regression model outputs the adjusted load predicted value.

[0022] According to an embodiment of the present application, the comparison is made between the load predicted value and a preset safety threshold. If the predicted value is greater than the preset safety threshold, an alarm is triggered, specifically:

[0023] Use the load predicted value corrected by the linear regression model as the core input, and read the preset safety threshold from the configuration file;

[0024] Make a real-time comparison between the predicted value and the threshold at each time point;

[0025] The system scans the predicted values at all time points to identify whether there are continuous overlimits or instantaneous peaks;

[0026] According to the degree of overlimit and the duration, different levels of warning signals are triggered.

[0027] According to an embodiment of the present application, the warning signals include:

[0028] Primary warning: When the predicted value is close to the threshold but does not exceed it, prompt the operator to pay attention and observe;

[0029] Intermediate warning: When the predicted value first exceeds the threshold, require an immediate check of the load distribution or the operating environment;

[0030] Advanced warning: When the predicted value significantly exceeds the limit and continues to rise, force the crane to pause operation and lock the control handle.

[0031] According to an embodiment of the present application, the environmental data includes: wind speed data and tilt angle data.

[0032] According to an embodiment of the present application, it further includes:

[0033] Real-time monitor the wind speed data. If the wind speed data is greater than 14 m / s, trigger a first-level warning, and the first-level warning is to limit the amplitude and it is recommended to pause non-essential operations;

[0034] If the wind speed data is greater than 17 m / s, trigger a second-level warning, and the second-level warning is to lock the slewing mechanism and forcefully adjust the boom in the downwind direction;

[0035] If the wind speed data is greater than 21 m / s, trigger a third-level warning, and the third-level warning is to cut off the power supply and start the wind prevention plan.

[0036] According to an embodiment of the present application, real-time monitor the tilt angle data. If the tilt angle data is greater than 0.5°, trigger a first-level alarm, and the first-level alarm is to prompt to check the foundation stability;

[0037] If the tilt angle data is greater than 1.0°, trigger a second-level alarm, and the second-level alarm is to close the brake and lock the turntable pin;

[0038] If the tilt angle data is greater than 2.0°, trigger a third-level alarm, and the third-level alarm is to trigger an emergency stop and isolate the equipment power supply.

[0039] A computer-readable storage medium, on which a program is stored, characterized in that when the program is executed by a processor, the steps in the method are implemented.

[0040] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, the steps in the method are implemented.

[0041] Due to the adoption of the above technical solution, the beneficial effects obtained by this application are as follows:

[0042] This application utilizes the powerful non-linear modeling ability of the LSTM model, which can capture the complex patterns of the load changing over time, thereby providing more accurate long-term trend predictions. By learning historical data, LSTM can discover potential risk points in advance, providing a scientific basis for operators to take preventive measures. Compared with LSTM, the linear regression model is simple and fast to calculate, suitable for real-time data processing, reducing the computational burden. Combining short-term data and environmental factors for real-time adjustment enables the system to respond quickly when the environmental conditions change, enhancing the system's dynamic adaptability. Through reasonable weight allocation, comprehensively considering the impacts of various input features, it ensures that the final prediction result is closer to the actual situation. Timely comparison of the predicted value with the safety threshold can issue a warning before exceeding the limit, guiding the operator to take corresponding measures to avoid the occurrence of safety accidents. Triggering different levels of early warnings according to different levels of over-limit conditions can neither overly interfere with normal operations nor take emergency measures when necessary, ensuring on-site safety. Through the prompt of the warning signal, it helps managers better grasp the equipment status, formulate a reasonable maintenance plan, and extend the service life of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The drawings described herein are used to provide a further understanding of this application and constitute a part of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0044] Figure 1 It is a schematic flow chart of a tower crane safety monitoring method based on the Internet of Things provided by an embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] In order to more clearly illustrate the overall concept of this application, the following will be described in detail by way of examples in combination with the drawings of the specification.

[0046] Many specific details are set forth in the following description in order to provide a thorough understanding of this application. However, this application may be implemented in other ways different from those described herein. Therefore, the protection scope of this application is not limited by the specific embodiments disclosed below. It should be noted that, without conflict, the embodiments of this application and the features in each embodiment may be combined with each other.

[0047] In this application, unless otherwise clearly specified and defined, the first feature being "on" or "under" the second feature may be that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0048] Embodiment 1

[0049] As Figure 1 shown, a safety monitoring method for tower cranes based on the Internet of Things includes:

[0050] Based on the long-term load data, prediction is performed through the LSTM model to obtain a prediction result.

