CN115424106A

CN115424106A - Trolley detection method and device, electronic equipment and storage medium

Info

Publication number: CN115424106A
Application number: CN202211086102.0A
Authority: CN
Inventors: 陈鹏飞; 传江; 林刚; 余萍
Original assignee: Midea Group Co Ltd
Current assignee: Midea Group Co Ltd
Priority date: 2022-09-06
Filing date: 2022-09-06
Publication date: 2022-12-02

Abstract

The invention relates to the technical field of manufacturing industry, and provides a trolley detection method, a trolley detection device, electronic equipment and a storage medium, wherein the trolley detection method comprises the following steps: acquiring original characteristic information of a trolley to be detected, wherein the original characteristic information comprises parameter information collected by the trolley to be detected in a process of detecting a target product; determining target characteristic information based on the original characteristic information, wherein the target characteristic information comprises the residual characteristic information after the original characteristic information is subjected to exception processing; and obtaining a target detection result aiming at the trolley to be detected based on the target characteristic information and a preset detection model. The method can realize the purpose of timely and intelligently detecting whether the trolley has abnormity, saves time and labor, and is efficient and accurate.

Description

Trolley detection method and device, electronic equipment and storage medium

Technical Field

The invention relates to the technical field of manufacturing industry, in particular to a trolley detection method, a trolley detection device, electronic equipment and a storage medium.

Background

In the production process of air conditioner, assembly line terminal has the operation room, is provided with the platform truck of whether inspection air conditioner has the problem in the operation room to the platform truck lasts work in the air conditioner production process, but receives influences such as environment, equipment loss, and the great condition of deviation appears easily in the platform truck inspection air conditioner. Therefore, it is important to detect whether the trolley is abnormal.

In the related art, a detection method for a trolley generally includes that a worker installs 1 normal air conditioner on a plurality of trolleys respectively, manually collects parameters such as power and pressure during the process of checking the air conditioner by each trolley, calculates average values of corresponding parameters respectively, and finally determines the trolley with a larger deviation degree of the average values from a preset average value as an abnormal trolley.

However, when a worker inspects the trolley on site, the air conditioner needs to be installed manually, data is acquired manually in the operation process after installation, and the air conditioner needs to be removed manually after operation is finished, so that the whole detection process not only depends on human factors excessively, but also wastes time and labor, and the efficiency of detecting the trolley is low and the accuracy is not high.

Disclosure of Invention

The present invention has been made to solve at least one of the problems occurring in the related art. Therefore, the trolley detection method provided by the invention can be used for timely and intelligently detecting whether the trolley is abnormal or not in a mode of automatically acquiring information, performing exception handling and detecting a model, and is time-saving, labor-saving, efficient and accurate.

The invention further provides a trolley detection device.

The invention further provides the electronic equipment.

The invention also proposes a non-transitory computer-readable storage medium.

The invention also proposes a computer program product.

According to the trolley detection method provided by the embodiment of the first aspect of the invention, the trolley detection method comprises the following steps:

acquiring original characteristic information of a trolley to be detected, wherein the original characteristic information comprises parameter information collected by the trolley to be detected in a process of detecting a target product;

determining target characteristic information based on the original characteristic information, wherein the target characteristic information comprises the characteristic information left after the original characteristic information is subjected to exception processing;

and obtaining a target detection result aiming at the trolley to be detected based on the target characteristic information and a preset detection model.

According to the trolley detection method, the original characteristic information of the trolley to be detected is obtained, the original characteristic information comprises parameter information collected by the trolley to be detected in the process of detecting a target product, the target characteristic information is determined from the original characteristic information in a mode of carrying out exception processing on the original characteristic information, and then a target detection result for the trolley to be detected is obtained further based on the target characteristic information and a preset detection model; whether the unusual purpose of timely and intellectual detection system platform truck exists is realized through the mode of automatic acquisition information, exception handling and model detection with this, and not only labour saving and time saving, it is high-efficient accurate moreover, combine model study and detection technology can further ensure the accuracy of testing result simultaneously to the accuracy nature and the reliability that the platform truck detected have also effectively been improved.

According to an embodiment of the present invention, the determining target feature information based on the original feature information includes:

performing row-column-row operation on the column data of the column field where the parameter information is located in the original characteristic information, and determining the characteristic information to be processed of the trolley to be detected;

performing exception handling on the feature information to be processed based on a preset exception handling rule;

determining the target characteristic information based on the characteristic information obtained by exception handling;

the preset exception handling rule comprises the steps of rejecting line data with a line field missing rate exceeding a line missing rate threshold value, filling row data with a row field missing rate lower than a row missing rate threshold value, rejecting line data which does not meet a preset distribution rule, rejecting line data which does not meet a preset temperature correlation, rejecting line data which does not meet a preset temperature range and rejecting line data which does not meet a preset power correlation number aiming at the characteristic information to be processed.

According to an embodiment of the present invention, the performing exception handling on the feature information to be processed further includes:

acquiring the target quantity of target products belonging to the same preset equipment part code, wherein the preset equipment part code is the hardware parameter model of the same target product;

and determining that the target number is smaller than a preset number threshold, and eliminating the line data of the preset equipment part codes corresponding to the target number in the characteristic information to be processed.

According to an embodiment of the present invention, the determining the target feature information based on the feature information obtained by exception handling includes:

performing characteristic amplification processing on the characteristic information obtained by the abnormal processing, and determining a plurality of pieces of characteristic information after the amplification processing;

and screening the plurality of characteristic information based on the correlation of the preset characteristic information and the quantity of the preset characteristic information to determine the target characteristic information.

According to an embodiment of the present invention, the preset detection model includes different preset detection submodels, and the obtaining of the target detection result for the to-be-detected trolley based on the target feature information and the preset detection model includes:

inputting the target characteristic information into the preset detection model to obtain a preset number of detection results of the trolley to be detected, which are output by the preset detection model, under different preset equipment part codes and different preset detection submodels;

and obtaining a target detection result aiming at the trolley to be detected based on the abnormal detection result in the preset number of detection results.

