Deviation correction method and system for container portal crane

Abstract

The invention relates to the technical field of automatic operation and maintenance, and provides a correction method and system for a container portal crane. The method comprises the following steps: retrieving crane deviation rectification operation and maintenance frequency according to a preset historical window, and generating a deviation rectification characteristic frequency coefficient; inputting a monitoring computing power calculation branch, and generating a target monitoring computing power; activating a monitoring sensing module, monitoring the crane in real time, and generating a monitoring data source; carrying out multi-dimensional offset detection in combination with an offset detection algorithm; and according to the correction registration decision channel, generating a correction scheme and implementing the correction scheme. The technical problems that in the actual operation process of the container portal crane, due to the fact that the container portal crane depends on a traditional sensor and a fixed algorithm, external environment changes and equipment abrasion are ignored, deviation correction is inaccurate, and response is not timely are solved; the correction accuracy and efficiency of the container portal crane are improved through automatic monitoring and intelligent decision making, the risk of operation interruption caused by deviation is reduced, and the reliability and safety of equipment are enhanced.

Description

Technical Field

[0001] The present application relates to the field of operation and maintenance management technology, specifically to the field of automated operation and maintenance technology, and in particular to a deviation correction method and system for a container gantry crane. Background Art

[0002] With the continuous development of global trade, the application of container gantry cranes in ports, docks and logistics centers has become increasingly important. However, in the actual operation of the crane, due to the limitations of traditional sensors and fixed algorithms, the problem of deviation correction has become a key factor restricting its performance improvement. Traditional deviation correction methods often ignore the real-time changes of external environmental factors such as wind, humidity, etc., as well as the wear and tear of the equipment itself. Therefore, in actual operation, the accuracy and response speed of deviation correction often fail to meet the requirements, which not only affects the operating efficiency of the crane, but also may bring safety risks. Summary of the invention

[0003] The present application provides a method and system for correcting the deviation of a container gantry crane, aiming to solve the technical problems of inaccurate correction and untimely response caused by reliance on traditional sensors and fixed algorithms while ignoring changes in the external environment and wear of the equipment itself during actual operation of the container gantry crane.

[0004] In view of the above problems, the present application provides a deviation correction method and system for a container gantry crane.

[0005] The first aspect disclosed in the present application provides a method for correcting a container gantry crane, the method comprising: retrieving the correction operation and maintenance frequency of the container gantry crane according to a predetermined historical window to generate a correction feature frequency coefficient; inputting the correction feature frequency coefficient into a monitoring computing power calculation branch to generate a target monitoring computing power; performing feature activation on a crane monitoring sensor module according to the target monitoring computing power to obtain a feature monitoring sensor module; monitoring the container gantry crane in real time according to the feature monitoring sensor module to generate a crane monitoring data source, wherein the crane monitoring data source includes a crane operation monitoring data stream, a crane environment monitoring data stream and a crane track monitoring data stream; performing multi-dimensional offset detection on the container gantry crane according to a crane offset detection algorithm in combination with the crane monitoring data source to generate a crane offset detection result; performing a correction decision on the container gantry crane according to a crane correction registration decision channel in combination with the crane offset detection result to generate a crane correction plan, and performing correction operation and maintenance on the container gantry crane according to the crane correction plan.

[0006] Another aspect disclosed in the present application provides a correction system for a container gantry crane, the system comprising: a frequent coefficient generation component, the frequent coefficient generation component is used to retrieve the correction operation and maintenance frequency of the container gantry crane according to a predetermined historical window, and generate a correction feature frequent coefficient; a monitoring computing power generation component, the monitoring computing power generation component is used to input the correction feature frequent coefficient into a monitoring computing power calculation branch, and generate a target monitoring computing power; a feature activation component, the feature activation component is used to perform feature activation on a crane monitoring sensor module according to the target monitoring computing power to obtain a feature monitoring sensor module; a real-time monitoring component, the real-time monitoring component is used to monitor the container gantry crane in real time according to the feature monitoring sensor module. A container gantry crane is configured to generate a crane monitoring data source, wherein the crane monitoring data source includes a crane operation monitoring data stream, a crane environment monitoring data stream and a crane track monitoring data stream; a multi-dimensional offset detection component, wherein the multi-dimensional offset detection component is used to perform multi-dimensional offset detection on the container gantry crane according to a crane offset detection algorithm in combination with the crane monitoring data source, and generate a crane offset detection result; a correction operation and maintenance component, wherein the correction operation and maintenance component is used to perform a correction decision on the container gantry crane according to a crane correction registration decision channel in combination with the crane offset detection result, generate a crane correction plan, and perform correction operation and maintenance on the container gantry crane according to the crane correction plan.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: The above-mentioned method for correcting the deviation of a container gantry crane retrieves the frequency of the crane's deviation correction operation and maintenance based on historical data, thereby obtaining the correction feature frequency coefficient. Subsequently, this coefficient is used to calculate the required target monitoring computing power, and the corresponding monitoring sensor module is activated accordingly. These modules can collect key data such as the crane's operation, environment, and track in real time. After obtaining these data, the deviation of the crane is accurately determined through the offset detection algorithm. After that, combined with the correction decision channel, a specific correction plan is generated and implemented immediately to ensure the stable operation of the crane. This method greatly improves the accuracy and efficiency of the correction, and provides a strong guarantee for the safe operation of container gantry cranes.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic diagram of a flow chart of a deviation correction method for a container gantry crane in one embodiment; Figure 2 The present invention is an architecture diagram of a deviation correction system for a container gantry crane in one embodiment.

[0011] Explanation of the accompanying drawings: frequent coefficient generation component 1, monitoring computing power generation component 2, feature activation component 3, real-time monitoring component 4, multi-dimensional offset detection component 5, deviation correction operation and maintenance component 6. DETAILED DESCRIPTION

[0012] The embodiments of the present application provide a correction method and system for a container gantry crane, thereby solving the technical problem of inaccurate correction and untimely response caused by reliance on traditional sensors and fixed algorithms while ignoring changes in the external environment and wear of the equipment itself during actual operation of the container gantry crane.

[0013] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0014] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.

[0015] Embodiment 1 like Figure 1 As shown, the present application provides a deviation correction method for a container gantry crane, the method comprising: Retrieve the deviation correction operation and maintenance frequency of container gantry cranes according to the predetermined historical window and generate the deviation correction feature frequency coefficient; With the continuous development of global trade, ports, logistics centers and other places are increasingly dependent on container gantry cranes, and the requirements for their performance and safety are also increasing accordingly. In this context, more accurate correction methods are particularly important.

[0016] In an embodiment of the present application, the system terminal determines a suitable historical time window according to actual needs, and this time window can cover enough correction operation and maintenance data to reflect the historical situation of crane correction operation and maintenance. In the determined historical window, the correction operation and maintenance records of the container gantry crane are retrieved by querying the historical database. These records contain detailed information such as the time, cause, corrective measures taken, and effect after correction of each correction. Subsequently, the retrieved correction operation and maintenance records are counted, and the total number of correction operation and maintenance in the historical window and the average interval time of each correction are calculated. After that, the correction feature frequency coefficient is calculated based on the statistical correction operation and maintenance frequency, combined with factors such as the use frequency of the crane and the total operation time of the crane. This coefficient is the ratio of the number of corrections to the total operation time of the crane. When calculating the correction feature frequency coefficient, the system terminal comprehensively considers other factors that may affect the correction operation and maintenance frequency, such as weather conditions, work intensity, equipment aging, etc., and matches these factors with an initial weight value. These initial weight values ​​are derived based on historical data and experience. Then, according to actual conditions, the initial weight is dynamically adjusted. For example, if the frequency of deviation correction increases when weather conditions are bad, then the weight of the weather condition factor is increased. After the weight adjustment is completed, for each influencing factor, the system terminal matches the actual state according to the existing quantitative standard, and then multiplies the matching result by the adjusted weight to obtain the corresponding correction coefficient. The value range of this correction coefficient is between 0 and 1. For example, when the wind is too strong, it may cause unstable operation of the crane, thereby increasing the demand for deviation correction, then the matched wind correction coefficient is close to 1; conversely, if the wind is very small, the impact on the deviation correction demand is not significant, and the correction coefficient is close to 0. Finally, the obtained correction coefficient is multiplied by the calculated deviation correction feature frequency coefficient to obtain the corrected deviation correction feature frequency coefficient. This coefficient reflects the deviation correction demand and trend of the crane in the past period of time. Through this step, the operating status and deviation correction demand of the crane can be understood more accurately, providing an important basis for subsequent monitoring and deviation correction operations.

