Automated-device evaluation optimization method and system

Abstract

Provided in the present invention are an automated-device evaluation optimization method and system. The automated-device evaluation optimization method comprises the steps of: obtaining determination information of a device and a corresponding initial sensing signal; capturing a periodic signal from the initial sensing signal; performing calculation on the periodic signal to obtain at least one initial characteristic parameter; performing calculation on the determination information and the at least one initial characteristic parameter to obtain corresponding initial weight information; and establishing an evaluation model on the basis of the at least one initial characteristic parameter and the initial weight information. An automated-device evaluation optimization system executes the method, so as to monitor a device in real time, precisely analyze the operating status of the device and perform anomaly detection.

Description

Method and system for optimizing automated equipment evaluation TECHNICAL FIELD

[0001] The present invention relates to an evaluation method and system, in particular to a method and system for optimizing automated equipment evaluation. BACKGROUND

[0002] In the past, before / during / after the operation of a production line, it is necessary to detect and maintain the equipment used by the production line. Such detection and maintenance mostly requires manpower to confirm whether the detection index is abnormal or to find out the cause of the abnormality to ensure that each equipment is in a state that can produce, but it is impossible to know the actual situation of the equipment. When the production actually starts, since the equipment is not in a 100% good state, it is necessary to rely on the real-time observation of the equipment by the on-site personnel to ensure that the equipment state can normally operate. This method may not be able to determine in real time which part of the equipment is problematic. Generally speaking, it is only when the production capacity is poor, the production line is blocked, or the defect rate is increased that the on-site personnel confirms which link is problematic. In this way, a large amount of manpower and a long waiting time are required, not only increasing the time and labor costs, but also failing to efficiently grasp the equipment state.

[0003] The precision equipment in the production line is complex and has many types. If the state of different equipment is to be monitored, a corresponding solution needs to be proposed to meet the monitoring needs of the equipment. Similarly, the engineers who maintain the equipment need to be familiar with various solutions, or different engineers are needed to maintain different equipment sections. This method still cannot achieve integrated management.

[0004] Furthermore, if the data of the monitoring equipment is imported into an AI model, a large amount of monitoring data and events need to be accumulated for a long time to achieve the purpose of monitoring and evaluation. A large amount of data needs to be cleaned according to different factory processes. Different products, processes, parameters, and actions will have different data. Each type of data needs a professional to construct a corresponding analysis model, which results in each analysis model being exclusively for a specific production line and cannot be used universally, and the construction cost will be very high.

[0005] Therefore, the present invention provides a method and system for optimizing automated equipment evaluation. The obtained sensing signals are operated, and the generated feature parameters are input into an evaluation model to obtain corresponding evaluation results, thereby obtaining the overall health of the equipment and whether there is an abnormal situation, and predicting whether a maintenance program needs to be performed. The method and system can be applied to all periodic equipment monitoring, have high universality, greatly reduce the construction cost, and effectively manage the monitoring situation.

[0006] SUMMARY

[0007] The main purpose of the present application is to provide an automatic equipment evaluation optimization method, which automatically tracks the sensing signals of the equipment, evaluates the equipment status in real time, and accurately grasps the current situation and future situation of the equipment.

[0008] Another purpose of the present application is to provide an automatic equipment evaluation optimization system, which can be evaluated by a monitoring module after the feature parameters are obtained from the sensing signals by an operation processing module, thereby achieving the purpose of real-time monitoring.

[0009] In order to achieve the above purpose, an embodiment of the present application discloses an automatic equipment evaluation optimization method, comprising the steps of: obtaining judgment information of the equipment and corresponding initial sensing signals; extracting periodic signals from the initial sensing signals; calculating the periodic signals to obtain at least one initial feature parameter; calculating the judgment information and the at least one initial feature parameter to obtain corresponding initial weight information; and establishing an evaluation model with the at least one initial feature parameter and the initial weight information.

[0010] In a preferred embodiment, the steps include: extracting sensing signals; calculating the sensing signals to obtain at least one corresponding feature parameter; and inputting the at least one feature parameter into the evaluation model to generate a corresponding evaluation result.