According to an embodiment of the present invention, the obtaining a target detection result for the to-be-detected trolley based on an abnormal detection result in the preset number of detection results includes:

determining the proportion of abnormal detection results in the preset number of detection results;

determining that the occupation ratio exceeds a first preset occupation ratio, and acquiring a target detection result that the trolley to be detected is an abnormal trolley;

determining that the occupation ratio is between a second preset occupation ratio and the first preset occupation ratio, and acquiring a target detection result of the to-be-detected trolley as a risk trolley;

and determining that the occupation ratio is lower than the second preset occupation ratio, and acquiring a target detection result that the trolley to be detected is a normal trolley.

According to an embodiment of the present invention, after the obtaining of the target detection result that the to-be-detected trolley is a risk trolley, the method further includes:

acquiring target coding information of the risk trolley;

and sending early warning information to the user terminal based on the target coding information, wherein the early warning information is used for reminding the user terminal of corresponding rechecking personnel to carry out on-site evaluation on the risk trolley.

According to one embodiment of the invention, the training method of the preset detection model comprises the following steps:

acquiring a plurality of different initial detection submodels for carrying out abnormity detection on the trolley to be detected;

training each initial detection submodel for a preset number of times based on the target characteristic information, and determining a plurality of intermediate detection results of a plurality of intermediate detection submodels after each training;

sending the plurality of intermediate detection results to a user terminal, and receiving a rechecking result fed back by the user terminal and aiming at the plurality of intermediate detection results;

and determining the preset detection submodel and the preset detection model corresponding to the preset detection submodel based on the rechecking result and the intermediate detection submodel after each model parameter is updated.

According to a second aspect of the present invention, a dolly detection apparatus comprises:

the system comprises an acquisition module, a detection module and a processing module, wherein the acquisition module is used for acquiring original characteristic information of a trolley to be detected, and the original characteristic information comprises parameter information collected by the trolley to be detected in a process of detecting a target product;

the determining module is used for determining target characteristic information based on the original characteristic information, wherein the target characteristic information comprises the residual characteristic information after the original characteristic information is subjected to exception processing;

and the detection module is used for obtaining a target detection result aiming at the trolley to be detected based on the target characteristic information and a preset detection model.

According to the trolley detection device provided by the embodiment of the invention, the original characteristic information of the trolley to be detected is obtained firstly, and the original characteristic information comprises the parameter information collected by the trolley to be detected in the process of detecting a target product, so that the target characteristic information is determined from the original characteristic information in a mode of carrying out exception processing on the original characteristic information, and then a target detection result aiming at the trolley to be detected is obtained further based on the target characteristic information and a preset detection model; with this through the mode of automatic acquisition information, exception handling and model detection, realize in time and whether the intellectual detection system platform truck has unusual purpose, labour saving and time saving not only, it is high-efficient accurate moreover, combine model study and detection technique can further ensure the accuracy of testing result simultaneously to the accuracy nature and the reliability that the platform truck detected have also effectively been improved.

One or more technical solutions in the embodiments of the present invention have at least one of the following technical effects: the method comprises the steps that firstly, original characteristic information of a trolley to be detected is obtained, and the original characteristic information comprises parameter information collected by the trolley to be detected in the process of detecting a target product, so that the target characteristic information is determined from the original characteristic information in a mode of carrying out exception processing on the original characteristic information, and then a target detection result aiming at the trolley to be detected is obtained further on the basis of the target characteristic information and a preset detection model; with this through the mode of automatic acquisition information, exception handling and model detection, realize in time and whether the intellectual detection system platform truck has unusual purpose, labour saving and time saving not only, it is high-efficient accurate moreover, combine model study and detection technique can further ensure the accuracy of testing result simultaneously to the accuracy nature and the reliability that the platform truck detected have also effectively been improved.

Further, firstly, determining the characteristic information to be processed of the trolley to be detected in a mode of performing row-column-row operation on the column data of the column field where the parameter characteristic in the original characteristic information is located; performing exception processing on the feature information to be processed based on a preset exception processing rule, and determining target feature information based on the feature information obtained by exception processing; by combining the column transfer and abnormal data processing technology, the accuracy and pertinence of abnormal data processing are improved, and a foundation is laid for the accuracy of subsequent trolley abnormality detection.

Furthermore, the data of the row where the relevant preset equipment part codes are located is removed from the data to be processed in a mode of determining that the target quantity of the target products of the same preset equipment part code in the characteristic information to be processed is smaller than a preset quantity threshold value, so that the flexibility and the comprehensiveness of processing abnormal data are improved.

Furthermore, the target characteristic information meeting the requirements of the correlation of the preset characteristic information and the quantity of the preset characteristic information is determined by performing characteristic amplification processing on the characteristic information obtained by abnormal processing and then performing correlation analysis, so that the content richness of the target characteristic information is improved, and a sufficient basis is provided for subsequent model learning.

And further, determining that the trolley to be detected is an abnormal trolley, a normal trolley or a risk trolley by judging the relation between the occupation ratio of the abnormal detection results in the preset number of detection results and the first preset occupation ratio and the second preset occupation ratio, so that the high efficiency and the accuracy of the abnormal detection are effectively improved by combining a comprehensive voting system mode.

Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the related arts, the drawings used in the description of the embodiments or the related arts will be briefly introduced below, it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.

Fig. 1 is a schematic flow chart diagram of a trolley detection method provided in an embodiment of the present invention;

FIG. 2 is a schematic overall flow chart of a method for detecting a trolley according to an embodiment of the present invention;

fig. 3 is a schematic structural diagram of a trolley detection device provided in an embodiment of the present invention;

fig. 4 is a schematic physical structure diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

In the production process of air conditioner, assembly line terminal has the operation room, is provided with the platform truck of whether inspection air conditioner has the problem in the operation room to the platform truck lasts work in the air conditioner production process, but receives influences such as environment, equipment loss, and the great condition of deviation appears easily in the platform truck inspection air conditioner. In the existing trolley detection method, when a plurality of reports of a trolley are detected within a period of time, a worker passively detects whether the trolley is abnormal or not; and the worker's verification method is, install 1 air conditioner that is verified to be normal on about 10 dollies respectively, measure parameter values such as power, pressure, calculate the mean value of the corresponding parameter value, and the deviation ratio of each dollies, and then judge whether each dollies need to stop using, maintain.

