Steel crane beam safety rating method, device, equipment and medium

Abstract

The invention relates to a steel crane beam safety rating method, device, equipment and medium, and belongs to the technical field of industrial building structure diagnosis and treatment. The method comprises the steps that at least one influence factor of any component is obtained; at least one influence factor of the component is preprocessed, the preprocessed influence factor is processed through a decision tree algorithm in the trained rating model, and at least one rating decision result of the component is obtained; performing voting processing on the at least one rating decision result through a random forest algorithm in the trained rating model, and obtaining a safety rating result of the component according to the voting result; and based on the safety rating results of the multiple components of the steel crane beam, obtaining the safety rating result of the steel crane beam. By means of the scheme, the structural diagnosis and treatment data processing flow of the steel crane beam is more efficient, dependence on personnel experience is reduced, the safety of the steel crane beam structure is monitored in real time, and safety accidents are effectively prevented.

Description

Technical Field

[0001] The invention relates to the field of industrial building structure diagnosis and treatment, and in particular to a steel crane beam safety rating method, device, equipment and medium. Background Art

[0002] Traditional industrial building structure diagnosis and treatment methods have technical development bottlenecks and are in urgent need of reform. Currently, most of the inspections of steel crane beams are often completed by on-site investigation and measurement by inspection experts. The tasks are heavy, the inspection methods are diverse, and the structures are complex. It is easy to encounter difficulties such as long construction period, large personnel input, and strong dependence on experience. In addition, the data processing is highly dependent on experience, there is no standard processing standard and process, and the inspection methods are one-time, which can only reflect the component status at the time of inspection, and cannot monitor the safety of the structure in real time, and cannot prevent safety accidents. Summary of the invention

[0003] To solve the above problems, the present application provides a steel crane beam safety rating method, device, equipment and medium.

[0004] A first aspect of an embodiment of the present application provides a method for safety rating of a steel crane beam, wherein the steel crane beam comprises a plurality of components, including: Obtain at least one influencing factor of any component; the influencing factors include: steel structure damage information, steel crane beam structure information, steel crane beam size information, steel crane beam material strength, steel crane beam corrosion level, steel crane beam displacement deformation information and steel crane beam weld defect information; Preprocessing at least one influencing factor of the component to obtain the preprocessed influencing factor; Processing the preprocessed influencing factors by using a decision tree algorithm in a trained rating model to obtain at least one rating decision result of the component; Voting at least one rating decision result using the random forest algorithm in the trained rating model to obtain a voting result, and obtaining a safety rating result of the component according to the voting result; Based on the safety rating results of the multiple components of the steel crane beam, a safety rating result of the steel crane beam is obtained.

[0005] Optionally, preprocessing at least one influencing factor of the component to obtain the preprocessed influencing factor, and processing the preprocessed influencing factor by a decision tree algorithm in a trained rating model to obtain at least one rating decision result of the component specifically includes: Encoding the feature data of at least one influencing factor of the steel crane beam to obtain the encoded feature data, assigning values ​​and marking rating labels to the encoded feature data to obtain the encoded feature data and the rating labels corresponding to the encoded feature data; The decision tree algorithm in the trained rating model is used to perform judgment processing on the encoded feature data and the rating labels corresponding to the encoded feature data to obtain a rating decision result.

[0006] Optionally, historical raw data of the steel crane beam is obtained; wherein the historical raw data includes: at least one influencing factor of each component of the steel crane beam; Performing data cleaning on the historical raw data to obtain target feature data; A dataset for constructing a rating model based on the target feature data of each component; Dividing the data set into a training set and a test set; Training the initial rating model using the training set to obtain a trained rating model; The trained rating model is verified by using a test set to obtain a trained rating model.

[0007] Optionally, the performing data cleaning on the historical raw data to obtain target feature data specifically includes: Classify the characteristic data in the influencing factors of each component to obtain the classified characteristic data; The classified feature data is encoded to obtain the encoded feature data, and the encoded feature data is assigned values ​​and marked with rating labels to obtain the target feature data.

[0008] Optionally, it also includes: Matching the corresponding rating standard value according to the numerical value of the characteristic data; wherein the rating standard value is determined according to the characteristic data of components of different safety levels in a preset appraisal table; Obtaining the security level corresponding to the rating standard value; The security level is encoded as a rating label corresponding to the feature data.

[0009] Optionally, it also includes: In the case where there are missing values ​​in the influencing factors, the missing values ​​of the historical original data are filled by a data interpolation algorithm based on random forest to obtain optimized historical original data.

