Engineering machinery fault prediction method and system, engineering machinery and electronic equipment

Abstract

The invention discloses an engineering machinery fault prediction method and system, engineering machinery and electronic equipment, and relates to the technical field of engineering machinery. According to the method, after obtained operation data and carrying data of the engineering machinery are input into a fault prediction model obtained through pre-training based on historical operation data, historical carrying data and historical fault data of the engineering machinery, a fault prediction result of the engineering machinery is obtained through the fault prediction model. Therefore, the problem that the prediction accuracy is not high due to the fact that fault prediction is conducted on the engineering machinery only through the operation data of the engineering machinery is solved, fault prediction is conducted on the basis of the operation data and the carrying data of the engineering machinery, and accurate prediction of faults of the engineering machinery is achieved.

Description

Technical Field

[0001] The present application relates to the field of vehicle technology, and more specifically, to a method and system for predicting engineering machinery faults, engineering machinery, and electronic equipment. Background Art

[0002] Construction machinery is a general term for various mechanical equipment used in engineering construction. In engineering construction, they replace manpower, improve production efficiency and reduce labor intensity. Construction machinery is a necessary tool for developing industry and enhancing productivity. If its failures are not discovered and resolved in time, it will seriously affect the progress of engineering construction. Take dump trucks as an example. As a key part of social logistics, their performance and reliability directly affect the operating efficiency of the entire logistics system. Therefore, any sudden failure of dump trucks will not only affect the customer's operating plan, but may also bring additional challenges to the company's emergency maintenance and parts inventory management.

[0003] Data analysis shows that before a specific fault occurs, some operating data of construction machinery will show obvious abnormalities. Therefore, by monitoring these abnormal data, it is possible to predict the specific faults that are about to occur in construction machinery, and then perform necessary maintenance and care in advance, thereby avoiding the downtime of construction machinery caused by faults and ensuring the progress of construction projects.

[0004] At present, the prediction method for engineering machinery failure mainly includes the following steps: selecting an appropriate data set, converting the raw data into feature vectors, selecting an appropriate prediction model, and performing failure prediction. However, due to the limitation of only using the operation data of engineering machinery for failure prediction, the data type is single, which makes the prediction model unable to capture enough information to accurately predict failures. Summary of the invention

[0005] In view of this, the embodiments of the present application are committed to providing a method, system, engineering machinery and electronic equipment for predicting engineering machinery faults, so as to solve the problem of low prediction accuracy caused by simply predicting engineering machinery faults based on the operating data of the engineering machinery. By predicting faults based on the operating data and transportation data of the engineering machinery, accurate prediction of engineering machinery faults can be achieved.

[0006] In a first aspect, the present invention provides a method for predicting engineering machinery faults, comprising:

[0007] Obtain operation data and transportation data of construction machinery;

[0008] Inputting the operation data and the transport data into a fault prediction model to obtain a fault prediction result of the engineering machinery output by the fault model;

[0009] The fault prediction model is a model trained based on historical indicator data of the engineering machinery, and the historical indicator data includes: historical operation data, historical transportation data and historical fault data.

[0010] In a possible implementation, the training method of the fault prediction model includes:

[0011] Obtaining the historical indicator data;

[0012] Preprocessing the historical indicator data to obtain a feature vector;

[0013] Based on the feature vector, an integrated learning model is trained with a naive Bayes model, a CatBoost model, a SVM model and a random forest model as base models and a logistic regression model as a meta-model to obtain the fault prediction model.

[0014] In a possible implementation manner, the preprocessing of the historical indicator data includes:

[0015] Removing outliers from the historical indicator data;

[0016] Filling missing values ​​in the historical indicator data after removing outliers;

[0017] Performing fault data labeling on the historical indicator data after filling the missing values, so as to divide the historical indicator data into normal sample data and fault sample data;

[0018] Feature engineering is performed on the historical indicator data after the fault data is labeled to obtain the feature vector.

[0019] In a possible implementation manner, the performing feature engineering on the historical indicator data after the fault data is labeled to obtain the feature vector further includes:

[0020] Based on the historical indicator data after the fault data is marked, the correlation analysis between the fault and operation, and the fault and transportation of the engineering machinery is performed to determine the historical indicator data whose correlation with the fault exceeds a preset value.

