Abnormal index determination method, apparatus and device, and computer readable storage medium

Abstract

The invention provides an abnormal index determination method and device, equipment and a storage medium, and the method comprises the steps: carrying out the evaluation of target parameter data based on a deep learning model, and obtaining a first evaluation result; wherein the target parameter data is any one kind of parameter data in various kinds of parameter data of the target equipment; evaluating the target parameter data based on a statistical model to obtain a second evaluation result; evaluating the target parameter data based on the feature extraction model to obtain a third evaluation result; fusing the first evaluation result, the second evaluation result and the third evaluation result according to a preset weight to obtain a target evaluation result; and if the target evaluation result exceeds the target confidence interval, determining the parameter item corresponding to the target parameter data as an abnormal index. By applying the technical scheme of the invention, the problems that the representativeness of data features extracted by a single feature extraction model is insufficient, and the monitoring result of abnormal indexes is inaccurate can be solved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, device, equipment and computer-readable storage medium for determining abnormal indicators. Background Art

[0002] With the rapid development of Internet of Things technology, industrial robot equipment plays an increasingly important role in production lines. As the scale of production continues to increase, the number of robot equipment continues to increase, making robot maintenance more difficult. In related technologies, a single feature extraction model is usually used to analyze the data of robot equipment to monitor abnormal indicators of robot equipment. However, the data features extracted by the single feature extraction model are not representative enough, and the monitoring results of abnormal indicators are not accurate.

[0003] Application Contents

[0004] In view of the above problems, the present application provides a method, device, equipment and computer-readable storage medium for determining abnormal indicators, which are used to solve the problem that the data features extracted by the single feature extraction model in the related art are not representative enough and the monitoring results of abnormal indicators are not accurate.

[0005] According to a first aspect of an embodiment of the present application, a method for determining an abnormal indicator is provided, the determination method comprising: evaluating target parameter data based on a deep learning model to obtain a first evaluation result; wherein the target parameter data is any one of a plurality of parameter data of a target device; evaluating the target parameter data based on a statistical model to obtain a second evaluation result; evaluating the target parameter data based on a feature extraction model to obtain a third evaluation result; fusing the first evaluation result, the second evaluation result and the third evaluation result according to preset weights to obtain a target evaluation result; if the target evaluation result exceeds a target confidence interval, determining that the parameter item corresponding to the target parameter data is an abnormal indicator.

[0006] In an optional manner, the target parameter data is evaluated based on a deep learning model, and before obtaining a first evaluation result, the method includes: determining a data collection frequency and a data collection duration according to a parameter type of the target device; obtaining a plurality of initial parameter data of the target device based on the data collection frequency and the data collection duration; and preprocessing the plurality of initial parameter data to obtain a plurality of preprocessed parameter data.

[0007] In an optional manner, the deep learning model includes: a temporal convolutional network and a long short-term memory network; the target parameter data is evaluated based on the deep learning model to obtain a first evaluation result, including: feature extraction of the target parameter data based on the temporal convolutional network to obtain temporal local features; the temporal local features are input into the long short-term memory network so that the long short-term memory network predicts the first evaluation result.

[0008] In an optional manner, the target parameter data is evaluated based on a statistical model to obtain a second evaluation result, including: obtaining a residual value based on the target parameter data and a model fitting prediction value; wherein the model fitting prediction value is obtained by training the statistical model based on a variety of historical parameter data; and the residual value is evaluated to obtain a second evaluation result.

[0009] In an optional manner, the target parameter data is evaluated based on the feature extraction model to obtain a third evaluation result, including: extracting statistical features of the target parameter data based on the feature extraction model to obtain multiple time-domain statistical feature values; evaluating each of the time-domain statistical feature values ​​to obtain multiple sub-evaluation values; and obtaining a third evaluation result based on the multiple sub-evaluation values.

[0010] In an optional manner, the determination method also includes: determining the alarm display area of ​​the monitoring interface according to the target device corresponding to the abnormal indicator; determining the target alarm level according to the alarm threshold range corresponding to the target evaluation result; and displaying the target alarm color in the alarm display area based on the target alarm level.