[0017] Inputting the correction feature frequency coefficient into the monitoring computing power calculation branch to generate the target monitoring computing power; In one embodiment, the system terminal determines the degree and frequency of the crane's correction requirements based on the correction feature frequency coefficient calculated previously. Subsequently, the correction feature frequency coefficient is input into the monitoring computing power calculation branch, which calculates the required monitoring computing power, i.e., the target monitoring computing power, based on the size and change trend of the coefficient, combined with the internal monitoring computing power calculation function. This process is to ensure that the system terminal has sufficient computing power to monitor the status of the crane in real time, promptly detect and handle possible deviations, thereby improving the safety and efficiency of the crane's operation, so as to better perform the crane's correction and maintenance work.

[0018] Furthermore, the present application provides a monitoring computing power calculation function, including: The monitoring computing power calculation branch includes a monitoring computing power calculation function, and the monitoring computing power calculation function is: ; Among them, TMC represents the target monitoring computing power, TMA represents the preset benchmark monitoring computing power, TMB represents the correction feature frequency coefficient, TMC represents the preset correction feature frequency threshold, BCW represents the preset weight coefficient, and 0<BCW<1.

[0019] Preferably, the system terminal encapsulates the pre-built monitoring computing power calculation function and integrates it with the input end and the output end to generate a monitoring computing power calculation branch. Specifically, the input end is used to transfer the correction feature frequency coefficient input by the system terminal to the monitoring computing power calculation function. The output end is used to output the target monitoring computing power calculated by the monitoring computing power calculation function to the system terminal. The monitoring computing power calculation function is as follows: ; in, It is the target monitoring power. This is the final monitoring power value that needs to be calculated, which is used to monitor the status of the crane in real time to ensure that possible deviations can be discovered and handled in time. It is the preset baseline monitoring hashrate. This is a baseline value that represents the level of monitoring hashrate required under normal circumstances. is the correction feature frequency coefficient. This coefficient reflects the degree and frequency of the crane's correction needs. It is generated based on the correction operation and maintenance frequency in historical data and is used to measure the current crane's correction status. is the preset weight coefficient. This is a value between 0 and 1 that is used to adjust the importance or influence of TMB in the final calculation. When the system terminal calculates the correction feature frequency coefficient based on the actual operation status and historical data of the crane, Then, the correction feature frequent coefficient TMB and the preset weight coefficient Multiply them together to get an adjusted correction factor value. Then, add this adjusted correction factor value to 1 and then add it to the preset benchmark monitoring computing power. Multiply them together to get the target monitoring computing power TMC. This calculation process comprehensively considers the benchmark monitoring needs of the crane and the current correction needs. By adjusting the weight coefficient, it can more flexibly adapt to the requirements of monitoring computing power in different situations, thereby improving the accuracy and efficiency of monitoring.

[0020] Performing feature activation on the crane monitoring sensor module according to the target monitoring computing power to obtain a feature monitoring sensor module; In one embodiment, the target monitoring computing power obtained by the system terminal represents the optimal monitoring computing power required under the current operating state of the crane. Subsequently, according to this target monitoring computing power, the monitoring sensor module of the crane is feature activated. Feature activation is a process of adjusting the working state of the crane monitoring sensor module according to the computing power requirements, which aims to enable the crane monitoring sensor module to collect data on the operation of the crane with higher accuracy and efficiency. Specifically, the system terminal first initializes the monitoring sensor module of the crane to ensure that it is in a normal working state, laying the foundation for the subsequent activation steps. Subsequently, according to the requirements of the target monitoring computing power, the sampling rate of the crane monitoring sensor module is adjusted. Increasing the sampling rate means that the crane monitoring sensor module can collect the operating data of the crane more frequently, thereby capturing the dynamic changes of the crane more accurately. Otherwise, it will be suspended, reducing the power consumption of the crane monitoring sensor module and extending its service life. Afterwards, the measurement range and accuracy of the crane monitoring sensor module are optimized according to the target monitoring computing power. By expanding the measurement range, the crane monitoring sensor module can cover a wider range of parameter changes; by improving the measurement accuracy, the crane monitoring sensor module can provide more accurate data and reduce errors. Then, the data transmission mechanism of the crane monitoring sensor module is optimized to improve the data transmission speed and ensure that the system terminal can obtain the data collected by the crane monitoring sensor module in real time. At the same time, the stability of data transmission is enhanced to avoid data loss or damage and ensure the integrity and accuracy of the data. Finally, the sensor module is feature activated again in combination with the requirements of the target monitoring computing power and the parameters obtained after the above optimization. This includes activating the functional submodules of the crane monitoring sensor module, setting the corresponding parameters, etc., so that the sensor module has the best performance and work under the target monitoring computing power. Through these feature activation operations, the sensor module can more accurately capture the operating status and offset of the crane, providing strong support for subsequent deviation correction and maintenance work. After completing the feature activation, we get the feature monitoring sensor module. This module has the best working state under the target monitoring computing power, can provide crane operation data in real time and accurately, and help the system terminal to promptly discover and deal with possible offset problems, thereby improving the safety and efficiency of crane operation.

[0021] According to the characteristic monitoring sensor module, the container gantry crane is monitored in real time to generate a crane monitoring data source, wherein the crane monitoring data source includes a crane operation monitoring data stream, a crane environment monitoring data stream and a crane track monitoring data stream; In one embodiment, the system terminal generates a rich crane monitoring data source based on the real-time monitoring of the container gantry crane by the feature monitoring sensor module. These data sources cover the monitoring data streams of three aspects: crane operation, environment and track. Specifically, the crane operation monitoring data stream reflects the various operating parameters and status information of the crane during operation, helping the system terminal to understand the operating efficiency and safety performance of the crane. The crane environment monitoring data stream provides relevant data of the environment in which the crane is located, such as temperature, humidity, wind force, etc., providing a basis for the system terminal to evaluate the impact of the environment on the operation of the crane. In terms of the crane track monitoring data stream, the feature monitoring sensor module can capture and generate the corresponding track images and wheel images of the container gantry crane in real time when it is running. These image data not only intuitively show the real-time status of the track and wheel, but also help the system terminal to promptly detect abnormalities, damage or wear of the wheel on the track. By analyzing these image data, the safety and stability of the crane track can be more accurately evaluated, thereby ensuring the smooth operation of the crane. In summary, these monitoring data sources together constitute a comprehensive monitoring of the container gantry crane, providing strong data support for subsequent analysis and decision-making.

[0022] According to the crane offset detection algorithm, in combination with the crane monitoring data source, multi-dimensional offset detection is performed on the container gantry crane to generate a crane offset detection result; In one embodiment, the system terminal performs multi-dimensional offset detection on the container gantry crane based on the crane offset detection algorithm and the previously generated crane monitoring data source. This process involves comprehensive analysis and processing of data from multiple aspects such as crane operation, environment and track. Specifically, the crane offset detection algorithm can use the operation parameters and status information in the operation monitoring data stream to analyze whether there is abnormal offset during the operation of the crane and generate an operation offset vector set. At the same time, the environmental data in the environmental monitoring data stream is also taken into consideration to evaluate whether environmental factors affect the offset of the crane and generate a normal environmental deviation vector set. Similarly, the track image and wheel image in the crane track monitoring data stream provide an intuitive and important basis for offset detection. By analyzing these image data, the algorithm can identify abnormalities or damage on the track and the wear of the wheel, thereby determining whether the crane has an offset risk and generating a rail gnawing feature recognition result. Subsequently, the generated multiple results are summarized to obtain the final crane offset detection result. This result can accurately reflect the offset of the container gantry crane and provide a strong basis for subsequent correction measures. In summary, the crane offset detection algorithm and monitoring data source were used to conduct comprehensive and accurate offset detection of the container gantry crane, ensuring the safe and stable operation of the crane.