Obviously, the existing trolley detection method finds the problem of the trolley untimely, and when the problem is found, the problem of the trolley is serious; moreover, whether the trolley has problems needs to be checked and verified in a manual field, namely, the procedures of checking one trolley, installing and operating household appliances such as an air conditioner and dismantling the household appliances need to be performed, a complete procedure needs about 7 minutes, more than 1 hour is needed for detecting one trolley with problems, time and labor are wasted, and the efficiency is low.

Based on this, the invention provides a trolley detection method, a trolley detection device, an electronic device and a storage medium, wherein an execution main body of the trolley detection method is a terminal device, and the terminal device can be other electronic devices such as a Personal Computer (PC), a portable device, a notebook Computer, a smart phone, a tablet Computer and a portable wearable device. The invention is not limited to the specific form of the terminal device. It is understood that the execution subject of the trolley detection method may also be a server. The following describes a trolley detection method, a trolley detection device, an electronic device and a storage medium provided by the invention with reference to fig. 1 to 4, and the following method embodiment takes an execution subject as a terminal device as an example.

Fig. 1 is a schematic flow chart of a trolley detection method provided by the present invention, and as shown in fig. 1, the trolley detection method includes the following steps:

and 110, acquiring original characteristic information of the trolley to be detected, wherein the original characteristic information comprises parameter information collected by the trolley to be detected in the process of detecting a target product.

The trolley to be detected can be a trolley for inspecting target products produced on a production line, and the number of the trolleys to be detected can be 1 or more. It is not particularly limited herein; the target product can be an air conditioner, and the number of the target products can be 1 or more. And is not particularly limited herein. In addition, the original characteristic information exists in the form of row fields in the table, the row fields may include, but are not limited to, temperature, current, voltage, power, and the like, and the column fields corresponding to the row fields may include, but are not limited to, an identification number (id) of the target product, target code information of the trolley to be detected, and a preset equipment code, which is a hardware parameter model of the same target product.

Specifically, the original characteristic information of the trolley to be detected is obtained by the terminal device from a Manufacturing Execution system (MES system), and the MES system is used for recording state information of power, pressure, current, temperature and the like collected in the process of running the target product to be detected by the trolley to be detected in the house and accumulating and summarizing the state information into a field in a form of a table; when the fields read from the MES system need to be processed, the fields can be cleaned, so that the original characteristic information of the trolley to be detected is obtained. For example, the field may include 4 lines of data including a preset equipment part code, an id of a target product, target code information and state information of the trolley to be detected, and the state information includes two columns of values of a detection item and a corresponding detection item, where the two columns of values for the detection item and the corresponding detection item may include, but are not limited to, 20 lines of data, and the 20 lines of data may include a three-phase current, a three-phase voltage, a temperature of the air inlet, a temperature of the air outlet, cooling power and cooling air pressure of each step of cooling, and a three-phase current, a three-phase voltage, a temperature of the air inlet, a temperature of the air outlet, heating power and cooling air pressure of each step of heating; the field is analyzed, and the data of the field is processed into a preset equipment part code, an id of a target product, and temperature, power, air pressure, voltage, current and the like corresponding to target code information of the trolley to be detected in a certain refrigerating step link or a certain heating step link, so that the original characteristic information of the trolley to be detected is determined.

It can be understood that the trolley to be detected can be 1, a plurality of or all trolleys in the running room; the operation room can be a place where the trolley to be detected inspects the target product and is a closed room.

And step 120, determining target characteristic information based on the original characteristic information, wherein the target characteristic information comprises the characteristic information remained after the original characteristic information is subjected to exception processing.

Specifically, the terminal device analyzes which data in the original characteristic information can be used and which data cannot be used, for example, analyzes usable data and abnormal data in the data acquired by links such as a refrigeration step 1, a refrigeration step 2, a refrigeration step 3, a heating step 4, a heating step 5 and a heating step 6, wherein the key point of analyzing data abnormality is to determine abnormal data caused by data acquisition link errors in the original characteristic information, based on the abnormal data screening of the data acquisition link, a corresponding reference value range is preset and a mapping relation is established for the three-phase current, the three-phase voltage, the temperature of an air inlet, the temperature of an air outlet and the air pressure of the air outlet acquired by each step link, and then determines data which is not matched with the mapping relation in the original characteristic information as abnormal data based on the mapping relation, that is, the abnormal data can include but is not limited to data and/or column data which are not matched with the mapping relation in the original characteristic information; and then performing exception processing on the exception data determined in the original characteristic information so as to determine target characteristic information.

And step 130, obtaining a target detection result for the trolley to be detected based on the target characteristic information and a preset detection model.

Specifically, considering that the influence of abnormal data on subsequent model processing is large, and in order to improve the detection accuracy for the trolley to be detected, it may be determined that the preset detection model includes at least two abnormal detection algorithms, and each abnormal detection algorithm may determine whether the trolley to be detected is abnormal based on the target characteristic information, so that a plurality of detection results may be obtained, and the plurality of detection results are summarized for analysis, so as to determine the target detection result. For example, if the number of the detection results judged to be abnormal is far greater than the number of the detection results judged to be normal in the plurality of detection results, the trolley to be detected can be determined to be an abnormal trolley; on the contrary, if the number of the detection results judged to be abnormal in the plurality of detection results is far smaller than the number of the detection results judged to be normal, the trolley to be detected can be determined to be a normal trolley.