[0011] In a preferred embodiment, the sensing signals and the initial sensing signals are selected from vibration signals, current signals, pressure signals, torque signals, or combinations thereof.

[0012] In a preferred embodiment, the at least one feature parameter and the at least one initial feature parameter are selected from amplitude, frequency, phase, wavelet, standard deviation, stability, or combinations thereof.

[0013] In a preferred embodiment, the evaluation result includes operation analysis information, abnormal detection information, and operation trend information of the equipment.

[0014] In order to achieve another purpose, an embodiment of the present application discloses an automatic equipment evaluation optimization system, comprising: an operation processing module, which obtains sensing signals of the equipment and calculates at least one feature parameter according to the sensing signals; and a monitoring module, which is signal connected with the operation processing module, receives and evaluates the at least one feature parameter to obtain a corresponding evaluation result.

[0015] In a preferred embodiment, a display interface is signal connected with the monitoring module to display the evaluation result, the display interface includes an equipment overall score area, an equipment state management area, and an equipment operation trend management area, the equipment overall score area displays operation analysis information of the equipment, the equipment state management area displays abnormal detection information, and the equipment operation trend management area displays operation trend information.

[0016] In a preferred embodiment, a monitoring device is connected to the operation processing module for detecting the device to obtain the corresponding sensing signal, wherein the sensing signal is selected from a vibration signal, a current signal, a pressure signal, a torque signal, or a combination thereof.

[0017] In a preferred embodiment, the at least one characteristic parameter is selected from an amplitude, a frequency, a phase, a wavelet, a standard deviation, a stability, or a combination thereof.

[0018] In a preferred embodiment, the operation processing module obtains the judgment information of the device and the initial sensing signal corresponding thereto, and performs operation thereon, and extracts a periodic signal from the initial sensing signal, and obtains the at least one initial characteristic parameter and the initial weight information corresponding thereto from the judgment information and the periodic signal, and establishes an evaluation model according to the at least one initial characteristic parameter and the initial weight information, and transmits the evaluation model to the monitoring module, so that the monitoring module evaluates the sensing signal through the evaluation model to obtain the corresponding evaluation result.

[0019] The present application has the advantages of monitoring the health condition of the machine in real time and determining whether there is an abnormal situation, and can be applied to all periodic device monitoring, has high universality, greatly reduces the installation cost, and effectively manages the monitoring situation. BRIEF DESCRIPTION OF DRAWINGS

[0020] FIG. 1A is a flowchart of a method according to an embodiment of the present application;

[0021] FIG. 1B is a partial flowchart of a method according to an embodiment of the present application;

[0022] FIG. 2 is a schematic diagram of a system according to an embodiment of the present application;

[0023] FIG. 3A is a schematic diagram of an implementation flow according to an embodiment of the present application;

[0024] FIG. 3B is a schematic diagram of an implementation flow according to an embodiment of the present application;

[0025] FIG. 3C is a schematic diagram of an implementation flow according to an embodiment of the present application;

[0026] FIG. 3D is a schematic diagram of an implementation flow according to an embodiment of the present application;

[0027] FIG. 3E is a schematic diagram of an implementation flow according to an embodiment of the present application;

[0028] FIG. 3F is a schematic diagram of an implementation flow according to an embodiment of the present application;

[0029] FIG. 3G is a schematic diagram of an implementation flow according to an embodiment of the present application;

[0030] FIG. 3H is a schematic diagram of an implementation flow according to an embodiment of the present application;

[0031] FIG. 3I is a flowchart of an embodiment of the present application; and

[0032] FIG. 3J is a flowchart of an embodiment of the present application. DETAILED DESCRIPTION

[0033] To make the above and / or other purposes, effects, features of the present application more apparent, a preferred embodiment is specifically described below:

[0034] Please refer to FIG. 1A to FIG. 1B, which are a flowchart of a method and a partial flowchart of an embodiment of the present application. As shown in the figure, the automatic equipment evaluation optimization method of the present application, the steps include:

[0035] Step S1: obtaining judgment information of equipment and corresponding initial sensing signals;

[0036] Step S2: extracting periodic signals from the initial sensing signals;

[0037] Step S3: calculating the periodic signals to obtain at least one initial feature parameter;

[0038] Step S4: calculating the judgment information and the at least one initial feature parameter to obtain corresponding initial weight information; and

[0039] Step S5: establishing an evaluation model with the at least one initial feature parameter and the initial weight information.