[0021] In a possible implementation manner, the performing feature engineering on the historical indicator data after the fault data is labeled to obtain the feature vector further includes:

[0022] Performing undersampling processing on the normal sample data;

[0023] Oversampling is performed on the fault sample data.

[0024] In a possible implementation, the filling of missing values ​​in the historical indicator data after removing outliers includes:

[0025] Performing missing value preprocessing on the historical indicator data after removing outliers;

[0026] The missing value preprocessing includes: deleting attribute columns in the historical indicator data whose missing values ​​exceed a preset number; deleting data rows in the historical indicator data that lack preset key operating data, wherein the preset key operating data is operating data related to the fault of the engineering machinery;

[0027] Based on the data attributes of the historical indicator data after the missing value preprocessing, the missing values ​​of the historical indicator data are filled.

[0028] In a possible implementation, the filling missing values ​​of the historical indicator data based on the data attributes of the historical indicator data after the missing value preprocessing includes:

[0029] Using the mode, filling missing values ​​of the historical indicator data whose data attributes are discrete variables;

[0030] Using the average value, the missing values ​​of the historical indicator data of the continuous variable whose data attribute is a normal distribution are filled;

[0031] The missing values ​​of the historical indicator data of the continuous variable whose data attribute is non-normal distribution are filled by using the median value.

[0032] In a second aspect, the present invention provides a construction machinery fault prediction system, comprising:

[0033] A collection unit, used to obtain operation data and transportation data of the construction machinery;

[0034] A processing unit, configured to input the operation data and the transport data into a fault prediction model to obtain a fault prediction result of the engineering machinery output by the fault model;

[0035] The fault prediction model is a model trained based on historical indicator data of the engineering machinery, and the historical indicator data includes: historical operation data, historical transportation data and historical fault data.

[0036] In a third aspect, the present invention provides an engineering machine, comprising an engineering machine body and the engineering machine fault prediction system provided in the second aspect of the present invention.

[0037] In a fourth aspect, the present invention provides an electronic device, the electronic device comprising:

[0038] processor;

[0039] a memory for storing instructions executable by the processor;

[0040] The processor is used to execute the method provided by the first aspect of the present invention.

[0041] According to the engineering machinery fault prediction method provided by the present invention, the operation data and the transport data of the engineering machinery are first obtained, and then the operation data and the transport data of the engineering machinery are input into the fault prediction model pre-trained based on the historical operation data, the historical transport data and the historical fault data of the engineering machinery, so as to obtain the fault prediction result of the engineering machinery output by the fault model. That is, by using the fault prediction model pre-trained based on the historical operation data, the historical transport data and the historical fault data of the engineering machinery, on the one hand, by enriching the data types used for fault prediction, the amount of fault data related to the engineering machinery is increased, and the classification performance and generalization ability of the model are improved, thereby improving the accuracy of the engineering machinery fault prediction based on the fault prediction model, and on the other hand, when predicting the fault of the engineering machinery, the correlation between the fault of the engineering machinery and the load strength of the engineering machinery can be considered, and the accuracy of the prediction result is also improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0043] Figure 1 Shown is a flow chart of a method for predicting engineering machinery faults provided by an embodiment of the present invention.

[0044] Figure 2 Shown is a schematic diagram of fault data annotation provided by an embodiment of the present invention.

[0045] Figure 3 Shown is a flow chart of a missing value filling process provided by an embodiment of the present invention.

[0046] Figure 4 The figure shows the relationship between the maximum temperature, the minimum temperature and the fault of the battery pack of the new energy dump truck provided by the embodiment of the present invention.

[0047] Figure 5 Shown is a graph showing the relationship between the total transport tonnage and the number of brake system failures of a new energy dump truck provided by an embodiment of the present invention.

[0048] Figure 6The figure shows a flow chart of performing fault prediction on a dump truck using the engineering machinery fault prediction method provided by an embodiment of the present invention.

[0049] Figure 7 Shown is a structural diagram of a construction machinery fault prediction system provided by an embodiment of the present invention.