[0011] In an optional manner, the determination method also includes: determining the target evaluation range of each warning parameter item corresponding to each preset warning type; obtaining at least one target warning type according to the target evaluation range of the parameter evaluation results corresponding to each abnormal indicator; wherein a plurality of the preset warning types include the target warning type; and displaying the at least one target warning type and the maintenance recommendations corresponding to each of the target warning types on a monitoring interface.

[0012] According to a second aspect of an embodiment of the present application, a device for determining an abnormal indicator is provided, the determination device comprising: a first evaluation module, which evaluates target parameter data based on a deep learning model to obtain a first evaluation result; wherein the target parameter data is any one of a plurality of parameter data of a target device; a second evaluation module, which evaluates the target parameter data based on a statistical model to obtain a second evaluation result; a third evaluation module, which evaluates the target parameter data based on a feature extraction model to obtain a third evaluation result; a fusion module, which fuses the first evaluation result, the second evaluation result and the third evaluation result according to preset weights to obtain a target evaluation result; a determination module, which determines that the parameter item corresponding to the target parameter data is an abnormal indicator if the target evaluation result exceeds a target confidence interval.

[0013] According to a third aspect of an embodiment of the present application, an electronic device is provided, comprising: a controller; a memory storing one or more programs, wherein when the one or more programs are executed by the controller, the controller implements the operation of the determination method: evaluating target parameter data based on a deep learning model to obtain a first evaluation result; wherein the target parameter data is any one of a plurality of parameter data of a target device; evaluating the target parameter data based on a statistical model to obtain a second evaluation result; evaluating the target parameter data based on a feature extraction model to obtain a third evaluation result; fusing the first evaluation result, the second evaluation result and the third evaluation result according to preset weights to obtain a target evaluation result; if the target evaluation result exceeds a target confidence interval, determining that the parameter item corresponding to the target parameter data is an abnormal indicator.

[0014] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, wherein a computer program is stored in the storage medium, wherein the computer program includes at least one executable instruction, and when the executable instruction is run on a determination device / electronic device, the determination device / device performs the operation of the determination method: evaluating the target parameter data based on a deep learning model to obtain a first evaluation result; wherein the target parameter data is any one of a plurality of parameter data of a target device; evaluating the target parameter data based on a statistical model to obtain a second evaluation result; evaluating the target parameter data based on a feature extraction model to obtain a third evaluation result; fusing the first evaluation result, the second evaluation result and the third evaluation result according to preset weights to obtain a target evaluation result; if the target evaluation result exceeds a target confidence interval, determining that the parameter item corresponding to the target parameter data is an abnormal indicator.

[0015] In the embodiment of the present application, the target parameter data is evaluated based on a deep model to obtain a first evaluation result, and the evaluation is performed from the local features of the target data; the target parameter data is evaluated based on a statistical model to obtain a second evaluation result, and the evaluation is performed from the overall change trend of the target data; the target parameter is evaluated based on a feature extraction model to obtain a third evaluation result, and the evaluation is performed from the time domain change characteristics of the target data, and the first evaluation result, the second evaluation result and the third evaluation result are fused according to preset weights to obtain a target evaluation result. If the target result exceeds the target confidence interval, the parameter item corresponding to the target parameter data is determined to be an abnormal indicator. Different models are used to evaluate from different feature dimensions of the target parameter data, so that the target evaluation result obtained can be more accurate, thereby making the result of determining the target parameter data as an abnormal indicator more accurate.

[0016] The above description is only an overview of the technical solution of the embodiment of the present application. In order to more clearly understand the technical means of the embodiment of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the embodiment of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are only used to illustrate the embodiments and are not to be considered as limiting the present application. In addition, the same reference symbols are used to represent the same components throughout the accompanying drawings. In the accompanying drawings:

[0018] Figure 1 It is a schematic diagram of an implementation environment involved in this application.

[0019] Figure 2 A flow chart of a method for determining an abnormality indicator according to an exemplary embodiment of the present application is shown.

[0020] Figure 3 A flow chart of another method for determining an abnormality indicator according to an exemplary embodiment of the present application is shown.