[0051] As described above, the process of predicting through the LSTM model based on the long-term load data is mainly to accurately predict the change in the lifting weight of the tower crane in the future. LSTM, that is, Long Short-Term Memory network, is a special recurrent neural network (RNN) that can learn long-term dependencies and is particularly suitable for processing complex data such as time series.

[0052] In this application, first, the lifting weight data of the tower crane per minute in the past period (for example, within 24 hours) will be collected to form a time series data set. These data reflect the actual working load conditions of the tower crane at different time points. Then, the LSTM model is used to train these historical data, aiming to let the model learn to identify and understand the patterns and rules of the change in the lifting weight over time.

[0053] Once the model training is completed, it can be used to predict the future change in the lifting weight. Specifically, given the latest load data as input, the LSTM model can output the predicted value of the lifting weight in a future time period (for example, within the next 15 minutes). This prediction is crucial for early detection of potential overloading risks because it allows the operator to take preventive measures before actual problems occur, such as adjusting the operation plan or stopping certain operations, thus avoiding the occurrence of safety accidents.

[0054] In addition, to improve the prediction accuracy and reduce the computational burden, a method of combining LSTM with a linear regression model is also used. Under this method, LSTM is mainly used to capture long-term trends, while linear regression is used to correct the prediction results in real time, especially when encountering rapid environmental changes (such as a sudden increase in wind speed). This method not only improves the overall prediction accuracy but also enhances the system's response speed and robustness.

[0055] For example, imagine a construction site where a tower crane conducts hoisting operations every day. To ensure safety and prevent potential overloading risks, the system needs to predict the future changes in the crane's lifting weight.

[0056] First, the system collects the lifting weight data for each minute within the past 24 hours, which means there will be a total of 1440 data points (since there are 1440 minutes in a day). These data reflect the actual working load conditions of the tower crane at different time points, including weight changes under various operating conditions.

[0057] Next, these historical data are used to train the LSTM model. LSTM is a deep learning algorithm that can learn and remember long-term sequence data patterns and is particularly suitable for dealing with time series prediction problems like this. Through training, the LSTM model learns to identify and understand the patterns and regularities of the lifting weight changing over time.

[0058] Once the model is trained, it can be used to predict the changes in the lifting weight within a certain future time period. For example, if you want to predict the changes in the lifting weight of the tower crane within the next 15 minutes, you can input the data from the past 24 hours into the trained LSTM model. The model will output the predicted lifting weight values for each minute within the next 15 minutes based on the patterns and regularities learned previously.

[0059] In this example, assume that the LSTM model predicts that the lifting weight of the tower crane will gradually increase within the next 15 minutes, rising from the current 6.2 tons to 7.7 tons. This prediction result can help the operator understand the future working load trend of the crane in advance and take necessary measures, such as adjusting the operation plan or suspending certain operations, to avoid possible overloading situations and thus improve the construction safety.

[0060] It is worth noting that in practical applications, to further improve the prediction accuracy and real-time response ability, this document also mentions combining a linear regression model to correct the prediction results of LSTM in real time. This method can reduce the computational complexity while ensuring the prediction accuracy and enhance the overall performance of the system. However, the main focus here is on the basic process of using long-term load data for prediction through the LSTM model.

[0061] Based on the short-term load data, environmental data, and the said prediction results, perform real-time adjustment through a linear regression model to obtain the load prediction value.

[0062] As described above, the process of using a linear regression model for real-time adjustment based on short-term load data, environmental data, and the prediction results obtained through the LSTM model is to improve the accuracy and response speed of load prediction. This process combines long-term trend prediction (provided by LSTM) and short-term precise adjustment (achieved through linear regression) to better adapt to the rapidly changing conditions at the construction site.

[0063] First, the system collects the hoisting weight data within the current 1 hour as short-term data, and obtains real-time environmental data such as wind speed and the inclination angle of the crane. These short-term data reflect the specific changes in the working load of the tower crane within a short period of time, while the environmental data provides information on external factors affecting the operating state of the crane.

[0064] Then, combine these short-term data and environmental data with the future load prediction value previously obtained by the LSTM model. Here, the LSTM model provides a basic prediction of the change in hoisting weight within a future period of time (such as the next 15 minutes). However, due to the sudden changes in the environment and operating conditions at the construction site, relying solely on the prediction of LSTM may not be sufficient to handle all situations.