[0050] Figure 8 Shown is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0051] Unless otherwise defined, the technical terms or scientific terms used in the embodiments of this specification shall have the common meanings understood by persons with ordinary skills in the field to which this specification belongs. The words "first", "second" and similar words used in the embodiments of this specification do not indicate any order, quantity or importance, but are only used to avoid confusion of constituent elements.

[0052] Unless the context requires otherwise, throughout the specification, "plurality" means "at least two", and "including" is interpreted as an open, inclusive meaning, that is, "including, but not limited to". In the description of the specification, the terms "one embodiment", "some embodiments", "exemplary embodiments", "examples", "specific examples" or "some examples" are intended to indicate that a particular feature, structure, material or characteristic associated with the embodiment or example is included in at least one embodiment or example of the specification. The schematic representation of the above terms does not necessarily refer to the same embodiment or example.

[0053] The following will be combined with the drawings in the embodiments of this specification to clearly and completely describe the technical solutions in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0054] At present, the fault prediction method for construction machinery mainly includes the following steps: obtaining raw data, converting raw data into feature vectors, inputting feature vectors into pre-trained prediction models, and obtaining prediction results output by the prediction model. Among them, the prediction model is usually trained through historical operation data and fault data. Therefore, the raw data is usually the operation data of the construction machinery. However, it is understandable that simply using the operation data of construction machinery will result in few fault prediction features due to the scarcity of vehicle condition information, and the prediction model cannot capture enough information to accurately predict faults. In addition, insufficient features will also cause the model to be too simple, with weak generalization ability, and it is difficult to adapt to unseen data situations, so it is difficult to be practical.

[0055] The present invention aims to solve the above-mentioned problem by inputting the acquired operation data and transportation data of the engineering machinery into a fault prediction model pre-trained based on the historical operation data, historical transportation data and historical fault data of the engineering machinery, thereby obtaining a fault prediction result for the engineering machinery output by the fault model, and realizing the association between the fault prediction of the engineering machinery and the transportation data of the engineering machinery, thereby improving the accuracy of the prediction result.

[0056] It should be noted that the engineering machinery fault prediction method provided by the present invention is used for fault prediction of engineering machinery with loading and unloading functions, and is mainly used for fault prediction of new energy dump trucks.

[0057] Furthermore, the engineering machinery fault prediction method provided by the present invention is executed on an electronic device, which may be a controller configured on a dump truck, or an intelligent terminal such as a laptop computer or a smart phone remotely connected to the controller of the dump truck, or a server on the network side.

[0058] Based on the above, see Figure 1 , Figure 1 : is a flow chart of a method for predicting a fault of an engineering machine provided by an embodiment of the present invention. The process of the method for predicting a fault of an engineering machine provided by this embodiment may include:

[0059] S100: Acquire operation data and transportation data of the construction machinery.

[0060] Specifically, operating data refers to the data generated after the construction machinery is started and operated. Taking the new energy dump truck as an example, its operating data includes: fuel cell voltage, total battery pack current, battery pack temperature, battery status and gear status, etc., which can be collected through the T-box of the new energy dump truck; carrying data refers to the data uploaded by the construction machinery after load carrying. Taking the new energy dump truck as an example, its carrying data includes: gross weight of the dump truck, tare weight of the dump truck and number of lifts of the dump truck, etc., which can be collected through the remote control terminal or control module of the new energy dump truck.

[0061] S110, inputting the operation data and the transport data into a fault prediction model to obtain a fault prediction result for the construction machinery output by the fault model.

[0062] Among them, the fault prediction model is a model trained based on the historical indicator data of construction machinery, and the historical indicator data includes: historical operation data, historical transportation data and historical fault data.

[0063] Specifically, the fault data of new energy dump trucks include: fault code, fault location, fault level and fault start and end time, etc.