[0040] As shown in step S1, the judgment information of the equipment and the corresponding initial sensing signals are obtained to track the actual operation of each action of the equipment. In an embodiment, the initial sensing signals can be vibration signals, current signals, pressure signals, torque signals or combinations thereof. The initial sensing signals can select different types of sensing signals according to different equipment operation modes, and are not limited herein.

[0041] As shown in step S2, the periodic signals are extracted from the initial sensing signals, wherein the periodic signals are signals obtained according to each cycle of the operation of the equipment. For example, when the equipment is a mechanical arm, the cycle is the start of each action to the end of each action, and each cycle is a repeated action, so the signals obtained in each cycle are also repeated, but are not limited herein.

[0042] As shown in step S3, the initial characteristic parameters are obtained by performing operation according to the periodic signal. In an embodiment, the initial characteristic parameters can be amplitude, frequency, phase, wavelet, standard deviation, stability or combination thereof. Similarly, when the initial sensing signal is a vibration signal, fast Fourier transform (FFT) can be used to perform operation analysis to generate amplitude, frequency and phase information.

[0043] As shown in step S4, the initial weight information corresponding to the initial characteristic parameters is obtained by performing operation according to the judgment information and the initial characteristic parameters. That is, different weights are determined according to the initial characteristic parameters corresponding to each round of action, so that the evaluation result can be more accurate, but not limited thereto.

[0044] As shown in step S5, the evaluation model is established according to the initial characteristic parameters and the initial weight information. In this way, different weight distributions can be provided according to the characteristic parameters actually focused on each round of action, so that the evaluation model is more accurate.

[0045] Further, the implementation method of the evaluation model of an embodiment of the present application can include the following steps:

[0046] Step S6: obtaining a sensing signal;

[0047] Step S7: performing operation on the sensing signal to obtain at least one characteristic parameter; and

[0048] Step S8: inputting the at least one characteristic parameter into the evaluation model to generate a corresponding evaluation result.

[0049] As shown in step S6, the corresponding sensing signal is obtained during device operation. In an embodiment, the sensing signal can be a vibration signal, a current signal, a pressure signal, a torque signal or a combination thereof. Different types of sensing signals can be selected according to different device operation modes, so not limited thereto.

[0050] As shown in step S7, the characteristic parameters are obtained by performing operation according to the sensing signal. In an embodiment, the characteristic parameters can be amplitude, frequency, phase, wavelet, standard deviation, stability or combination thereof. When the sensing signal is a vibration signal, fast Fourier transform (FFT) can be used to perform operation analysis to generate amplitude, frequency and phase information.

[0051] As shown in step S8, the characteristic parameters obtained by the foregoing calculation are input into the evaluation model to obtain the evaluation result. In an embodiment, the evaluation model can give corresponding weights according to the characteristic parameters, and a score can be generated. The evaluation result can be presented in the form of numbers, words, symbols or charts, but not limited thereto.

[0052] Referring to FIG. 2, it is a schematic diagram of a system of an embodiment of the present application. As shown, the automated equipment evaluation optimization system of the present application comprises: a calculation processing module 1, a monitoring module 2, a display interface 3 and a monitoring device 4, wherein the monitoring module 2 is signal connected with the calculation processing module 1 and the display interface 3 respectively, and the monitoring device 4 is signal connected with the calculation processing module 1, and the details are as follows:

[0053] The calculation processing module 1, after obtaining the sensing signal of the equipment D, performs calculation according to the sensing signal to obtain the corresponding characteristic parameter, in an embodiment, the characteristic parameter is selected from the group consisting of amplitude, frequency, phase, wavelet, standard deviation, stability or combination thereof, but is not limited thereto.