[0065] Therefore, a linear regression model is introduced for real-time adjustment. The linear regression model uses the current short-term data (i.e., the hoisting weight data within the most recent 1 hour), environmental data (such as wind speed and inclination angle), and the basic prediction value given by the LSTM model as inputs. In this way, linear regression can fine-tune the prediction of LSTM, taking into account the latest actual situation on-site, so as to provide a more accurate load prediction.

[0066] Specifically, if the construction site suddenly encounters strong winds or other factors that may cause load changes, the linear regression model can quickly respond according to the latest short-term data and environmental data, and correct the prediction result of LSTM. The purpose of doing this is to ensure that the final load prediction value is as close as possible to the actual situation, enabling the operator to take timely measures to avoid potential safety risks.

[0067] In summary, this method combines the advantages of the two models: LSTM is good at capturing complex patterns in long time series, while linear regression can quickly respond to short-term changes. The combination of the two not only improves the prediction accuracy but also enhances the real-time response ability of the system. This makes the operation of the tower crane safer and more reliable, and can effectively operate in a complex and changeable working environment.

[0068] For example, imagine that at a construction site, a tower crane is performing its daily lifting operations. To ensure the safety of the operation, the system needs to monitor and predict the load changes of the crane in real time. Here, a specific example is used to illustrate how to make real-time adjustments through a linear regression model based on short-term load data, environmental data, and the prediction results of the LSTM model, so as to obtain a more accurate load prediction value.

[0069] Suppose the current time is exactly 10:00 am. The system has collected the lifting weight data within the one-hour period from 9:00 am to 10:00 am as short-term data, and obtained real-time environmental data, including the current wind speed and the tilt angle of the crane. At the same time, the LSTM model has previously predicted the lifting weight within the next 15 minutes (i.e., from 10:00 to 10:15) based on the data of the past 24 hours.

[0070] Based on this, the system inputs this information into the linear regression model for real-time adjustment. Specifically, the linear regression model takes into account the following factors: Firstly, the average value of the lifting weight recorded every minute within the one-hour period from 9:00 am to 10:00 am. Secondly, the current wind speed and the tilt angle of the crane. Finally, the basic prediction value for the next 15 minutes given by the LSTM model.

[0071] For example, if the LSTM model predicts that in the next 1st minute (i.e., from 10:00 to 10:01), the lifting weight of the crane will be 6.2 tons. However, considering that the actual average value of the lifting weight within the recent one hour is 5.5 tons, the relatively high current wind speed may affect the lifting weight, making the actual value slightly lower than the predicted value. At the same time, if the crane has a certain tilt angle, this may also affect its load-bearing capacity. Therefore, the linear regression model will fine-tune the predicted value of the LSTM according to these actual situations.

[0072] In this case, the linear regression model may calculate a corrected predicted value, such as 5.97 tons, which means that considering the current specific conditions, it is expected that the actual lifting weight of the crane in the next first minute is closer to this value, rather than the 6.2 tons directly predicted by the LSTM.

[0073] In this way, based on the long-term trend prediction provided by the LSTM, the linear regression model can use the latest short-term data and environmental information to immediately adjust the prediction, thereby improving the accuracy and response speed of the prediction. This is particularly important for the rapidly changing situations at the construction site, as it can help the operators timely understand the real working state of the crane and take necessary safety measures, such as adjusting the operation plan or suspending certain operations, to avoid potential safety risks.

[0074] Based on the predicted load value, compare it with a preset safety threshold. If the predicted value is greater than the preset safety threshold, a warning is triggered.

[0075] As described above, the process of comparing the predicted load value with the preset safety threshold is a crucial step in ensuring the safe operation of tower cranes. This process aims to prevent situations that may lead to safety accidents, such as overloading, tilting, or excessive wind speed, through a real-time monitoring and warning system.

[0076] First, the system obtains a predicted load value according to the methods mentioned above (combining the long-term trend prediction of the LSTM model and the real-time adjustment of the linear regression model). This predicted value reflects the expected working load of the tower crane within a certain future time period, such as the change in the lifting weight within the next 15 minutes.

[0077] Next, the system compares this predicted value with a pre-set safety threshold. The safety threshold is determined based on the design specifications of the tower crane, safety standards, and specific conditions at the construction site. These thresholds include, but are not limited to, the maximum allowable lifting weight, the maximum tilting angle, and the maximum ambient wind speed, etc. They represent the limit values for the safe operation of the tower crane under various conditions.

