Forklift accessory fatigue life prediction method and system using transfer learning

Abstract

The present application relates to the technical field of mechanical life prediction, and particularly relates to a forklift accessory fatigue life prediction method and system using transfer learning. A fatigue frequent point list and a frequent stress direction list are obtained; the accessory structure topology, the accessory material parameters, the fatigue frequent point list and the frequent stress direction list are used as constraints to collect a plurality of one-to-one corresponding first accessory operation working condition time sequence data and first identification fatigue life labels, and a fatigue life prediction base model is trained; the target forklift accessory model, the fatigue frequent point list and the frequent stress direction list are used to collect a plurality of one-to-one corresponding second accessory operation working condition time sequence data and second identification fatigue life labels, transfer learning is performed on the fatigue life prediction base model, a target forklift accessory fatigue life prediction model is obtained, and a forklift accessory fatigue life prediction task is executed. The present application solves the problems of insufficient fatigue life prediction accuracy and low efficiency caused by insufficient fatigue monitoring samples.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mechanical life prediction, in particular to a forklift accessory fatigue life prediction method and system using transfer learning. BACKGROUND

[0002] In forklift operation, the accessory is a key component to realize its multifunctionality, which is divided into carrying type, clamping type, adjusting type, etc., and is suitable for various special working conditions. Its fatigue life is crucial to the safety and stability of operation, so accurate prediction of the fatigue life of the forklift accessory is an important demand in the industry. In the current fatigue life prediction of forklift accessories, one means is to collect historical fatigue data, train a fatigue life prediction model for prediction. However, due to uncertain accessory scheduling and large load fluctuations, it is difficult to obtain sufficient and effective fatigue monitoring samples, resulting in insufficient precision and low efficiency of the trained prediction model in the prior art for modeling and predicting the fatigue life of the forklift accessory, which is difficult to meet the demand for fatigue life prediction of the forklift accessory. SUMMARY

[0003] The present application provides a forklift accessory fatigue life prediction method and system using transfer learning to solve the technical problems of insufficient precision and low efficiency of the trained prediction model due to insufficient fatigue monitoring samples in the prior art.

[0004] The technical solution of the present application to solve the above technical problems is as follows:

[0005] In a first aspect, the present application provides a forklift accessory fatigue life prediction method using transfer learning, comprising: loading operation monitoring data of a target forklift accessory, performing fatigue position frequent pattern analysis to obtain a fatigue frequent point list; based on the operation monitoring data, traversing the fatigue frequent point list, performing stress direction frequent pattern analysis to obtain a frequent stress direction list; taking the accessory structure topology, accessory material parameters, the fatigue frequent point list and the frequent stress direction list as constraints, collecting a plurality of one-to-one first accessory operation working condition time series data and first identification fatigue life labels, training a fatigue life prediction base model; taking the target forklift accessory model, the fatigue frequent point list and the frequent stress direction list, collecting a plurality of one-to-one second accessory operation working condition time series data and second identification fatigue life labels, performing transfer learning on the fatigue life prediction base model to obtain a target forklift accessory fatigue life prediction model, and performing a forklift accessory fatigue life prediction task.

[0006] Optionally, the work monitoring data of the target forklift accessory is loaded, a fatigue position frequent pattern analysis is performed, and a fatigue frequent point list is obtained, including: collecting fatigue detection record data with accessory material parameters as constraints, performing fatigue symptom type frequent pattern analysis, and obtaining a frequent fatigue symptom type set; extracting forklift accessory defect monitoring data from the work monitoring data, and counting a forklift accessory position set and a frequent fatigue type trigger frequency set in which at least any one of the frequent fatigue symptom type set appears; and adding a forklift accessory position in which the frequent fatigue type trigger frequency in the frequent fatigue type trigger frequency set is greater than or equal to a trigger frequency threshold to the fatigue frequent point list.

[0007] In the method, the forklift accessory defect monitoring data is extracted from the work monitoring data, and a forklift accessory position set and a frequent fatigue type trigger frequency set in which at least any one of the frequent fatigue symptom type set appears are counted, including: extracting first group defect monitoring data to Qth group defect monitoring data from the forklift accessory defect monitoring data, where Q represents the number of groups of forklift accessory defect monitoring data; integrating the first group defect monitoring data to the Qth group defect monitoring data according to positions to obtain first position defect monitoring data set to Nth position defect monitoring data set, where N represents the number of positions; deleting sets in which the number of defect monitoring data is less than or equal to a data number threshold to obtain a plurality of remaining position defect monitoring data sets; and traversing the plurality of remaining position defect monitoring data sets, counting a forklift accessory position set and a frequent fatigue type trigger frequency set in which at least any one of the frequent fatigue symptom type set appears based on the frequent fatigue symptom type set.

[0008] In the method, the forklift accessory defect monitoring data is extracted from the work monitoring data, and a forklift accessory position set and a frequent fatigue type trigger frequency set in which at least any one of the frequent fatigue symptom type set appears are counted, including: extracting first group defect monitoring data to Qth group defect monitoring data from the forklift accessory defect monitoring data, where Q represents the number of groups of forklift accessory defect monitoring data; integrating the first group defect monitoring data to the Qth group defect monitoring data according to positions to obtain first position defect monitoring data set to Nth position defect monitoring data set, where N represents the number of positions; deleting sets in which the number of defect monitoring data is less than or equal to a data number threshold to obtain a plurality of remaining position defect monitoring data sets; and traversing the plurality of remaining position defect monitoring data sets, counting a forklift accessory position set and a frequent fatigue type trigger frequency set in which at least any one of the frequent fatigue symptom type set appears based on the frequent fatigue symptom type set.

[0009] Optionally, based on the job monitoring data, the list of fatigue frequent points is traversed, stress direction frequent pattern analysis is performed, and a list of frequent stress directions is obtained, including: extracting a first fatigue frequent position from the list of fatigue frequent points; based on the first fatigue frequent position, extracting a first set of fatigue frequent position stress directions from the job monitoring data; based on an angle deviation threshold, performing clustering on the first set of fatigue frequent position stress directions to obtain a plurality of clusters of first fatigue frequent position stress directions; taking a plurality of first fatigue frequent position centroid stress directions of the plurality of clusters of first fatigue frequent position stress directions as a first set of reconstructed fatigue frequent position stress directions, and simultaneously taking the number of clusters of the plurality of clusters of first fatigue frequent position stress directions as a plurality of initial trigger frequencies of the plurality of first fatigue frequent position centroid stress directions, and adding the plurality of initial trigger frequencies to the first set of reconstructed fatigue frequent position stress direction initial trigger frequencies; until the Lth set of reconstructed fatigue frequent position stress directions and the Lth set of reconstructed fatigue frequent position stress direction initial trigger frequencies are obtained; based on the first set of reconstructed fatigue frequent position stress directions and the first set of reconstructed fatigue frequent position stress direction initial trigger frequencies, until the Lth set of reconstructed fatigue frequent position stress directions and the Lth set of reconstructed fatigue frequent position stress direction initial trigger frequencies, stress direction frequent pattern analysis is performed on the first fatigue frequent position; until stress direction frequent pattern analysis is completed on all fatigue frequent points, and a list of frequent stress directions is output.

[0010] Optionally, under the constraints of the tool structure topology, tool material parameters, the fatigue frequent point list and the frequent force direction list, a plurality of one-to-one corresponding first tool operation working condition time series data and first identification fatigue life labels are collected, a fatigue life prediction base model is trained, including: based on the tool structure topology and the tool material parameters, a first level constraint is constructed; based on the fatigue frequent point list and the frequent force direction list, a second level constraint is constructed; a set of samples of tools to be analyzed is loaded, wherein the set of samples of tools to be analyzed has the same to-be-analyzed structure topology, to-be-analyzed material parameters, to-be-analyzed fatigue frequent point list, to-be-analyzed frequent force direction list, tool operation fatigue frequent point force time series data and tool operation environment monitoring time series information; when the to-be-analyzed structure topology has a structure topology similarity with the tool structure topology greater than or equal to a structure topology similarity threshold, and the to-be-analyzed material parameters are the same as the to-be-analyzed material parameters, it is considered that the first level constraint is met; when the first level constraint is met, and the point position distribution similarity of the fatigue frequent point list and the to-be-analyzed fatigue frequent point list is greater than or equal to a point position distribution similarity threshold, and the force direction distribution similarity of the frequent force direction list and the to-be-analyzed frequent force direction list is greater than or equal to a force direction distribution similarity threshold, it is considered that the second level constraint is met; when the second level constraint is met, the tool operation fatigue frequent point force time series data and the tool operation environment monitoring time series information are set as the first tool operation working condition time series data, and the set of samples of tools to be analyzed is subjected to fatigue symptom duration mode analysis to obtain the first identification fatigue life label.