[0023] Furthermore, the present application provides a method for performing multi-dimensional offset detection on the container gantry crane in combination with the crane monitoring data source according to the crane offset detection algorithm to generate a crane offset detection result, and the method further includes: Performing standardization processing on the crane operation monitoring data stream to generate a crane monitoring standard data stream; Performing monitoring position clustering according to the crane monitoring standard data stream to obtain multiple position monitoring data partitions; Preferably, the system terminal performs standardization processing on the crane operation monitoring data stream, mainly to ensure that the operation monitoring data from different sources and different formats can be uniformly compared and analyzed. In this step, the crane operation monitoring data stream is first cleaned to remove duplicate, erroneous or invalid data. This includes checking missing values, outliers and data that do not meet the specifications in the data, and performing corresponding processing, such as filling missing values ​​and deleting abnormal data. Subsequently, the data is converted according to actual needs for subsequent standardization processing. For example, the temperature data is converted from Celsius to Fahrenheit, and the time data is converted from string format to timestamp format. Before standardization processing, the system terminal performs feature selection, that is, selects data features that have an important impact on the crane operation status and environmental characteristics. This helps to reduce the dimension of the data and improve the efficiency of subsequent processing. After that, for each selected feature, its mean and standard deviation in the entire data set are calculated. The mean indicates the center position of the data, while the standard deviation indicates the degree of dispersion of the data. Then the data of each selected feature is converted using Z-score standardization to generate a crane monitoring standard data stream. After this annotation conversion, the data will be scaled to a distribution with a mean of 0 and a standard deviation of 1. Standardization processing helps to eliminate dimensional differences and noise interference between data and improve the reliability and accuracy of data.

[0024] After obtaining the standard data stream for crane monitoring, the system terminal performs monitoring location clustering. Clustering is a data analysis technique that groups similar data points into different categories or clusters. In this monitoring location clustering, monitoring location clustering means grouping the monitoring data in the standard data stream according to its location characteristics to form multiple location monitoring data partitions. Each partition represents the operating status and environmental characteristics of the crane at different locations, providing a clear data structure and basis for subsequent monitoring and data analysis. Specifically, the system terminal first extracts location-related data features from the standard data stream for crane monitoring, which can accurately reflect the operating location of the crane. These data features are all numerical and have been standardized so that the distance can be correctly calculated in the K-means algorithm. Subsequently, the system terminal determines the K value in the K-means algorithm based on actual business needs and the approximate distribution of the data. K represents the number of clusters to be formed, that is, the number of location monitoring data partitions. Afterwards, K data points are randomly selected as the initial cluster centroids, which will be used as the initial cluster centers. Then, according to the distance between each data point and each centroid, the data point is assigned to the nearest cluster. For each cluster, recalculate the mean of all data points within it and use the mean as the new centroid. Repeat the above steps until the position of the centroid no longer changes significantly. This process aims to optimize the cluster division so that the data points within each cluster are as similar as possible, while the data points between different clusters are as different as possible. Then, based on the cluster division results, the data points in the crane monitoring standard data stream are divided according to the clusters they belong to, forming multiple location monitoring data partitions. Each partition represents the operating status and environmental characteristics of the crane in a specific location or area. In summary, the crane operation monitoring data stream is standardized to generate a unified monitoring standard data stream, and then based on these standard data, the monitoring data is divided into multiple partitions according to location characteristics through a clustering algorithm to better understand and analyze the operation of the crane.

[0025] Extracting a target position monitoring data partition corresponding to a target monitoring position according to the multiple position monitoring data partitions, wherein the target monitoring position is any monitoring position of the container gantry crane; Loading a predetermined operation plan of the container gantry crane, matching a target operation plan of the target monitoring position based on the predetermined operation plan, and performing standard processing on the target operation plan to generate a standard operation partition of the target position; Preferably, the system terminal extracts data for a specific target monitoring position according to the obtained multiple position monitoring data partitions. The target monitoring position here refers to any monitoring point of interest on the container gantry crane. By accurately locating this position, detailed monitoring data at this position can be obtained to form a target position monitoring data partition. Subsequently, the system terminal obtains the predetermined operation plan of the container gantry crane from the operation plan library. These plans are pre-set and describe how the crane should operate in different positions or different working conditions. After that, according to the specific information of the target monitoring position such as position coordinates, working environment, etc., the predetermined operation plan is screened and matched. Then the format, unit and description method of the target operation plan are analyzed to determine the content that needs to be standardized, and the specifications and standards for standardized processing are formulated based on the content that needs to be standardized to ensure the consistency and accuracy of the processing process. Then, the units of the parameters in the target operation plan are unified to ensure that all parameters use the same measurement unit, and the description of the operation process is standardized to eliminate redundant and unclear expressions, and then the operation requirements are clarified to ensure that each step has clear guidance and requirements. Finally, the standardized target operation plan is integrated into a complete partition, namely the target location standard operation partition, as a reference for the standard operation of the target monitoring location. In summary, by extracting the data of the target monitoring location and combining it with the predetermined operation plan, a target location standard operation partition that reflects the standard operation status of the location is generated. This provides a basis for subsequent comparison of the differences between actual operations and standard operations and identification of potential problems or anomalies.

[0026] Performing operation offset analysis on the target location monitoring data partition according to the target location standard operation partition to generate an operation offset vector set; Performing an offset depth evaluation according to the operation offset vector set to generate an operation offset depth coefficient; Preferably, the system terminal performs an operation offset analysis on the target location monitoring data partition according to the target location standard operation partition generated previously. The operation offset here refers to the difference or degree of deviation between the actual operation state and the standard operation state. Specifically, the system terminal aligns the time series of the target location standard operation partition and the target location monitoring data partition for point-by-point comparison. Subsequently, the standard parameter values ​​and indicator values ​​in the target location standard operation partition are compared one by one with the actual monitoring values ​​in the target location monitoring data partition. The difference between the actual value and the standard value of each parameter or indicator is recorded, including the direction and magnitude of the difference. After that, the operation steps in the standard operation process are compared with the operation records in the actual monitoring data. Identify whether there are omitted, added or reversed operation steps in the actual operation, and record these deviations. Then, based on the comparison results, the operation offset vector is constructed. Each vector contains the parameter where the offset occurs, the direction of the offset (positive or negative), the size of the offset (specific value or degree), and the time and step position of the offset, ensuring that each offset vector clearly and accurately reflects the difference between the actual operation and the standard operation. Finally, all constructed offset vectors are integrated into a set, namely the operation offset vector set. This set can intuitively reflect in what aspects and to what extent the actual work deviates from the standard work. For example, if the actual value of a certain work parameter is higher or lower than the standard value, the work process will be delayed or accelerated unnecessarily.