[0021] Figure 4 A flow chart of another method for determining an abnormality indicator according to an exemplary embodiment of the present application is shown.

[0022] Figure 5 A flow chart of another method for determining an abnormality indicator according to an exemplary embodiment of the present application is shown.

[0023] Figure 6 A flow chart of another method for determining an abnormality indicator according to an exemplary embodiment of the present application is shown.

[0024] Figure 7 A flow chart of another method for determining an abnormality indicator according to an exemplary embodiment of the present application is shown.

[0025] Figure 8 A flow chart of another method for determining an abnormality indicator according to an exemplary embodiment of the present application is shown.

[0026] Fig. 9 A schematic structural diagram of an embodiment of a device for determining an abnormality indicator provided by the present application is shown.

[0027] Fig.10 A schematic structural diagram of an embodiment of the electronic device of the present application is shown. DETAILED DESCRIPTION

[0028] Here, exemplary embodiments will be described in detail, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are only examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.

[0029] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0030] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0031] The term "multiple" as used in this application refers to two or more than two. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.

[0032] With the rapid development of Internet of Things technology, industrial robot equipment plays an increasingly important role in production lines. As the scale of production continues to increase, the number of robot equipment is also increasing, and the maintenance of robot equipment has become more difficult. In related technologies, a single feature extraction model is usually used to analyze the robot data for robot maintenance to monitor the abnormal indicators of the machine. However, the single feature model analyzes from a certain dimension of the data features, so the extracted data features are not representative enough. When the analyzed data represents an abnormality, the robot equipment may not be abnormal, and the monitoring of abnormal indicators is not accurate.

[0033] See also Figure 1 , Figure 1 It is a schematic diagram of an implementation environment involved in this application. Figure 1 The implementation environment shown is specifically a monitoring system for a robot device, including: a robot device 110, a gateway 120, a data acquisition platform 130, a predictive maintenance system 140 and a monitoring alarm interface 150. The robot device 110 can be a welding robot, and the various parameter data of the welding robot include: TCP (tool centre position) coordinates, current values ​​and voltage values, etc. The gateway 120 transmits the various parameter data to the data acquisition platform for storage. The data acquisition platform 130 can be a kafka platform. The predictive parameter platform 140 can analyze and evaluate any parameter data to determine whether the parameter data is an abnormal indicator. If it is an abnormal indicator, the abnormal indicator is displayed on the monitoring alarm interface 150 and a warning signal is issued.

[0034] Figure 2 A flowchart of a method for determining an abnormal indicator according to an exemplary embodiment of the present application is shown. The method at least includes S210 to S250, which are described in detail as follows:

[0035] S210: Evaluate the target parameter data based on the deep learning model to obtain a first evaluation result.

[0036] Among them, the monitoring system usually monitors multiple robot devices on the industrial production line at the same time, and monitors multiple parameter data of the target device at the same time. The robot device can be a welding robot, and the welding robot includes multiple welding axes. Exemplarily, the welding axes of the welding robot can be 6. The target device is any device among the multiple robot devices, and the target parameter data is any parameter data among the multiple parameter data of the target device. The multiple parameter data include: motor torque, current, voltage, each welding axis axis value and TCP coordinates, etc. The deep learning model can capture the local features, short-term dependencies and long-term dependencies in the target parameter data, and the local features of the target parameter data can be evaluated, so as to predict the first evaluation result; the first evaluation result can be characterized by the predicted first confidence factor, and whether the target parameter is an abnormal indicator can be determined by judging whether the first confidence factor is within the confidence interval.

[0037] S220: Evaluate the target parameter data based on the statistical model to obtain a second evaluation result.

[0038] Among them, the statistical model is a model that predicts and evaluates the target parameter data based on time series. In the actual production process, when the production task changes, it can capture the seasonal changes in the data. For example, the failure rate of the target equipment in a specific period of time undergoes periodic fluctuations. The target parameter data can be evaluated based on the data change trend to obtain a second evaluation result. The second evaluation result can be characterized by the predicted second confidence factor, and whether the target parameter is an abnormal indicator can be determined by judging whether the second confidence factor is within the confidence interval.