[0011] Optionally, under the constraints of the target forklift tool model, the fatigue frequent point list and the frequent force direction list, a plurality of one-to-one corresponding second tool operation working condition time series data and second identification fatigue life labels are collected, and transfer learning is performed on the fatigue life prediction base model to obtain a target forklift tool fatigue life prediction model, the fatigue life prediction base model being a model integrating the outputs of a plurality of topologically different sub-models, including: obtaining a plurality of fatigue life prediction base sub-models of the fatigue life prediction base model; constructing a fully connected neural network, and connecting the plurality of fatigue life prediction base sub-models in parallel to the input layer of the fully connected neural network; taking the second tool operation working condition time series data as the input of the plurality of fatigue life prediction base sub-models, and taking the second identification fatigue life label as the output of the fully connected neural network, performing transfer learning on the fatigue life prediction base model to obtain the target forklift tool fatigue life prediction model.

[0012] In a second aspect, the present application provides a forklift tool fatigue life prediction system using transfer learning, comprising:

[0013] a frequent mode analysis module, configured to load work monitoring data of a target forklift accessory, perform fatigue position frequent mode analysis, and obtain a fatigue frequent point list;

[0014] a stress direction analysis module, configured to traverse the fatigue frequent point list based on the work monitoring data, perform stress direction frequent mode analysis, and obtain a frequent stress direction list;

[0015] a basic model training module, configured to collect a plurality of first accessory work condition time sequence data and first label fatigue life corresponding to each other and train a fatigue life prediction basic model by taking the accessory structure topology, accessory material parameters, the fatigue frequent point list and the frequent stress direction list as constraints;

[0016] a target model training module, configured to collect a plurality of second accessory work condition time sequence data and second label fatigue life corresponding to each other by taking the fatigue frequent point list and the frequent stress direction list, and perform transfer learning on the fatigue life prediction basic model to obtain a target forklift accessory fatigue life prediction model and perform a forklift accessory fatigue life prediction task.

[0017] By implementing the present application, the work monitoring data of the target forklift accessory can be loaded, the fatigue position frequent mode analysis can be performed, and the fatigue frequent point list can be obtained, the key positions prone to fatigue can be accurately located, the differential analysis of all positions of the accessory can be avoided, the redundant workload of subsequent data processing and model training can be reduced, and the efficiency can be improved.

[0018] By implementing the present application, the work monitoring data can be loaded, the fatigue position frequent mode analysis can be performed, and the fatigue frequent point list can be obtained, the key positions prone to fatigue can be accurately located, the differential analysis of all positions of the accessory can be avoided, the redundant workload of subsequent data processing and model training can be reduced, and the efficiency can be improved.

[0019] By implementing the present application, the work monitoring data can be loaded, the fatigue position frequent mode analysis can be performed, and the fatigue frequent point list can be obtained, the key positions prone to fatigue can be accurately located, the differential analysis of all positions of the accessory can be avoided, the redundant workload of subsequent data processing and model training can be reduced, and the efficiency can be improved.

[0020] By implementing the present application, the target forklift accessory model, the fatigue frequent point list and the frequent stress direction list, a plurality of one-to-one second accessory operation working condition time sequence data and a second identification fatigue life label are collected, the fatigue life prediction base model is migrated learning, the target forklift accessory fatigue life prediction model is obtained, the forklift accessory fatigue life prediction task is executed, the problem that the monitoring sample is few due to the uncertain scheduling and the large load fluctuation of the target accessory is solved, the existing knowledge of the base model is utilized by means of the migration learning, the dependence on a large number of samples of the target accessory is reduced, the modeling difficulty and cost are reduced, and finally the target model is completely adapted to the specific model of the forklift accessory, can accurately predict the fatigue life, provides a scientific basis for the maintenance and replacement of the accessory, reduces the operation interruption and safety accidents caused by fatigue failure.

[0021] In summary, by implementing the present application, the problem that the monitoring sample is few due to the uncertain scheduling and the large load fluctuation of the forklift accessory is solved by means of migration learning, a precise and efficient forklift accessory fatigue life prediction model is trained to realize the rapid and accurate prediction of the fatigue life of the forklift accessory, further reliable technical support is provided for the whole life cycle management of the forklift accessory, which helps to plan the maintenance strategy in advance, reduces the equipment failure risk, and improves the safety and stability of the forklift operation. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 A flowchart of a forklift accessory fatigue life prediction method using migration learning is provided for the present application;

[0023] Figure 2 A structural diagram of a forklift accessory fatigue life prediction system using migration learning is provided for the present application.

[0024] In the drawings, the components represented by the respective reference numerals are as follows:

[0025] The frequent pattern analysis module 11, the stress direction analysis module 12, the base model training module 13, and the target model training module 14. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0027] In the description of the present application, the terms "first", "second" are only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more features. In the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise explicitly specified.

[0028] In the description of the present application, the term "for example" is used to indicate "as an example, illustration or explanation". Any embodiment described as "for example" in the present application is not necessarily interpreted as more preferred or more advantageous than other embodiments. The following description is given in order to enable any person skilled in the art to implement and use the present application. In the following description, details are listed for the purpose of explanation. It should be understood that those skilled in the art can recognize that the present application can be implemented without using these specific details. In other examples, well-known structures and processes will not be described in detail to avoid unnecessary details making the description of the present application obscure. Therefore, the present application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope consistent with the principles and characteristics disclosed.

[0029] Embodiment one, as Figure 1 shown, the present application provides a forklift accessory fatigue life prediction method and system using transfer learning, comprising:

[0030] S100: loading the operation monitoring data of the target forklift accessory, performing fatigue position frequent pattern analysis, and obtaining a fatigue frequent point list;

[0031] S200: based on the operation monitoring data, traversing the fatigue frequent point list, performing stress direction frequent pattern analysis, and obtaining a frequent stress direction list;

[0032] S300: taking the accessory structure topology, accessory material parameters, the fatigue frequent point list and the frequent stress direction list as constraints, collecting a plurality of one-to-one corresponding first accessory operation working condition time series data and first identification fatigue life labels, training a fatigue life prediction base model;

[0033] S400: taking the target forklift accessory model, the fatigue frequent point list and the frequent stress direction list, collecting a plurality of one-to-one corresponding second accessory operation working condition time series data and second identification fatigue life labels, performing transfer learning on the fatigue life prediction base model, obtaining a target forklift accessory fatigue life prediction model, and performing a forklift accessory fatigue life prediction task.

[0034] In step S100 of the embodiment of the present application, the working monitoring data of the target forklift accessory is loaded, the fatigue position frequent pattern analysis is performed, and a fatigue frequent point list is obtained, including:

[0035] With the accessory material parameters as constraints, fatigue detection record data is collected, fatigue symptom type frequent pattern analysis is performed, and a frequent fatigue symptom type set is obtained.

[0036] From the working monitoring data, forklift accessory defect monitoring data is extracted, and a forklift accessory position set and a frequent fatigue type trigger frequency set that appear at least any one of the frequent fatigue symptom type set are counted.

[0037] The forklift accessory position with a frequent fatigue type trigger frequency greater than or equal to a trigger frequency threshold in the frequent fatigue type trigger frequency set is added to the fatigue frequent point list.

[0038] In the embodiment of the present application, the core purpose of step S100 is to accurately screen out the key positions with high fatigue problems, i.e., the fatigue frequent points, from the massive working data of the target forklift accessory, to provide focused core research objects for subsequent stress direction analysis and model training, to avoid low efficiency and resource waste caused by indiscriminate analysis, and to ensure the pertinence and accuracy of subsequent fatigue life prediction from the source.

[0039] First, fatigue detection record data is collected with the accessory material parameters as constraints, fatigue symptom type frequent pattern analysis is performed, and a frequent fatigue symptom type set is obtained.

[0040] First, the key material attributes of the target forklift accessory, i.e., material parameters, need to be extracted, which directly determine the fatigue failure mode of the accessory. The material parameters that need to be determined include but are not limited to basic material categories, mechanical property parameters, and process characteristic parameters.

[0041] The basic material categories include high-strength steel, aluminum alloy, engineering plastic, and composite material. The mechanical property parameters include fatigue limit, tensile strength, yield strength, elastic modulus, and fracture toughness. The process characteristic parameters include heat treatment state, surface treatment method, and forming process.

