Crane state on-line monitoring system and method

Abstract

The invention belongs to the technical field of cranes, and provides a crane state online monitoring system and method. The method comprises the following steps: analyzing and obtaining the current working state of the crane based on key operation parameters of the crane, wherein the working state comprises a non-lifting state and a lifting state; determining a fault analysis model corresponding to the working state, and switching the fault analysis model into an activated state; determining a plurality of crane state data corresponding to the activated fault analysis model, inputting each crane state data into the activated fault analysis model, and obtaining a fault type and a fault probability predicted by the fault analysis model; and generating and outputting an alarm signal corresponding to the fault type of which the fault probability is higher than the probability threshold. Compared with a traditional single model monitoring mode, the model size of the fault analysis model set in different working states is smaller, the smaller model needs fewer resources and is higher in processing speed during data processing and operation, and then the fault monitoring and early warning speed is increased.

Description

Technical Field

[0001] The present invention relates to the technical field of cranes, and in particular to an online monitoring system and method for crane status. Background Art

[0002] Online monitoring of crane status is a system that uses sensors, data collection and transmission, data analysis and processing technologies to grasp the working status of the crane in real time and prevent failures and accidents.

[0003] There have been some studies in the prior art involving the use of models to identify, predict and alarm abnormalities of cranes. For example, CN112418460A discloses a fault diagnosis method for engineering vehicles, including: correlating the operating data of the vehicle to be diagnosed with the operating data of a sample faulty vehicle, and outputting a correlation degree to be diagnosed; when the correlation degree to be diagnosed is not less than a target correlation degree, determining the characteristics of the fault to be diagnosed based on the operating data of the vehicle to be diagnosed; inputting the characteristics of the fault to be diagnosed into a fault diagnosis model, and outputting a diagnosis result; wherein the fault diagnosis model is trained using the sample fault characteristics of the sample faulty vehicle as samples and using the predetermined fault types corresponding to the sample fault characteristics as sample labels.

[0004] The above-mentioned existing technology uses a single model to identify and predict various types of crane faults. Under this fault monitoring method, the model used needs to be trained with a variety of data, resulting in the final model size being generally large. An excessively large model size will lead to a decrease in the response rate of the model, which is not conducive to more timely warning of early faults or impending faults.

[0005] Therefore, how to further improve the fault monitoring and early warning rate of cranes is a technical problem that needs to be solved or improved. Summary of the invention

[0006] In this regard, the present invention provides a method, system, electronic equipment, computer storage medium and computer program product for online monitoring of crane status to solve the above technical problems.

[0007] The present invention discloses a method for online monitoring of crane status, which comprises the following steps: Analyzing the current working state of the crane based on the key operating parameters of the crane, wherein the working state includes a non-lifting state and a lifting state; determining a fault analysis model corresponding to the working state, and switching it to an activated state; Determine a number of crane status data corresponding to the activated fault analysis model, input each of the crane status data into the activated fault analysis model, and obtain the fault type and fault probability predicted by the fault analysis model; An alarm signal corresponding to the fault type for which the fault probability is higher than a probability threshold is generated and output.

[0008] In some embodiments, the fault analysis model corresponding to the lifting state includes a plurality of fault analysis sub-models, each of which corresponds to a lifting sub-state; wherein the lifting sub-state includes at least one of a lifting state, a descending state, a translation state, and a rotation state; According to the historical lifting data of the crane for this operation object, the order of each lifting sub-state and the corresponding average time consumption are obtained according to the statistics of the historical lifting data. When the ratio of the actual time consumption of the lifting sub-state being executed to the average time consumption reaches a preset ratio, the fault analysis sub-model corresponding to the lifting sub-state with the next sequence number is controlled to be activated.

[0009] In some embodiments, the preset ratio is determined by: Extracting all time-consuming data corresponding to the currently executed lifting sub-state from the historical lifting data, calculating the fluctuation range of all the time-consuming data, and obtaining the preset ratio according to the fluctuation range matching; The fluctuation range is the variance or standard deviation of all the time-consuming data.

[0010] In some embodiments, the fault analysis model is constructed based on a convolutional neural network, which includes at least a convolutional layer, a pooling layer, and a fully connected layer; Before inputting each crane status data into the activated fault analysis model, each crane status data is normalized to map it to a specific interval.