[0039] S230: Evaluate the target parameter data based on the feature extraction model to obtain a third evaluation result.

[0040] Among them, the feature extraction model extracts various change features of the target parameter data from the target parameter, and the third evaluation result can be characterized by the predicted third confidence factor. Each change feature corresponds to a third confidence factor and a confidence interval. If any confidence factor exceeds the confidence interval, the target parameter is determined to be an abnormal indicator.

[0041] S240: The first evaluation result, the second evaluation result and the third evaluation result are integrated according to preset weights to obtain a target evaluation result.

[0042] Among them, the preset weights include: a first preset weight, a second preset weight and a third preset weight, which are respectively the proportions of the evaluation results predicted by the deep learning model, the statistical model and the feature extraction model. Exemplarily, the target evaluation result f can be obtained using the following calculation formula:

[0043] f=a1x1+a2x2+a3x3

[0044] Among them, x1 represents the first evaluation result, x2 represents the first evaluation result, x3 represents the first evaluation result, a1 represents the first preset weight, a2 represents the second preset weight, and a3 represents the third preset weight.

[0045] S250: If the target evaluation result exceeds the target confidence interval, it is determined that the parameter item corresponding to the target parameter data is an abnormal indicator.

[0046] Among them, by judging whether the target evaluation result exceeds the target confidence interval and considering the results predicted by different evaluation models, the target evaluation result is used as the evaluation basis for the final abnormal index, so that the target parameter data can be rated as an abnormal index with higher accuracy.

[0047] The first evaluation result is obtained by evaluating the target parameter data based on the deep model, and the evaluation is performed from the local features of the target data; the second evaluation result is obtained by evaluating the target parameter data based on the statistical model, and the evaluation is performed from the overall change trend of the target data; the third evaluation result is obtained by evaluating the target parameter based on the feature extraction model, and the evaluation is performed from the time domain change characteristics of the target data, and the first evaluation result, the second evaluation result and the third evaluation result are fused according to preset weights to obtain the target evaluation result. If the target result exceeds the target confidence interval, the parameter item corresponding to the target parameter data is determined to be an abnormal indicator. Different models are used to evaluate from different feature dimensions of the target parameter data, which can make the target evaluation result more accurate, thereby making the result of determining the target parameter data as an abnormal indicator more accurate.

[0048] In other embodiments of the present application, how to obtain various parameter data is described in detail. Figure 3 FIG. 1 is a flow chart showing another method for determining an abnormality indicator according to an exemplary embodiment of the present application. Figure 2 Before S110, it also includes at least S310 to S330, which are described in detail as follows:

[0049] S310: Determine the data collection frequency and data collection duration according to the parameter type of the target device;

[0050] Among them, there are many types of parameters of the target device. When the target device is a welding robot, the parameter types collected can be various angular and torque data. These types of parameters require a sufficiently high collection frequency to reflect the changing trend and slight changes of the parameters, so as to make the predicted target evaluation results more accurate. For example, the maximum speed of the robot's motor is 4000rpm, and the period is 66.67Hz. The signal acquisition frequency must be more than twice the signal generation frequency to restore the integrity of the signal. If the data of each parameter is collected in real time, the amount of data collected will be very large, which will bring a very large communication burden to the gateway and cannot meet the needs of data transmission. It is possible to obtain a sufficient amount of data to reflect the changing trend of the target parameters by reducing the collection frequency and increasing the data collection time. For example, through calibration, it can be obtained that the collection time can be set to 4 hours and the collection frequency can be set to 10Hz to obtain a sufficient amount of data. The amount of data obtained at this data collection frequency and data collection time can be analyzed to obtain a more accurate target evaluation result.

[0051] S320: Acquire various initial parameter data of the target device based on the data collection frequency and the data collection duration;

[0052] S330: Preprocessing the various initial parameter data to obtain various preprocessed parameter data

[0053] The directly acquired initial parameter data includes data collection and cleaning to scale the initial parameter data, so that the obtained various parameter data can be directly input into each model for prediction.