[0010] Optionally, the steel structure damage information includes: defect information or damage information of the steel structure; The structural information of the steel crane beam includes: load information of the steel crane; The size information of the steel crane beam includes: the cross-sectional size information of the flange of the steel crane beam; The material strength of the steel crane beam includes: tensile strength, measured Leeb hardness value, average Leeb hardness value and Vickers hardness; The corrosion level of the steel crane beam includes: the measured value of the cross-sectional dimensions of the corroded part, the average value of the cross-sectional dimensions of the corroded part, the measured value of the cross-sectional dimensions of the uncorroded part, the average value of the cross-sectional dimensions of the uncorroded part, the designed value of the cross-sectional dimensions of the uncorroded part, the measured value of the corrosion damage amount and the average value of the corrosion damage amount; The displacement deformation information of the steel crane beam includes: longitudinal deviation, lateral deviation, inclination value, measurement height, inclination rate, span, mid-span relative support height difference, deflection, deformation rate and truss beam vertical deformation; The weld defect information of the steel crane beam includes: vertical distance from the detection surface to the defect, defect indication length, plate thickness and refraction angle.

[0011] A second aspect of an embodiment of the present application provides a steel crane beam safety rating device, comprising: An initial data acquisition module is used to acquire at least one influencing factor of at least one component; the influencing factor includes: steel structure damage information, steel crane beam structure information, steel crane beam size information, steel crane beam material strength, steel crane beam corrosion level, steel crane beam displacement deformation information and steel crane beam weld defect information; A preprocessing module, used for preprocessing at least one influencing factor of the component to obtain the preprocessed influencing factor; A decision module, used to process the pre-processed influencing factors through a decision tree algorithm in a trained rating model to obtain a rating decision result; The rating module is used to vote on at least one rating decision result through the random forest algorithm in the trained rating model to obtain a voting result, and obtain the safety rating result of the component according to the voting result.

[0012] A third aspect of an embodiment of the present application provides an electronic device, including a memory and a processor, wherein: The memory is used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in a steel crane beam safety rating method of any of the above solutions.

[0013] In a fourth aspect of an embodiment of the present application, a computer-readable storage medium is used to store a computer-readable program or instruction, which, when executed by a processor, can implement the steps of a steel crane beam safety rating method of any of the above-mentioned schemes.

[0014] By applying the technical solution provided in the embodiment of the present application, through integrating the influencing factors of multiple components of the steel crane beam, it is proposed to conduct safety assessment of the reliability of the steel crane beam based on the influencing factors of the components combined with the decision tree algorithm and the random forest algorithm, so that the structural diagnosis and treatment data processing flow of the steel crane beam is more efficient, the dependence on personnel experience is reduced, and real-time monitoring of the safety of the steel crane beam structure is realized, effectively preventing the occurrence of safety accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0016] Figure 1 is a flow chart of the steps of a steel crane beam safety rating method provided in an embodiment of the present invention; Figure 2 It is a flowchart of the steps of preprocessing component information and rating decision processing provided in an embodiment of the present invention; Figure 3 is a flowchart of the steps of the rating model training process provided in an embodiment of the present invention; Figure 4 is a schematic diagram of a decision tree structure provided in an embodiment of the present invention; Figure 5 is a flowchart of the steps of data cleaning provided in an embodiment of the present invention; Figure 6 is a flowchart of the steps of determining a rating label provided in an embodiment of the present invention; Figure 7 is a structural block diagram of a steel crane beam safety rating device provided in an embodiment of the present invention; Figure 8 It is a hardware structure block diagram of an electronic device provided in each embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Reference Figure 1As shown, a flow chart of the steps of a steel crane beam safety rating method is shown, which can be applied to the diagnosis and treatment of industrial building structures. The steel crane beam includes multiple components, such as Figure 1 As shown, the following steps may be specifically included: Step S101, obtaining at least one influencing factor of any component; the influencing factors include: steel structure damage information, steel crane beam structure information, steel crane beam size information, steel crane beam material strength, steel crane beam corrosion level, steel crane beam displacement deformation information and steel crane beam weld defect information; In this embodiment, steel structure damage information, steel crane beam structural information, steel crane beam size information, steel crane beam material strength, steel crane beam corrosion level, steel crane beam displacement deformation information and steel crane beam weld defect information are obtained through on-site investigation and / or inspection of the steel crane beam.

[0019] Step S102, preprocessing at least one influencing factor of the component to obtain a preprocessed influencing factor; In this embodiment, preprocessing at least one influencing factor of the component may include: The characteristic data in the influencing factors of the components are cleaned, duplicate records in the data set are checked and deleted, errors in the data, such as input errors and measurement errors, are identified and corrected to ensure the accuracy of the data; the format and dimension of similar characteristic data are unified.

[0020] The cleaned feature data is converted into data encoding to facilitate subsequent rating model processing.