[0054] In an embodiment, the calculation processing module 1 obtains the judgment information of the equipment D and the initial sensing signal corresponding thereto, performs calculation thereon, and extracts the periodic signal therefrom, and obtains the initial characteristic parameter and the corresponding initial weight information from the judgment information and the periodic signal, and establishes and transmits the evaluation model to the monitoring module 2 according to the initial characteristic parameter and the initial weight information, wherein the initial sensing signal is composed of a plurality of repeated periodic signals, and each period is the beginning of each round of action to the end of each round of action of the equipment D, but is not limited thereto.

[0055] The monitoring module 2 receives and evaluates according to the characteristic parameter to obtain the corresponding evaluation result, in an embodiment, the monitoring module 3 calculates and compares the sensing signal through the evaluation model to obtain the corresponding evaluation result.

[0056] The display interface 3 is used to display the evaluation result, in an embodiment, the display interface 3 comprises an equipment overall score area 31, an equipment state management area 32 and an equipment operation trend management area 33, the equipment overall score area 31 displays the operation analysis information of the equipment D, the equipment state management area 32 displays the abnormal detection information, and the equipment operation trend management area 33 displays the operation trend information, but is not limited thereto.

[0057] In an embodiment, the operation analysis information can be equipment health degree, equipment stability, vibration exceeding times, frequency exceeding times, actual running condition, comprehensive score, but is not limited thereto.

[0058] In an embodiment, the abnormal detection information can be abnormality of a production line arm, abnormality of automatic welding, abnormality of forging die, abnormality of stamping burr, abnormality of lathe machining, abnormality of drilling tool, and further, the fault position or circuit can be judged according to the abnormal information, but is not limited thereto.

[0059] In one embodiment, the operation trend information can be estimated from the real-time monitored signals to predict the future sensing signals, and the maintenance program can be scheduled in advance to improve the service life of the equipment, but not limited thereto.

[0060] The monitoring device 4 is used to detect the equipment D, and after obtaining the sensing signals of the equipment D, the sensing signals are transmitted to the operation processing module 1 for operation. In one embodiment, the monitoring device 4 can use different sensors according to the process characteristics to achieve accurate analysis and monitoring, wherein the monitoring device 4 can be a vibration sensor, a current sensor, a pressure sensor, or a combination thereof. For example, if the process is related to mechanical behavior, such as a robot arm, a stamping forging die, a machine tool; if the process is directly related to current, such as automatic welding; if the process is directly related to pressure, such as automatic dispensing, gluing, and injection molding, but not limited thereto.

[0061] In one embodiment, the sensing signals are selected from vibration signals, current signals, pressure signals, torque signals, or combinations thereof. The monitoring device 4 can be an additional device provided on the equipment D, or can be the original monitoring device of the equipment D, but not limited thereto.

[0062] In order to more clearly illustrate the operation process of one embodiment of the present application, the following is described:

[0063] In one embodiment, the automatic equipment evaluation optimization method and system are applicable to industrial manufacturing equipment, such as semiconductor crystal growth, semiconductor manufacturing, semiconductor packaging, robot arms, mechanical industry applications, panel manufacturing, PCB industry, or automobile manufacturing, etc. In these fields, a large number of periodic production equipment will be used, such as machining machines, robot arms, CNC machine tools, stamping forging machine equipment, automatic welding arm equipment, and injection molding machine equipment. Therefore, the periodic production equipment will have certain actions, time, and processes when operating or manufacturing. The method and system establish an evaluation model by using these repetitive actions, but not limited thereto.

[0064] When the process starts, the monitoring device 4 acquires the sensing signal of the device D. For example, the sensing signal is a vibration signal. The change of the vibration can be used to represent the state change of the device D when it is actually running. If the vibration signal changes when the same action is performed, it usually indicates that the device D has a problem, and thus a force is generated. This force can be caused by various factors on the device D, such as imbalance, poor centering, bearing problems, etc. The vibration on the device D is a mixed vibration, which can be a combination of vibration signals with different frequencies, amplitudes, and phases. The amplitude can be observed by three indicators, namely displacement, velocity, and acceleration, as shown in Table 1.

[0065] Table 1: Related parameters of the vibration signal.