[0011] In some embodiments, generating and outputting an alarm signal corresponding to the fault type whose fault probability is higher than a probability threshold comprises: For the fault types whose fault probabilities are higher than the probability threshold, taking them as target fault types, generating the alarm signals corresponding to the target fault types, and determining the output order of the alarm signals according to the fault hazard levels corresponding to the target fault types; The alarm signals are outputted according to the output sequence.

[0012] The present invention also provides a crane status online monitoring system, the system comprising a model selection module, a fault analysis module, and an early warning module; The model selection module is used to analyze the current working state of the crane based on the key operating parameters of the crane, wherein the working state includes a non-lifting state and a lifting state; determine a fault analysis model corresponding to the working state, and switch it to an active state; The fault analysis module is used to determine a number of crane status data corresponding to the activated fault analysis model, input each of the crane status data into the activated fault analysis model, and obtain the fault type and fault probability predicted by the fault analysis model; The early warning module is used to generate and output an alarm signal corresponding to the fault type whose fault probability is higher than a probability threshold.

[0013] The present invention also discloses an electronic device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the computer program code implements the method described in any of the preceding items when executed.

[0014] The present invention also discloses a computer storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program code is executed, the method described in any of the preceding items is implemented.

[0015] The present invention also discloses a computer program product. The computer program product is prepackaged with computer program codes. When the computer program codes are executed, the method described in any of the above items is implemented.

[0016] The beneficial effect of the present invention is that: through the subdivision of working states and the use of targeted fault analysis models, the data processing volume and complexity of fault analysis models under different working states are greatly reduced compared to the traditional single model monitoring method, making the model smaller in size. Smaller models require fewer resources and faster processing speeds during data processing and calculation, thereby improving fault monitoring and early warning rates. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a flow chart of a method for online monitoring of crane status disclosed in an embodiment of the present invention; Figure 2 is a schematic diagram of the structure of a fault analysis model disclosed in an embodiment of the present invention; Figure 3 It is a structural schematic diagram of a crane status online monitoring system disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following is a description of the implementation of the present application by specific specific embodiments. People familiar with the technology can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0020] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0021] In the prior art, there is already a fault monitoring method that uses a single model to identify and predict various types of crane faults. The model used in this fault monitoring method needs to be trained using a variety of data, resulting in a generally large model size. An overly large model size will reduce the model's response rate, which is not conducive to more timely warning of early or impending faults. Therefore, how to further improve the fault monitoring and warning rate of cranes is a technical problem that needs to be solved or improved.

[0022] like Figure 1 As shown, an embodiment of the present invention discloses a method for online monitoring of crane status, the method comprising the following steps: S10, obtaining the current working state of the crane based on the key operating parameters of the crane, wherein the working state includes a non-lifting state and a lifting state; determining a fault analysis model corresponding to the working state, and switching it to an activated state.

[0023] In this step, there are several key operating parameters in the operation of the crane, such as lifting weight, lifting height, operating speed, working range, etc. Through real-time monitoring and analysis of these parameters, it is possible to determine whether the crane is currently in a non-lifting state or a lifting state. For example, when the lifting weight is zero and the lifting height does not change, and other moving mechanisms do not move, it can be determined that the crane is in a non-lifting state; and when the lifting weight is greater than zero and the lifting mechanism or other related mechanisms have corresponding movements, it is determined to be in a lifting state.

[0024] Since the crane has different fault types in different working states, the present invention constructs different fault analysis models for non-lifting state and lifting state respectively. After determining the current working state of the crane, the corresponding fault analysis model is switched to the activated state so that targeted fault analysis can be performed later. This method avoids using a single model to handle all fault types, thereby reducing the complexity and volume of the model. Specifically, if it is determined to be a non-lifting state, a fault analysis model specifically for the non-lifting state is activated. This model only needs to handle the possible fault types that may occur in this state, such as electrical system faults, abnormal initial state of mechanical components, etc.; and when it is determined to be a lifting state, a fault analysis model specifically for the non-lifting state is activated. This model can handle the possible fault types that may occur in this state, such as lifting mechanism faults, electrical system faults, structural component faults, etc. The fault types of the crane in different working states are generally divided into electrical system faults and mechanical system faults, as shown in the following table: S20, determining a number of crane status data corresponding to the activated fault analysis model, inputting each of the crane status data into the activated fault analysis model, and obtaining the fault type and fault probability predicted by the fault analysis model.