[0027] After having the set of operation offset vectors, the system terminal conducts a depth evaluation of these offsets and generates an operation offset depth coefficient. This coefficient can comprehensively evaluate the severity of the operation offset, and it takes into account multiple factors, such as the duration of the offset, the size of the offset, and the impact on the operation efficiency or safety. Specifically, the system terminal assigns weights to each offset parameter based on the importance of the offset parameter, the impact on the operation efficiency or safety, and other factors, and based on relevant professional knowledge. For each offset vector, the initial offset depth value is weighted and calculated according to the offset size and the corresponding weight, and the initial offset depth value is added with a sign according to the offset direction to obtain the standard offset depth value. For example, if it is a forward offset, the preliminary offset depth value remains unchanged, and conversely, if it is a reverse offset, a negative sign is added to the preliminary offset depth value. Subsequently, it is determined whether the duration of the offset has an impact on the operation offset depth. If it has an impact, the standard offset depth value is multiplied by the time length to obtain the offset depth value. After that, all the offset depth values ​​obtained are weighted averaged to obtain the final operation offset depth coefficient. Through this coefficient, the operation offset situation can be more comprehensively understood, providing a basis for subsequent improvement and optimization. In summary, by performing operation offset analysis on the target location monitoring data partition, the differences between the actual operation and the standard operation can be identified, and these differences can be quantified to generate an operation offset vector set. Furthermore, by performing a deep evaluation on these offsets, the operation offset depth coefficient can be obtained, thereby gaining a more comprehensive understanding of the operation offset situation.

[0028] If the operation offset depth coefficient is greater than / equal to an offset depth coefficient threshold, the operation offset vector set is used as a first crane offset identification result, and the first crane offset identification result is added to the crane offset detection result.

[0029] Preferably, the operation offset depth coefficient is an important indicator for evaluating the degree of crane operation offset. When the calculated operation offset depth coefficient is greater than or equal to the preset offset depth coefficient threshold, it means that the crane's operation offset has reached a serious level that requires attention. In this case, the system terminal uses the previously generated operation offset vector set as the first crane offset identification result, which can show in detail which parameters or operations of the crane have offsets, as well as the specific degree and direction of the offset. Subsequently, this first crane offset identification result is added to the set of crane offset detection results. This set is used to record all offset conditions that occur during the operation of the crane, which is convenient for subsequent analysis and processing. In this way, the crane's operation offset can be discovered in a timely manner, and appropriate measures can be taken to correct it, thereby ensuring the safety and efficiency of the crane's operation.

[0030] Further, the present application provides that, according to the crane offset detection algorithm, in combination with the crane monitoring data source, multi-dimensional offset detection is performed on the container gantry crane to generate a crane offset detection result, and the method also includes: Obtaining a retrieval constraint object, wherein the retrieval constraint object includes a container gantry crane and a plurality of cranes of the same model as the container gantry crane; Perform normal environment monitoring sample retrieval according to the retrieval constraint object to generate a crane normal environment monitoring sample library; Preferably, the system terminal obtains a specific retrieval constraint object, which mainly includes container gantry cranes and multiple cranes of the same model. This step is to ensure that the retrieval scope is accurate and specific, so that high-quality data related to the target crane can be collected. Subsequently, the system terminal uses the retrieval constraint object to search in the existing equipment data warehouse. The retrieval content mainly includes environmental parameters of the crane under a normal working environment. Among them, the normal environment refers to the environmental conditions of the crane during normal operation and operation. Afterwards, the retrieved raw data is cleaned to remove duplicate, erroneous or abnormal data. Then, the cleaned data is organized into structured samples to form a normal environment monitoring sample library for the crane. This sample library will become an important reference for subsequent analysis, comparison and evaluation of the operating status of the crane, which will help to promptly discover the abnormal status of the crane, so as to take corresponding maintenance and management measures.

[0031] The normal environment monitoring sample library of the crane is transformed according to a reference vector transformation operator to generate an environment monitoring reference vector set, wherein the reference vector transformation operator includes standardization processing, sample partitioning, and partition concentration interval calculation; Preferably, the system terminal uses the reference vector conversion operator to convert the normal environment monitoring sample library of the crane, and the process of generating the environmental monitoring reference vector set is actually a series of processing and conversion of the original sample data, so as to better extract and analyze the normal environmental status information of the crane. Among them, the reference vector conversion operator includes standardization processing, sample partitioning, and partition concentration interval calculation. Standardization processing is to eliminate the dimensional differences between different parameters or indicators so that all data are in the same comparable range. Sample partitioning is to divide the standardized sample data into different subsets to describe the normal environmental state of the crane in more detail. Partition concentration interval calculation is to calculate the central tendency and discreteness of the sample data in each partition as a component of the environmental monitoring reference vector. Specifically, standardization processing is a key step, which can adjust different parameters or indicators in the sample data to the same dimension and range, eliminate the dimensional differences between different parameters, and make subsequent analysis and comparison more accurate and reliable. The system terminal first calls the standardization processing sub-algorithm in the reference vector conversion operator. In this sub-algorithm, the reference vector conversion operator calculates the mean and standard deviation of each monitoring parameter, such as temperature, humidity, pressure, etc. Then, using a similar method to the above-mentioned method of generating the standard data stream for crane monitoring, the Z-score standardization method is used, combined with the calculation of the mean and standard deviation, to convert the monitoring parameter value of each sample into a standard score. After standardization, all monitoring parameters will be converted to values ​​with a mean of 0 and a standard deviation of 1. Subsequently, the standardized crane normal environment monitoring sample library is transferred to the sample partitioning sub-algorithm, which uses the K-means clustering algorithm to divide the standardized crane normal environment monitoring sample library into K partitions through similar steps as mentioned above. The samples in each partition have similar environmental state characteristics. After that, the partition concentration interval calculation is performed on the K partitions. For each partition, the mean, median, mode and other statistics representing the central tendency are calculated. At the same time, the standard deviation, interquartile range and other statistics representing the degree of dispersion are calculated. These statistics are then sorted to obtain the statistical characteristics of each partition. These statistical characteristics will serve as an important part of the environmental monitoring benchmark vector for subsequent state monitoring and fault diagnosis. Through this series of processing and transformation, the final generated environmental monitoring benchmark vector set will be able to better reflect the state characteristics of the crane under normal conditions, providing strong support for subsequent state recognition and anomaly detection.

[0032] Performing standardization processing on the crane environment monitoring data stream to generate a crane environment monitoring vector set; Performing deviation identification and evaluation on the crane environment monitoring vector set according to the environment monitoring reference vector set, generating a normal environment deviation vector set and a normal environment deviation coefficient; Preferably, according to the environmental monitoring data stream of the crane, the system terminal still uses the Z-score standardization method for standardization processing. This step is mainly to eliminate the dimensional differences between the data so that all data can be within the same comparable range. Through standardization processing, the original environmental monitoring data stream can be converted into a crane environmental monitoring vector set, which is easier to analyze and compare in the future. Subsequently, using the aforementioned method similar to the generation of the operation offset vector set and the operation offset depth coefficient, the previously generated environmental monitoring reference vector set is used to perform deviation identification and evaluation on the crane environmental monitoring vector set. This evaluation process is to compare the difference between the real-time monitoring data and the reference data, so as to identify and record whether the crane deviates from the normal environmental state during operation. Through this step, a normal environmental deviation vector set can be generated, and a normal environmental deviation coefficient can be calculated. Among them, the normal environmental deviation vector set reflects the deviation between the real-time monitoring data of the crane and the reference data. The normal environmental deviation coefficient is a quantitative representation of the degree of deviation. In summary, by standardizing the environmental monitoring data stream and using the reference vector set for deviation identification and evaluation, a vector set and a deviation coefficient reflecting the deviation of the crane environmental state are generated. This provides important data support for subsequent crane status monitoring and fault diagnosis.

[0033] If the normal environment deviation coefficient is greater than / equal to the normal environment deviation threshold, the normal environment deviation vector set is used as the second crane offset identification result, and the second crane offset identification result is added to the crane offset detection result.

[0034] Preferably, if the calculated normal environment deviation coefficient exceeds the preset normal environment deviation threshold, it means that the current environmental state of the crane has obviously deviated from the normal state. In this case, the system terminal uses the calculated normal environment deviation vector set as the second crane offset identification result, which reflects the specific deviation between the current and normal environmental states of the crane. Subsequently, this offset identification result is added to the offset detection result of the crane for subsequent analysis and processing. This helps to promptly detect abnormal conditions of the crane, take appropriate measures for adjustment and maintenance, and ensure the normal operation and operational safety of the crane.