[0042] Then, the above parameters are converted into quantifiable screening rules to form a “constraint condition list” to ensure that only fatigue detection records matching the material characteristics of the target accessory are collected. For example, if the target accessory material is “Q355 high-strength steel, fatigue limit 320 MPa, and welded forming”, the constraint rule is set as: only collect fatigue detection records of accessories with material Q355 steel, fatigue limit within 300-340 MPa (allowing ±5% error), and forming process containing welding.

[0043] Then, the fatigue detection record data meeting the constraint condition is subjected to fatigue position frequent pattern analysis

[0044] The fatigue symptom description in the data is standardized and classified according to “failure mode + failure feature”, so as to avoid statistical deviation caused by description difference. The failure mode (fatigue symptom type) is, for example, fatigue crack, fatigue wear, fatigue fracture, etc.; and the failure feature (fatigue symptom description) is, for example, linear crack at the welding position, excessive wear at the tool connecting pin shaft, sudden fracture of the tool bearing beam, etc.

[0045] A corresponding standardized symptom type label is matched for each fatigue detection record, that is, one record corresponds to one core symptom type, such as “welding position crack” mapped to “fatigue crack”.

[0046] The total number of occurrences and the occurrence frequency of each standardized symptom type label in the data set are counted. Then, the standardized symptom type labels are sorted in descending order of occurrence frequency, and a “frequent threshold” is set. The frequent threshold can be set according to the data volume, for example, the occurrence frequency is greater than or equal to 15%, and the symptom types meeting the definition of “frequent” are screened out.

[0047] Then, the fatigue detection record data collected is subjected to statistical analysis, and the fatigue symptom types meeting the frequent threshold are identified, and the high-frequency types are integrated to form a “frequent fatigue symptom type set”.

[0048] In step S100 of the embodiment of the present application, the forklift tool defect monitoring data is extracted from the operation monitoring data, and the forklift tool position set and the frequent fatigue type trigger frequency set in which at least any one of the frequent fatigue symptom types in the frequent fatigue symptom type set appears are counted, including:

[0049] The first group of defect monitoring data to the Qth group of defect monitoring data is extracted from the forklift tool defect monitoring data, where Q represents the number of groups of forklift tool defect monitoring data.

[0050] The first group of defect monitoring data to the Qth group of defect monitoring data is integrated according to the position, and the first position defect monitoring data set to the Nth position defect monitoring data set is obtained, where N represents the number of positions.

[0051] The sets of defect monitoring data whose number is less than or equal to the data number threshold are deleted, and a plurality of retained position defect monitoring data sets are obtained.

[0052] The plurality of retained position defect monitoring data sets are traversed, and the forklift tool position set and the frequent fatigue type trigger frequency set in which at least any one of the frequent fatigue symptom types in the frequent fatigue symptom type set appears are counted based on the frequent fatigue symptom type set.

[0053] In the embodiments of the present application, the purpose of the above subdivision step in step S100 is to accurately associate "frequent fatigue symptoms" with "specific occurrence positions" from the defect monitoring data of the forklift accessory, and to quantify the frequency of high-frequency fatigue symptoms at different positions, to provide "position + frequency" double data support for subsequent screening of fatigue frequent points, and to remove invalid information through data screening to ensure the reliability and pertinence of the statistical results.

[0054] Firstly, a first group of defect monitoring data to a Qth group of defect monitoring data need to be extracted from the defect monitoring data of the forklift accessory, wherein Q represents the number of groups of defect monitoring data of the forklift accessory. Specifically, a suitable grouping basis needs to be selected according to the collection logic of the defect monitoring data. For example, if the monitoring data is collected according to the daily operation cycle, the data is grouped by "day", and 1 day of monitoring data is 1 group, and a total of Q groups (Q is the total number of days) are extracted; if it is divided according to "sensor collection batch", the single collection data of each monitoring sensor is 1 group, and Q is the total number of sensor collection batches.

[0055] Then, the "defect monitoring data of the forklift accessory" separated from the operation monitoring data is split into independent data sets according to the above grouping dimensions, and is sequentially identified as "first group of defect monitoring data", "second group of defect monitoring data",..., "Qth group of defect monitoring data", to ensure that the boundaries of each group of data are clear and non-repetitive. The defect monitoring data of the forklift accessory includes defect type, defect position, occurrence time and other information.

[0056] Further, the first group of defect monitoring data to the Qth group of defect monitoring data need to be integrated according to the position to obtain a first position defect monitoring data set to an Nth position defect monitoring data set, wherein N represents the number of positions.

[0057] Firstly, the key structure positions of the target forklift accessory need to be standardizedly labeled, covering all possible defect positions. For example, for a carrying accessory such as a push-pull device, possible defect positions include the slide plate connection point, the push-pull arm root, and the guide wheel support; for a clamping accessory such as a paper roll clamp, possible defect positions include the middle part of the clamping arm, the connection between the clamping arm and the drive shaft, and the buffer pad mounting position; for an adjusting accessory such as a distance-adjusting fork, possible defect positions include the fork body telescopic guide rail, the hydraulic cylinder connection flange, and the fork distance adjusting gear box.

[0058] The first group to the Qth group of defect monitoring data are traversed, and for each defect record, the corresponding standardized position label is matched according to the "occurrence position" information. All defect records under the same label, regardless of which group they come from, are aggregated to form a "first position defect monitoring data set" (corresponding to the first position label) to an "Nth position defect monitoring data set" (corresponding to the Nth position label), and N is the total number of standardized position labels.

[0059] Further, the set of defect monitoring data whose data quantity is less than or equal to the data quantity threshold value needs to be deleted, and a plurality of remaining position defect monitoring data sets are obtained.

[0060] The data quantity threshold value can be set according to the total data quantity, the position quantity and industry experience, for example: if the total defect record quantity is 1000, the total number of key positions of the implement is 20 (N=20), the data quantity threshold value can be set as “data quantity≥20”, so as to ensure that the sample quantity of each remaining data set is sufficient to reflect the defect law of the position; if the structure of part of the positions is simple and the probability of defect occurrence is low, the data quantity threshold value can be appropriately reduced, such as≥10.

[0061] The defect record quantity contained in each data set in the “first position to Nth position defect monitoring data set” is counted, the data set whose data quantity is less than or equal to the data quantity threshold value is deleted, and the remaining data set is the “remaining position defect monitoring data set”.

[0062] In step S100 of the embodiment of the present application, the plurality of remaining position defect monitoring data sets are traversed, the forklift implement position set and the frequent fatigue type trigger frequency set in which at least any one frequent fatigue symptom type in the frequent fatigue symptom type set appears are counted based on the frequent fatigue symptom type set, including:

[0063] The first frequent fatigue symptom type to the Qth frequent fatigue symptom type is extracted from the frequent fatigue symptom type set, wherein Q represents the number of frequent fatigue symptom types;

[0064] The first remaining position defect type set is extracted from the plurality of remaining position defect monitoring data sets;

[0065] The trigger frequency of the first frequent fatigue symptom type in the first remaining position defect type set is counted and set as the first trigger frequency;

[0066] The trigger frequency of the Qth frequent fatigue symptom type in the first remaining position defect type set is counted and set as the Qth trigger frequency;

[0067] The first trigger frequency to the Qth trigger frequency is summed and set as the first remaining position frequent fatigue type trigger frequency;

[0068] When the first remaining position frequent fatigue type trigger frequency is not equal to 0, the first remaining position frequent fatigue type trigger frequency is added to the frequent fatigue type trigger frequency set, and the first remaining position is added to the forklift implement position set.

[0069] In the embodiments of the present application, the purpose of the above-mentioned subdivision step in step S100 is to perform refined “frequent fatigue symptom-position-frequency” correlation statistics on the screened effective position defect data, i.e., the retained position defect monitoring data set, accurately identify the tool positions that have appeared at least one high-frequency fatigue symptom, and quantify the total frequency of the positions appearing high-frequency fatigue symptoms, thereby providing accurate and quantifiable core basis for finally determining the “fatigue frequent point” and avoiding omissions or deviations caused by single symptom statistics.

[0070] First, each frequent fatigue symptom type needs to be extracted from the set of frequent fatigue symptom types in order, and sequentially named as the first frequent fatigue symptom type to the Qth frequent fatigue symptom type, where Q represents the number of frequent fatigue symptom types.

[0071] Then, a specific “retained position” is selected from the retained position defect monitoring data set, i.e., the first retained position, which is an effective position screened by data volume, such as the middle of the clamp arm. The “defect types” corresponding to all defect records of the position, such as cracks, deformations, wear, looseness, etc., are extracted, and these defect types are integrated to form a “first retained position defect type set”, which provides an analysis object for subsequent statistics of the occurrence of high-frequency symptoms at the position.