[0064] In this embodiment, the operation data and transport data of the engineering machinery are obtained, and then the operation data and transport data of the engineering machinery are input into the fault prediction model pre-trained based on the historical operation data, historical transport data and historical fault data of the engineering machinery, so as to obtain the fault prediction result of the engineering machinery output by the fault model. Among them, on the one hand, by pre-training the fault prediction model based on the historical operation data, historical transport data and historical fault data of the engineering machinery, it is realized that by enriching the data types used for fault prediction, the amount of fault data related to the engineering machinery is increased, thereby improving the classification performance and generalization ability of the model, that is, improving the accuracy of the fault prediction of the engineering machinery based on the fault prediction model, and on the other hand, when predicting the fault of the engineering machinery, the correlation between the fault of the engineering machinery and the load strength of the engineering machinery can be considered, which also improves the accuracy of the prediction result.

[0065] Based on the content of the above embodiment, the training method of the fault prediction model includes:

[0066] Get historical indicator data;

[0067] Preprocess the historical indicator data to obtain the feature vector;

[0068] Based on the feature vector, the naive Bayes model, CatBoost model, SVM model and random forest model are trained as the base models, and the logistic regression model is used as the meta-model to obtain the fault prediction model.

[0069] In this embodiment, historical indicator data is obtained and then preprocessed to obtain a feature vector, so that a fault prediction module based on ensemble learning can use the feature vector as data input and use the ensemble learning model to learn historical dump truck data, thereby discovering the relationship between dump truck failures and dump truck operation data and transportation data.

[0070] Furthermore, after the current dump truck operation data and transport data are input into the fault prediction model as input data, the model can predict the possibility of a specific fault of the dump truck occurring in the future.

[0071] It is understandable that the model training process is a process of learning specific noise and details in the training data, and cannot learn the generalization rules of the training data. At the same time, a single model may not be able to optimize multiple indicators or meet multiple constraints at the same time. Therefore, a single model is prone to overfitting on the training data. Using a single model for fault prediction has the problems of limited generalization ability and low flexibility. For example, the decision tree model has problems such as overfitting, limited ability to handle continuous variables, and difficulty in multi-dimensional expansion. The existence of these problems is likely to affect the correctness and versatility of the fault prediction model.

[0072] Based on this, in this embodiment, by training the naive Bayes model, CatBoost (categorical boosting) model, SVM (Support Vector Machine) model and random forest model as the base model, and the logistic regression model as the meta-model integrated learning model, a fault prediction model is obtained, so that the trained fault prediction model integrates the unique advantages of different models in different dimensions, and through the characteristics of multiple models, the overfitting problem that may occur in a single model is effectively reduced, and the stability and credibility of the prediction results are enhanced. That is, by integrating the prediction capabilities of multiple models, not only the prediction errors that may occur in a single model are reduced, but also a higher prediction performance in a complex working environment is guaranteed.

[0073] In some possible embodiments, Table 1 shows a comparison of performance results of predicting specific faults of new energy dump trucks based on the same balanced data set using a single model such as a naive Bayes model, a CatBoost model, and an integrated learning model provided by an embodiment of the present invention.

[0074] Table 1 Performance comparison of each model based on a unified balanced dataset

[0075] Model Name Accuracy Accuracy Recall F1 score AUC value Naive Bayes 0.905 0.9058 0.904 0.9049 0.9187 CatBoost 0.925 0.926 0.923 0.9248 0.9295 SVM 0.907 0.9127 0.9 0.9063 0.9116 Logistic Regression 0.891 0.8902 0.892 0.8911 0.9117 Random Forest 0.9095 0.9034 0.917 0.9101 0.9161 K Nearest Neighbors 0.8865 0.8916 0.88 0.8858 0.9018 Neural Networks 0.9155 0.9109 0.921 0.9159 0.9141 Gradient Boosting Machine 0.9185 0.9147 0.923 0.9188 0.921 Ensemble Learning 0.9387 0.9385 0.9357 0.9385 0.9379

[0076] It can be seen from Table 1 that, compared with other single models, the fault prediction model provided by the embodiment of the present invention has obvious advantages in both accuracy and precision when performing fault prediction.

[0077] It is understandable that by preprocessing the collected data such as removing outliers, the data quality can be effectively improved, thereby improving the accuracy of the training model.