[0035] Furthermore, the present application provides a method for performing multi-dimensional offset detection on the container gantry crane in combination with the crane monitoring data source according to the crane offset detection algorithm to generate a crane offset detection result, and the method further includes: Activate a pre-built crane rail gnawing detection channel, wherein the crane rail gnawing detection channel includes an enhanced pre-processing model, a crane rail gnawing identification model, a crane rail gnawing evaluation model, and a crane rail gnawing verification model; Optionally, the system terminal activates a pre-built crane rail gnawing detection channel, which means that a series of models specifically used to detect crane rail gnawing problems have been started. This detection channel includes several key links: first, the raw data is optimized and processed by the enhanced preprocessing model to improve the data quality; then, the crane rail gnawing identification model is used to conduct in-depth analysis of the processed data to accurately identify whether the crane has rail gnawing; then, the crane rail gnawing evaluation model is used to quantitatively evaluate the identification results to determine the severity of the rail gnawing; then, the crane rail gnawing verification model is used to verify the results of the previous steps to ensure the accuracy of the detection results. Specifically, the enhanced preprocessing model includes a histogram equalization algorithm, and the crane track monitoring data stream is image enhanced according to the histogram equalization algorithm. Histogram equalization is an image enhancement algorithm that enhances the contrast of the image by stretching the pixel intensity distribution range. If the image contrast in the crane track monitoring data stream is low, making it difficult to distinguish details, the histogram equalization algorithm is used to enhance the image. This algorithm can improve the brightness distribution of the image, making the image clearer and easier to analyze. The crane rail gnawing verification model includes a rail gnawing depth evaluation threshold, and the results generated by the crane rail gnawing identification model and the crane rail gnawing evaluation model are verified according to the rail gnawing depth evaluation threshold. The rail gnawing depth evaluation threshold is determined based on the crane's design parameters, operating experience and safety standards, and is a numerical limit.

[0036] For the crane rail gnawing recognition model, the crane rail gnawing recognition model is obtained through training of the system terminal. Specifically, the system terminal collects a large amount of data about the operation status of the crane, especially the status data of the contact part with the rail. These data include the wear of the rail, the operation trajectory of the crane, the speed, load and other information. Subsequently, these raw data are preprocessed to remove noise and outliers, and the data with rail gnawing phenomenon are marked for subsequent model training. After that, based on the data preprocessing, the features related to the crane rail gnawing are further extracted. These features include the depth, width, shape and so on of the rail wear. Through feature extraction and selection, the features that have an important influence on the rail gnawing recognition are screened out to improve the recognition accuracy of the model. Then, according to the nature of the problem and the characteristics of the data, the neural network is selected to construct the initial crane rail gnawing recognition model. During the model training process, the system terminal uses the extracted data as the training set, and optimizes the initial crane rail gnawing recognition model by continuously adjusting the parameters and structure of the initial crane rail gnawing recognition model, so that the initial crane rail gnawing recognition model can accurately identify whether the crane has rail gnawing phenomenon. When the training reaches the maximum number of iterations, the crane rail gnawing recognition model is output.

[0037] For the crane rail gnawing evaluation model, the system terminal extracts features related to rail gnawing evaluation from a large amount of data collected previously. These features include rail wear, crane operating status, environmental factors, etc. Subsequently, the system terminal still selects the neural network to construct the initial crane rail gnawing evaluation model. And using the same method, the initial crane rail gnawing evaluation model is trained with the extracted data to obtain the crane rail gnawing evaluation model.

[0038] After obtaining the enhanced preprocessing model, crane rail gnawing identification model, crane rail gnawing evaluation model and crane rail gnawing verification model, the system terminal integrates the enhanced preprocessing model, crane rail gnawing identification model, crane rail gnawing evaluation model and crane rail gnawing verification model to form a crane rail gnawing detection channel. The crane rail gnawing detection channel can work in coordination with the four models to form a complete crane rail gnawing detection process, providing strong support for timely discovery and resolution of rail gnawing problems.

[0039] Inputting the crane track monitoring data stream into the enhanced preprocessing model to obtain track enhanced monitoring data; Inputting the vehicle rail enhanced monitoring data into the crane rail gnawing recognition model to generate rail gnawing feature recognition results; Inputting the rail gnawing feature recognition result into the crane rail gnawing evaluation model to obtain a rail gnawing depth evaluation coefficient; Inputting the rail gnawing feature recognition result and the rail gnawing depth evaluation coefficient into the crane rail gnawing verification model, wherein the crane rail gnawing verification model includes a rail gnawing depth evaluation threshold; Preferably, the system terminal inputs the crane track monitoring data stream into the enhanced preprocessing model, processes it through the internal histogram equalization algorithm, improves the clarity of the data and the extractability of the features, and thus obtains track enhanced monitoring data. Subsequently, the obtained track enhanced monitoring data is passed to the crane rail gnawing recognition model. The crane rail gnawing recognition model will identify and analyze the features in the data based on the learned content, determine whether there is a rail gnawing phenomenon, and generate a rail gnawing feature recognition result. After that, the rail gnawing feature recognition result is input into the crane rail gnawing evaluation model. The crane rail gnawing evaluation model will further evaluate the severity of the rail gnawing and calculate the rail gnawing depth evaluation coefficient, which can quantify the depth of the rail gnawing. Finally, the rail gnawing feature recognition result and the rail gnawing depth evaluation coefficient are input into the crane rail gnawing verification model. The crane rail gnawing verification model will verify the transmitted recognition and evaluation results according to the internal preset rail gnawing depth evaluation threshold to ensure the accuracy and reliability of the results. Through the above steps, the crane's rail gnawing situation can be comprehensively evaluated, providing strong data support for subsequent maintenance and management.

[0040] If the rail gnawing depth evaluation coefficient is greater than / equal to the rail gnawing depth evaluation threshold, the rail gnawing feature recognition result is used as the third crane offset recognition result, and the third crane offset recognition result is added to the crane offset detection result.

[0041] Preferably, if the calculated rail gnawing depth evaluation coefficient exceeds the preset rail gnawing depth evaluation threshold, it means that the rail gnawing phenomenon of the crane has reached a relatively serious level. In this case, the system terminal uses the rail gnawing feature recognition result as the third crane offset recognition result, which describes in detail the specific characteristics and conditions of the crane rail gnawing. Then, this third crane offset recognition result is added to the existing crane offset detection result, so as to have a more comprehensive understanding and grasp of the overall offset status of the crane. Doing so helps to promptly discover and deal with the crane's rail gnawing problem, ensuring the normal operation and operational safety of the crane.

[0042] Furthermore, the present application provides a method for performing multi-dimensional offset detection on the container gantry crane in combination with the crane monitoring data source according to the crane offset detection algorithm to generate a crane offset detection result, and the method further includes: The crane monitoring data source also includes a crane temperature monitoring stream, a crane vibration monitoring stream and a crane noise monitoring stream; Based on the crane operation monitoring data stream, the crane environment monitoring data stream, the crane temperature monitoring stream, the crane vibration monitoring stream and the crane noise monitoring stream, combined with a pre-built crane fault prediction model, a fault prediction is performed to generate a crane fault prediction result, wherein the crane fault prediction result includes a predicted fault risk coefficient; Preferably, in addition to the crane operation monitoring data stream, the crane environment monitoring data stream and the crane track monitoring data stream, the system terminal also extracts the crane temperature monitoring stream, the vibration monitoring stream and the noise monitoring stream from the crane monitoring data source. These data provide the system terminal with real-time status information of the crane in multiple aspects. Subsequently, fault prediction is performed by combining these monitoring data streams with a pre-built crane fault prediction model. This model can predict the risk of future failure of the crane based on current and past monitoring data. The prediction results are presented in the form of a predicted failure risk coefficient, which can quantify the possibility of a crane failure. In this way, the system terminal can promptly discover potential failure hazards of the crane, provide an important basis for prevention and maintenance work, and ensure the safe and stable operation of the crane.