[0054] In other embodiments of the present application, how the deep learning model evaluates to obtain the first evaluation result is described in detail. Figure 4 A flowchart of another method for determining abnormal indicators according to an exemplary embodiment of the present application is shown. The deep learning model includes: a temporal convolutional network and a long short-term memory network; Figure 2 S210 in the example includes S410 to S420, which are described in detail as follows:

[0055] S410: Extract features of target parameter data based on a temporal convolutional network to obtain temporal local features.

[0056] Among them, the deep learning model can be a TCN-LSTM algorithm model, which consists of two parts: a TCN algorithm and a LSTM algorithm. Specifically, the TCN algorithm includes a temporal convolutional network. After the target parameter data is input into the TCN algorithm, the temporal convolutional network can extract temporal local features, and the temporal local features can characterize the changes in local features of the target parameter data as the time series changes.

[0057] S420: Inputting the temporal local features into the long short-term memory network so that the long short-term memory network predicts a first evaluation result.

[0058] Among them, the LSTM algorithm includes a long short-term memory network, which can specifically include an input gate, a forget gate and an output gate. The local temporal features are input through the input gate, and irrelevant information is forgotten through the forget gate. The output gate generates a first confidence factor, i.e., a first evaluation result, based on the learned data pattern and the local temporal features passing through the forget gate. By judging whether the first confidence factor exceeds the confidence interval, it can be judged whether the target parameter data is an abnormal indicator.

[0059] In other embodiments of the present application, how the statistical learning model evaluates to obtain the second evaluation result is described in detail. Figure 5 A flowchart of another method for determining an abnormality indicator according to an exemplary embodiment of the present application is shown. Figure 2 The S220 in the system includes S510 to S520, which are described in detail as follows:

[0060] S510: Obtaining a residual value according to the target parameter data and the model fitting prediction value; wherein the model fitting prediction value is obtained by training a statistical model based on a variety of historical parameter data;

[0061] The parameter types of the historical parameter data and the target parameter data are the same. Before using the statistical model to evaluate the second evaluation result, a variety of historical parameter data will be used to train the statistical model, and the model fitting value will be predicted based on the historical parameter data, so as to obtain the model fitting prediction value based on the historical parameter data and the target parameter data value to calculate the residual value.

[0062] S520: Evaluate the residual value to obtain a second evaluation result.

[0063] Among them, the degree of abnormality of the target parameter data can be determined according to the residual value, and the residual value can be converted into a second confidence factor, that is, a second evaluation result, which characterizes whether the target parameter is an abnormal indicator. By judging whether the second confidence factor exceeds the confidence interval, it can be determined whether the target parameter is an abnormal indicator.

[0064] In other embodiments of the present application, how the deep learning model evaluates to obtain the first evaluation result is described in detail. Figure 6 FIG. 1 is a flow chart showing another method for determining an abnormality indicator according to an exemplary embodiment of the present application. Figure 2 The S230 in the system includes S610 to S630, which are described in detail as follows:

[0065] S610: Extract statistical features of target parameter data based on the feature extraction model to obtain multiple time-domain statistical feature values.

[0066] Among them, the time domain statistical characteristic values ​​are values ​​that characterize the changing trends of various characteristics within a specified time. For example, the time domain statistical characteristic values ​​include: mean, variance, standard deviation, peak value, peak-to-peak value, skewness, kurtosis, waveform index, pulse indicator and margin indicator, etc. For example, the mean can reveal the trend of sensor drift or performance degradation, helping to predict and prevent potential failures or wear. Variance and standard deviation help monitor changes in system stability. High variance or continued growth in standard deviation may indicate wear or failure of machine parts. Peak and peak-to-peak values ​​can reveal extreme mechanical shocks or sudden events, helping to quickly identify mechanical failures, collisions or overload events. Skewness and kurtosis can reveal changes in signal distribution, helping to detect non-uniform wear, load imbalance or low-probability extreme vibration events. Form factor, pulse factor and margin factor help identify changes in signal waveforms, transient impact events and the dynamic range of the system, thereby promptly detecting uneven operation or damage to the mechanical system.

[0067] S620: Evaluate each time-domain statistical characteristic value to obtain multiple sub-evaluation values.