[0078] Based on this, in an optional embodiment, the historical indicator data is preprocessed, including:

[0079] Remove outliers from historical indicator data;

[0080] Fill missing values ​​in historical indicator data after removing outliers;

[0081] Perform fault data labeling on the historical indicator data after filling the missing values, so as to divide the historical indicator data into normal sample data and fault sample data;

[0082] Perform feature engineering on the historical indicator data after annotating the fault data to obtain the feature vector.

[0083] Specifically, by removing outliers, the original data that is obviously erroneous in the historical indicator data can be removed to ensure the accuracy of the data; by filling in missing values, the missing data in the historical indicator data caused by system failures, sensor recording errors, etc. can be supplemented to improve the integrity of the data; by marking fault data, the historical indicator data can be divided into normal sample data and fault sample data; by performing feature engineering on the historical indicator data, for example: performing simple operations such as adding two features together to obtain new features, using conversion strategies such as logarithmic or power functions to generate nonlinear features from original features, using polynomial feature generation to convert original features into polynomial features, and using one-hot features to convert discrete variables into new one-hot features, etc., effective features for fault prediction can be mined, thereby improving the performance of the fault prediction model when training the fault prediction model based on the feature vector.

[0084] In some possible embodiments, the acquired historical indicator data may be compared with the parameter range recorded in the technical manual of the engineering machinery to determine the abnormal values ​​that need to be removed from the historical indicator data.

[0085] In some possible embodiments, Figure 2 As shown, when fault data is labeled, the operation data and transportation data of the engineering machinery in a specified time period T before the failure occurs can be marked as fault sample data, while other data are normal sample data, so that the fault prediction model can detect possible faults of the engineering machinery in the specified time period T in advance.

[0086] Based on the content of the above embodiment, missing value filling is performed on the historical indicator data after removing outliers, including:

[0087] Preprocess missing values ​​on historical indicator data after removing outliers;

[0088] The missing value preprocessing includes: deleting attribute columns in the historical indicator data whose missing values ​​exceed a preset number; deleting data rows in the historical indicator data that lack preset key operating data, where the preset key operating data is operating data related to faults of engineering machinery;

[0089] Based on the data attributes of the historical indicator data after missing value preprocessing, the missing values ​​of the historical indicator data are filled.

[0090] It is understandable that missing values ​​are due to system failures, sensor recording errors, etc., which result in the information that should have been recorded not being correctly recorded by the system.

[0091] Specifically, before filling the missing values, the missing value preprocessing is performed on the historical indicator data after removing the outliers, that is, the attribute column of the historical indicator data is first analyzed, and when there are too many missing values, the column is directly deleted; then the row data is analyzed. If a row lacks key operating indicator information, such as average power consumption, voltage, current or temperature, which is highly correlated with the failure of construction machinery, this row of data is deleted, and then based on the data attributes of the historical indicator data after the missing value preprocessing, the missing values ​​of the historical indicator data are filled, thereby reducing the data processing amount of the missing value filling process and avoiding the waste of computing power.

[0092] Based on the content of the above embodiment, the missing value filling is performed on the historical indicator data based on the data attributes of the historical indicator data after the missing value preprocessing, including:

[0093] Use the mode to fill missing values ​​in historical indicator data whose data attributes are discrete variables;

[0094] Use the average value to fill missing values ​​in the historical indicator data of continuous variables with normal distribution as data attributes;

[0095] Use the median value to fill missing values ​​in historical indicator data of continuous variables with non-normal distribution data attributes.

[0096] In some optional embodiments, the specific process of filling missing values ​​of historical indicator data is as follows: Figure 3 As shown. By analyzing the historical indicator data Figure 3 The missing value filling process shown makes the data used to train the fault prediction model more accurate and complete, thereby improving the accuracy and reliability of the model.

[0097] It is understandable that before a specific fault occurs, some operation data of the engineering machinery will show obvious abnormalities, that is, some operation data of the engineering machinery has a certain correlation with the fault. Accordingly, by analyzing the correlation between the carrying data of the engineering machinery and the fault of the engineering machinery, the relationship between the carrying data and the fault can be further determined, thereby determining the operation data and carrying data with a strong correlation with the fault.