[0043] For the crane fault prediction model, the system terminal extracts features related to crane faults from a large amount of crane operation data previously acquired. For example, crane temperature features, crane vibration features, and crane noise features. Then collect a large amount of historical fault data. Subsequently, continue to use the construction of the initial crane fault prediction model, and use the same method to train the constructed initial crane fault prediction model with the extracted features and the collected historical fault data to obtain the crane fault prediction model. This crane fault prediction model will be used for subsequent fault predictions and generate a predicted fault risk coefficient.

[0044] If the predicted failure risk coefficient satisfies the failure risk constraint, the crane failure prediction result is used as the fourth crane offset identification result, and the fourth crane offset identification result is added to the crane offset detection result.

[0045] Preferably, if the predicted failure risk coefficient calculated by the crane failure prediction model satisfies the preset failure risk constraint condition, it means that the crane has a certain failure risk. In this case, the system terminal uses the crane failure prediction result as the fourth crane offset identification result, because the predicted failure can be regarded as an offset of the crane in the operating state. Subsequently, the system terminal adds this fourth crane offset identification result to the existing crane offset detection result, so as to have a more comprehensive understanding and grasp of the overall offset status of the crane. In this way, the potential failure risk of the crane can be discovered in time, and corresponding measures can be taken for prevention or maintenance to ensure the normal operation and operation safety of the crane.

[0046] According to the crane correction registration decision channel, a correction decision is made for the container gantry crane in combination with the crane offset detection result, a crane correction plan is generated, and correction operation and maintenance of the container gantry crane is performed according to the crane correction plan.

[0047] In one embodiment, the system terminal performs a detailed analysis of the offset detection results of the container gantry crane based on the pre-built crane correction and registration decision channel. Through this analysis, it can be determined whether the crane is offset and the degree of offset. Subsequently, the crane correction and registration decision channel calls the correction and registration recognition network and the crane correction record library to generate a specific correction plan based on the results of the analysis. This plan will clearly indicate how to correct the crane to ensure that it returns to normal. After that, the crane is operated and maintained according to the correction plan, including adjusting the position, angle and other parameters of the crane until its offset problem is corrected. In this way, it can be ensured that the container gantry crane can operate safely and efficiently.

[0048] Furthermore, the present application provides a method for making a correction decision for the container gantry crane according to a crane correction registration decision channel and combining the crane offset detection result to generate a crane correction plan, and the method also includes: Extracting target crane deviation correction scene features according to the crane deviation detection results, wherein the target crane deviation correction scene features include any one crane deviation recognition result in the crane deviation detection results; Preferably, after obtaining the crane offset detection results, the system terminal extracts specific target crane offset correction scene features from these results. These target crane offset correction scene features include any crane offset identification results in the detection results, which can directly reflect the offset of the crane in a specific scene. By extracting these features, it is possible to more accurately understand under what circumstances the crane is prone to offset and the specific manifestations of the offset. In this way, a more targeted correction plan can be formulated to improve the efficiency and accuracy of the correction.

[0049] The crane correction registration decision channel includes a crane correction record library and a correction registration recognition network, wherein the crane correction record library includes multiple groups of crane correction records of the same model, and each group of crane correction records of the same model includes sample crane correction scene features and sample crane correction decisions; Preferably, the crane correction registration decision channel is a channel that integrates multiple information and technologies, including a crane correction record library and a correction registration recognition network, which is designed to assist the system terminal in performing correction operations on container gantry cranes. Among them, the crane correction record library is a core component, which stores a large number of correction records of cranes of the same model. Each set of correction records records in detail the offset characteristics of the sample crane in a specific scenario, as well as the correction decisions taken for these offsets. These records provide valuable experience and reference for the crane correction registration decision channel, helping to better understand the offset of the crane in different scenarios and formulate effective correction plans accordingly. The correction registration recognition network is another key element in this decision channel. It uses a neural network to identify and analyze the target crane correction scene features extracted from the crane offset detection results. By comparing and matching with the sample features in the crane correction record library, it automatically recommends appropriate correction decisions to provide intelligent support for the system terminal.

[0050] The system terminal collects a large number of historical correction records of cranes, including scene features of each correction, correction measures taken, and the effect after correction, and classifies this information, and aggregates the same category. Subsequently, the sorted correction records are classified and stored according to the crane model to form a crane correction record library. Each group of correction records is then marked and indexed to facilitate subsequent query and matching operations. After that, the scene features in the crane correction records are analyzed to extract the key features that have an important impact on the correction decision. These features are then encoded and converted into a format that can be processed by the neural network for subsequent matching and recognition. Then, the initial correction registration recognition network is constructed, and the feature data extracted from the historical correction records and the corresponding correction decisions are used as training data to train and optimize the initial correction registration recognition network. By continuously adjusting the parameters and structure of the initial correction registration recognition network, the recognition accuracy and matching efficiency of the initial correction registration recognition network for the crane correction scene features are improved. Until the maximum number of training times is reached, the trained correction registration recognition network is output. Finally, the trained correction registration recognition network is integrated with the crane correction record library to form a complete crane correction registration decision channel. The decision channel is then tested and verified using independent data that is not used for training to ensure its reliability and effectiveness in practical applications. If it meets the expected requirements, the crane correction registration decision channel is output.

[0051] In general, the crane correction and registration decision channel provides strong technical support for the crane's correction operation and maintenance by integrating the functions of the crane correction record library and the correction and registration recognition network, thereby improving the accuracy and efficiency of correction.

[0052] Based on the correction registration recognition network, the correction scene feature of the target crane and the correction scene features of each sample crane in the crane correction record library are subjected to correction scene feature similarity recognition to generate a plurality of sample correction registration recognition coefficients; Preferably, based on the correction and registration recognition network inside the crane correction and registration decision channel, the system terminal identifies the correction scene features of the target crane and compares it with the correction scene features of each sample crane in the crane correction record library. This comparison process is essentially looking for similarity, that is, identifying which scenes in historical records the correction scene of the target crane is closest to. Through this recognition process, the correction and registration recognition network generates a plurality of sample correction and registration recognition coefficients, which represent the degree of similarity between the correction scene of the target crane and each sample scene. These coefficients not only provide the system terminal with an in-depth understanding of the correction scene of the target crane, but also provide an important reference basis for subsequent correction decisions.

[0053] Extracting an optimal deflection correction and registration recognition coefficient according to the plurality of sample deflection correction and registration recognition coefficients; Decision extraction is performed on the crane correction record library according to the optimal correction registration recognition coefficient to generate a target scene correction decision, and the target scene correction decision is added to the crane correction plan.

[0054] Preferably, the system terminal further analyzes and extracts the optimal correction and registration recognition coefficient based on the obtained multiple sample correction and registration recognition coefficients, that is, the largest one among these sample correction and registration recognition coefficients. This optimal coefficient represents the highest similarity between the correction scene of the target crane and a sample scene in the crane correction record library. Subsequently, according to this optimal correction and registration recognition coefficient, the corresponding correction decision is searched in the crane correction record library. This decision is the correction measure taken for the scene that is most similar to the target scene in the historical records. Once this decision is found, the system terminal extracts the corresponding correction decision as the target scene correction decision, and adds it to the crane correction plan. In this way, the system terminal can formulate a suitable correction plan for the current target crane correction scene based on historical experience and data.

[0055] In summary, the embodiments of the present application have at least the following technical effects: The embodiment of the present application retrieves the frequency of the crane's deviation correction operation and maintenance according to a preset historical window, and generates a deviation correction feature frequent coefficient. Subsequently, these coefficients are input into the monitoring computing power calculation branch to generate the target monitoring computing power. After that, the crane monitoring sensor module is feature activated according to the target monitoring computing power to obtain a feature monitoring sensor module for real-time monitoring of the crane. During the monitoring process, a crane monitoring data source is generated, including an operation monitoring data stream, an environmental monitoring data stream, and a vehicle track monitoring data stream. Based on these data, a multi-dimensional deviation detection is performed in combination with the crane offset detection algorithm to generate an offset detection result. Then, according to the offset detection result, a deviation correction decision is made through the deviation correction registration decision channel, a deviation correction plan is generated, and deviation correction operation and maintenance is performed. These technical effects jointly solve the technical problems of inaccurate deviation correction and untimely response caused by ignoring external environmental changes and equipment wear during actual operation of container gantry cranes due to relying on traditional sensors and fixed algorithms, and realizing the improvement of the accuracy and efficiency of the deviation correction of container gantry cranes through automated monitoring and intelligent decision-making, reducing the risk of operation interruption caused by offset, and enhancing the reliability and safety of the equipment.