The trolley detection method provided by the embodiment of the invention comprises the steps of firstly obtaining original characteristic information of a trolley to be detected, determining target characteristic information from the original characteristic information in a mode of carrying out exception processing on the original characteristic information because the original characteristic information comprises parameter information collected by the trolley to be detected in a process of detecting a target product, and further obtaining a target detection result aiming at the trolley to be detected based on the target characteristic information and a preset detection model; whether the unusual purpose of timely and intellectual detection system platform truck exists is realized through the mode of automatic acquisition information, exception handling and model detection with this, and not only labour saving and time saving, it is high-efficient accurate moreover, combine model study and detection technology can further ensure the accuracy of testing result simultaneously to the accuracy nature and the reliability that the platform truck detected have also effectively been improved.

It can be understood that, in consideration of inevitable abnormal data which is not beneficial to subsequent detection when the MES system collects data, for example, when the air conditioner runs a plurality of steps such as cooling or heating test, the heating test and the cooling test are usually different in frequency, so that the temperature distribution in the running room is uneven, and the accuracy of the collected data is affected, an abnormal processing rule can be preset to perform abnormal processing on the original characteristic information. Based on this, the specific implementation process of step 120 may include:

firstly, performing row-column conversion operation on row data of a row field where parameter information in original characteristic information is located, and determining characteristic information to be processed of a trolley to be detected; performing exception processing on the characteristic information to be processed based on a preset exception processing rule; then, based on the feature information obtained by the exception processing, target feature information is determined.

The preset exception handling rule comprises the steps of rejecting line data with a line field loss rate exceeding a line loss rate threshold, filling row data with a row field loss rate lower than a row loss rate threshold, rejecting line data which does not meet a preset distribution rule, rejecting line data which does not meet a preset temperature correlation, rejecting line data which does not meet a preset temperature range and rejecting line data which does not meet a preset power correlation value aiming at feature information to be processed.

Specifically, since the original characteristic information includes 5 lines of data including preset equipment code, id of the target product, target code information of the to-be-detected trolley, detection item and numerical value of the corresponding detection item, and 20 lines of data of the two lines of data after the detection item and the numerical value of the corresponding detection item, the two lines of data after the 5 lines of data can be subjected to column-to-row operation, that is, the two lines of data are spread into one line, so as to obtain a result obtained after the column-to-row operation, and determine the result obtained after the column-to-row operation as characteristic information to be processed of the to-be-detected trolley, the characteristic information to be processed exists in a form of the line of data in a table, and under three dimensions of the preset equipment code, id of the target product, and the target code information of the to-be-detected trolley, three-phase current, three-phase voltage, temperature of the air inlet, temperature of the air outlet, cooling power and cooling air pressure of each cooling step, and three-phase current, three-phase voltage, temperature of the air inlet, temperature of the line of the air outlet, heating power, and air pressure, and the like of each heating step are multiple pieces of characteristic information, namely, one piece of data; the numerical values of all the parameters collected when different trolleys detect different target products are visually displayed.

At this time, based on a preset exception handling rule, for the feature information to be processed, removing the line data of which the column field loss rate exceeds a column loss rate threshold, for example, removing the line data of which the column loss rate exceeds 80%; filling the row data with the row field missing rate lower than the row missing rate threshold, for example, 5 row data in 20 row data corresponding to a certain preset equipment part code are empty and 15 row data are not empty, and at this time, filling the corresponding column field of the 5 row data with the average value of each column in the 15 row data; removing line data which do not meet a preset distribution rule, for example, respectively making violin graphs based on the collected three-phase current, three-phase voltage, temperature, power and air pressure, selecting upper and lower boundaries of a region in which data are intensively distributed in the violin graphs as upper and lower boundaries of a screening value, and removing data distributed outside the upper and lower boundaries of the screening value; removing column data which do not meet the preset temperature correlation, for example, based on the fact that the temperature of an air inlet is higher than that of an air outlet when an air conditioner is used for refrigerating and the temperature of the air inlet is lower than that of the air outlet when the air conditioner is used for heating, determining a covariance matrix according to the temperature of the air inlet and the temperature of the air outlet acquired in the refrigerating and heating step links, calculating a temperature difference correlation coefficient of the air inlet and the air outlet based on the covariance matrix, and if the difference value between the current C of one of three-phase currents in the refrigerating link and the temperature difference correlation coefficient of the air inlet and the air outlet corresponding to the refrigerating link is small enough, removing column data corresponding to the current C; removing column data which do not meet a preset temperature range, for example, taking the temperature of the air conditioner in the last step of refrigeration as a stable refrigeration temperature, determining a plurality of stable refrigeration temperatures corresponding to a plurality of air conditioners in the running room, then taking the average value of the stable refrigeration temperatures as a refrigeration temperature average value, similarly, taking the temperature of the air conditioner in the last step of heating as a stable heating temperature, determining a plurality of stable heating temperatures corresponding to a plurality of air conditioners in the running room, then taking the average value of the stable heating temperatures as a heating temperature average value, and then removing column data of which the temperature value is outside the range of plus or minus 16% of the heating average temperature or the cooling average temperature of the air conditioner coded by the same preset equipment; and (3) eliminating line data which do not meet the preset power correlation, for example, calculating the ratio of power to temperature for a line of data to be screened, and if the ratio is out of the range of plus or minus 8% of the preset average ratio of the air conditioner with the same preset equipment part code, eliminating the line data of the power corresponding to the ratio.

At this time, it is determined that the number of the feature information satisfies the preset feature quantity requirement and the correlation requirement with respect to the feature information obtained by the exception handling, and at this time, the feature information obtained by the exception handling may be determined as the target feature information, that is, the target feature information may be a plurality of features satisfying the preset feature quantity requirement and the correlation requirement under the preset equipment part code, the identification number id of the target product, and the target code information of the trolley to be detected, for example, about 30 features may be included.