[0025] In this step, different fault analysis models require different input data. For the activated fault analysis model, the corresponding crane status data is collected. These data may include electrical parameters (such as voltage, current, temperature), mechanical parameters (such as vibration, stress, wear degree), etc. monitored in real time by various sensors. For example, if the fault analysis model under the lifting state is activated, in addition to conventional parameters such as lifting weight and lifting height, it is also necessary to collect data closely related to the lifting operation, such as the current of the lifting motor, the tension of the wire rope, and the temperature of the brake.

[0026] The collected corresponding status data is input into the activated fault analysis model. Based on the existing training data and algorithms, the model analyzes and processes the input data to predict the possible fault types and the probability of each fault. For example, the model may output results such as "wire rope broken, probability 80%", "motor overheating, probability 30%", etc. These results provide a basis for subsequent early warning and processing.

[0027] S30, generating and outputting an alarm signal corresponding to the fault type whose fault probability is higher than a probability threshold.

[0028] In this step, the probability of each fault predicted by the fault analysis model is compared with the set probability threshold. For fault types with a fault probability higher than the probability threshold, the system generates a corresponding alarm signal and outputs it. The alarm signal can be presented in a variety of ways, such as sound and light alarms, SMS notifications, APP push, etc., so that relevant personnel can be informed of possible faults in a timely manner and take corresponding measures, such as shutdown inspection and maintenance, so as to achieve timely warning of early faults or impending faults.

[0029] In addition, the above probability thresholds are determined based on actual experience, the importance of the crane, the severity of the consequences of the failure, etc. For example, for some failures that may lead to serious safety accidents, the threshold may be set relatively low (such as 70%) to ensure that even if the probability of the failure is not particularly high, an early warning can be issued in time; while for some failures with less impact, the threshold can be set higher (such as 90%).

[0030] The above-mentioned crane status online monitoring method of the present invention, through the subdivision of the working status and the use of targeted fault analysis models, is compared with the traditional single model monitoring method. The data processing volume and complexity of the fault analysis model under different working conditions are greatly reduced, making the model smaller in size. The smaller model requires fewer resources and has a faster processing speed during data processing and calculation, thereby improving the fault monitoring and early warning rate.

[0031] In some embodiments, the fault analysis model corresponding to the lifting state includes a plurality of fault analysis sub-models, each of which corresponds to a lifting sub-state; wherein the lifting sub-state includes at least one of a lifting state, a descending state, a translation state, and a rotation state; According to the historical lifting data of the crane for this operation object, the order of each lifting sub-state and the corresponding average time consumption are obtained according to the statistics of the historical lifting data. When the ratio of the actual time consumption of the lifting sub-state being executed to the average time consumption reaches a preset ratio, the fault analysis sub-model corresponding to the lifting sub-state with the next sequence number is controlled to be activated.

[0032] In the embodiment of the present invention, the lifting state of the crane can be further subdivided into a plurality of lifting sub-states, such as lifting state, descending state, translation state, rotation state, etc. Due to the different lifting sub-states, their operating characteristics, stress conditions and possible fault types are different. For example, the lifting state mainly focuses on the faults related to the lifting mechanism, such as the problems of the wire rope, drum, brake, etc. in the process of lifting heavy objects; while the descending state focuses more on the faults in the descending speed control, braking reliability, etc. Therefore, the present invention further constructs a special fault analysis sub-model for each lifting sub-state, so that the faults at different stages can be monitored and predicted more targetedly.

[0033] At the same time, the present invention will also collect the historical lifting data of the crane for this operation object, which includes the order of each lifting sub-state in each lifting operation and the time spent in each lifting sub-state. For example, the crane is unloading containers on a cargo ship. Based on the statistics of the crane's unloading data for the first 30 and 50 times, it is found that it is generally in the lifting state first, which takes an average of 2 minutes, followed by the translation state, which takes an average of 3 minutes, etc. When the ratio of the actual time spent in a certain lifting sub-state to the average time reaches a preset ratio (for example, 80%), it means that the lifting sub-state is about to end. At this time, the fault analysis sub-model corresponding to the lifting sub-state of the next sequence number is activated, so that the problem of fault monitoring "blank period" caused by activating the corresponding fault analysis sub-model after entering the next lifting sub-state can be avoided.