[0098] Furthermore, by performing correlation analysis on the specified faults and historical indicator data of the engineering machinery, data with strong correlation with the faults can be selected as features for fault prediction, thereby further improving the accuracy of the fault prediction model based on feature vector training.

[0099] Based on this, in an optional embodiment, feature engineering is performed on the historical indicator data after the fault data is labeled, and before the feature vector is obtained, the following is also included:

[0100] Based on the historical indicator data after the fault data is marked, the correlation analysis between the fault and operation, and the fault and transportation of the construction machinery is carried out to determine the historical indicator data whose correlation with the fault exceeds the preset value.

[0101] Specifically, taking dump trucks as an example, by analyzing the correlation between the loading and failure of dump trucks, such as examining the relationship between the failure of the dump truck's brake system and the size of a single load and the total load within a specified time period, it was found that there is a correlation between the failure of the dump truck's brake system and the load intensity of the dump truck.

[0102] In some optional embodiments, such as Figure 4 The figure shows the relationship between the maximum and minimum temperatures of the battery pack of a new energy dump truck and the fault. Figure 5 is the relationship between the total transport tonnage of new energy dump trucks and the number of brake system failures. Figure 4 It can be seen that when the temperature of the battery pack is too high or too low, it will cause the dump truck to malfunction. Figure 5 It can be seen that as the total tonnage of transportation increases, the number of failures also increases.

[0103] Therefore, by analyzing the correlation between failure and operation of engineering machinery, as well as failure and transportation based on historical indicator data, it is possible to determine indicators that are strongly correlated with failures, thereby improving the accuracy of the model. After introducing transportation data for fault prediction, the accuracy of fault prediction can be improved compared to fault prediction based solely on operation data.

[0104] It is understandable that the amount of fault sample data is much smaller than that of normal sample data. If it is not processed, the trained fault prediction model will have problems such as poor classification performance and weak generalization ability.

[0105] Based on this, in an optional embodiment, feature engineering is performed on the historical indicator data after the fault data is labeled, and before the feature vector is obtained, the following is also included:

[0106] Under-sampling normal sample data;

[0107] Oversampling is performed on the fault sample data.

[0108] In this embodiment, by under-sampling the normal operating sample data and over-sampling the fault sample data, the data volume of the normal operating sample data and the fault sample data can be made to reach the same order of magnitude, thereby facilitating the subsequent training of the fault prediction model.

[0109] In some optional embodiments, the normal operation sample data may be subjected to random under-sampling processing, and the fault sample data may be subjected to SMOTE (Synthetic Minority Over-sampling Technique) over-sampling processing.

[0110] In this embodiment, by preprocessing the historical indicator data through steps such as abnormal data processing, missing value processing, fault data labeling, correlation analysis between dump truck faults and dump truck indicator data, solving data imbalance problems and data feature engineering, the obtained feature vector for dump truck fault prediction is accurate and complete, laying a solid data foundation for subsequent fault prediction.

[0111] In summary, taking a new energy dump truck as an example, the engineering machinery fault prediction method provided by the embodiment of the present invention is adopted, and the process of fault prediction for the dump truck can be adopted as follows: Figure 6 The process shown in Figure 6 It can be seen that the engineering machinery fault prediction method provided by the embodiment of the present invention makes full use of the data advantages of new energy dump trucks, normalizes and integrates the dump truck operation data, dump truck carrying data and dump truck fault data, and analyzes the relationship between the dump truck fault and the operation and carrying data of the dump truck to reveal the intrinsic connection between various types of data and dump truck faults; after obtaining the original data related to the dump truck, the data is subjected to steps such as outlier removal, missing value filling, data analysis, fault data annotation, feature selection and unbalanced data processing, and the original dump truck data is converted into a feature vector that can be used for fault prediction, thereby laying a good data foundation for subsequent fault prediction; by using the naive Bayes model, CatBoost model, SVM model and random forest model as the base model and the logistic regression model as the meta-model as the fault prediction model, the integrated learning model integrates the unique advantages of different models in different dimensions, and through the comprehensive characteristics of multiple models, effectively reduces the overfitting problem that may occur in a single model, enhances the stability and credibility of the prediction results, and greatly improves the accuracy of dump truck fault prediction. That is, by inputting the current operating data and carrying data of the dump truck, it is possible to predict what kind of failure the dump truck will have in the future, thereby better guiding the customer's operation plan and the company's maintenance plan.