[0021] Step S103, processing the preprocessed influencing factors by using a decision tree algorithm in the trained rating model to obtain at least one rating decision result of the component; In this embodiment, the decision tree algorithm in the trained rating model can be used to confirm the first-level branch nodes, the second-level branch nodes, and the third-level branch nodes of the decision tree of the steel crane beam component; The first-level branch nodes are impact factors; The second-level tree branches are steel structure damage information, steel crane beam structure information, steel crane beam size information, steel crane beam material strength, steel crane beam corrosion level, steel crane beam displacement deformation information and steel crane beam weld defect information; The third-level tree branches are the specific values ​​or descriptions of steel structure damage information, steel crane beam structure information, steel crane beam size information, steel crane beam material strength, steel crane beam corrosion level, steel crane beam displacement deformation information, and steel crane beam weld defect information; According to the first-level branch nodes, the second-level branch nodes and the third-level branch nodes of the decision tree of the steel crane beam component and the decision tree algorithm, the input influencing factor data is processed to obtain a rating decision result of at least one influencing factor.

[0022] In one embodiment, there are more than one specific numerical values ​​or description contents of steel structure damage information, steel crane beam structural information, steel crane beam size information, steel crane beam material strength, steel crane beam corrosion level, steel crane beam displacement deformation information and steel crane beam weld defect information. A rating decision result is generated based on different specific numerical values ​​or description contents, so there will be one or more influencing factors.

[0023] Step S104: voting on at least one rating decision result by using the random forest algorithm in the trained rating model to obtain a voting result, and obtaining a safety rating result of the component according to the voting result; In this embodiment, at least one rating decision result is obtained based on a decision tree, and multiple rating decision results are voted on by a random forest in a trained rating model to obtain a safety rating result of the component. The overfitting defect of the decision tree can be solved by combining the random forest algorithm.

[0024] Step S105: obtaining a safety rating result of the steel crane beam based on the safety rating results of the multiple components of the steel crane beam.

[0025] In this embodiment, multiple influencing factors of any component of the steel crane beam are obtained, and the multiple influencing factors are preprocessed to facilitate subsequent rating model processing. The preprocessed influencing factors are processed by the decision tree algorithm in the trained rating model to obtain at least one rating decision result of the component. The at least one rating decision result is voted on by the random forest algorithm in the trained rating model to obtain a voting result, and a safety rating result of the component is obtained based on the voting result; the safety rating results of the multiple components of the steel crane beam are combined to obtain the safety rating result of the steel crane beam.

[0026] By integrating the influencing factors of multiple components of steel crane beams, it is proposed to conduct safety assessment of the reliability of steel crane beams based on the influencing factors of the components combined with the decision tree algorithm and the random forest algorithm, which makes the structural diagnosis and treatment data processing process of steel crane beams more efficient, reduces the dependence on personnel experience, realizes real-time monitoring of the safety of the steel crane beam structure, and effectively prevents the occurrence of safety accidents.

[0027] In one embodiment, referring to Figure 2 As shown, a flowchart of the steps of preprocessing component information and rating decision processing is shown, such as Figure 2 As shown, the following steps may be specifically included: Step S201, encoding the feature data of at least one influencing factor of the steel crane beam to obtain the encoded feature data, assigning values ​​and marking rating labels to the encoded feature data to obtain the encoded feature data and the rating labels corresponding to the encoded feature data; Step S202: using the decision tree algorithm in the trained rating model to determine and process the encoded feature data and the rating labels corresponding to the encoded feature data, to obtain a rating decision result.

[0028] The technical solution of the embodiment of the present application obtains a clear data structure and clear data relationship through preprocessing. When the data is clear, the decision tree and random forest models can better learn the patterns and laws in the data, thereby improving the accuracy of the prediction.

[0029] In one embodiment, referring to Figure 3 As shown, a flowchart of the steps of the rating model training process is shown, as Figure 3 As shown, the following steps may be specifically included: Step S301, obtaining historical original data of the steel crane beam; wherein the historical original data includes: at least one influencing factor of each component of the steel crane beam; Step S302, clean the historical raw data to obtain target feature data; perform data conversion and / or group similar data on the cleaned feature data, specify corresponding codes for different data categories, specify the units of the detection data, and convert different data into unified units and organize them into a data table, as shown in Table 1.

[0030] Table 1

[0031] Step S303: constructing a data set of a rating model according to the target feature data of each component; Step S304, dividing the data set into a training set and a test set; Step S305: training the initial rating model using the training set to obtain a trained rating model; In this implementation, the information increment corresponding to the pivot point of the decision tree in the initial rating model after classification is: , Where: D is the current dataset; a is the feature currently being considered. V is the number of different values ​​of feature a. It is a sub-dataset whose feature a value is v.

[0032] Ent(D) is the information entropy of data set D, and its calculation formula is: , in, is the proportion of samples of the kth category in the data set D, and K is the total number of sample categories.

[0033] In one embodiment, the information entropy value of the root node is calculated based on the training data. For example, there are 8 Class A safety ratings and 9 Class B safety ratings. Then the information entropy of the root node is Ent(D)=-1*(8 / 17)*log(8 / 17)-(9 / 17)log(9 / 17)=0.998.