[0034] In some embodiments, the preset ratio is determined by: Extracting all time-consuming data corresponding to the currently executed lifting sub-state from the historical lifting data, calculating the fluctuation range of all the time-consuming data, and obtaining the preset ratio according to the fluctuation range matching; The fluctuation range is the variance or standard deviation of all the time-consuming data.

[0035] In an embodiment of the present invention, the size of the preset ratio is determined by the fluctuation of the historical time consumption corresponding to the currently executed lifting sub-state, that is, the larger the fluctuation amplitude of the historical time consumption is, the smaller the preset ratio is set, that is, the earlier the fault analysis sub-model corresponding to the next numbered lifting sub-state is activated; otherwise, the smaller the preset ratio is set, that is, the later the fault analysis sub-model corresponding to the next numbered lifting sub-state is activated.

[0036] Specifically, after determining the lifting sub-state currently being executed (such as lifting state, descending state, etc.), all the time-consuming data corresponding to the lifting sub-state are screened out from these historical data. For example, if the current state is lifting, the time-consuming data of all lifting states in the historical operation are extracted, and these data reflect the time spent on each lifting in the operation for the current operation object. On this basis, the present invention further calculates the fluctuation range of these time-consuming data, for example, by calculating the variance or standard deviation to obtain the fluctuation range, and the fluctuation range is used to measure the discrete degree of these time-consuming data.

[0037] If the fluctuation range (variance or standard deviation) is small, it means that the time consumption of the lifting sub-state is relatively stable, and the time consumption of each operation is close to the average value. At this time, a relatively high preset ratio can be set, such as 90%. This means that when the actual time consumption reaches 90% of the average time consumption, the lifting sub-state is considered to be close to completion, and the fault analysis sub-model corresponding to the next lifting sub-state is activated.

[0038] If the fluctuation is large, it means that the time consumption of the lifting sub-state is unstable and the time consumption of each operation varies greatly. In order to avoid activating the next fault analysis sub-model too early or too late due to time consumption fluctuations, it is necessary to set a relatively low preset ratio, such as 80%. That is, when the actual time consumption reaches 80% of the average time consumption, the next model is activated to adapt to the large fluctuation of time consumption and ensure timely switching of fault monitoring when the lifting sub-state may change.

[0039] By determining the preset ratio according to the fluctuation amplitude of historical data, the present invention can more flexibly and accurately dynamically activate the corresponding fault analysis sub-model according to the time-consuming characteristics of different lifting sub-states, thereby improving the real-time and effectiveness of crane fault monitoring.

[0040] In some embodiments, the fault analysis model is constructed based on a convolutional neural network, which includes at least a convolutional layer, a pooling layer, and a fully connected layer; Before inputting each crane status data into the activated fault analysis model, each crane status data is normalized to map it to a specific interval.

[0041] In the embodiment of the present invention, the input data is various status data related to the crane failure, such as vibration signals, temperature data, electrical parameters, etc. These data need to be pre-processed according to different types and converted into a format suitable for CNN processing. If the data is a one-dimensional time series, such as the change of vibration signals over time, it can be organized into a one-dimensional array; if it is a two-dimensional image data, such as a temperature distribution image obtained by a sensor array, it is kept in a two-dimensional matrix form. For multi-channel data, such as vibration signals collected at multiple locations at the same time, it can be input as multi-channel data, similar to the RGB channels of a color image.

[0042] Before the above input data is input into the fault analysis model, it needs to be normalized. Map the data to a specific interval, such as [0,1] or [-1,1]. For example, for temperature data with a value range of [0,100], the formula Normalize, where is the original data, and are the minimum and maximum values ​​of this group of data respectively.

[0043] like Figure 2 As shown, the fault analysis model of the present invention is constructed based on a convolutional neural network, which at least includes a convolutional layer, a pooling layer, and a fully connected layer. Among them: The convolution layer is the core part of CNN. It performs convolution operations on the input data through convolution kernels to extract local features in the data. The convolution kernel is a learnable weight matrix, such as 3×3 or 5×5. For one-dimensional data, the convolution kernel is a one-dimensional vector. For example, when processing vibration signals, a convolution kernel with 3 time steps can capture the change pattern of the signal in a short period of time. Multiple convolution kernels with different weights work in parallel, and each convolution kernel extracts a specific feature, such as vibration patterns of different frequencies, specific temperature change trends, etc.