[0078] If the predicted load value of the system exceeds these preset safety thresholds, it indicates that the crane may be in an unsafe working state at some future moment. For example, if the predicted lifting weight exceeds the maximum allowable lifting weight of the crane, or the predicted tilting angle exceeds the safe range, or the ambient wind speed exceeds the safety operation standard, then the system will trigger the warning mechanism.

[0079] Once the warning is triggered, the system will immediately notify relevant personnel to take measures. This is usually achieved through a mobile application (APP), which can display the monitored data in real time and issue an alarm when potential risks are detected. Staff can receive specific warning messages through the APP, such as warnings of overweight lifting weight, abnormal tilting angle, or excessive ambient wind speed. In this way, the operator can timely understand the problem and take corresponding actions to avoid accidents, such as stopping the current operation, adjusting the lifting plan, or taking other necessary safety measures.

[0080] This warning mechanism based on the comparison of the predicted value with the safety threshold greatly improves the safety and reliability of the tower crane operation, enabling potential risks to be identified and handled before they actually occur.

[0081] For example, assume that at a construction site, a tower crane is performing a hoisting operation. The system has obtained the predicted load weight value within the next 15 minutes according to the method described above (combining the long-term trend prediction of the LSTM model and the real-time adjustment of the linear regression model).

[0082] In this example, the maximum allowable load weight of the tower crane is set at 6 tons as a safety threshold. Through algorithm prediction, the system predicts that within the next 5th minute, the load weight of the tower crane will reach 6.7 tons. This means that according to the current data analysis and model prediction, the load weight of the crane at the 5th minute will exceed the pre-set safety threshold (6 tons).

[0083] When this predicted value (6.7 tons) is compared with the pre-set safety threshold (6 tons), it is found that the predicted value is greater than the safety threshold. At this time, the system will automatically trigger the warning mechanism. The warning signal will not only be displayed on the operation interface of the crane, but also be sent to relevant management personnel or on-site operators through a mobile application (APP). For example, the APP may pop up a notification informing the operator: "It is predicted that in 5 minutes, your load weight will reach 6.7 tons, exceeding the safety limit. Please take corresponding measures."

[0084] After receiving the warning, the operator can immediately take actions to avoid potential safety risks. Possible measures include suspending the current hoisting operation, checking whether the cargo weight actually exceeds the safety range, redistributing the load, or adjusting the operation plan to ensure that the maximum allowable load is not exceeded. This warning mechanism based on the comparison of the predicted value and the safety threshold enables the operator to take preventive measures before potential problems occur, thus effectively improving the safety of the construction site and reducing the risk of accidents.

[0085] According to an embodiment of the present application, based on the long-term load data, prediction is performed through the LSTM model to obtain a prediction result, specifically:

[0086] The load sensor continuously records the load weight data of the tower crane, converts the original signal into a digital signal, stores it in the form of a time series, and forms a continuous historical data set;

[0087] The original data is subjected to filtering, denoising, and normalization processing to eliminate outliers and environmental interference;

[0088] The LSTM model processes the input sequence layer by layer according to the time step. The LSTM model outputs the load prediction result within a certain future time window based on the statistical law of long-term data.

[0089] As described above, the process of making predictions through the LSTM model based on long-term load data involves several key steps, aiming to accurately predict the changes in the lifting load of tower cranes in the future using historical data. The following is a detailed explanation:

[0090] First, the load sensor continuously records the lifting load data of the tower crane during operation. These original signals represent the changes in the actual load borne by the crane and are converted into digital signals for further processing and analysis. These digital signals are arranged in chronological order to form a continuous historical dataset. This dataset contains the lifting load values per minute of the crane over a certain period in the past (e.g., 24 hours), reflecting the changing trend of its working load.

[0091] Next, to improve data quality and ensure prediction accuracy, the raw data needs to go through a series of preprocessing steps. First is the filtering and denoising process, which aims to remove data fluctuations caused by sensor errors or environmental interference to obtain a smoother and more reliable data sequence. Subsequently is the normalization process, through which the data scale is adjusted so that data of different magnitudes can be compared on the same scale, and at the same time, the influence of outliers can be eliminated. The purpose of this is to ensure that the data input into the model is as clean and representative as possible, reducing unnecessary error sources.