[0034] Assume that the first-level tree nodes are the size information of steel crane beams. The size information has three categories: A, B, and C. There are 6 A, 5 B, and 6 C. Among them, three of "Class A" have Class A safety rating, and three have Class B safety rating. Then Ent(Class A) = - (3 / 6log3 / 6 + 3 / 6log3 / 6) = 1; Among “Class b”, one has Class A safety rating and four have Class B safety rating, then Ent (Class b) = - (1 / 5log1 / 5 + 4 / 5log4 / 5) = 0.722; There are four Class A safety ratings and two Class B safety ratings in “Class c”, so Ent (Class c) = - (4 / 6log4 / 6 + 2 / 6log2 / 6) = 0.918; Considering that the probabilities of the three types of size information are 6 / 17, 5 / 17, and 6 / 17 respectively, if the first-level tree node selects "size information", the entropy value of this level after size information classification is: (6 / 17*1+5 / 17*0.722+6 / 17*0.918)=0.889.

[0035] Then if the size information is selected as the first-level branch, the corresponding information increment is Gain (size information): 0.998-0.889=0.109.

[0036] In the same way, the information increments of other influencing factors can be calculated, and the influencing factor with the largest information increment is used as the first-level branch node, and then the lower-level nodes are determined based on the information increments of each influencing factor.

[0037] By repeating in the same way, we can get the trained decision tree. The structure of the decision tree taking size information as an example is as follows: Figure 4 shown.

[0038] Calculate the information gain of different influencing factors, take the influencing factor with the largest information gain as the first-level branch node, and then determine the second-level branch nodes and the third-level branch nodes in turn. After determining the branch nodes, perform cyclic training through the training data to obtain the trained decision tree.

[0039] The decision results output by the decision tree are then input into the random forest model for cyclic training to obtain the trained random forest, and the trained decision tree and random forest are determined as the trained rating model.

[0040] Step S306: Verify the trained rating model using a test set to obtain a trained rating model.

[0041] The technical solution of the embodiment of the present application improves the consistency and accuracy of the data by preprocessing the original data such as classification and coding. The preprocessing of the cleaned data can make its structure clearer and easier to understand and interpret. The processed data can improve the calculation accuracy of the rating model.

[0042] The rating model is trained through training sets and test sets to determine the optimal algorithm and parameter combination, and ultimately determine the optimal model parameters of the rating model.

[0043] In one embodiment, referring to Figure 5 As shown, a flowchart of the steps of data cleaning is shown, such as Figure 5 As shown, the following steps may be specifically included: Step S501, classifying the characteristic data in the influencing factors of each component to obtain the classified characteristic data; Step S502: Encode the classified feature data to obtain the encoded feature data, assign values ​​and mark rating labels on the encoded feature data to obtain target feature data.

[0044] In this embodiment, the characteristic data in the influencing factors of each component are classified, wherein the influencing factors are shown in the data categories in Table 1. The influencing factors are first classified to obtain basic information, structural effects, defects or damage, and dimensions, and then the influencing factors in the data categories, that is, the characteristic data, are classified into corresponding data categories. The classified characteristic data are encoded, standardized and coded, and different characteristic data are converted into a unified unit to facilitate subsequent rating model processing.

[0045] In one embodiment, referring to Figure 6 As shown, a flowchart of the steps for determining a rating label is shown, such as Figure 6 As shown, the following steps may be specifically included: Step S601, matching the corresponding rating standard value according to the numerical value of the characteristic data; wherein the rating standard value is determined according to the characteristic data of components of different safety levels in a preset appraisal table; in this embodiment, the preset appraisal table can be the national standard "Industrial Building Reliability Appraisal Standard".

[0046] Step S602: Obtain the security level corresponding to the rating standard value; Step S603: Encode the security level into a rating label corresponding to the feature data.

[0047] By determining the rating labels, subsequent rating models can accurately distinguish the security levels of feature data.

[0048] In one embodiment, a method for filling missing values ​​is provided, which may specifically include the following steps: In the case where the influencing factor has missing values, the missing values ​​of the historical original data are filled by a data interpolation algorithm based on random forest to obtain optimized historical original data. In an embodiment, the data interpolation algorithm may be a MissForest (missing data filling method based on random forest) algorithm.

[0049] The completeness of the data can be improved by filling in missing values. A complete data set can avoid analytical bias caused by missing data and enable the rating model to use more information for learning and prediction.