[0044] The pooling layer follows the convolution layer and is used to reduce the dimension of the feature map output by the convolution layer, reduce the amount of data, and retain the main features. The pooling operation can choose maximum pooling or average pooling. The pooling operation can not only reduce the data dimension, reduce the amount of calculation and model parameters, but also improve the robustness of the model to a certain extent, and has stronger adaptability to slight changes in data.

[0045] After feature extraction and dimensionality reduction in the convolution layer and the pooling layer, the obtained feature map is expanded into a one-dimensional vector and then input into the fully connected layer. The fully connected layer can determine what kind of fault state the crane is currently in based on the vibration, temperature, electrical and other characteristics of the crane extracted previously. In addition, the fully connected layer in the present invention is preferably a multi-layer fully connected layer, that is, a network is composed of multiple fully connected layers, and the output of each fully connected layer is used as the input of the next fully connected layer. In a layer-by-layer progressive manner, the features are continuously abstracted and combined. For example, the first fully connected layer may learn some relatively basic feature combinations, and the subsequent fully connected layers learn more advanced and abstract features on this basis to improve the accuracy of fault classification.

[0046] In addition, an output layer should also be included. The number of neurons in the output layer depends on the specific fault analysis task. For example, if n different types of crane faults are to be identified, the output layer has n neurons. The activation function of the output layer is selected according to the task type. For multi-classification problems, the Softmax function is preferably used. The Softmax function converts the output of each neuron into a probability value. The sum of these probability values ​​is 1, representing the probability that the input data belongs to each fault type. For example, after the output of the three neurons in the output layer is processed by the Softmax function, [0.1, 0.3, 0.6] is obtained, indicating that the probability that the input data belongs to the three fault types is 10%, 30%, and 60%, respectively.

[0047] In some embodiments, generating and outputting an alarm signal corresponding to the fault type whose fault probability is higher than a probability threshold comprises: For the fault types whose fault probabilities are higher than the probability threshold, taking them as target fault types, generating the alarm signals corresponding to the target fault types, and determining the output order of the alarm signals according to the fault hazard levels corresponding to the target fault types; The alarm signals are outputted according to the output sequence.

[0048] In the embodiment of the present invention, the present invention selects those fault types whose fault probabilities are higher than the probability threshold as target fault types that need to be focused on and processed. For example, if the probability threshold is set to 50%, if the fault analysis model predicts that the probability of "wire rope breakage" is 60% and the probability of "motor overheating" is 70%, these two fault types will be determined as target fault types.

[0049] Different fault types have different degrees of harm to crane operation and personnel safety, so the system will set a corresponding fault hazard level for each target fault type. The fault hazard level can be divided according to the severity of the consequences that may be caused by the fault, for example, it can be divided into three levels: high, medium and low. "Wire rope breakage" may cause heavy objects to fall, causing serious casualties and equipment damage, and its hazard level can be set to "high"; while "motor overheating" may not cause extremely serious consequences if handled in time, and the hazard level can be set to "medium". According to the fault hazard level corresponding to each target fault type, the output order of each alarm signal is determined. The higher the hazard level of the fault, the higher the priority of its alarm signal output, which can ensure that the operator first pays attention to the most serious fault situation, takes corresponding measures in time, and avoids more serious accidents. For example, if the two target fault types "wire rope breakage" (high hazard level) and "motor overheating" (medium hazard level) appear at the same time, the system will first output the alarm signal of "wire rope breakage" and then output the alarm signal of "motor overheating".

[0050] The embodiment of the present invention further provides a crane status online monitoring system, the system comprising a model selection module, a fault analysis module, and an early warning module; The model selection module is used to analyze the current working state of the crane based on the key operating parameters of the crane, wherein the working state includes a non-lifting state and a lifting state; determine a fault analysis model corresponding to the working state, and switch it to an active state; The fault analysis module is used to determine a number of crane status data corresponding to the activated fault analysis model, input each of the crane status data into the activated fault analysis model, and obtain the fault type and fault probability predicted by the fault analysis model; The early warning module is used to generate and output an alarm signal corresponding to the fault type whose fault probability is higher than a probability threshold.

[0051] This system corresponds to the aforementioned method, and its beneficial technical effects will not be described in detail here.

[0052] An embodiment of the present invention further discloses an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the computer program code implements the method described in any of the preceding items when executed.

[0053] The embodiment of the present invention further discloses a computer storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program code is executed, the method as described in any of the preceding items is implemented.