[0092] Finally, the preprocessed data is fed into the LSTM model for analysis. LSTM is a deep learning algorithm particularly suitable for processing time series data and can identify and remember data patterns over a long time span. In this application, the LSTM model processes the input time series data layer by layer according to the time step based on the changing pattern of the lifting load in the past period. In this way, LSTM can capture the complex dynamic characteristics of the lifting load changing over time and output the load prediction results within a certain future time window. For example, it may predict the changes in the lifting load per minute within the next 15 minutes. This prediction is crucial for early detection of potential overloading risks as it allows the operator to take preventive measures before problems occur, thus avoiding safety accidents.

[0093] According to an embodiment of the present application, based on the short-term load data, environmental data, and the prediction results, real-time adjustment is performed through a linear regression model to obtain a load prediction value, specifically:

[0094] Obtain the real-time lifting load data in the short term from the load sensor;

[0095] Synchronously collect the real-time data of the environmental sensor;

[0096] Taking the future load change trend predicted by the LSTM model as the benchmark input, the prediction result captures the complex patterns of long-term historical data;

[0097] The linear regression model takes the short-term load data, environmental data, and the LSTM prediction result as input features, and performs weighted calculations through preset weight allocation to comprehensively evaluate the contribution of each factor to the final predicted value;

[0098] According to the changes in the current environment, the weight ratio is dynamically adjusted, and the linear regression model outputs the adjusted load prediction value.

[0099] As described above, the process of obtaining a more accurate load prediction value through real-time adjustment by the linear regression model based on the short-term load data, environmental data, and the prediction result of the LSTM model is a comprehensive analysis method that combines long-term trends and short-term changes. The following is a detailed explanation:

[0100] First, the system obtains the real-time suspended load weight data in the short term (e.g., the most recent 1 hour) from the load sensor. These short-term data reflect the specific changes in the working load of the tower crane in a short period of time and provide the latest operation status information.

[0101] At the same time, the system also synchronously collects real-time data from environmental sensors, such as wind speed and the inclination angle of the crane. These environmental data provide an important basis for understanding the impact of the current operating environment on the crane operation. For example, strong winds may affect the stability of the crane, and a large inclination angle may be a sign of potential risks.

[0102] Next, the prediction of the future load change trend by the LSTM model is used as the benchmark input. Through learning historical data, the LSTM model can capture the complex patterns and long-term trends of the suspended load weight changing over time. This means that the prediction result provided by the LSTM model not only considers the past data patterns but also predicts the general trend in a future period.

[0103] Then, the linear regression model is used to integrate the short-term load data, environmental data, and the prediction result of the LSTM model. In the linear regression model, these different input features are assigned specific weights, which reflect their different degrees of contribution to the final load prediction value. For example, if the current wind speed increases significantly, the weight of the wind speed factor may increase accordingly because this indicates that the environmental conditions have a greater impact on the crane operation.

[0104] Specifically, according to the changes in the current environment, the linear regression model can dynamically adjust the weight ratios of various input features. This means that when the environment or operating conditions change, the model can respond flexibly and re-evaluate the importance of various factors, thereby providing an adjusted load prediction value. This adjustment ensures that the prediction is not only based on the trends of historical data but also can quickly respond to the immediate changes in actual operations.

[0105] In summary, this method utilizes the powerful prediction ability of the LSTM model for long-term trends and combines the sensitivity of the linear regression model to short-term fluctuations and environmental changes. The purpose of this is to more accurately reflect the actual working state of the tower crane, help operators promptly understand potential risks, and take necessary preventive measures to improve the safety of the construction site.

[0106] According to an embodiment of the present application, based on the load prediction value, it is compared with a preset safety threshold. If the prediction value is greater than the preset safety threshold, an early warning is triggered. Specifically:

[0107] Take the load prediction value corrected by the linear regression model as the core input, and read the preset safety threshold from the configuration file;

[0108] The prediction value and the threshold are compared in real time at each time point;

[0109] The system scans the prediction values at all time points to identify whether there are continuous overlimit or instantaneous peaks;

[0110] According to the degree of overlimit and the duration, different levels of early warning signals are triggered.

[0111] As described above, the process of comparing the load prediction value with the preset safety threshold aims to ensure the safety of the tower crane operation through real-time monitoring and trigger corresponding early warnings when potential risks are detected. The following is a detailed explanation:

[0112] First, the system takes the load prediction value corrected by the linear regression model as the core input. This means that using the method mentioned above (combining the long-term trend prediction of the LSTM model and the real-time adjustment of the linear regression model), the system has obtained the prediction value of the change in the hoisting weight of the tower crane for a period of time in the future (for example, the next 15 minutes).