It can be understood that, for a plurality of air conditioners in an operating room, in addition to the fact that the temperature distribution in the operating room affects the information of each parameter of the air conditioners in the detection process, the following conditions can also cause temperature changes: when the shape of the trolley to be detected is changed from a metal plate to a large machine or from a small machine to a large machine, the heating power and pressure are obviously increased; when the air conditioner is unstable in production, the temperature can be reduced; aiming at the condition that the power, the pressure and the temperature change when the trolley is powered on and started after power failure; when all trolleys in the running room are in a refrigerating state for 15 minutes, the temperature in the running room is increased; therefore, it is necessary to perform temperature compensation on the original feature information, that is, to remove the line data that does not satisfy the preset temperature range and to remove the line data that does not satisfy the preset power correlation.

The trolley detection method provided by the embodiment of the invention comprises the steps of firstly determining the characteristic information to be processed of the trolley to be detected in a row-column-row-column operation mode of the column data of the column field where the parameter characteristic is located in the original characteristic information; performing exception processing on the feature information to be processed based on a preset exception processing rule, and finally determining target feature information based on the feature information obtained by exception processing; by combining the column transfer and abnormal data processing technology, the accuracy and pertinence of abnormal data processing are improved, and a foundation is laid for the accuracy of subsequent trolley abnormality detection.

It can be understood that, when the number of the trolleys in the operating room is less than the number of the produced air conditioners, a situation that a small number of air conditioners are insufficient to cover a plurality of trolleys, so that a problem is detected, at this time, the row data where the codes of the preset equipment parts corresponding to the small number of air conditioners are produced can be deleted. Based on this, the exception handling is carried out on the characteristic information to be processed, and the method further comprises the following steps:

firstly, acquiring the target quantity of target products belonging to the same preset equipment part code; and further determining that the target quantity is smaller than a preset quantity threshold value, and eliminating the data of the data where the preset equipment part codes corresponding to the target quantity in the characteristic information to be processed are located.

Specifically, in order to ensure accuracy and fairness of trolley detection, the target quantity of target products belonging to the same preset equipment code can be acquired based on the characteristic information to be processed, whether the quantity of the trolleys to be detected in the running room is enough to cover the target quantity or not is judged, 3 times of the quantity of the trolleys of the production line of the air conditioners with the target quantity can be set at the target quantity to be produced and used as the minimum quantity of the air conditioners with the preset equipment code to be judged, and if the quantity of the air conditioners to be detected comprises 20 trolleys and belongs to the same preset equipment code and is smaller than 60, the row data of the preset equipment code in the characteristic information to be processed is eliminated.

According to the trolley detection method provided by the embodiment of the invention, the data of the row where the relevant preset equipment part codes are located is removed from the data to be processed in a mode of determining that the target number of the target products of the same preset equipment part code in the characteristic information to be processed is smaller than the threshold value of the preset number, so that the flexibility and the comprehensiveness of processing abnormal data are improved.

It can be understood that when the number and/or the feature correlation of the feature information obtained by performing the abnormal data processing on the original feature information do not meet the preset requirement, the target feature information may be determined by performing the first feature amplification and the second feature correlation analysis. Based on the above, the target feature information is determined based on the feature information obtained by exception handling, and the implementation process may include:

firstly, performing characteristic amplification treatment on characteristic information obtained by exception treatment, and determining a plurality of pieces of characteristic information after the amplification treatment; and further screening the plurality of characteristic information based on the correlation of the preset characteristic information and the quantity of the preset characteristic information to determine the target characteristic information.

Specifically, when it is determined that the number of the feature information does not satisfy the preset feature information number and/or the preset feature information correlation with respect to the feature information obtained by the exception handling, the feature information obtained by the exception handling may be subjected to feature amplification processing, for example, a plurality of feature information may be determined by calculating a mean value, a median, a range, a standard deviation, and the like with respect to the refrigeration power of each refrigeration step, and correlation analysis may be performed step by step with respect to the plurality of features, where feature information with low correlation may be screened out from the plurality of feature information by determining a covariance matrix, and sorted by the size of a correlation coefficient, and finally, feature information satisfying the preset feature correlation and the preset feature information number with respect to a preset equipment code, an identification number id of a target product, and target code information of a to-be-detected dolly, is determined, that is, target feature information is determined.

According to the trolley detection method provided by the embodiment of the invention, the target characteristic information meeting the requirements of the correlation of the preset characteristic information and the quantity of the preset characteristic information is determined by performing characteristic amplification processing on the characteristic information obtained by abnormal processing and then performing correlation analysis, so that the content richness of the target characteristic information is improved, and a sufficient basis is provided for subsequent model learning.

It can be understood that, in order to improve the accuracy and efficiency of the detection of the trolley, the detection can be performed on the trolley to be detected for detecting the air conditioners belonging to the same preset equipment part code, and the final detection result can be determined by using different detection results respectively detected by a plurality of algorithms. Based on this, in the case that the preset detection model includes different preset detection submodels, the specific implementation process of step 130 may include:

firstly, inputting target characteristic information into a preset detection model to obtain a preset number of detection results of a trolley to be detected, which are output by the preset detection model, under different preset equipment part codes and different preset detection submodels; and further obtaining a target detection result aiming at the trolley to be detected based on abnormal detection results in the preset number of detection results.

Specifically, for the condition that the preset detection model comprises different preset detection submodels, target characteristic information can be input into the preset detection model and is detected through the different preset detection submodels, namely, each preset detection submodel is used for judging whether the trolley to be detected is abnormal or not according to the corresponding data such as power, current, voltage, temperature of an air inlet, temperature of an inlet and an outlet and the like aiming at the air conditioner belonging to the same preset equipment piece code, so that the detection result of whether the trolley to be detected of the air conditioner which is detected by each preset detection submodel and is coded by each preset equipment piece is abnormal or not is obtained, the detection results of the preset number are determined, and the preset number is determined according to the total number of the preset equipment piece codes and the total number of the preset detection submodels; and then, acquiring a target detection result aiming at the trolley to be detected based on abnormal detection results in the preset number of detection results, so as to determine whether the trolley to be detected is in a normal state, an abnormal state or a risk state currently.

It is understood that each preset detection submodel may contain one anomaly detection algorithm, and each anomaly detection algorithm may be determined by training a large number of existing anomaly detection algorithms.