[0068] Among them, each time-domain statistical eigenvalue can obtain a corresponding sub-confidence factor and sub-confidence interval. The sub-confidence factor corresponding to each time-domain eigenvalue is the sub-evaluation value. By judging whether each sub-confidence factor is in the corresponding confidence interval, it is judged whether the corresponding time-domain eigenvalue is abnormal, thereby determining whether the target parameter is an abnormal indicator.

[0069] S630: Obtain a third evaluation result according to the plurality of sub-evaluation values.

[0070] Among them, when any sub-confidence factor exceeds the sub-confidence interval, the target parameter is determined to be an abnormal indicator, and the third confidence factor of the final target parameter, that is, the third evaluation result, is obtained by mapping the size of each sub-evaluation value, so that the first evaluation result and the second evaluation result can be fused on the same calculation dimension.

[0071] In other embodiments of the present application, how to issue an early warning for abnormal indicators is described in detail. Figure 7 A flowchart of another method for determining an abnormality indicator according to an exemplary embodiment of the present application is shown. The method further includes S710 to S720, which are described in detail as follows:

[0072] S710: Determine an alarm display area of ​​the monitoring interface according to the target device corresponding to the abnormal indicator.

[0073] Among them, each target device has a corresponding monitoring area in the monitoring interface. By determining the abnormal indicators, the display area of ​​the target device with abnormal indicators in the monitoring interface can be determined, so that the display area is used as an alarm display area to enable relevant personnel to know the target device with warning.

[0074] S720: Determine the target alarm level according to the alarm threshold range corresponding to the target evaluation result.

[0075] Among them, different target evaluation results have different alarm threshold ranges. By determining the target alarm level, the severity of the failure of the target equipment can be determined.

[0076] S730: Based on the target alarm level, display the target alarm color in the alarm display area.

[0077] Among them, the target alarm level can be: zero level, level one and level two; the target alarm color can be: blue, yellow and red. For example, if the target alarm level is 0, that is, normal, the alarm display area displays blue; if the target alarm level is level one alarm, the alarm display area displays yellow; if the target alarm level is level two alarm, the alarm display area displays red; so that relevant staff can know the severity of the fault according to different alarm colors.

[0078] In other embodiments of the present application, how to determine the target warning type is described in detail. Figure 8 A flowchart of another method for determining an abnormal indicator according to an exemplary embodiment of the present application is shown. The determination method further includes S810 to S830, which are described in detail as follows:

[0079] S810: Determine the target evaluation range of each warning parameter item corresponding to each preset warning type.

[0080] Among them, the preset warning type is the fault type that will cause an alarm or the maintenance type that requires maintenance, the warning parameter item is the parameter item corresponding to the abnormal indicator corresponding to the warning type, each warning type has corresponding warning parameter items, and the target evaluation range of the warning parameter item, that is, the range value of the target evaluation result.

[0081] S820: Obtain at least one target warning type according to the target evaluation range of the parameter evaluation results corresponding to each abnormal indicator.

[0082] Among them, the multiple preset warning types include target warning types; by evaluating the target evaluation range of the parameter evaluation results corresponding to each abnormal indicator, it is possible to determine which fault types or maintenance types have occurred in the target equipment.

[0083] S830: Display at least one target warning type and maintenance suggestions corresponding to each target warning type on a monitoring interface.

[0084] Among them, by displaying at least one target warning type and corresponding maintenance suggestions on the monitoring interface, relevant staff can check, repair and maintain at least one target warning type to improve the service life of the target equipment. For example, if the 3-axis motor torque and 3-axis shaft angle of the target welding robot are both abnormal indicators, then at least one target warning type includes: motor damage, bearing wear, etc., and maintenance suggestions are given and displayed on the monitoring interface.

[0085] Fig. 9 The schematic diagram of the structure of an embodiment of a device for determining abnormal indicators provided by the present application is shown. Fig. 9 As shown, the device 900 includes: a first evaluation module 910 , a second evaluation module 920 , a third evaluation module 930 , a fusion module 940 and a determination module 950 .