[0072] Further, the triggering frequency of the first frequent fatigue symptom type in the first retained position defect type set needs to be calculated and set as the first triggering frequency. Optionally, the calculation method can be first triggering frequency = (the number of times of the occurrence of a certain frequent fatigue symptom type in the defect type set of the position) ÷ (the total number of defect records of the defect type set of the position) x 100%.

[0073] Taking “the first frequent fatigue symptom type” fatigue crack as an example, the number of times of its occurrence in the “first retained position defect type set”, such as the defect type set of the middle of the clamp arm, is counted. Assuming that the total number of defect records is 50 and fatigue cracks occur 25 times, then the first triggering frequency = 25 / 50 x 100% = 50%.

[0074] In the same way, the triggering frequencies of the second frequent fatigue symptom type to the Qth frequent fatigue symptom type at the position are counted in turn, and the second triggering frequency to the Qth triggering frequency are obtained.

[0075] Then, the first triggering frequency to the Qth triggering frequency obtained in the above step are summed up. For example, the three triggering frequencies of the above-mentioned middle of the clamp arm are 50%, 20%, and 10% respectively, and then the first retained position frequent fatigue type triggering frequency = 50% + 20% + 10% = 80%.

[0076] Finally, when the first retention position frequent fatigue type trigger frequency is not equal to 0, the first retention position frequent fatigue type trigger frequency is added to the frequent fatigue type trigger frequency set, and the first retention position is added to the forklift accessory position set.

[0077] Further, the forklift accessory positions with frequent fatigue type trigger frequencies greater than or equal to the trigger frequency threshold in the frequent fatigue type trigger frequency set need to be added to the fatigue frequent point list.

[0078] The trigger frequency threshold is combined with industry standards, accessory use scenarios, and historical failure data to determine a specific trigger frequency threshold, such as 30%.

[0079] The forklift accessory position set and the frequent fatigue type trigger frequency set obtained in the early stage are sorted to establish a one-to-one correspondence, such as using a table or dictionary to store "middle of clamping arm, 80%" and "fork body telescopic guide rail, 25%".

[0080] The correspondence is traversed, and the trigger frequency of each position is compared with the trigger frequency threshold one by one. The positions with trigger frequencies greater than or equal to the trigger frequency threshold are included in the fatigue frequent point list, such as the middle of the clamping arm with a trigger frequency of 80%; positions below the trigger frequency threshold are excluded, such as the fork body telescopic guide rail with a trigger frequency of 25%, and the fatigue frequent point list is finally formed.

[0081] In step S200 of the embodiments of the present application, based on the work monitoring data, the fatigue frequent point list is traversed to perform stress direction frequent pattern analysis and obtain a frequent stress direction list, including:

[0082] A first fatigue frequent position is extracted from the fatigue frequent point list.

[0083] Based on the first fatigue frequent position, a first group of fatigue frequent position stress directions is extracted from the work monitoring data.

[0084] Based on an angle deviation threshold, clustering is performed on the first group of fatigue frequent position stress directions to obtain multiple clusters of first fatigue frequent position stress directions.

[0085] Multiple first fatigue frequent position centroid stress directions of the multiple clusters of first fatigue frequent position stress directions are taken as a first group of reconstructed fatigue frequent position stress directions, and the number of clusters of the multiple clusters of first fatigue frequent position stress directions is taken as multiple initial trigger frequencies of the multiple first fatigue frequent position centroid stress directions, which are added to the initial trigger frequencies of the first group of reconstructed fatigue frequent position stress directions.

[0086] Until the Lth group of reconstructed fatigue frequent position stress directions and the Lth group of reconstructed fatigue frequent position stress direction initial trigger frequencies are obtained.

[0087] Based on the first set of reconstructed fatigue frequent position force direction and the first set of reconstructed fatigue frequent position force direction initial trigger frequency, the first fatigue frequent position is subjected to force direction frequent pattern analysis until the Lth set of reconstructed fatigue frequent position force direction and the Lth set of reconstructed fatigue frequent position force direction initial trigger frequency.

[0088] Until the force direction frequent pattern analysis on all fatigue frequent points is completed, a frequent force direction list is output.

[0089] In the embodiments of the present application, the purpose of step S200 is to accurately extract and analyze the force direction characteristics of each position in the operation for the "fatigue frequent points" determined in S100, to screen out the high-frequency force direction of each fatigue frequent point, to provide key "fatigue frequent position-force direction" correlation data for subsequent construction of a fatigue life prediction model, to reveal the internal relationship between force direction and fatigue failure, and to improve the prediction accuracy of the fatigue life prediction model.

[0090] First, the first position needs to be extracted from the fatigue frequent point list in order as the first fatigue frequent position, for example, the "middle of the clamping arm" is selected as the first fatigue frequent position, that is, the first analysis object.

[0091] Then, based on the first fatigue frequent position, the first set of fatigue frequent position force direction is extracted from the operation monitoring data. The operation monitoring data contains the force monitoring information of each position of the implement, which is usually collected by force sensors and angle sensors installed at key positions of the implement, and records the force size and direction angle at different times. Here, the screening condition is set as "monitoring position = first fatigue frequent position", such as "monitoring position = middle of clamping arm", and all force direction data meeting the condition are extracted from the operation monitoring data. The force direction data includes force direction angle values such as 0°, 45°, 90° at different operation scenarios and different time points, and these data are integrated into a "first set of fatigue frequent position force direction" data set.

[0092] Further, a reasonable angle deviation threshold such as ±5°, ±10° is set according to the operation scenario of the implement and the mechanical analysis requirement, that is, when the angle difference between two force directions is ≤ the angle deviation threshold, it is considered as "similar direction" and can be classified into the same cluster. For example, if the angle deviation threshold is set as ±5°, the force directions of 30°, 33° and 35° can be classified into one cluster.

[0093] The clustering algorithm suitable for angle data, such as K-means clustering, density clustering DBSCAN, etc., is used to cluster the "first group of fatigue frequent position stress direction" data. Taking K-means as an example, the stress directions with similar angles are automatically divided into multiple clusters by the algorithm, and finally the "multiple cluster first fatigue frequent position stress direction" is obtained, such as cluster 1: 25°-35°, cluster 2: 85°-95°, cluster 3: 175°-185°.

[0094] Further, the multiple first fatigue frequent position centroid stress directions of the multiple cluster first fatigue frequent position stress direction are set as the first group of reconstructed fatigue frequent position stress directions, and the number of clusters in the multiple cluster first fatigue frequent position stress direction is set as the multiple initial trigger frequencies of the multiple first fatigue frequent position centroid stress directions, which are added to the initial trigger frequencies of the first group of reconstructed fatigue frequent position stress directions.

[0095] First, the centroid stress direction needs to be calculated. For each stress direction cluster, the average value of the angle of all stress directions in the cluster, i.e. the "centroid direction", is calculated as the representative stress direction of the cluster. For example, cluster 1 (25°-35°) contains angle data [25°, 28°, 30°, 32°, 35°], and its centroid direction = (25+28+30+32+35) / 5 = 30°, which is one of the "first fatigue frequent position centroid stress directions".

[0096] Then, the number of clusters needs to be counted, i.e. the number of stress direction data included in each cluster, which is the number of clusters in the cluster. This number directly reflects the frequency of occurrence of the corresponding centroid direction as the "initial trigger frequency". For example, cluster 1 contains 5 data, and its corresponding initial trigger frequency = 5.

[0097] Finally, the "centroid stress direction" of all clusters is integrated into the "first group of reconstructed fatigue frequent position stress directions", such as [30°, 90°, 180°], and the corresponding "initial trigger frequency" is integrated into the "first group of reconstructed fatigue frequent position stress direction initial trigger frequency", such as [5, 8, 3], to ensure that the elements of the two data sets correspond one by one, such as 30° corresponding to 5, 90° corresponding to 8, and 180° corresponding to 3.

[0098] The above operations are repeated for the "second group of fatigue frequent position stress direction" and the "third group of fatigue frequent position stress direction" in turn, and the "second group of reconstructed fatigue frequent position stress direction and initial trigger frequency" and the "third group of reconstructed fatigue frequent position stress direction and initial trigger frequency" are obtained respectively, until the Lth group of reconstructed fatigue frequent position stress direction and the Lth group of reconstructed fatigue frequent position stress direction initial trigger frequency are obtained.