[0113] Next, the system reads the preset safety threshold from the configuration file. These safety thresholds are preset according to the design specifications of the tower crane, safety standards, and specific conditions of the construction site, representing the limit values for the safe operation of the crane. For example, the maximum allowable hoisting weight may be set to 6 tons.

[0114] Then, the system will perform a real-time comparison of the predicted values at each time point with the corresponding safety thresholds point by point. This means that for the predicted values of each future minute, the system will check whether it exceeds the safety threshold at the corresponding time point.

[0115] Based on this, the system will conduct a comprehensive scan of the predicted values at all time points to identify whether there are situations of continuous overlimit or instantaneous peaks. If it is found that the predicted value at any time point exceeds the safety threshold, whether it is continuously exceeded or an instantaneous peak, the system will mark these abnormal situations.

[0116] Finally, according to the degree of overlimit and the duration, the system will trigger warning signals at different levels. For example, if it only slightly exceeds the threshold briefly, a lower-level warning may be triggered; while if it significantly exceeds the threshold and lasts for a long time, a higher-level emergency alarm will be triggered. This hierarchical warning mechanism helps the operator accurately judge the severity of the problem and take appropriate measures to deal with it. For example, after receiving the warning, the operator can decide whether to immediately stop the current operation, check the load distribution, or take other necessary safety measures to avoid potential safety accidents.

[0117] This method ensures that the tower crane can operate safely in a complex construction environment. By warning of potential risks in a timely manner, it greatly improves the safety and reliability of the operation.

[0118] According to an embodiment of the present application, the warning signal includes:

[0119] Primary warning: When the predicted value is close to the threshold but does not exceed it, it prompts the operator to pay attention and observe.

[0120] Intermediate warning: When the predicted value first exceeds the threshold, it requires an immediate check of the load distribution or the operating environment.

[0121] High-level warning: When the predicted value significantly exceeds the limit and continues to rise, it forcibly suspends the operation of the crane and locks the control handle.

[0122] As described above, when the load predicted value of the system is close to but does not exceed the preset safety threshold, the system will issue a primary warning. The main purpose of this level of warning is to remind the operator to pay attention to the current operating conditions and closely observe the working state of the crane. The primary warning prompts the operator that although there is no immediate danger at present, they need to remain vigilant because the working load is approaching the safety limit. This warning helps the operator make preparations in advance and take preventive measures to avoid potential risks.

[0123] If the predicted load value exceeds the safety threshold for the first time, the system will trigger a medium-level warning. The medium-level warning requires the operator to immediately check the load distribution of the crane or the operating environment. This means that there may be overloading or other factors affecting the normal operation of the crane, such as excessive wind speed or tilt caused by uneven ground. The medium-level warning emphasizes the importance of immediate action. By checking, it can be determined whether it is necessary to adjust the load or improve the operating environment conditions, so as to quickly resume normal operation and prevent accidents.

[0124] The high-level warning is triggered when the predicted load value not only significantly exceeds the safety threshold but also continues to rise. In this case, to ensure absolute safety, the system will forcefully suspend all operating activities of the crane and lock the control handle to prevent further operations. The high-level warning indicates the existence of serious safety hazards, which may pose a direct threat to the equipment and personnel. At this time, all related operations must be immediately stopped until the problem is completely solved. Such strict measures help to minimize the risk of accidents and protect the safety of the lives and property of the on-site personnel.

[0125] According to an embodiment of the present application, the environmental data includes: wind speed data and tilt angle data.

[0126] As mentioned above, the wind speed refers to the speed of the air relative to the ground. In an open environment such as a construction site, the wind speed can significantly affect the operating safety of a tower crane. Strong winds may generate additional lateral forces on the objects lifted by the crane, increasing the risk of the crane losing stability. Therefore, real-time monitoring of the wind speed is crucial for evaluating whether the current operating conditions are suitable for lifting operations. Through the wind speed sensors installed on the crane, the system can continuously collect wind speed information and transmit this data to the central control system for analysis and decision-making.

[0127] The tilt angle refers to the angle change of the tower crane relative to the horizontal plane. Under normal operating conditions, the crane should remain vertical to ensure its structural stability and operating safety. However, due to uneven ground, foundation settlement, or the influence of external forces (such as wind), the crane may tilt. If this tilt is not controlled, it may lead to serious safety accidents, such as the overturning of the crane. By installing high-precision tilt sensors, the system can real-time monitor the tilt state of the crane and feed the data back to the control system so as to take timely measures to adjust or stop the operation, thereby avoiding potential risks.