[0050] In one embodiment, the steel structure damage information provided includes: defect information or damage information of the steel structure; The structural information of the steel crane beam includes: load information of the steel crane; The size information of the steel crane beam includes: the cross-sectional size information of the flange of the steel crane beam; The material strength of the steel crane beam includes: tensile strength, measured Leeb hardness value, average Leeb hardness value and Vickers hardness; The corrosion level of the steel crane beam includes: the measured value of the cross-sectional dimensions of the corroded part, the average value of the cross-sectional dimensions of the corroded part, the measured value of the cross-sectional dimensions of the uncorroded part, the average value of the cross-sectional dimensions of the uncorroded part, the designed value of the cross-sectional dimensions of the uncorroded part, the measured value of the corrosion damage amount and the average value of the corrosion damage amount; The displacement deformation information of the steel crane beam includes: longitudinal deviation, lateral deviation, inclination value, measurement height, inclination rate, span, mid-span relative support height difference, deflection, deformation rate and truss beam vertical deformation; The weld defect information of the steel crane beam includes: vertical distance from the detection surface to the defect, defect indication length, plate thickness and refraction angle.

[0051] By integrating the characteristic data of influencing factors, it is proposed to evaluate the reliability of steel crane beams based on the characteristic data of components combined with decision tree algorithm and random forest algorithm. This makes the structural diagnosis and treatment data processing process more efficient, reduces the dependence on personnel experience, realizes real-time monitoring of the safety of steel crane beam structures, and effectively prevents the occurrence of safety accidents.

[0052] In one embodiment, the data preprocessing process includes: feature design, data sorting and missing value filling.

[0053] In the feature design process, if a certain influencing factor requires multiple inspections and monitoring of different parts of the same component to obtain the corresponding situation analysis, in order to reflect all data characteristics of multiple inspections and monitoring, this influencing factor will participate in the model construction with the mean and variance of its multiple physical quantities as characteristic values, such as the tensile strength of the component; if a certain influencing factor needs to be compared by inspecting and monitoring two parts of the same component to obtain the corresponding situation analysis, and does not care about the inspection and monitoring order, in order to reflect the comparison situation and the actual measurement value, this influencing factor will be modeled with the minimum value and difference of the two measurements of its physical quantity as the characteristic value, such as the actual measured value of the side length of the angle steel; if a certain influencing factor is distinguished by category, and all categories need to be combined to analyze the corresponding influencing factor situation, in order to reflect the joint determination of multiple types, this influencing factor should be re-set for the combination of multiple types of characteristics, and modeled with single categories and multi-type combination categories as features, such as steel structure defects and damage.

[0054] In the data sorting process, the data is standardized and coded. The specific processing method is to specify the corresponding codes for different data categories, specify the units of test data, and convert different data into unified units. Then the different original test reports are sorted into data tables.

[0055] In the process of filling missing values, some data are often missing in the original data, which is handled in two categories. According to the above influencing factor table, 1) if the data of the "3 defects and damages" (as shown in Table 1) feature is not recorded, based on experience, this means that the structure does not have this type of defects and damage. 2) The other 8 types of influencing factors are all filled with missing values ​​using the MissForest algorithm, which can automatically detect and fill in missing data. The MissForest algorithm here is a packaged algorithm model and has not been modified by itself.

[0056] In one embodiment, the data is divided into a training set and a test set by a support vector machine algorithm, and the decision tree algorithm and the random forest algorithm are used to train the test set data to determine the optimal algorithm and parameter combination, and finally the random forest algorithm is determined to be the best. The training set accounts for 70% of the total data and is mainly used for the establishment of machine learning models, parameter selection and cross-validation; the test set accounts for 30% of the total data and is mainly used to verify the classification accuracy of the machine learning model. The 10-fold cross validation divides the training set into 10 parts, selects 9 of them for training, and the remaining 1 is used as a cross-validation test model accuracy. According to the performance of the test set prediction results, we choose the random forest algorithm.

[0057] In one embodiment, a random forest algorithm is used to perform model training on the preprocessed data to ultimately determine the safety rating of the steel crane beam.

[0058] In the process of using the random forest algorithm, automatic parameter adjustment is performed during the training process. According to the rating requirements of steel crane beams, the parameter ranges of the important parameters in the random forest algorithm are given. Then, within each parameter range, values ​​are taken at certain intervals, and different parameter values ​​are combined to form different parameter combinations. Then, in the test set, the optimal parameter combination is selected based on the accuracy of the results through cross-validation. This approach allows the optimal parameter combination of the algorithm to be non-fixed. As new data enters, the model can continuously optimize its parameter combination to make the current prediction result optimal.

[0059] When processing data, we used three different algorithms: decision tree, random forest and MissForest to ensure the precision and accuracy of data processing. These algorithms have different focuses in data preprocessing, model selection, model training and model application.