[0099] Further, the force direction frequent pattern analysis of the first fatigue frequent position is performed based on the first set of reconstructed force direction of fatigue frequent positions and the first set of initial trigger frequencies of the force direction of fatigue frequent positions until the Lth set of reconstructed force direction of fatigue frequent positions and the Lth set of initial trigger frequencies of the force direction of fatigue frequent positions. The purpose of this step is to integrate the reconstruction data of all batches of the position, to count the total trigger frequencies of the force direction of each centroid, to screen the high-frequency force direction, and to determine the core force direction feature of the fatigue frequent point.

[0100] The force direction frequent pattern analysis herein has the same analysis logic as the fatigue position frequent pattern analysis and the fatigue symptom type frequent pattern analysis, that is, the total trigger frequencies are summarized, the frequency threshold is set, and the high-frequency force direction of the position is screened according to the frequency threshold. The specific analysis process will not be described here.

[0101] Finally, the screened high-frequency force direction, such as 30° and 90°, is taken as the frequent force direction of the first fatigue frequent position, and the force direction frequent pattern analysis of all positions in the “fatigue frequent point list” is sequentially completed, the high-frequency force direction of all positions is integrated, and the final “frequent force direction list” is formed. The frequent force direction list contains one-to-one corresponding data of fatigue frequent positions and high-frequency force directions. For example, the middle part of the clamp arm is 30° and 90°, the telescopic guide rail of the fork body is 45° and 135°, and the flange connected to the hydraulic cylinder is 0° and 180°.

[0102] In step S300 of the embodiment of the present application, a fatigue life prediction base model is trained by collecting a plurality of one-to-one corresponding first tool operation working condition time sequence data and first labeled fatigue life labels, taking the tool structure topology, tool material parameters, the fatigue frequent point list and the frequent force direction list as constraints, including:

[0103] A first-level constraint is constructed based on the tool structure topology and the tool material parameters.

[0104] A second-level constraint is constructed based on the fatigue frequent point list and the frequent force direction list.

[0105] A tool sample set to be analyzed is loaded, wherein the tool sample set to be analyzed has the same to-be-analyzed structure topology, to-be-analyzed material parameters, to-be-analyzed fatigue frequent point list, to-be-analyzed frequent force direction list, tool operation fatigue frequent point force time sequence data and tool operation environment monitoring time sequence information.

[0106] When the to-be-analyzed structure topology has a structure topology similarity greater than or equal to a structure topology similarity threshold with the tool structure topology, and the to-be-analyzed material parameters are the same as the to-be-analyzed material parameters, it is considered that the first-level constraint is met.

[0107] When the primary constraint is met, and the point position distribution similarity between the fatigue frequent point list and the to-be-analyzed fatigue frequent point list is greater than or equal to the point position distribution similarity threshold, and the stress direction distribution similarity between the frequent stress direction list and the to-be-analyzed frequent stress direction list is greater than or equal to the stress direction distribution similarity threshold, it is considered that the secondary constraint is met.

[0108] When the secondary constraint is met, the tool fatigue frequent point stress timing data and the tool operation environment monitoring timing information are set as the first tool operation working condition timing data, and the mode of fatigue symptom occurrence is analyzed for the to-be-analyzed tool sample set to obtain the label of the first identified fatigue life.

[0109] In the embodiments of the present application, the purpose of step S300 is to screen out sample data highly matched with the core characteristics of the target forklift tool through double constraints, train a fatigue life prediction basic model with strong generalization ability and high prediction accuracy based on these high-quality data, lay a reliable model foundation for subsequent transfer learning to adapt to the target tool, and solve the prediction deviation problem caused by poor sample adaptability in traditional modeling.

[0110] Firstly, based on the tool structure topology and the tool material parameters, a primary constraint is constructed

[0111] The tool structure topology is the key structural features of the extracted target tool, including structural types such as the “sliding plate-pull arm” connection structure of the carrying type puller and the “double clamp arm-drive shaft” symmetrical structure of the clamping type paper roll clamp; key component sizes such as the length of the clamp arm and the diameter of the connecting shaft; load transmission paths such as the transmission route of stress from the contact end to the rack, etc., forming a “target structure topology parameter set”.

[0112] The tool material parameters are the material parameters determined in S100, including basic material categories, mechanical property parameters, process characteristic parameters, etc., forming a “target material parameter set”.

[0113] Then, the “target structure topology parameter set” and the “target material parameter set” are used as hard conditions for screening samples, and the rule is expressed as: “the structure topology of the to-be-analyzed tool sample needs to be highly similar to the target tool structure topology, and the material parameters need to be completely consistent with the target tool material parameters”.

[0114] Further, based on the fatigue frequent point list and the frequent stress direction list, a secondary constraint is constructed. That is, the above two lists are used as supplementary screening conditions, and the rule can be expressed as: “the fatigue frequent point position distribution of the to-be-analyzed sample needs to be highly similar to the target tool, and the frequent stress direction distribution of each fatigue frequent point needs to be highly similar to the target tool”.

[0115] Further, a set of samples of tools to be analyzed needs to be loaded, which can be derived from enterprise historical operation data, industry public database, laboratory simulation test data, etc. Each sample of tool to be analyzed should have the following six types of core data to ensure matching with the constraint conditions and training requirements, i.e. analyzed structure topology, analyzed material parameters, analyzed fatigue frequent point list, analyzed frequent stress direction list, tool operation fatigue frequent point stress time series data, and tool operation environment monitoring time series information.

[0116] Among them, the analyzed structure topology is the structural characteristic parameter of the sample tool, which corresponds to the target structure topology parameter set. The analyzed material parameter is the material characteristic parameter of the sample tool, which corresponds to the target material parameter set. The tool operation fatigue frequent point stress time series data is the stress data of each fatigue frequent point changing with time during the operation of the sample tool, such as stress size and direction angle per second. The tool operation environment monitoring time series information is the data of environmental parameters changing with time during the operation of the sample tool, such as temperature, humidity, and operation load fluctuation range.

[0117] Then, the analyzed structure topology of the sample tool to be analyzed is compared with the “target structure topology parameter set” of the target tool in terms of feature, including structure type matching degree, key component size deviation rate, and load transfer path coincidence degree, etc. The structure topology similarity is obtained by weighted calculation. A structure topology similarity threshold is set, such as 85%. When the structure topology similarity of the sample tool to be analyzed is greater than or equal to the structure topology similarity threshold, it is determined that the structure topology is matched.

[0118] Next, the “analyzed material parameters” of the sample tool to be analyzed are compared with the “target material parameter set” of the target tool item by item, such as judging whether the material is the same, whether the fatigue limit deviation is less than or equal to 3%, whether the heat treatment method is consistent, etc. When all parameters are completely matched, or the deviation of key parameters is within the allowable range, it is determined that the material parameters are consistent.

[0119] Only when the sample tool to be analyzed meets both “structure topology similarity greater than or equal to structure topology similarity threshold” and “material parameters consistent”, it is considered to meet the first level constraint; otherwise, the sample tool to be analyzed is excluded.

[0120] Further, when the sample tool to be analyzed meets the first level constraint, and the point distribution similarity of the fatigue frequent point list and the analyzed fatigue frequent point list is greater than or equal to the point distribution similarity threshold, and the stress direction distribution similarity of the frequent stress direction list and the analyzed frequent stress direction list is greater than or equal to the stress direction distribution similarity threshold, it is considered to meet the second level constraint.

[0121] Among them, the point distribution similarity judgment is specifically to calculate the point coincidence rate.

[0122] Specifically, the "to-be-analyzed fatigue frequent point list" of the tool sample to be analyzed is compared with the "fatigue frequent point list" of the target tool, and the proportion of the number of coinciding positions in the total number of positions in the target list is calculated. For example, if the fatigue frequent point list has 3 positions and the to-be-analyzed fatigue frequent point list coincides with 2 positions, the coincidence rate is approximately 66.7%, and this proportion is the "point distribution similarity". Set the point distribution similarity threshold, such as ≥80%, and when the point distribution similarity of the tool sample to be analyzed is ≥ the point distribution similarity threshold, it is determined that the fatigue frequent point distribution matches.

[0123] The stress direction distribution similarity determination is specifically to calculate the direction matching degree one by one according to the position.

[0124] Specifically, for each coinciding fatigue frequent point, the "to-be-analyzed frequent stress direction" of the tool sample to be analyzed at the position is compared with the "frequent stress direction" of the target tool at the position, and the proportion of the number of direction angle deviations ≤ the angle threshold in the total number of target directions is calculated. For example, the angle threshold can be ±5°. For example, if there are 2 stress directions at the target position and the tool sample to be analyzed matches 1, the direction matching degree at this position is 50%. The average of the direction matching degrees of all coinciding positions is taken as the "stress direction distribution similarity" of the tool sample to be analyzed.