[0056] Embodiment 2 Based on the same inventive concept as the deviation correction method of a container gantry crane in the aforementioned embodiment, Figure 2 As shown, the present application provides a deviation correction system for a container gantry crane, the system comprising: Frequent coefficient generation component 1: The frequent coefficient generation component 1 is used to retrieve the deviation correction operation and maintenance frequency of the container gantry crane according to a predetermined historical window, and generate a deviation correction feature frequent coefficient; Monitoring computing power generation component 2: the monitoring computing power generation component 2 is used to input the correction feature frequency coefficient into the monitoring computing power calculation branch to generate the target monitoring computing power; Feature activation component 3: the feature activation component 3 is used to perform feature activation on the crane monitoring sensor module according to the target monitoring computing power to obtain a feature monitoring sensor module; Real-time monitoring component 4: the real-time monitoring component 4 is used to monitor the container gantry crane in real time according to the characteristic monitoring sensor module, and generate a crane monitoring data source, wherein the crane monitoring data source includes a crane operation monitoring data stream, a crane environment monitoring data stream and a crane track monitoring data stream; Multidimensional offset detection component 5: the multidimensional offset detection component 5 is used to perform multidimensional offset detection on the container gantry crane according to the crane offset detection algorithm and in combination with the crane monitoring data source, and generate a crane offset detection result; Correction operation and maintenance component 6: The correction operation and maintenance component 6 is used to make a correction decision on the container gantry crane according to the crane correction registration decision channel and the crane offset detection result, generate a crane correction plan, and perform correction operation and maintenance on the container gantry crane according to the crane correction plan.

[0057] Furthermore, the monitoring computing power generation component 2 is also used to execute the following method: The monitoring computing power calculation branch includes a monitoring computing power calculation function, and the monitoring computing power calculation function is: ; Among them, TMC represents the target monitoring computing power, TMA represents the preset benchmark monitoring computing power, TMB represents the correction feature frequency coefficient, TMC represents the preset correction feature frequency threshold, BCW represents the preset weight coefficient, and 0<BCW<1.

[0058] Furthermore, the multi-dimensional offset detection component 5 is also used to perform the following method: Performing standardization processing on the crane operation monitoring data stream to generate a crane monitoring standard data stream; Performing monitoring position clustering according to the crane monitoring standard data stream to obtain multiple position monitoring data partitions; Extracting a target position monitoring data partition corresponding to a target monitoring position according to the multiple position monitoring data partitions, wherein the target monitoring position is any monitoring position of the container gantry crane; Loading a predetermined operation plan of the container gantry crane, matching a target operation plan of the target monitoring position based on the predetermined operation plan, and performing standard processing on the target operation plan to generate a standard operation partition of the target position; Performing operation offset analysis on the target location monitoring data partition according to the target location standard operation partition to generate an operation offset vector set; Performing an offset depth evaluation according to the operation offset vector set to generate an operation offset depth coefficient; If the operation offset depth coefficient is greater than / equal to an offset depth coefficient threshold, the operation offset vector set is used as a first crane offset identification result, and the first crane offset identification result is added to the crane offset detection result.

[0059] Furthermore, the multi-dimensional offset detection component 5 is also used to perform the following method: Obtaining a retrieval constraint object, wherein the retrieval constraint object includes a container gantry crane and a plurality of cranes of the same model as the container gantry crane; Perform normal environment monitoring sample retrieval according to the retrieval constraint object to generate a crane normal environment monitoring sample library; The normal environment monitoring sample library of the crane is transformed according to a reference vector transformation operator to generate an environment monitoring reference vector set, wherein the reference vector transformation operator includes standardization processing, sample partitioning, and partition concentration interval calculation; Performing standardization processing on the crane environment monitoring data stream to generate a crane environment monitoring vector set; Performing deviation identification and evaluation on the crane environment monitoring vector set according to the environment monitoring reference vector set, generating a normal environment deviation vector set and a normal environment deviation coefficient; If the normal environment deviation coefficient is greater than / equal to the normal environment deviation threshold, the normal environment deviation vector set is used as the second crane offset identification result, and the second crane offset identification result is added to the crane offset detection result.

[0060] Furthermore, the multi-dimensional offset detection component 5 is also used to perform the following method: Activate a pre-built crane rail gnawing detection channel, wherein the crane rail gnawing detection channel includes an enhanced pre-processing model, a crane rail gnawing identification model, a crane rail gnawing evaluation model, and a crane rail gnawing verification model; Inputting the crane track monitoring data stream into the enhanced preprocessing model to obtain track enhanced monitoring data; Inputting the vehicle rail enhanced monitoring data into the crane rail gnawing recognition model to generate rail gnawing feature recognition results; Inputting the rail gnawing feature recognition result into the crane rail gnawing evaluation model to obtain a rail gnawing depth evaluation coefficient; Inputting the rail gnawing feature recognition result and the rail gnawing depth evaluation coefficient into the crane rail gnawing verification model, wherein the crane rail gnawing verification model includes a rail gnawing depth evaluation threshold; If the rail gnawing depth evaluation coefficient is greater than / equal to the rail gnawing depth evaluation threshold, the rail gnawing feature recognition result is used as the third crane offset recognition result, and the third crane offset recognition result is added to the crane offset detection result.

[0061] Furthermore, the multi-dimensional offset detection component 5 is also used to perform the following method: The crane monitoring data source also includes a crane temperature monitoring stream, a crane vibration monitoring stream and a crane noise monitoring stream; Based on the crane operation monitoring data stream, the crane environment monitoring data stream, the crane temperature monitoring stream, the crane vibration monitoring stream and the crane noise monitoring stream, combined with a pre-built crane fault prediction model, a fault prediction is performed to generate a crane fault prediction result, wherein the crane fault prediction result includes a predicted fault risk coefficient; If the predicted failure risk coefficient satisfies the failure risk constraint, the crane failure prediction result is used as the fourth crane offset identification result, and the fourth crane offset identification result is added to the crane offset detection result.

[0062] Furthermore, the deviation correction operation and maintenance component 6 is also used to perform the following method: Extracting target crane deviation correction scene features according to the crane deviation detection results, wherein the target crane deviation correction scene features include any one crane deviation recognition result in the crane deviation detection results; The crane correction registration decision channel includes a crane correction record library and a correction registration recognition network, wherein the crane correction record library includes multiple groups of crane correction records of the same model, and each group of crane correction records of the same model includes sample crane correction scene features and sample crane correction decisions; Based on the correction registration recognition network, the correction scene feature of the target crane and the correction scene features of each sample crane in the crane correction record library are subjected to correction scene feature similarity recognition to generate a plurality of sample correction registration recognition coefficients; Extracting an optimal deflection correction and registration recognition coefficient according to the plurality of sample deflection correction and registration recognition coefficients; Decision extraction is performed on the crane correction record library according to the optimal correction registration recognition coefficient to generate a target scene correction decision, and the target scene correction decision is added to the crane correction plan.