[0128] According to an embodiment of the present application, it further includes:

[0129] Real-time monitor the wind speed data. If the wind speed data is greater than 14 m / s, trigger a first-level warning, and the first-level warning is a restricted range, and it is recommended to suspend non-essential operations;

[0130] If the wind speed data is greater than 17 m / s, a secondary warning is triggered. The secondary warning is to lock the slewing mechanism and forcibly adjust the boom in the downwind direction;

[0131] If the wind speed data is greater than 21 m / s, a tertiary warning is triggered. The tertiary warning is to cut off the power supply and activate the wind prevention plan.

[0132] As described above, when the wind speed data monitored in real time exceeds 14 meters per second (m / s), the system will trigger a primary warning. This level of warning is mainly to remind the operator that the current wind speed conditions may have a certain impact on the operation, and it is recommended to suspend non-essential operation activities. This warning is intended to limit the operating range of the crane and reduce the risks that may be caused by strong winds, but it does not forcibly stop all operations. It gives the operator the opportunity to assess the current situation and make appropriate adjustments to ensure safety.

[0133] If the wind speed further increases and exceeds 17 meters per second, the system will trigger a secondary warning. In this case, in addition to warning the operator, the system will also automatically lock the slewing mechanism, that is, prevent the rotating part of the crane from continuing to move. At the same time, the system will recommend or forcibly adjust the direction of the boom so that it is placed downwind. This is done to reduce the pressure of the wind on the boom and reduce the risk of the crane tipping over. In this way, the stability of the equipment can be maintained as much as possible even at higher wind speeds.

[0134] When the wind speed reaches or exceeds 21 meters per second, the system will trigger the most serious tertiary warning. At this time, in order to maximize the safety of personnel and equipment, the system will cut off the power supply and completely stop all operations of the crane. In addition, the wind prevention plan will be activated, which may include a series of pre-established safety measures, such as strengthening the crane structure and ensuring the stability of all components, to resist the upcoming strong wind. This level of warning indicates that there is an extremely high risk and immediate action must be taken to avoid possible major accidents.

[0135] Through these three levels of warning mechanisms, the system can automatically respond according to different wind speed conditions, providing gradually upgraded safety measures, from the initial warning to the final emergency shutdown, to ensure the safe operation of the tower crane under various weather conditions. This hierarchical warning strategy not only improves the safety of operation but also enables the operator to make more informed decisions in the face of changing environmental conditions.

[0136] According to an embodiment of the present application, the inclination angle data is monitored in real time. If the inclination angle data is greater than 0.5°, a primary alarm is triggered. The primary alarm is to prompt an inspection of the foundation stability;

[0137] If the tilt angle data is greater than 1.0°, a secondary alarm is triggered, and the secondary alarm is to close the brake and lock the turntable pin;

[0138] If the tilt angle data is greater than 2.0°, a tertiary alarm is triggered, and the tertiary alarm is to trigger an emergency stop and isolate the equipment power supply.

[0139] As described above, when the tilt angle monitored in real time exceeds 0.5 degrees, the system will trigger a primary alarm. At this time, although the degree of tilt is not yet sufficient to immediately endanger safety, it already indicates that there may be some factors affecting the foundation stability, such as uneven ground settlement or problems with the support structure, etc. The primary alarm is mainly to prompt the operator to check the foundation stability of the crane to ensure there are no potential safety hazards. This step is very important because it can help detect problems in advance and avoid the situation from deteriorating.

[0140] If the tilt angle further increases and exceeds 1.0 degree, the system will trigger a secondary alarm. In this case, in addition to issuing a warning, the system will also automatically execute more stringent measures, such as closing the brake and locking the turntable pin. These measures are to prevent the crane from continuing to move or rotate, thereby reducing the risk caused by further tilting. By locking key components, the occurrence of accidents can be effectively avoided, and at the same time, it provides a safe working condition for maintenance personnel to investigate and solve problems.

[0141] When the tilt angle reaches or exceeds 2.0 degrees, the system will trigger the most serious tertiary alarm. This is an extremely dangerous situation, indicating that the crane may be about to overturn or other serious accidents may occur. To maximize the safety of on-site personnel and equipment, the system will automatically trigger an emergency stop procedure and isolate the equipment power supply. This means that all operations will be forced to stop to prevent any operations that may cause accidents from continuing. This extreme measure aims to quickly eliminate the risk source until the problem is completely solved.