[0125] Set the similarity threshold, such as 75%: when the stress direction distribution similarity of the tool sample to be analyzed is ≥ the stress direction distribution similarity threshold, it is determined that the stress direction distribution matches.

[0126] Only when the sample meets the first-level constraint and simultaneously meets the "point distribution similarity ≥ the point distribution similarity threshold" and the "stress direction distribution similarity ≥ the stress direction distribution similarity threshold", it is considered to meet the second-level constraint; otherwise, the sample is rejected.

[0127] Further, when the second-level constraint is met, the tool operation fatigue frequent point stress timing data and the tool operation environment monitoring timing information are set as the first tool operation working condition timing data, the to-be-analyzed tool sample set is subjected to fatigue symptom occurrence duration mode analysis, and the first identification fatigue life label is obtained.

[0128] For the sample that meets the second-level constraint, the "tool operation fatigue frequent point stress timing data" and the "tool operation environment monitoring timing information" are integrated to form the input feature data of the fatigue life prediction basic model, that is, the first tool operation working condition timing data - this data contains the core dynamic factors affecting tool fatigue life.

[0129] Then, the "fatigue symptom occurrence duration mode analysis" is performed on the "to-be-analyzed tool sample set", that is, the duration from the use of the sample tool to the first occurrence of any one of the "frequent fatigue symptom type set" is counted, that is, the fatigue life duration, the mode of the duration data under multiple same working conditions is taken, that is, the duration value with the highest occurrence frequency, and the mode is the "first fatigue life identification label". Here, the mode is used to reduce the influence of single abnormal data on the accuracy of the label.

[0130] Further, the fatigue life prediction base model needs to be trained.

[0131] Considering the use of the fatigue life prediction base model, a multi-topology sub-model integration and output mean fusion method can be used to build the fatigue life prediction base model. The fatigue life prediction base model includes three sub-models, namely, an LSTM sub-model, a CNN-LSTM sub-model, and a GRU sub-model.

[0132] The LSTM sub-model is adapted to tool fatigue frequent point stress time series data, captures long-term dependence of time series data, such as the influence of stress in different operation stages on fatigue accumulation, and has a structure of an input layer, an LSTM layer, a full connection layer, and an output layer.

[0133] The CNN-LSTM sub-model is adapted to the integrated fatigue frequent point stress and time series environment data, extracts local features through CNN, and captures long-term rules through LSTM, and has a structure of an input layer, a CNN layer, an LSTM layer, a full connection layer, and an output layer.

[0134] The GRU sub-model is adapted to the integrated operation working condition time series data, simplifies the network structure to improve the training efficiency, and is complementary to the LSTM, and has a structure of an input layer, a GRU layer, a full connection layer, and an output layer.

[0135] After the three sub-models output the prediction values, the arithmetic mean value is taken as the final output of the base model.

[0136] In the parameter setting of the fatigue life prediction base model, the optimizer uses Adam, the learning rate is set to 0.001, the loss function uses mean square error (MSE), the batch size is 32, L2 regularization is used with a coefficient of 0.0001, and the training is stopped when the loss of the validation set does not decrease for five consecutive rounds.

[0137] The training samples of the fatigue life prediction base model are derived from the first tool operation working condition time series data and the first fatigue life identification label. The sample size is not less than 1200 groups, of which the training set is 80% and the validation set is 20%. The maximum number of training rounds is set to 100 rounds.

[0138] The convergence criterion of the fatigue life prediction base model is that the MSE of the training set is less than or equal to 0.001 and the MSE of the verification set is less than or equal to 0.0015.

[0139] In step S400 of the embodiment of the present application, the target forklift accessory model, the fatigue frequent point list and the frequent stress direction list are used to collect a plurality of one-to-one second accessory working condition time sequence data and second fatigue life identification labels, and the fatigue life prediction base model is subjected to transfer learning to obtain a target forklift accessory fatigue life prediction model. The fatigue life prediction base model is an output mean integrated model of a plurality of topologically different sub-models, and includes:

[0140] A plurality of fatigue life prediction base sub-models of the fatigue life prediction base model are obtained.

[0141] A fully connected neural network is constructed, and the plurality of fatigue life prediction base sub-models are connected in parallel to the input layer of the fully connected neural network.

[0142] The second accessory working condition time sequence data is used as the input of the plurality of fatigue life prediction base sub-models, and the second fatigue life identification label is used as the output of the fully connected neural network, and the fatigue life prediction base model is subjected to transfer learning to obtain the target forklift accessory fatigue life prediction model.

[0143] In the embodiment of the present application, the purpose of step S400 is to adapt the fatigue life prediction base model trained in step S300 to a specific target forklift accessory through transfer learning, to obtain a target forklift accessory fatigue life prediction model with stronger pertinence and higher prediction accuracy, to solve the problem of prediction deviation of the base model for specific accessory models, and to retain the general fatigue law learned by the base model.

[0144] Firstly, a plurality of fatigue life prediction base sub-models of the fatigue life prediction base model need to be obtained, that is, all sub-models constituting the integrated model, such as LSTM sub-models, CNN-LSTM sub-models and GRU sub-models, are extracted from the fatigue life prediction base model trained in step S300, and the pre-training parameters of each sub-model are retained.

[0145] Then, a fully connected neural network needs to be constructed, and the plurality of fatigue life prediction base sub-models are connected in parallel to the input layer of the fully connected neural network.

[0146] Specifically, a fully connected network containing 2-3 hidden layers is designed, for example, the number of input layer neurons = the number of basic sub-models, such as 3 basic sub-models, the input layer has 3 neurons, the number of neurons in the first hidden layer = 16, the number of neurons in the second hidden layer = 8, the number of neurons in the output layer = 1, the output target attribute fatigue life prediction value, and the activation function is ReLU.

[0147] Then, the outputs of the extracted plurality of fatigue life prediction basic sub-models are respectively connected to the input layer of the newly constructed fully connected neural network, forming a "basic sub-model parallel connection → fully connected network integration" transfer learning architecture, retaining the feature extraction capability of the basic sub-models, and adapting to the target attribute characteristics through the fully connected network.

[0148] Further, the second attribute operation working condition time series data is used as the input of the plurality of fatigue life prediction basic sub-models, and the second fatigue life identification label is used as the output of the fully connected neural network, and the fatigue life prediction basic model is subjected to transfer learning to obtain the target forklift attribute fatigue life prediction model.

[0149] Among them, the "second attribute operation working condition time series data" as input is collected under the constraints of target forklift attribute model, fatigue frequent point list and frequent stress direction list, and contains stress time series data and operation environment time series information of the fatigue frequent point of the model attribute. The "second fatigue life identification label" as output is the life label obtained by analyzing the mode of the target model attribute sample when the fatigue symptom appears.

[0150] In the implementation process, the "second attribute operation working condition time series data" is simultaneously input into a plurality of fatigue life prediction basic sub-models, and each sub-model outputs a preliminary prediction value, which is used as the input of the fully connected neural network.

[0151] Then, taking the "second fatigue life identification label" as the target, the parameters of the fully connected neural network are optimized through back propagation to minimize the mean square error of the prediction value and the label. After training, the overall model composed of "basic sub-model parallel connection + fully connected network" is the "target forklift attribute fatigue life prediction model", which can be directly used for fatigue life prediction of the model attribute.

[0152] In summary, by implementing the forklift attribute fatigue life prediction method provided in the embodiment, the following technical effects can be achieved:

[0153] Through transfer learning, the architecture of the basic sub-model parallel connection + full connection network integration is constructed, the general knowledge learned by the basic sub-model is reused, the dependence on a large number of monitoring samples of the target attribute is greatly reduced, the problem that the monitoring samples are few due to uncertain scheduling and large load fluctuation of the forklift attribute is solved, and a precise and efficient forklift attribute fatigue life prediction model is trained.

[0154] Through performing 'fatigue position frequent pattern analysis' on the target forklift attribute operation monitoring data, the key positions with high fatigue are accurately screened to form a fatigue frequent point list, thereby avoiding indiscriminate analysis of all positions of the attribute, reducing the time and labor cost of data processing, and improving the efficiency of the overall prediction process.

[0155] 3. Based on the fatigue frequent point list,'stress direction frequent pattern analysis' is further performed, the main stress directions of each fatigue-prone position are extracted from the operation monitoring data to form a frequent stress direction list, thereby avoiding the prediction deviation caused by ignoring the key stress characteristics in the traditional method, and improving the accuracy of the model.