[0054] The embodiment of the present invention further discloses a computer program product. The computer program product is prepackaged with computer program code. When the computer program code is executed, the method described in any of the above items is implemented.

[0055] The device embodiments described above are merely illustrative, wherein the designed modules may or may not be physically separated, i.e., may be located in one place, or may be distributed to multiple locations. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. Those of ordinary skill in the art may understand and implement the present invention without creative effort.

[0056] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0057] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for online monitoring of crane status, characterized in that: The method comprises the following steps: Analyzing the current working state of the crane based on the key operating parameters of the crane, wherein the working state includes a non-lifting state and a lifting state; determining a fault analysis model corresponding to the working state, and switching it to an activated state; Determine a number of crane status data corresponding to the activated fault analysis model, input each of the crane status data into the activated fault analysis model, and obtain the fault type and fault probability predicted by the fault analysis model; An alarm signal corresponding to the fault type for which the fault probability is higher than a probability threshold is generated and output.

2. A crane status online monitoring method according to claim 1, characterized in that: The fault analysis model corresponding to the lifting state includes a plurality of fault analysis sub-models, each of which corresponds to a lifting sub-state; wherein the lifting sub-state includes at least one of a lifting state, a descending state, a translation state, and a rotation state; According to the historical lifting data of the crane for this operation object, the order of each lifting sub-state and the corresponding average time consumption are obtained according to the statistics of the historical lifting data. When the ratio of the actual time consumption of the lifting sub-state being executed to the average time consumption reaches a preset ratio, the fault analysis sub-model corresponding to the lifting sub-state with the next sequence number is controlled to be activated.

3. A crane status online monitoring method according to claim 2, characterized in that: The preset ratio is determined by: Extracting all time-consuming data corresponding to the currently executed lifting sub-state from the historical lifting data, calculating the fluctuation range of all the time-consuming data, and obtaining the preset ratio according to the fluctuation range matching; The fluctuation range is the variance or standard deviation of all the time-consuming data.

4. The method for online monitoring of crane status according to claim 1, characterized in that: The fault analysis model is constructed based on a convolutional neural network, which at least includes a convolutional layer, a pooling layer, and a fully connected layer; Before inputting each crane status data into the activated fault analysis model, each crane status data is normalized to map it to a specific interval.

5. The method for online monitoring of crane status according to claim 1, characterized in that: The generating and outputting an alarm signal corresponding to the fault type whose fault probability is higher than a probability threshold comprises: For the fault types whose fault probabilities are higher than the probability threshold, taking them as target fault types, generating the alarm signals corresponding to the target fault types, and determining the output order of the alarm signals according to the fault hazard levels corresponding to the target fault types; The alarm signals are outputted according to the output sequence.

6. A crane status online monitoring system, characterized in that: The system includes a model selection module, a fault analysis module, and an early warning module; The model selection module is used to analyze the current working state of the crane based on the key operating parameters of the crane, wherein the working state includes a non-lifting state and a lifting state; determine a fault analysis model corresponding to the working state, and switch it to an active state; The fault analysis module is used to determine a number of crane status data corresponding to the activated fault analysis model, input each of the crane status data into the activated fault analysis model, and obtain the fault type and fault probability predicted by the fault analysis model; The early warning module is used to generate and output an alarm signal corresponding to the fault type whose fault probability is higher than a probability threshold.

7. The crane status online monitoring system according to claim 6, characterized in that: The fault analysis model corresponding to the lifting state includes a plurality of fault analysis sub-models, each of which corresponds to a lifting sub-state; wherein the lifting sub-state includes at least one of a lifting state, a descending state, a translation state, and a rotation state; According to the historical lifting data of the crane for this operation object, the order of each lifting sub-state and the corresponding average time consumption are obtained according to the statistics of the historical lifting data. When the ratio of the actual time consumption of the lifting sub-state being executed to the average time consumption reaches a preset ratio, the fault analysis sub-model corresponding to the lifting sub-state with the next sequence number is controlled to be activated.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the computer program code implements the method according to any one of claims 1 to 5 when executed.

9. A computer storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program code is executed, the method according to any one of claims 1 to 5 is implemented.

10. A computer program product, characterized in that: The computer program product is prepackaged with computer program codes, and when the computer program codes are executed, the method according to any one of claims 1 to 5 is implemented.

Images