[0112] A construction machinery fault prediction system provided in an embodiment of the present invention is introduced below. The construction machinery fault prediction system described below can be considered as a module architecture for implementing a construction machinery fault prediction method provided in an embodiment of the present invention; the content described below can be cross-referenced with the above.

[0113] See also Figure 7 , Figure 7: is a structural block diagram of a shift control system provided by an embodiment of the present invention, the system may include:

[0114] The acquisition unit 10 is used to acquire the operation data and transportation data of the construction machinery;

[0115] The processing unit 20 is used to input the operation data and the transport data into the fault prediction model to obtain the fault prediction result of the engineering machinery output by the fault model;

[0116] Among them, the fault prediction model is a model trained based on the historical indicator data of construction machinery, and the historical indicator data includes: historical operation data, historical transportation data and historical fault data.

[0117] Optional training methods for fault prediction models include:

[0118] Get historical indicator data;

[0119] Preprocess the historical indicator data to obtain the feature vector;

[0120] Based on the feature vector, the naive Bayes model, CatBoost model, SVM model and random forest model are trained as the base models, and the logistic regression model is used as the meta-model to obtain the fault prediction model.

[0121] Optionally, preprocess the historical indicator data, including:

[0122] Remove outliers from historical indicator data;

[0123] Fill missing values ​​in historical indicator data after removing outliers;

[0124] Perform fault data labeling on the historical indicator data after filling the missing values, so as to divide the historical indicator data into normal sample data and fault sample data;

[0125] Perform feature engineering on the historical indicator data after annotating the fault data to obtain the feature vector.

[0126] Optionally, feature engineering is performed on the historical indicator data after the fault data is labeled, and before obtaining the feature vector, the following is also included:

[0127] Based on the historical indicator data after the fault data is marked, the correlation analysis between the fault and operation, and the fault and transportation of the construction machinery is carried out to determine the historical indicator data whose correlation with the fault exceeds the preset value.

[0128] Optionally, feature engineering is performed on the historical indicator data after the fault data is labeled, and before obtaining the feature vector, the following is also included:

[0129] Under-sampling normal sample data;

[0130] Oversampling is performed on the fault sample data.

[0131] Optionally, fill missing values ​​in the historical indicator data after removing outliers, including:

[0132] Preprocess missing values ​​on historical indicator data after removing outliers;

[0133] The missing value preprocessing includes: deleting attribute columns in the historical indicator data whose missing values ​​exceed a preset number; deleting data rows in the historical indicator data that lack preset key operating data, where the preset key operating data is operating data related to faults of engineering machinery;

[0134] Based on the data attributes of the historical indicator data after missing value preprocessing, the missing values ​​of the historical indicator data are filled.

[0135] Optionally, based on the data attributes of the historical indicator data after missing value preprocessing, missing value filling is performed on the historical indicator data, including:

[0136] Use the mode to fill missing values ​​in historical indicator data whose data attributes are discrete variables;

[0137] Use the average value to fill missing values ​​in the historical indicator data of continuous variables with normal distribution as data attributes;

[0138] Use the median value to fill missing values ​​in historical indicator data of continuous variables with non-normal distribution data attributes.

[0139] Optionally, an embodiment of the present invention further provides an engineering machinery, comprising an engineering machinery body and an engineering machinery fault prediction system as provided in any of the aforementioned embodiments.

[0140] Specifically, the engineering machinery may be a new energy dump truck or the like.