[0156] Embodiment two, as shown in Figure 2 the same inventive concept as the forklift attribute fatigue life prediction method provided in embodiment one, the present application embodiment also provides a forklift attribute fatigue life prediction system using transfer learning, comprising:

[0157] The frequent pattern analysis module 11 is used to load the operation monitoring data of the target forklift attribute, perform fatigue position frequent pattern analysis, and obtain a fatigue frequent point list.

[0158] The stress direction analysis module 12 is used to traverse the fatigue frequent point list based on the operation monitoring data, perform stress direction frequent pattern analysis, and obtain a frequent stress direction list.

[0159] The basic model training module 13 is used to collect a plurality of first attribute operation working condition time series data and first identification fatigue life labels corresponding to the fatigue frequent point list and the frequent stress direction list as constraints, and train a fatigue life prediction basic model.

[0160] The target model training module 14 is used to collect a plurality of second attribute operation working condition time series data and second identification fatigue life labels corresponding to the fatigue frequent point list and the frequent stress direction list, and perform transfer learning on the fatigue life prediction basic model to obtain a target forklift attribute fatigue life prediction model, and perform a forklift attribute fatigue life prediction task.

[0161] Further, the frequent pattern analysis module 11 includes the following execution steps:

[0162] collecting fatigue detection record data with the tool material parameters as constraints, performing fatigue symptom type frequent pattern analysis, and obtaining a frequent fatigue symptom type set;

[0163] extracting forklift tool defect monitoring data from the work monitoring data, and counting forklift tool position sets and frequent fatigue type trigger frequency sets in which at least any one of the frequent fatigue symptom types in the frequent fatigue symptom type set appears;

[0164] adding forklift tool positions in which the frequent fatigue type trigger frequency is greater than or equal to a trigger frequency threshold in the frequent fatigue type trigger frequency set to the fatigue frequent point list.

[0165] In the method, the forklift tool defect monitoring data is extracted from the work monitoring data, and the forklift tool position sets and the frequent fatigue type trigger frequency sets in which at least any one of the frequent fatigue symptom types in the frequent fatigue symptom type set appears are counted.

[0166] extracting first defect monitoring data to Qth defect monitoring data from the forklift tool defect monitoring data, where Q represents the number of groups of forklift tool defect monitoring data;

[0167] integrating the first defect monitoring data to the Qth defect monitoring data according to positions to obtain first position defect monitoring data set to Nth position defect monitoring data set, where N represents the number of positions;

[0168] deleting sets in which the number of defect monitoring data is less than or equal to a data number threshold to obtain a plurality of remaining position defect monitoring data sets;

[0169] traversing the plurality of remaining position defect monitoring data sets, and counting the forklift tool position sets and the frequent fatigue type trigger frequency sets in which at least any one of the frequent fatigue symptom types in the frequent fatigue symptom type set appears based on the frequent fatigue symptom type set.

[0170] In the method, the forklift tool defect monitoring data is extracted from the work monitoring data, and the forklift tool position sets and the frequent fatigue type trigger frequency sets in which at least any one of the frequent fatigue symptom types in the frequent fatigue symptom type set appears are counted.

[0171] extracting first frequent fatigue symptom type to Qth frequent fatigue symptom type from the frequent fatigue symptom type set, where Q represents the number of frequent fatigue symptom types;

[0172] extracting first remaining position defect type set from the plurality of remaining position defect monitoring data sets;

[0173] counting the triggering frequency of the first frequent fatigue symptom type in the first set of remaining position defect types as a first triggering frequency;

[0174] counting the triggering frequency of the Qth frequent fatigue symptom type in the first set of remaining position defect types as a Qth triggering frequency;

[0175] adding the first triggering frequency to the Qth triggering frequency as a first remaining position frequent fatigue type triggering frequency;

[0176] when the first remaining position frequent fatigue type triggering frequency is not equal to 0, adding the first remaining position frequent fatigue type triggering frequency into the set of frequent fatigue type triggering frequencies, and adding the first remaining position into the set of forklift accessory positions.

[0177] Further, the stress direction analysis module 12 includes the following execution steps:

[0178] extracting a first fatigue frequent position from the fatigue frequent point list;

[0179] extracting a first set of fatigue frequent position stress directions from the work monitoring data based on the first fatigue frequent position;

[0180] performing clustering on the first set of fatigue frequent position stress directions based on an angle deviation threshold to obtain a plurality of clusters of first fatigue frequent position stress directions;

[0181] taking a plurality of first fatigue frequent position centroid stress directions of the plurality of clusters of first fatigue frequent position stress directions as a first set of reconstructed fatigue frequent position stress directions, and adding a number of cluster within the plurality of clusters of first fatigue frequent position stress directions into a plurality of initial triggering frequencies of the plurality of first fatigue frequent position centroid stress directions as a first set of reconstructed fatigue frequent position stress direction initial triggering frequencies;

[0182] until the Lth set of reconstructed fatigue frequent position stress directions and the Lth set of reconstructed fatigue frequent position stress direction initial triggering frequencies are obtained;

[0183] performing stress direction frequent pattern analysis on the first fatigue frequent position based on the first set of reconstructed fatigue frequent position stress directions and the first set of reconstructed fatigue frequent position stress direction initial triggering frequencies until the Lth set of reconstructed fatigue frequent position stress directions and the Lth set of reconstructed fatigue frequent position stress direction initial triggering frequencies;

[0184] until the stress direction frequent pattern analysis on all fatigue frequent points is completed, outputting a frequent stress direction list.

[0185] Further, the basic model training module 13 includes the following execution steps:

[0186] constructing a primary constraint based on the tool structure topology and the tool material parameters;

[0187] constructing a secondary constraint based on the fatigue frequent point list and the frequent force direction list;

[0188] loading a tool sample set to be analyzed, wherein the tool sample set to be analyzed has the same structure topology to be analyzed, material parameters to be analyzed, fatigue frequent point list to be analyzed, frequent force direction list to be analyzed, tool operation fatigue frequent point force time sequence data, and tool operation environment monitoring time sequence information;

[0189] when the structure topology similarity of the structure topology of the tool structure topology and the structure topology to be analyzed is greater than or equal to a structure topology similarity threshold, and the tool material parameters are the same as the material parameters to be analyzed, it is considered that the primary constraint is met;

[0190] when the primary constraint is met, and the point distribution similarity of the fatigue frequent point list and the fatigue frequent point list to be analyzed is greater than or equal to a point distribution similarity threshold, and the force direction distribution similarity of the frequent force direction list and the frequent force direction list to be analyzed is greater than or equal to a force direction distribution similarity threshold, it is considered that the secondary constraint is met;

[0191] when the secondary constraint is met, the tool operation fatigue frequent point force time sequence data and the tool operation environment monitoring time sequence information are set as the first tool operation working condition time sequence data, and the tool sample set to be analyzed is subjected to fatigue symptom duration mode analysis to obtain the label of the first identified fatigue life.

[0192] Further, the target model training module 14 includes the following execution steps:

[0193] obtaining a plurality of fatigue life prediction base sub-models of the fatigue life prediction base model;

[0194] constructing a fully connected neural network, and connecting the plurality of fatigue life prediction base sub-models in parallel to the input layer of the fully connected neural network;

[0195] taking the second tool operation working condition time sequence data as the input of the plurality of fatigue life prediction base sub-models, taking the label of the second identified fatigue life as the output of the fully connected neural network, performing transfer learning on the fatigue life prediction base model, and obtaining the target forklift tool fatigue life prediction model.

[0196] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0197] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0198] The present application is described in reference to the flowchart illustrations and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0199] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0200] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0201] Although preferred embodiments of the application have been described, those skilled in the art will appreciate that additional modifications and alterations can be made thereto without departing from the basic construction and / or teachings of the application.

[0202] Obviously, many modifications and variations of the present application are possible in light of the above teachings. It is, therefore, to be understood that within the scope of the appended claims, the application can be practiced otherwise than as specifically described.