[0063] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. Other embodiments are within the scope of the attached claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0064] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0065] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A deviation correction method for a container gantry crane. It is characterized in that The method comprises: Retrieve the deviation correction operation and maintenance frequency of container gantry cranes according to the predetermined historical window and generate the deviation correction feature frequency coefficient; Inputting the correction feature frequency coefficient into the monitoring computing power calculation branch to generate the target monitoring computing power; Performing feature activation on the crane monitoring sensor module according to the target monitoring computing power to obtain a feature monitoring sensor module; According to the characteristic monitoring sensor module, the container gantry crane is monitored in real time to generate a crane monitoring data source, wherein the crane monitoring data source includes a crane operation monitoring data stream, a crane environment monitoring data stream and a crane track monitoring data stream; According to the crane offset detection algorithm, in combination with the crane monitoring data source, multi-dimensional offset detection is performed on the container gantry crane to generate a crane offset detection result; According to the crane correction registration decision channel, a correction decision is made for the container gantry crane in combination with the crane offset detection result, a crane correction plan is generated, and correction operation and maintenance of the container gantry crane is performed according to the crane correction plan.

2. The method according to claim 1, It is characterized in that The monitoring computing power calculation branch includes a monitoring computing power calculation function, and the monitoring computing power calculation function is: ； Among them, TMC represents the target monitoring computing power, TMA represents the preset benchmark monitoring computing power, TMB represents the correction feature frequency coefficient, TMC represents the preset correction feature frequency threshold, BCW represents the preset weight coefficient, and 0<BCW<1.

3. The method according to claim 1, It is characterized in that According to the crane offset detection algorithm, in combination with the crane monitoring data source, a multi-dimensional offset detection is performed on the container gantry crane to generate a crane offset detection result, including: Performing standardization processing on the crane operation monitoring data stream to generate a crane monitoring standard data stream; Performing monitoring position clustering according to the crane monitoring standard data stream to obtain multiple position monitoring data partitions; Extracting a target position monitoring data partition corresponding to a target monitoring position according to the multiple position monitoring data partitions, wherein the target monitoring position is any monitoring position of the container gantry crane; Loading a predetermined operation plan of the container gantry crane, matching a target operation plan of the target monitoring position based on the predetermined operation plan, and performing standard processing on the target operation plan to generate a standard operation partition of the target position; Performing operation offset analysis on the target location monitoring data partition according to the target location standard operation partition to generate an operation offset vector set; Performing an offset depth evaluation according to the operation offset vector set to generate an operation offset depth coefficient; If the operation offset depth coefficient is greater than / equal to an offset depth coefficient threshold, the operation offset vector set is used as a first crane offset identification result, and the first crane offset identification result is added to the crane offset detection result.

4. The method according to claim 1, It is characterized in that According to the crane offset detection algorithm, in combination with the crane monitoring data source, a multi-dimensional offset detection is performed on the container gantry crane to generate a crane offset detection result, including: Obtaining a retrieval constraint object, wherein the retrieval constraint object includes a container gantry crane and a plurality of cranes of the same model as the container gantry crane; Perform normal environment monitoring sample retrieval according to the retrieval constraint object to generate a crane normal environment monitoring sample library; The normal environment monitoring sample library of the crane is transformed according to a reference vector transformation operator to generate an environment monitoring reference vector set, wherein the reference vector transformation operator includes standardization processing, sample partitioning, and partition concentration interval calculation; Performing standardization processing on the crane environment monitoring data stream to generate a crane environment monitoring vector set; Performing deviation identification and evaluation on the crane environment monitoring vector set according to the environment monitoring reference vector set, generating a normal environment deviation vector set and a normal environment deviation coefficient; If the normal environment deviation coefficient is greater than / equal to the normal environment deviation threshold, the normal environment deviation vector set is used as the second crane offset identification result, and the second crane offset identification result is added to the crane offset detection result.

5. The method according to claim 1, It is characterized in that According to the crane offset detection algorithm, in combination with the crane monitoring data source, a multi-dimensional offset detection is performed on the container gantry crane to generate a crane offset detection result, including: Activate a pre-built crane rail gnawing detection channel, wherein the crane rail gnawing detection channel includes an enhanced pre-processing model, a crane rail gnawing identification model, a crane rail gnawing evaluation model, and a crane rail gnawing verification model; Inputting the crane track monitoring data stream into the enhanced preprocessing model to obtain track enhanced monitoring data; Inputting the vehicle rail enhanced monitoring data into the crane rail gnawing recognition model to generate rail gnawing feature recognition results; Inputting the rail gnawing feature recognition result into the crane rail gnawing evaluation model to obtain a rail gnawing depth evaluation coefficient; Inputting the rail gnawing feature recognition result and the rail gnawing depth evaluation coefficient into the crane rail gnawing verification model, wherein the crane rail gnawing verification model includes a rail gnawing depth evaluation threshold; If the rail gnawing depth evaluation coefficient is greater than / equal to the rail gnawing depth evaluation threshold, the rail gnawing feature recognition result is used as the third crane offset recognition result, and the third crane offset recognition result is added to the crane offset detection result.

6. The method according to claim 1, It is characterized in that According to the crane offset detection algorithm, in combination with the crane monitoring data source, a multi-dimensional offset detection is performed on the container gantry crane to generate a crane offset detection result, including: The crane monitoring data source also includes a crane temperature monitoring stream, a crane vibration monitoring stream and a crane noise monitoring stream; Based on the crane operation monitoring data stream, the crane environment monitoring data stream, the crane temperature monitoring stream, the crane vibration monitoring stream and the crane noise monitoring stream, combined with a pre-built crane fault prediction model, a fault prediction is performed to generate a crane fault prediction result, wherein the crane fault prediction result includes a predicted fault risk coefficient; If the predicted failure risk coefficient satisfies the failure risk constraint, the crane failure prediction result is used as the fourth crane offset identification result, and the fourth crane offset identification result is added to the crane offset detection result.

7. The method according to claim 1, It is characterized in that According to the crane deviation correction registration decision channel, a deviation correction decision is made on the container gantry crane in combination with the crane deviation detection result, and a crane deviation correction plan is generated, including: Extracting target crane deviation correction scene features according to the crane deviation detection results, wherein the target crane deviation correction scene features include any one crane deviation recognition result in the crane deviation detection results; The crane correction registration decision channel includes a crane correction record library and a correction registration recognition network, wherein the crane correction record library includes multiple groups of crane correction records of the same model, and each group of crane correction records of the same model includes sample crane correction scene features and sample crane correction decisions; Based on the correction registration recognition network, the correction scene feature of the target crane and the correction scene features of each sample crane in the crane correction record library are subjected to correction scene feature similarity recognition to generate a plurality of sample correction registration recognition coefficients; Extracting an optimal deflection correction and registration recognition coefficient according to the plurality of sample deflection correction and registration recognition coefficients; Decision extraction is performed on the crane correction record library according to the optimal correction registration recognition coefficient to generate a target scene correction decision, and the target scene correction decision is added to the crane correction plan.

8. A deviation correction system for a container gantry crane. It is characterized in that The method for implementing any one of the deviation correction methods 1 to 7 for a container gantry crane comprises: Frequent coefficient generation component: retrieves the deviation correction operation and maintenance frequency of container gantry cranes according to the predetermined historical window, and generates the deviation correction feature frequent coefficient; Monitoring computing power generation component: inputting the correction feature frequency coefficient into the monitoring computing power calculation branch to generate the target monitoring computing power; Feature activation component: feature activation of the crane monitoring sensor module according to the target monitoring computing power to obtain a feature monitoring sensor module; Real-time monitoring component: monitors the container gantry crane in real time according to the characteristic monitoring sensor module, and generates a crane monitoring data source, wherein the crane monitoring data source includes a crane operation monitoring data stream, a crane environment monitoring data stream, and a crane track monitoring data stream; Multi-dimensional offset detection component: performing multi-dimensional offset detection on the container gantry crane according to the crane offset detection algorithm and in combination with the crane monitoring data source, and generating a crane offset detection result; Correction and operation and maintenance component: according to the crane correction and registration decision channel, combined with the crane offset detection result, a correction decision is made for the container gantry crane, a crane correction plan is generated, and correction and operation and maintenance are performed on the container gantry crane according to the crane correction plan.

Images