The trolley detection method provided by the embodiment of the invention obtains a preset number of detection results of the trolley to be detected under different preset equipment part codes and different preset detection submodels, and obtains a target detection result aiming at the trolley to be detected based on an abnormal detection result in the preset number of detection results. Therefore, the accuracy and reliability of the abnormal detection of the trolley are greatly improved by the mode of acquiring the detection result through different preset equipment part codes and different abnormal detection algorithms.

It can be understood that, considering that the comprehensive voting mechanism of the model can improve the stability and fairness of the final detection result, the target detection result can be obtained by setting the comprehensive voting mechanism. Based on this, based on the unusual testing result in the number of testing results of predetermineeing, obtain the target test result to waiting to detect the platform truck, its implementation process includes:

firstly, determining the proportion of abnormal detection results in a preset number of detection results; further determining that the occupation ratio exceeds a first preset occupation ratio, and acquiring a target detection result that the trolley to be detected is an abnormal trolley; determining that the occupation ratio is between the second preset occupation ratio and the first preset occupation ratio, and acquiring a target detection result of the to-be-detected trolley as a risk trolley; and determining that the occupation ratio is lower than a second preset occupation ratio, and acquiring a target detection result that the trolley to be detected is a normal trolley.

Specifically, each preset detection submodel may respectively determine whether the to-be-detected trolley of the air conditioner, for which each preset equipment part code is checked, is abnormal, that is, the corresponding determined detection result is a voting result for abnormality, risk and normality, so that it may be considered that the to-be-detected trolley obtains a K ticket, and K may be a product of the total number of the preset equipment part codes and the total number of the preset detection submodel. At the moment, determining the proportion of abnormal detection results from the preset number of detection results obtained from all preset detection submodels, namely the proportion of the abnormal detection results in the K ticket, wherein the proportion reflects the abnormal ticket rate of the trolley to be detected under all preset equipment part codes and all preset detection submodels, and if the proportion exceeds the first preset proportion, determining that the trolley to be detected is an abnormal trolley and needs to be stopped and maintained; if the occupation ratio is between the second preset occupation ratio and the first preset occupation ratio, determining that the trolley to be detected is a risk trolley and needs attention; and if the occupation ratio is lower than the second preset occupation ratio, determining that the trolley to be detected is a normal trolley and can be normally used subsequently. Further, the first preset proportion may be 20%, the second preset proportion may be 10%, if 20 preset equipment part codes exist, and 16 preset detection submodels exist, the trolley to be detected may obtain 320 tickets, wherein the abnormal ticket yield exceeds 320 × 20%, that is, when the abnormal ticket yield exceeds 64 tickets, it may be determined that the trolley to be detected is an abnormal trolley.

According to the trolley detection method provided by the embodiment of the invention, the trolley to be detected is determined to be an abnormal trolley, a normal trolley or a risk trolley by judging the relation between the occupation ratio of the abnormal detection results in the preset number of detection results and the first preset occupation ratio and the second preset occupation ratio, so that the high efficiency and the accuracy of the abnormal detection are effectively improved by combining a comprehensive voting system mode.

It can be understood that, for the risk trolley with the abnormal hidden danger, whether the risk trolley is abnormal or not can be evaluated again by combining the rechecking result of the field personnel. Based on this, after obtaining the target detection result that the trolley to be detected is the risk trolley, the trolley detection method provided by the embodiment of the invention further includes:

firstly, acquiring target coding information of a risk trolley; and further sending early warning information to the user terminal based on the target coding information of the risk trolley, wherein the early warning information is used for reminding the user terminal of corresponding rechecking personnel to carry out on-site evaluation on the risk trolley.

Specifically, for obtaining a target detection result that the trolley to be detected is a risk trolley, target coding information of the risk trolley can be obtained again, and the target coding information is used for positioning the risk trolley in the running room; sending early warning information to the user terminal based on the target coding information of the risk trolley to remind a rechecker corresponding to the target terminal to carry out site evaluation on the risk trolley in time, and determining to stop using and maintain the risk trolley if the site evaluation result is unavailable; and if the field evaluation result is available, determining that the risk trolley can be continuously used for subsequent air conditioner inspection.

According to the trolley detection method provided by the embodiment of the invention, the risk evaluation mode of the risk trolley is further combined with the rechecker, and whether the risk trolley is continuously used for subsequent air conditioner inspection or is stopped for maintenance is determined, so that the flexibility and the reliability of trolley abnormity detection are further improved.

It can be understood that, in order to improve the precision and accuracy of the preset detection model, the preset detection submodel may be determined in a manner of training a large number of anomaly detection algorithms through the target feature information. Based on this, the training method of the preset detection model comprises the following steps:

firstly, acquiring a plurality of different initial detection submodels for carrying out abnormity detection on a trolley to be detected; training each initial detection submodel for a preset number of times based on the target characteristic information, and determining a plurality of intermediate detection results of a plurality of intermediate detection submodels after each training; then, sending the plurality of intermediate detection results to the user terminal, and receiving a rechecking result fed back by the user terminal and aiming at the plurality of intermediate detection results; and finally, determining a preset detection submodel and the preset detection model corresponding to the preset detection submodel based on the rechecking result and the updated intermediate detection submodel of each model parameter.