[0086] A first evaluation module 910 evaluates the target parameter data based on the deep learning model to obtain a first evaluation result; wherein the target parameter data is any one of a plurality of parameter data of the target device;

[0087] A second evaluation module 920 evaluates the target parameter data based on a statistical model to obtain a second evaluation result;

[0088] A third evaluation module 930 evaluates the target parameter data based on the feature extraction model to obtain a third evaluation result;

[0089] A fusion module 940 is configured to fuse the first evaluation result, the second evaluation result and the third evaluation result according to a preset weight to obtain a target evaluation result;

[0090] Determination module 950 determines that the parameter item corresponding to the target parameter data is an abnormal indicator if the target evaluation result exceeds the target confidence interval.

[0091] The abnormality indicator determination device provided in the above embodiment and the abnormality indicator determination method provided in the above embodiment belong to the same concept, wherein the specific manner in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here.

[0092] Fig.10 A schematic diagram of the structure of an embodiment of the electronic device of the present application is shown, which shows a schematic diagram of the structure of a computer system suitable for implementing the electronic device of the embodiment of the present application. The specific embodiment of the present application does not limit the specific implementation of the electronic device.

[0093] See also Fig.10 As shown, the electronic device includes: a controller; a memory for storing one or more programs, and when the one or more programs are executed by the controller, the above-mentioned method for determining abnormal indicators is executed.

[0094] Please continue reading Fig.10 As shown, the computer system 1000 of the electronic device includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1002 or the program loaded from the storage part 1008 to the random access memory (RAM) 1003, such as executing the method in the above embodiment. In RAM 1003, various programs and data required for system operation are also stored. CPU 1001, ROM 1002 and RAM 1003 are connected to each other through a bus 1004. Input / output (I / O) interface 1005 is also connected to bus 1004.

[0095] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, a mouse, etc.; an output section 1007 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. A removable medium 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1010 as needed so that a computer program read therefrom is installed into the storage section 1008 as needed.

[0096] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section 1009, and / or installed from a removable medium 1011. When the computer program is executed by a central processing unit (CPU) 1001, various functions defined in the system of the present application are executed.

[0097] Another aspect of the present application also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the determination method described above is implemented. The computer-readable storage medium may be included in the electronic device described in the above embodiment, or may exist independently without being assembled into the electronic device.

[0098] Another aspect of the present application also provides a computer program product or a computer program, which includes at least one executable instruction. When the executable instruction is executed on an abnormal indicator determination device / electronic device, the abnormal indicator determination device / electronic device executes the abnormal indicator determination method as described above.

[0099] The executable instructions may be specifically used to enable the abnormal indicator determination device / equipment to perform the following operations:

[0100] Evaluate the target parameter data based on the deep learning model to obtain a first evaluation result; wherein the target parameter data is any one of a plurality of parameter data of the target device;

[0101] Evaluate the target parameter data based on the statistical model to obtain a second evaluation result;

[0102] Evaluate the target parameter data based on the feature extraction model to obtain a third evaluation result;

[0103] The first evaluation result, the second evaluation result and the third evaluation result are integrated according to the preset weights to obtain the target evaluation result;

[0104] If the target evaluation result exceeds the target confidence interval, the parameter item corresponding to the target parameter data is determined to be an abnormal indicator.

[0105] The computer-readable medium shown in the embodiment of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it. In the present application, a computer-readable signal medium can include a data signal propagated in a baseband or as a part of a carrier wave, wherein a computer-readable computer program is carried. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. A computer program contained on a computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0106] The flowchart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present application. Wherein, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0107] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. The names of these units do not, in some cases, constitute limitations on the units themselves.

[0108] According to one aspect of an embodiment of the present application, a computer system is also provided, including a central processing unit (CPU), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) or a program loaded from a storage portion into a random access memory (RAM), such as executing the method in the above embodiment. In RAM, various programs and data required for system operation are also stored. CPU, ROM and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0109] The following components are connected to the I / O interface: an input part including a keyboard, a mouse, etc.; an output part including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage part including a hard disk, etc.; and a communication part including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication part performs communication processing via a network such as the Internet. A drive is also connected to the I / O interface as needed. Removable media, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., are installed on the drive as needed so that the computer program read therefrom is installed into the storage part as needed.