[0060] First, in the data preprocessing stage, we use the physical quantities of steel structure defects and damage, displacement, material strength and dimensional measurement that affect the steel crane beam as the basis, and combine actual business experience to comprehensively analyze the safety and reliability characteristics of the components by comprehensively analyzing the fatigue cracks, track eccentricity, appearance quality and strain. Among the factors affecting the correct rating of the steel crane beam safety rating algorithm, the size of the crane beam is more important. We organize and summarize the data according to the feature design plan to form an input data table. For multiple measurement data, we preprocessed the mean, variance, minimum value, difference, etc., and classified the category data. To ensure the smooth progress of model selection and training, we ensured the integrity of the data, cleaned the data, unified the format and dimension of similar feature data, and avoided the situation where fuzzy categories or data ranges were used as feature data.

[0061] When dealing with missing values ​​of steel structure defects and damage in the data, we assumed that there were no defects and damages, and filled the corresponding missing values ​​by creating a new category. For the missing values ​​of other features, we used the MissForest algorithm because it is more realistic and helpful for model analysis, so we chose the MissForest algorithm to fill the missing values ​​of other features.

[0062] In terms of model selection, we used the preprocessed data set for model training. We used random sampling and 10-fold cross-validation to divide the data set into a training set and a test set. The training set accounted for 70% of the total data and was mainly used for the establishment of machine learning models, parameter selection and cross-validation; the test set accounted for 30% of the total data and was mainly used to verify the classification accuracy of the machine learning model. The 10-fold cross-validation divided the training set into 10 parts, selected 9 of them for training, and the remaining 1 was used as a cross-validation test model accuracy. By building models and verifying each type of parameter combination, we selected the best parameter combination.

[0063] As well as the application of feature importance, it can identify the contribution (importance) of existing data features to the correct classification of the model when training the data classification prediction model, and use a continuous variable with a threshold range of [0,1] to quantify the feature importance. The closer the feature is to 1, the more important it is for correct classification. Conversely, the closer the feature is to 0, the less important it is for correct classification, thereby achieving data-oriented sorting of the relevance of rating features of three typical components.

[0064] In summary, the three algorithms, decision tree, random forest and MissForest, have different focuses in data processing and model training. The decision tree method plays a role in feature selection and model interpretation with its intuitiveness and interpretability; random forest has outstanding performance in model training and prediction with its strong classification ability and anti-overfitting ability; and MissForest provides an effective solution in the data preprocessing stage with its advantage in handling missing values. By comprehensively using these three algorithms, we ensure the precision and accuracy of data processing and improve the prediction performance of the model.

[0065] In other embodiments, the statistical parameters of load effect and resistance are determined by statistically analyzing the measured data of steel structure crane beams and previous fatigue test data. These parameters include the mean and standard deviation of load effect and resistance, which are the basis for evaluating fatigue reliability. According to the calculation results, the recommended values ​​of fatigue target reliability index and the design expression of partial coefficient are given. These indexes and coefficients are key parameters for designing and evaluating the fatigue reliability of steel crane beams. Reliability is evaluated based on the safety and usability obtained from fatigue and deformation.

[0066] Reference Figure 7 As shown, a structural block diagram of a steel crane beam safety rating device 700 according to an embodiment of the present invention is shown. Figure 7 As shown, the device may specifically include the following modules: The initial data acquisition module 701 is used to acquire at least one influencing factor of at least one component; the influencing factor includes: steel structure damage information, steel crane beam structure information, steel crane beam size information, steel crane beam material strength, steel crane beam corrosion level, steel crane beam displacement deformation information and steel crane beam weld defect information; A preprocessing module 702 is used to preprocess at least one influencing factor of the component to obtain the preprocessed influencing factor; A decision module 703 is used to process the pre-processed influencing factors through a decision tree algorithm in the trained rating model to obtain a rating decision result; The rating module 704 is used to perform voting processing on at least one rating decision result through the random forest algorithm in the trained rating model to obtain a voting result, and obtain the safety rating result of the component according to the voting result.

[0067] In one embodiment, the preprocessing module 702 is specifically used to encode the feature data of at least one influencing factor of the steel crane beam, obtain the encoded feature data, assign values ​​and mark rating labels to the encoded feature data, and obtain the encoded feature data and the rating labels corresponding to the encoded feature data; The decision module 703 is specifically used to determine and process the encoded feature data and the rating labels corresponding to the encoded feature data through the decision tree algorithm in the trained rating model to obtain a rating decision result.

[0068] In one embodiment, it further includes: a training module for acquiring historical raw data of the steel crane beam; wherein the historical raw data includes: at least one influencing factor of each component of the steel crane beam; Performing data cleaning on the historical raw data to obtain target feature data; A dataset for constructing a rating model based on the target feature data of each component; Dividing the data set into a training set and a test set; Training the initial rating model using the training set to obtain a trained rating model; The trained rating model is verified by using a test set to obtain a trained rating model.

[0069] In one embodiment, the training module is specifically used to classify the characteristic data in the influencing factor of each component to obtain the classified characteristic data; The classified feature data is encoded to obtain the encoded feature data, and the encoded feature data is assigned values ​​and marked with rating labels to obtain the target feature data.