Claims

1. A method for predicting the fatigue life of forklift attachments using transfer learning, characterized in that, include: Load the operation monitoring data of the target forklift attachments, perform fatigue position frequent pattern analysis, and obtain a list of fatigue frequent points; Based on the work monitoring data, the list of frequent fatigue points is traversed, and the frequent force direction pattern analysis is performed to obtain the list of frequent force directions. Using the attachment structure topology, attachment material parameters, the list of frequent fatigue points, and the list of frequent force directions as constraints, collect several one-to-one corresponding time-series data of the first attachment's working conditions and labels that identify fatigue life, and train a basic model for fatigue life prediction. Using the target forklift attachment model, the list of frequent fatigue points, and the list of frequent force directions, collect several one-to-one corresponding second attachment working condition time series data and second label identifying fatigue life. Perform transfer learning on the fatigue life prediction basic model to obtain the target forklift attachment fatigue life prediction model and execute the forklift attachment fatigue life prediction task. Among them, constrained by the attachment structure topology, attachment material parameters, the list of frequent fatigue points, and the list of frequent force directions, several one-to-one corresponding time-series data of the first attachment's operating conditions and labels identifying the first fatigue life are collected to train a basic model for fatigue life prediction, including: Based on the attachment structure topology and the attachment material parameters, a first-level constraint is constructed; Based on the list of frequent fatigue points and the list of frequent force directions, construct a secondary constraint. Load the sample set of attachments to be analyzed, wherein the sample set of attachments to be analyzed has the same topology of the structure to be analyzed, parameters of the material to be analyzed, list of frequent fatigue points to be analyzed, list of frequent force directions to be analyzed, time series data of force on frequent fatigue points of attachments, and time series information of attachment operation environment monitoring. When the topological similarity between the structure topology to be analyzed and the topological similarity between the attachment structure topology is greater than or equal to the topological similarity threshold, and the material parameters to be analyzed are the same as the material parameters to be analyzed, it is considered that the first-level constraint is satisfied. When the first-level constraint is satisfied, and the similarity of the point distribution between the list of frequent fatigue points and the list of frequent fatigue points to be analyzed is greater than or equal to the point distribution similarity threshold, and the similarity of the force direction distribution between the list of frequent force directions and the list of frequent force directions to be analyzed is greater than or equal to the force direction distribution similarity threshold, it is considered that the second-level constraint is satisfied. When the secondary constraint is satisfied, the stress time series data of the frequent fatigue points of the attachment and the monitoring time series information of the attachment working environment are set as the first attachment working condition time series data. The fatigue symptom occurrence duration mode analysis is performed on the attachment sample set to be analyzed to obtain the first label identifying fatigue life. The process involves collecting time-series data of the second attachment's operating conditions and labels indicating fatigue life, using the target forklift attachment model, the list of frequent fatigue points, and the list of frequent force directions. Transfer learning is then applied to the fatigue life prediction baseline model to obtain the target forklift attachment fatigue life prediction model. This baseline model is an integrated model of the output averages of multiple sub-models with different topologies, including: Multiple fatigue life prediction sub-models are obtained from the fatigue life prediction basic model. The fatigue life prediction basic model includes three sub-models: LSTM sub-model, CNN-LSTM sub-model, and GRU sub-model. Construct a fully connected neural network and connect multiple basic sub-models for fatigue life prediction in parallel to the input layer of the fully connected neural network; Using the second attachment's operating condition time series data as input to multiple fatigue life prediction sub-models, and using the second label identifying fatigue life as the output of the fully connected neural network, transfer learning is performed on the fatigue life prediction base model to obtain the target forklift attachment fatigue life prediction model.

2. The method as described in claim 1, characterized in that, Load the target forklift attachment's operational monitoring data, perform frequent fatigue location pattern analysis, and obtain a list of frequent fatigue points, including: Using attachment material parameters as constraints, fatigue detection record data is collected, and frequent fatigue symptom type pattern analysis is performed to obtain a set of frequent fatigue symptom types. Extract forklift attachment defect monitoring data from the operation monitoring data, and statistically analyze the forklift attachment location set and the frequent fatigue symptom trigger frequency set that show at least one frequent fatigue symptom type in the frequent fatigue symptom type set. The forklift attachments whose frequent fatigue type trigger frequencies are greater than or equal to the trigger frequency threshold are added to the list of frequent fatigue points.

3. The method as described in claim 2, characterized in that, From the operation monitoring data, forklift attachment defect monitoring data is extracted, and the set of forklift attachment locations and the set of frequent fatigue symptom trigger frequencies that show at least one frequent fatigue symptom type in the frequent fatigue symptom type set are statistically analyzed, including: From the forklift attachment defect monitoring data, extract the first set of defect monitoring data up to the Qth set of defect monitoring data, where Q represents the number of sets of forklift attachment defect monitoring data; The first set of defect monitoring data is integrated according to location up to the Qth set of defect monitoring data to obtain the first location defect monitoring dataset up to the Nth location defect monitoring dataset, where N represents the number of locations; Delete sets whose defect monitoring data are less than or equal to the data quantity threshold to obtain several remaining defect monitoring datasets. Traverse the several retained location defect monitoring datasets, and based on the frequent fatigue symptom type set, statistically analyze the forklift attachment location set and the frequent fatigue type trigger frequency set that occur at least one frequent fatigue symptom type in the frequent fatigue symptom type set.

4. The method as described in claim 3, characterized in that, Traverse the aforementioned datasets of defect monitoring at various retention locations, and based on the set of frequent fatigue symptom types, statistically analyze the set of forklift attachment locations where at least one frequent fatigue symptom type appears in the set of frequent fatigue symptom types and the set of frequent fatigue type trigger frequencies, including: From the set of frequent fatigue symptom types, extract the first frequent fatigue symptom type up to the Qth frequent fatigue symptom type, where Q represents the number of frequent fatigue symptom types; From the aforementioned datasets of defect monitoring at several remaining locations, extract the first set of defect types at remaining locations; The trigger frequency of the first frequent fatigue symptom type in the first retention location defect type set is statistically analyzed and set as the first trigger frequency; The frequency of the Qth frequent fatigue symptom type in the first retention location defect type set is statistically analyzed and set as the Qth trigger frequency. The summation of the first trigger frequency up to the Qth trigger frequency is set as the first retention position frequent fatigue type trigger frequency; When the frequency of frequent fatigue type triggering at the first retention position is not equal to 0, the frequency of frequent fatigue type triggering at the first retention position is added to the set of frequent fatigue type triggering frequencies, and the first retention position is added to the set of forklift attachment positions.

5. The method as described in claim 1, characterized in that, Based on the aforementioned work monitoring data, the list of frequent fatigue points is traversed, and a frequent force direction pattern analysis is performed to obtain a list of frequent force directions, including: Extract the first frequent fatigue position from the list of frequent fatigue points; Based on the first frequently fatigued locations, the force directions of the first set of frequently fatigued locations are extracted from the work monitoring data; Based on the angle deviation threshold, clustering is performed on the force direction of the first group of frequent fatigue locations to obtain multiple clusters of the force direction of the first frequent fatigue locations. Take the multiple first fatigue frequent position centroid force directions of the multiple cluster first fatigue frequent position force directions and set them as the first group of reconstructed fatigue frequent position force directions. At the same time, set the number of clusters of the multiple cluster first fatigue frequent position force directions as multiple initial trigger frequencies of the multiple first fatigue frequent position centroid force directions and add them to the first group of reconstructed fatigue frequent position force directions initial trigger frequencies. Until the force direction of the frequently reconstructed fatigue location in the Lth group and the initial triggering frequency of the force direction of the frequently reconstructed fatigue location in the Lth group are obtained; Based on the force direction of the first group of reconstructed fatigue frequent locations and the initial trigger frequency of the force direction of the first group of reconstructed fatigue frequent locations, until the force direction of the Lth group of reconstructed fatigue frequent locations and the initial trigger frequency of the force direction of the Lth group of reconstructed fatigue frequent locations, a frequent force direction pattern analysis is performed on the first fatigue frequent location. Continue until the frequent stress direction pattern analysis is completed for all fatigue-prone points, and output a list of frequent stress directions.

6. A forklift attachment fatigue life prediction system employing transfer learning, characterized in that, The system is used to implement the forklift attachment fatigue life prediction method using transfer learning as described in any one of claims 1-5, including: The frequent pattern analysis module is used to load the operation monitoring data of the target forklift attachments, perform fatigue position frequent pattern analysis, and obtain a list of frequent fatigue points; The force direction analysis module is used to traverse the list of frequent fatigue points based on the operation monitoring data, perform frequent force direction pattern analysis, and obtain a list of frequent force directions. The basic model training module is used to collect several one-to-one corresponding time-series data of the first attachment's working conditions and labels that identify fatigue life, constrained by the attachment structure topology, attachment material parameters, the list of frequent fatigue points and the list of frequent force directions, and to train the basic model for fatigue life prediction. The target model training module is used to collect several corresponding second attachment operating condition time series data and second fatigue life labels based on the target forklift attachment model, the list of frequent fatigue points and the list of frequent force directions, and to perform transfer learning on the fatigue life prediction basic model to obtain the target forklift attachment fatigue life prediction model and execute the forklift attachment fatigue life prediction task.

Images