[0142] Through these three levels of alarm mechanisms, the system can automatically respond according to the change of the tilt angle, providing gradually upgraded safety measures, from the initial prompt to the final emergency stop, ensuring the highest safety of the tower crane in any situation. This method not only improves the safety of operation but also provides clear guidance for operators and maintenance teams so that they can make quick and effective responses when facing different levels of risks.

[0143] A computer-readable storage medium, on which a program is stored, characterized in that when the program is executed by a processor, the steps in the method are implemented.

[0144] An electronic device, comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein the steps in the method are implemented when the processor executes the program.

[0145] What is not described in this application can be implemented by adopting or referring to the prior art.

[0146] The various embodiments in this specification are described in a progressive manner. For the same or similar parts among the various embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments.

[0147] The above are only the embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. An IoT-based safety monitoring method for tower cranes, characterized in that, Including: Based on long-term load data, perform prediction through the LSTM model to obtain a prediction result; Based on short-term load data, environmental data, and the prediction result, perform real-time adjustment through a linear regression model to obtain a load prediction value; Based on the load prediction value, compare it with a preset safety threshold. If the prediction value is greater than the preset safety threshold, trigger a warning.

2. The method according to claim 1, wherein The performing prediction through the LSTM model based on long-term load data to obtain a prediction result is specifically: The load sensor continuously records the load weight data of the tower crane, converts the original signal into a digital signal, stores it in the form of a time series, and forms a continuous historical data set; The original data undergoes filtering, denoising, and normalization processing to eliminate outliers and environmental interference; The LSTM model processes the input sequence layer by layer according to the time step. The LSTM model outputs the load prediction result within a certain future time window based on the statistical law of long-term data.

3. The method according to claim 1, wherein The performing real-time adjustment through a linear regression model based on short-term load data, environmental data, and the prediction result to obtain a load prediction value is specifically: Obtain the real-time load weight data in the short term from the load sensor; Synchronously collect the real-time data of the environmental sensor; Use the future load change trend predicted by the LSTM model as the benchmark input. The prediction result captures the complex patterns of long-term historical data; The linear regression model takes short-term load data, environmental data, and the LSTM prediction result as input features, performs weighted calculation through preset weight allocation, and comprehensively evaluates the contribution of each factor to the final prediction value; Dynamically adjust the weight ratio according to the changes in the current environment, and the linear regression model outputs the adjusted load prediction value.

4. The method according to claim 1, characterized in that, The comparing the load prediction value with a preset safety threshold and triggering a warning if the prediction value is greater than the preset safety threshold is specifically: Use the load prediction value corrected by the linear regression model as the core input, and read the preset safety threshold from the configuration file; Perform real-time comparison of the prediction value and the threshold at each time point; The system scans the prediction values at all time points to identify whether there are continuous overlimits or instantaneous peaks; Trigger warning signals at different levels according to the degree of overlimit and the duration.

5. The method according to claim 4, characterized in that, The warning signals include: Primary warning: When the prediction value is close to the threshold but does not exceed it, prompt the operator to pay attention and observe; Intermediate warning: When the prediction value first exceeds the threshold, require an immediate inspection of the load distribution or the operating environment; Advanced warning: When the prediction value significantly exceeds the limit and continues to rise, force the crane to suspend operation and lock the control handle.

6. The method according to claim 5, characterized in that, The environmental data includes wind speed data and tilt angle data.

7. The method according to claim 6, wherein Also including: Real-time monitor the wind speed data. If the wind speed data is greater than 14 m / s, trigger a first-level warning, and the first-level warning is to limit the amplitude and it is recommended to suspend non-essential operations; If the wind speed data is greater than 17 m / s, trigger a second-level warning, and the second-level warning is to lock the slewing mechanism and forcefully adjust the boom in the downwind direction; If the wind speed data is greater than 21 m / s, trigger a third-level warning, and the third-level warning is to cut off the power supply and start the wind prevention plan.

8. The method according to claim 6, characterized in that Real-time monitor the tilt angle data. If the tilt angle data is greater than 0.5°, a first-level alarm is triggered, and the first-level alarm is to prompt to check the foundation stability; If the tilt angle data is greater than 1.0°, a second-level alarm is triggered, and the second-level alarm is to close the brake and lock the turntable pin; If the tilt angle data is greater than 2.0°, a third-level alarm is triggered, and the third-level alarm is to trigger an emergency stop and isolate the device power supply.

9. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by a processor, the steps in the method described in any one of claims 1-8 are implemented.

10. An electronic device, comprising a memory, a processor, and a program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, the steps in the method described in any one of claims 1-8 are implemented.

Images