[0070] In one embodiment, it further includes a rating label acquisition module, which is used to match the corresponding rating standard value according to the numerical value of the characteristic data; wherein the rating standard value is determined according to the characteristic data of components with different safety levels in a preset identification table; Obtaining the security level corresponding to the rating standard value; The security level is encoded as a rating label corresponding to the feature data.

[0071] In one embodiment, it also includes: a data optimization module, which is used to fill the missing values ​​of the historical original data through a random forest-based data interpolation algorithm to obtain optimized historical original data when there are missing values ​​in the influencing factors.

[0072] In one embodiment, the steel structure damage information includes: defect information or damage information of the steel structure; The structural information of the steel crane beam includes: load information of the steel crane; The size information of the steel crane beam includes: the cross-sectional size information of the flange of the steel crane beam; The material strength of the steel crane beam includes: tensile strength, measured Leeb hardness value, average Leeb hardness value and Vickers hardness; The corrosion level of the steel crane beam includes: the measured value of the cross-sectional dimensions of the corroded part, the average value of the cross-sectional dimensions of the corroded part, the measured value of the cross-sectional dimensions of the uncorroded part, the average value of the cross-sectional dimensions of the uncorroded part, the designed value of the cross-sectional dimensions of the uncorroded part, the measured value of the corrosion damage amount and the average value of the corrosion damage amount; The displacement deformation information of the steel crane beam includes: longitudinal deviation, lateral deviation, inclination value, measurement height, inclination rate, span, mid-span relative support height difference, deflection, deformation rate and truss beam vertical deformation; The weld defect information of the steel crane beam includes: vertical distance from the detection surface to the defect, defect indication length, plate thickness and refraction angle.

[0073] The steel crane beam safety rating device provided in the above embodiment can implement the technical solution described in the above embodiment of a steel crane beam safety rating method. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above embodiment of a steel crane beam safety rating method, which will not be repeated here.

[0074] like Figure 8 As shown, the present invention also provides an electronic device 800. The electronic device 800 includes a processor 801, a memory 802 and a display 803. Figure 8 Only some components of the electronic device 800 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0075] In some embodiments, the memory 802 may be an internal storage unit of the electronic device 800, such as a hard disk or memory of the electronic device 800. In other embodiments, the memory 802 may also be an external storage device of the electronic device 800, such as a plug-in hard disk, a smart memory card (SmartMediaCard, SMC), a secure digital (SecureDigital, SD) card, a flash card (FlashCard), etc. equipped on the electronic device 800.

[0076] Furthermore, the memory 802 may include both an internal storage unit of the electronic device 800 and an external storage device. The memory 802 is used to store application software installed in the electronic device 800 and various data.

[0077] In some embodiments, the processor 801 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 802, such as a steel crane beam safety rating method in the present invention.

[0078] In some embodiments, the display 803 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 803 is used to display information on the electronic device 800 and to display a visual user interface. The components 801-803 of the electronic device 800 communicate with each other via a system bus.

[0079] In some embodiments of the present invention, when the processor 801 executes the steel crane beam safety rating program in the memory 802, the steps in any of the steel crane beam safety rating method embodiments may be implemented: It should be understood that: when the processor 801 executes the steel crane beam safety rating program in the memory 802, in addition to the above functions, other functions can also be implemented. For details, please refer to the description of the corresponding method embodiment above.

[0080] Furthermore, the embodiment of the present invention does not specifically limit the type of the electronic device 800 mentioned, and the electronic device 800 may be a portable electronic device such as a mobile phone, a tablet computer, a personal digital assistant (PDA), a wearable device, a laptop computer, etc. Exemplary embodiments of portable electronic devices include but are not limited to portable electronic devices equipped with IOS, Android, Microsoft or other operating systems. The above-mentioned portable electronic devices may also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the electronic device 800 may not be a portable electronic device, but a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0081] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to perform a steel crane beam safety rating method provided by the above methods, the method comprising: Obtain at least one influencing factor of any component; preprocess at least one influencing factor of the component to obtain the preprocessed influencing factor; process the preprocessed influencing factor by the decision tree algorithm in the trained rating model to obtain at least one rating decision result of the component; vote on at least one rating decision result by the random forest algorithm in the trained rating model to obtain a voting result, and obtain a safety rating result of the component according to the voting result; obtain the safety rating result of the steel crane beam based on the safety rating results of multiple components of the steel crane beam.