Specifically, the obtaining of the Multiple different initial Detection sub-models for performing the anomaly Detection on the trolley to be detected may be to construct corresponding initial Detection sub-models Based on Multiple existing anomaly Detection algorithms for the anomaly Detection, where the Multiple existing anomaly Detection algorithms for the anomaly Detection may include, but are not limited to, a statistical 3 se seg code anomaly Detection algorithm, a Local anomaly Factor (LOF) algorithm, a k nearest neighbor (kNN) algorithm, an average k nearest neighbor (AvgkNN) algorithm, a Clustering-Based Local Outlier Factor (CBLOF) algorithm Based on a Local Outlier Factor, a Class-Based Support Vector machine (One-Class dominant vectors, OCSVM) algorithm, a Local correlation integral-Based fast Outlier Detection (LOCI) algorithm, a primary component analysis (primary components analysis, PCA) algorithm, a Minimum Covariance Determinant (MCD) algorithm, a Feature banding algorithm, an Angle-Based Outlier Detection (ABOD) algorithm, an isolated Forest (Isolation Forest) algorithm, a Histogram-Based Outlier Score (HBOS) algorithm, a random Outlier Detection (SOS) algorithm, a fully-connected auto-coding (AutoEncoder) algorithm, an AOM (AOM) algorithm, a mean Maximization (MOA) algorithm, a Single-target-generated-resistant Learning (SO-GAAL) algorithm, a multi-target-generated-resistant-Active Learning (Multiple-target-Active additive Learning), MO-GAAL) algorithm, an extremum boosting-based Outlier detection (XGBOD) algorithm, a Local Selective Combination (LSCP) algorithm of a Parallel set of outliers, and the like. And is not particularly limited herein.

The method comprises the following steps of further training each initial detection submodel for preset times based on target characteristic information, determining a plurality of intermediate detection results of the plurality of intermediate detection submodels after each training, sending the intermediate detection results obtained by each intermediate detection submodel to a user terminal, enabling the user terminal to recheck the intermediate detection results corresponding to rechecking personnel, and receiving the rechecking results fed back by the user terminal and aiming at the intermediate detection results, wherein the rechecking aims to determine the accuracy and the stability of each abnormal detection algorithm, and the model parameters of each intermediate detection submodel can be automatically updated after each training; and then selecting a part of detection models meeting preset conditions from a plurality of middle detection submodels after multiple times of training to determine the preset detection submodels, wherein the preset conditions are that the accuracy of a plurality of middle detection results of multiple times of training meets the preset accuracy requirement and also meets the stability requirement, each preset detection submodel is the middle detection submodel meeting the preset conditions and subjected to model parameter updating, and the preset detection models comprising different preset detection submodels are determined. The preset detection model may include a preset detection sub-model which is trained for multiple times and updated by model parameters for 16 anomaly detection algorithms, namely a statistical 3 segmentcode anomaly detection algorithm, an isolationnforest algorithm, a kNN algorithm, a LOF algorithm, a CBLOF algorithm, an OCSVM algorithm, a LOCI algorithm, a PCA algorithm, an MCD algorithm, a Feature Bagging algorithm, an ABOD algorithm, an Iforest algorithm, an HBOS algorithm, an SOS algorithm, an LSCP algorithm and a COPOD algorithm.

It can be understood that, because the original feature information acquired for the trolley in the operating room is valid only in the current week, that is, the original feature information is acquired once every monday to saturday for detection, and the detected abnormal trolley is repaired on the sunday, the original feature information acquired in the next week cannot reflect the last week, and thus the data size of model training is not large. Based on the method, the current preset detection model can be optimized by combining the original characteristic information acquired each time and the target detection result of the previous time, so that the target detection result is more accurate.

According to the trolley detection method provided by the embodiment of the invention, different preset detection submodels for determining the preset detection model are determined in a mode of training a large number of abnormal detection algorithms for multiple times by using the target characteristic information, so that the accuracy and the reliability of determining the preset detection model are improved.

Referring to the overall flowchart of the trolley detection shown in fig. 2, it can be understood that the overall process of the whole trolley detection method may include: the method comprises the steps of carrying out data warehouse technology (Extract-Transform-Load, ETL) on a large data number of data input values recorded by an MES system to obtain required original characteristic information, determining target characteristic information through data cleaning, column-to-row conversion, exception processing and characteristic amplification, carrying out exception detection based on preset detection models corresponding to 16 exception detection submodels and a comprehensive voting mechanism, determining a target detection result, carrying out iterative model optimization on the preset detection models in the application of the subsequent preset detection models, facilitating a subsequent scheduling system to call and display a preset result table, displaying the prediction result table on a billboard page, carrying out service field inspection and feedback of results, and finally combining the inspection result evaluated by a rechecker to achieve the purpose of model optimization. The specific abnormality detection process can be described with reference to the aforementioned method. And will not be described in detail herein.

When the trolley detection method is used, automatic early warning can be performed on abnormal trolleys every week, then the EAM fault maintenance module is driven, and relevant service personnel are prompted to perform predictive maintenance, so that the reliability, the effectiveness and the accuracy and the stability of whether the air conditioner is qualified or not in a running room are ensured.

The following describes the trolley detection device provided by the present invention, and the trolley detection device described below and the trolley detection method described above may be referred to in correspondence with each other.

Referring to fig. 3, a trolley detection device according to an embodiment of the present invention is shown, and in fig. 3, the trolley detection device 300 includes:

the acquiring module 310 is configured to acquire original feature information of the to-be-detected trolley, where the original feature information includes parameter information collected by the to-be-detected trolley in a process of inspecting a target product;

a determining module 320, configured to determine target feature information based on the original feature information, where the target feature information includes feature information remaining after exception handling is performed on the original feature information;

the detection module 330 is configured to obtain a target detection result for the to-be-detected trolley based on the target feature information and a preset detection model.

It can be understood that the determining module 320 may be specifically configured to perform row-to-row operation on the column data of the column field where the parameter information in the original feature information is located, and determine the feature information to be processed of the to-be-detected trolley; performing exception handling on the feature information to be processed based on a preset exception handling rule; determining target characteristic information based on the characteristic information obtained by exception handling; the preset exception handling rule comprises the steps of removing line data with a line field missing rate exceeding a line missing rate threshold value, filling row data with a row field missing rate lower than a row missing rate threshold value, removing line data which does not meet a preset distribution rule, removing line data which does not meet a preset temperature correlation, removing line data which does not meet a preset temperature range and removing line data which does not meet a preset power correlation for characteristic information to be handled.

It can be understood that the determining module 320 may be further configured to obtain a target number of target products belonging to the same preset equipment part code, where the preset equipment part code is a hardware parameter model of the same target product; and determining that the target quantity is smaller than a preset quantity threshold value, and eliminating the row data of the preset equipment part codes corresponding to the target quantity in the characteristic information to be processed.