[0141] Below, reference Figure 8 To describe the electronic device provided by the embodiment of the present application, the electronic device provided by the embodiment may include: at least one processor 100, at least one communication interface 200, at least one memory 300 and at least one communication bus 400;

[0142] In the embodiment of the present invention, the number of the processor 100, the communication interface 200, the memory 300, and the communication bus 400 is at least one, and the processor 100, the communication interface 200, and the memory 300 communicate with each other through the communication bus 400; obviously, Figure 8 The communication connections shown for the processor 100, the communication interface 200, the memory 300, and the communication bus 400 are merely optional;

[0143] Optionally, the communication interface 200 may be an interface of a communication module, such as an interface of a GSM module; the processor 100 may be a central processing unit CPU, or an application specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0144] The memory 300 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0145] The processor 100 is specifically used to execute the application program in the memory to implement the steps of the above-mentioned engineering machinery fault prediction method.

[0146] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0147] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0148] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0149] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0150] It should be understood that the qualifiers "first", "second", "third", "fourth", "fifth" and "sixth" used in the description of the embodiments of the present application are only used to more clearly explain the technical solutions and cannot be used to limit the scope of protection of the present application.

[0151] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for predicting engineering machinery failures, characterized in that: include: Obtain operation data and transportation data of construction machinery; Inputting the operation data and the transport data into a fault prediction model to obtain a fault prediction result of the engineering machinery output by the fault model; The fault prediction model is a model trained based on historical indicator data of the engineering machinery, and the historical indicator data includes: historical operation data, historical transportation data and historical fault data.

2. The method according to claim 1, characterized in that The training method of the fault prediction model comprises: Obtaining the historical indicator data; Preprocessing the historical indicator data to obtain a feature vector; Based on the feature vector, an integrated learning model is trained with a naive Bayes model, a CatBoost model, a SVM model and a random forest model as base models and a logistic regression model as a meta-model to obtain the fault prediction model.

3. The method according to claim 2, characterized in that The preprocessing of the historical indicator data includes: Removing outliers from the historical indicator data; Filling missing values ​​in the historical indicator data after removing outliers; Performing fault data labeling on the historical indicator data after filling the missing values, so as to divide the historical indicator data into normal sample data and fault sample data; Feature engineering is performed on the historical indicator data after the fault data is labeled to obtain the feature vector.

4. The method according to claim 3, characterized in that The performing of feature engineering on the historical indicator data after the fault data is labeled to obtain the feature vector further includes: Based on the historical indicator data after the fault data is marked, the correlation analysis between the fault and operation, and the fault and transportation of the engineering machinery is performed to determine the historical indicator data whose correlation with the fault exceeds a preset value.

5. The method according to claim 3, characterized in that: The performing of feature engineering on the historical indicator data after the fault data is labeled to obtain the feature vector further includes: Performing undersampling processing on the normal sample data; Oversampling is performed on the fault sample data.

6. The method according to claim 3, characterized in that The filling of missing values ​​for the historical indicator data after removing outliers includes: Performing missing value preprocessing on the historical indicator data after removing outliers; The missing value preprocessing includes: deleting attribute columns in the historical indicator data whose missing values ​​exceed a preset number; deleting data rows in the historical indicator data that lack preset key operating data, wherein the preset key operating data is operating data related to the fault of the engineering machinery; Based on the data attributes of the historical indicator data after the missing value preprocessing, the missing values ​​of the historical indicator data are filled.

7. The method according to claim 6, characterized in that The filling of missing values ​​in the historical indicator data based on the data attributes of the historical indicator data after the missing value preprocessing includes: Using the mode, filling missing values ​​of the historical indicator data whose data attributes are discrete variables; Using the average value, the missing values ​​of the historical indicator data of the continuous variable whose data attribute is a normal distribution are filled; The missing values ​​of the historical indicator data of the continuous variable whose data attribute is non-normal distribution are filled by using the median value.

8. A construction machinery fault prediction system, characterized in that: include: A collection unit, used to obtain operation data and transportation data of the construction machinery; A processing unit, configured to input the operation data and the transport data into a fault prediction model to obtain a fault prediction result of the engineering machinery output by the fault model; The fault prediction model is a model trained based on historical indicator data of the engineering machinery, and the historical indicator data includes: historical operation data, historical transportation data and historical fault data.

9. An engineering machine, characterized in that: It comprises an engineering machinery body and the engineering machinery fault prediction system as claimed in claim 8.

10. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing instructions executable by the processor; The processor is used to execute the method according to any one of claims 1 to 7.

Images