[0082] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0083] The above is a detailed introduction to the steel crane beam safety rating method, device, equipment and medium provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for technical personnel in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for safety rating of steel crane beams, characterized in that: Steel crane beams consist of multiple components, including: Obtain at least one influencing factor of any component; the influencing factors include: steel structure damage information, steel crane beam structure information, steel crane beam size information, steel crane beam material strength, steel crane beam corrosion level, steel crane beam displacement deformation information and steel crane beam weld defect information; Preprocessing at least one influencing factor of the component to obtain the preprocessed influencing factor; Processing the preprocessed influencing factors by using a decision tree algorithm in a trained rating model to obtain at least one rating decision result of the component; Voting at least one rating decision result using the random forest algorithm in the trained rating model to obtain a voting result, and obtaining a safety rating result of the component according to the voting result; Based on the safety rating results of the multiple components of the steel crane beam, a safety rating result of the steel crane beam is obtained.

2. A steel crane beam safety rating method according to claim 1, characterized in that: The preprocessing of at least one influencing factor of the component to obtain the preprocessed influencing factor, and processing the preprocessed influencing factor by a decision tree algorithm in a trained rating model to obtain at least one rating decision result of the component specifically includes: Encoding the feature data of at least one influencing factor of the steel crane beam to obtain the encoded feature data, assigning values ​​and marking rating labels to the encoded feature data to obtain the encoded feature data and the rating labels corresponding to the encoded feature data; The decision tree algorithm in the trained rating model is used to perform judgment processing on the encoded feature data and the rating labels corresponding to the encoded feature data to obtain a rating decision result.

3. A steel crane beam safety rating method according to claim 1 or 2, characterized in that: Also includes: Acquire historical original data of the steel crane beam; wherein the historical original data includes: at least one influencing factor of each component of the steel crane beam; Performing data cleaning on the historical raw data to obtain target feature data; A dataset for constructing a rating model based on the target feature data of each component; Dividing the data set into a training set and a test set; Training the initial rating model using the training set to obtain a trained rating model; The trained rating model is verified by using a test set to obtain a trained rating model.

4. A steel crane beam safety rating method according to claim 3, characterized in that: The data cleaning of the historical raw data to obtain target feature data specifically includes: Classify the characteristic data in the influencing factors of each component to obtain the classified characteristic data; The classified feature data is encoded to obtain the encoded feature data, and the encoded feature data is assigned values ​​and marked with rating labels to obtain the target feature data.

5. A steel crane beam safety rating method according to claim 3, characterized in that: Also includes: Matching the corresponding rating standard value according to the numerical value of the characteristic data; wherein the rating standard value is determined according to the characteristic data of components of different safety levels in a preset appraisal table; Obtaining the security level corresponding to the rating standard value; The security level is encoded as a rating label corresponding to the feature data.

6. A steel crane beam safety rating method according to claim 4 or 5, characterized in that: Also includes: In the case where there are missing values ​​in the influencing factors, the missing values ​​of the historical original data are filled by a data interpolation algorithm based on random forest to obtain optimized historical original data.

7. A method for safety rating of steel crane beams according to claim 1, 2, 4 or 5, characterized in that: The steel structure damage information includes: defect information or damage information of the steel structure; The structural information of the steel crane beam includes: load information of the steel crane; The size information of the steel crane beam includes: the cross-sectional size information of the flange of the steel crane beam; The material strength of the steel crane beam includes: tensile strength, measured Leeb hardness value, average Leeb hardness value and Vickers hardness; The corrosion level of the steel crane beam includes: the measured value of the cross-sectional dimensions of the corroded part, the average value of the cross-sectional dimensions of the corroded part, the measured value of the cross-sectional dimensions of the uncorroded part, the average value of the cross-sectional dimensions of the uncorroded part, the designed value of the cross-sectional dimensions of the uncorroded part, the measured value of the corrosion damage amount and the average value of the corrosion damage amount; The displacement deformation information of the steel crane beam includes: longitudinal deviation, lateral deviation, inclination value, measurement height, inclination rate, span, mid-span relative support height difference, deflection, deformation rate and truss beam vertical deformation; The weld defect information of the steel crane beam includes: vertical distance from the detection surface to the defect, defect indication length, plate thickness and refraction angle.

8. A steel crane beam safety rating device, characterized in that: include: An initial data acquisition module, used to acquire at least one influencing factor of at least one component; The influencing factors include: steel structure damage information, steel crane beam structure information, steel crane beam size information, steel crane beam material strength, steel crane beam corrosion level, steel crane beam displacement deformation information and steel crane beam weld defect information; A preprocessing module, used for preprocessing at least one influencing factor of the component to obtain the preprocessed influencing factor; A decision module, used to process the pre-processed influencing factors through a decision tree algorithm in a trained rating model to obtain a rating decision result; The rating module is used to vote on at least one rating decision result through the random forest algorithm in the trained rating model to obtain a voting result, and obtain the safety rating result of the component according to the voting result.

9. An electronic device, characterized in that: comprising a memory and a processor, wherein: The memory is used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the steel crane beam safety rating method described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the steel crane beam safety rating method described in any one of claims 1 to 7.

Images