[0110] The above content is only a preferred exemplary embodiment of the present application and is not intended to limit the implementation scheme of the present application. A person skilled in the art can easily make corresponding changes or modifications based on the main concept and spirit of the present application. Therefore, the scope of protection of the present application shall be based on the scope of protection required by the claims.

Claims

1. A method for determining an abnormality index, characterized in that: The determination method comprises: Evaluate the target parameter data based on the deep learning model to obtain a first evaluation result; wherein the target parameter data is any one of a plurality of parameter data of the target device; Evaluate the target parameter data based on a statistical model to obtain a second evaluation result; Evaluate the target parameter data based on the feature extraction model to obtain a third evaluation result; The first evaluation result, the second evaluation result and the third evaluation result are integrated according to a preset weight to obtain a target evaluation result; If the target evaluation result exceeds the target confidence interval, the parameter item corresponding to the target parameter data is determined to be an abnormal indicator.

2. The determination method according to claim 1, characterized in that: The step of evaluating the target parameter data based on the deep learning model to obtain the first evaluation result includes: Determine the data collection frequency and duration based on the parameter type of the target device; Based on the data collection frequency and the data collection duration, obtaining various initial parameter data of the target device; The multiple initial parameter data are preprocessed to obtain multiple preprocessed parameter data.

3. The determination method according to claim 1, characterized in that: The deep learning model includes: a temporal convolutional network and a long short-term memory network; the target parameter data is evaluated based on the deep learning model to obtain a first evaluation result, including: Extracting features of the target parameter data based on the temporal convolutional network to obtain temporal local features; The temporal local feature is input into the long short-term memory network so that the long short-term memory network predicts a first evaluation result.

4. The determination method according to claim 1, characterized in that: The step of evaluating the target parameter data based on the statistical model to obtain a second evaluation result includes: Obtaining a residual value according to the target parameter data and the model fitting prediction value; wherein the model fitting prediction value is obtained by training the statistical model based on a variety of historical parameter data; The residual value is evaluated to obtain a second evaluation result.

5. The determination method according to claim 1, characterized in that: The target parameter data is evaluated based on the feature extraction model to obtain a third evaluation result, including: Extract statistical features of the target parameter data based on the feature extraction model to obtain multiple time-domain statistical feature values; Evaluate each of the time-domain statistical characteristic values ​​to obtain a plurality of sub-evaluation values; A third evaluation result is obtained according to the plurality of sub-evaluation values.

6. The determination method according to claim 1, characterized in that: The determination method further comprises: Determine the alarm display area of ​​the monitoring interface according to the target device corresponding to the abnormal indicator; Determine the target alarm level according to the alarm threshold range corresponding to the target evaluation result; Based on the target alarm level, a target alarm color is displayed in the alarm display area.

7. The determination method according to claim 1, characterized in that: The determination method further comprises: Determine the target evaluation range of each warning parameter item corresponding to each preset warning type; According to the target evaluation range of the parameter evaluation results corresponding to each abnormal indicator, at least one target warning type is obtained; wherein the plurality of preset warning types include the target warning type; The at least one target warning type and the maintenance suggestions corresponding to each target warning type are displayed on a monitoring interface.

8. A device for determining an abnormality index, characterized in that: The determining device comprises: A first evaluation module evaluates the target parameter data based on the deep learning model to obtain a first evaluation result; wherein the target parameter data is any one of a plurality of parameter data of the target device; A second evaluation module evaluates the target parameter data based on a statistical model to obtain a second evaluation result; A third evaluation module evaluates the target parameter data based on a feature extraction model to obtain a third evaluation result; A fusion module, which fuses the first evaluation result, the second evaluation result and the third evaluation result according to a preset weight to obtain a target evaluation result; The determination module determines that the parameter item corresponding to the target parameter data is an abnormal indicator if the target evaluation result exceeds the target confidence interval.

9. An electronic device, characterized in that: include: Controller; A memory storing one or more programs, which, when executed by the controller, enables the controller to implement the determination method described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, which includes at least one executable instruction. When the executable instruction is executed on the determination device / electronic device, the determination device / equipment performs the operation of the determination method as described in any one of claims 1 to 7.

Images