It can be understood that the determining module 320 may be further configured to perform feature amplification processing on the feature information obtained by the exception handling, and determine a plurality of feature information after the feature amplification processing; and screening the plurality of characteristic information based on the correlation of the preset characteristic information and the quantity of the preset characteristic information to determine the target characteristic information.

It can be understood that the detection module 330 may be specifically configured to input the target feature information into the preset detection model, so as to obtain a preset number of detection results of the to-be-detected trolley output by the preset detection model under different preset equipment part codes and different preset detection submodels; and obtaining a target detection result aiming at the trolley to be detected based on the abnormal detection result in the preset number of detection results.

It can be understood that the detection module 330 may be further configured to determine a ratio of the abnormal detection results in the preset number of detection results; determining that the occupation ratio exceeds a first preset occupation ratio, and acquiring a target detection result that the trolley to be detected is an abnormal trolley; determining that the occupation ratio is between the second preset occupation ratio and the first preset occupation ratio, and acquiring a target detection result of the trolley to be detected as a risk trolley; and determining that the occupation ratio is lower than a second preset occupation ratio, and acquiring a target detection result that the trolley to be detected is a normal trolley.

It can be understood that the detection module 330 may be further configured to obtain target encoding information of the risk trolley; and sending early warning information to the user terminal based on the target coding information, wherein the early warning information is used for reminding the user terminal of corresponding rechecking personnel to carry out field evaluation on the risk trolley.

It can be understood that the trolley detection method provided by the embodiment of the present invention further includes a training module, configured to train a model, where the training method for presetting the detection model includes: acquiring a plurality of different initial detection submodels for carrying out abnormity detection on the trolley to be detected; training each initial detection submodel for a preset number of times based on the target characteristic information, and determining a plurality of intermediate detection results of a plurality of intermediate detection submodels after each training; sending the plurality of intermediate detection results to a user terminal, and receiving a rechecking result fed back by the user terminal aiming at the plurality of intermediate detection results; and determining a preset detection submodel and a preset detection model corresponding to the preset detection submodel based on the rechecking result and the intermediate detection submodel after each model parameter is updated.

The trolley detection device provided by the embodiment of the invention firstly obtains the original characteristic information of the trolley to be detected, and the original characteristic information comprises the parameter information collected by the trolley to be detected in the process of detecting a target product, so that the target characteristic information is determined from the original characteristic information in a mode of carrying out exception processing on the original characteristic information, and then a target detection result aiming at the trolley to be detected is obtained on the basis of the target characteristic information and a preset detection model; whether the unusual purpose of timely and intellectual detection system platform truck exists is realized through the mode of automatic acquisition information, exception handling and model detection with this, and not only labour saving and time saving, it is high-efficient accurate moreover, combine model study and detection technology can further ensure the accuracy of testing result simultaneously to the accuracy nature and the reliability that the platform truck detected have also effectively been improved.

Fig. 4 illustrates a physical structure diagram of an electronic device, and as shown in fig. 4, the electronic device 400 may include: a processor (processor) 410, a communication Interface 420, a memory (memory) 430 and a communication bus 440, wherein the processor 410, the communication Interface 420 and the memory 430 are communicated with each other via the communication bus 440. The processor 410 may call logic instructions in the memory 430 to perform the following method:

determining target characteristic information based on the original characteristic information, wherein the target characteristic information comprises the characteristic information which is left after the original characteristic information is subjected to exception processing;

In addition, the logic instructions in the memory 430 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

In another aspect, an embodiment of the present invention discloses a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, when the program instructions are executed by a computer, the computer can perform the methods provided by the above-mentioned method embodiments, for example, including:

In another aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented to perform the transmission method provided in the foregoing embodiments when executed by a processor, for example, the method includes:

The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.

Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. Based on such understanding, the above technical solutions substantially or contributing to the related art may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the embodiments or some parts of the embodiments.

Finally, it should be noted that the above embodiments are only for illustrating the present invention, and not for limiting the present invention. Although the present invention has been described in detail with reference to the embodiments, it should be understood by those skilled in the art that various combinations, modifications or equivalents may be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention, and the technical solution of the present invention is covered by the claims of the present invention.

Claims

1. A method for detecting a trolley, comprising:

2. The trolley detection method according to claim 1, wherein the determining target feature information based on the original feature information includes:

based on a preset exception handling rule, exception handling is carried out on the feature information to be processed;

3. The trolley detection method according to claim 2, wherein the performing exception handling on the feature information to be processed further comprises:

and determining that the target quantity is smaller than a preset quantity threshold value, and eliminating the row data of the preset equipment part codes corresponding to the target quantity in the characteristic information to be processed.

4. The method according to claim 2, wherein the determining the target feature information based on the feature information obtained by the anomaly processing includes:

5. The trolley detection method according to claim 1, wherein the preset detection model includes different preset detection submodels, and obtaining the target detection result for the trolley to be detected based on the target feature information and the preset detection model includes:

6. The trolley detection method according to claim 5, wherein the obtaining of the target detection result for the trolley to be detected based on the abnormal detection result in the preset number of detection results comprises:

and determining that the ratio is lower than the second preset ratio, and acquiring a target detection result that the trolley to be detected is a normal trolley.

7. The trolley detection method according to claim 6, wherein after the obtaining of the target detection result that the trolley to be detected is a risk trolley, the method further comprises:

acquiring target coding information of the risk trolley;

8. The trolley detection method according to any one of claims 1 to 7, wherein the training method of the preset detection model comprises:

9. A trolley detection device is characterized by comprising:

the determining module is used for determining target characteristic information based on the original characteristic information, wherein the target characteristic information comprises the characteristic information left after the original characteristic information is subjected to exception processing;

10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the trolley detection method according to any one of claims 1 to 8 when executing the program.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the dolly detection method according to any one of claims 1 to 8.

12. A computer program product comprising a computer program, wherein the computer program, when executed by a processor, implements a trolley detection method according to any one of claims 1 to 8.