[0066] In an embodiment, the vibration signal can be analyzed by using fast Fourier transform (FFT). The vibration spectrum characteristics can be used to analyze the damage location and the damage cause. For this purpose, the change of each frequency band of the spectrum generated by the action of the device D can be monitored. The width of the frequency band can be set by the user. This monitoring method is suitable for devices such as gearboxes, servo motors, slides, and screws. A database can be established based on the acquired sensing signal to quickly inspect the actual status of each component in the device.

[0067] In an embodiment, the sensing signal reflects the health status of each component on the device D. If a single component is abnormal but has not affected the overall vibration, it can cause the measured value of a single frequency band to be too large to generate an alarm. In this way, faster inspection and judgment can be achieved.

[0068] In an embodiment, different characteristic parameters can be used for judgment according to the process mode. Since the running action of the device D is extremely complex, the normal signal and the abnormal signal during running often overlap together. Therefore, the value is no longer the criterion for judging whether the running is abnormal. The initial value of a single value is obviously not sufficient.

[0069] In one embodiment, the characteristic parameters can be further calculated and presented in various data formats, such as Dynamic Similarity, F-Dynamic Similarity, Frequency Over, SOA Dynamic Similarity, Frequency Similarity, Amplitude Over, or a combination thereof.

[0070] Dynamic Similarity is the difference between the specified target motion specification and the real-time measured sensing signal pattern, and the total score can be calculated using 100 points as the benchmark.

[0071] F-Dynamic Similarity is the difference between the wave packet specification of the specified target motion and the measured pattern, and the total score can be calculated using 100 points as the benchmark.

[0072] Frequency Over is the entire time-frequency pattern extracted in a specific target motion time period, and a frequency spectrum specification is established for each frequency spectrum pattern to extract the total number of frequency spectrum to calculate the over-standard and qualified proportion.

[0073] SOA Dynamic Similarity is the entire time-frequency pattern extracted in a specific target motion time period, and a single frequency spectrum pattern is used to calculate the SOA value (Spectrum Overall value), and the dynamic pattern of the target motion is redrawn as the specification and the difference between the real-time measured pattern, and the total score can be calculated using 100 points as the benchmark.

[0074] Frequency Similarity is the entire time-frequency pattern extracted in a specific target motion time period, and a frequency spectrum specification is established for each frequency spectrum pattern, and the real-time frequency spectrum under the relative time sequence is compared, and after each frequency spectrum pattern is compared, the total score is calculated using 100 points as the benchmark, and the average of the total number of frequency spectrums is calculated using 100 points as the benchmark.

[0075] Amplitude Over is the establishment of a global amplitude threshold for a specific target motion, and the number of data points is used as the mother number to calculate the number of over-standard times of the real-time measured motion corresponding to the specification, and the percentage is calculated based on the total number of data.

[0076] Referring to FIGS. 3A-3B, which are evaluation result patterns of an embodiment of the present application, taking tracking of each action of the automated mechanical arm (running action recognition degree Pass) as an example, it can be seen from the figures that it can be divided into a device overall score area 31, a device state management area 32, and a device operation trend management area 33, wherein the device overall score area 31 displays the health degree score and the stability degree score of the corresponding device, the device state management area 32 displays the aforementioned various data (data and charts), the device state management area 32 displays the actual action situation of the plurality of actions in the machine, and the device operation trend management area 33 presents the various data trend situations of the machine in the time interval, and can also display the information of predictive maintenance.

[0077] Referring to FIGS. 3C-3D, which are evaluation result patterns of an embodiment of the present application, taking tracking of each action of the automated mechanical arm (running action recognition degree Pass) as an example, it can be seen from the figures that when each action includes action one and action two, the upper half gives scores for the health degree and the stability degree of action one and action two, so that the user can more intuitively master the situation of the device D, and the lower half is the sensing signal waveform diagram obtained by action one and action two, when the automated mechanical arm normally operates, the mechanical learning action (yellow line part) and the real-time sensing signal (white line part) are consistent.

[0078] Referring to FIGS. 3E-3F, which are evaluation result patterns of an embodiment of the present application, taking tracking of each action of the automated mechanical arm (mechanical action abnormality, impact surge) as an example, it can be seen from the figures that when each action includes action one and action two, from the overall dynamic similarity (Dynamic Similarity), it can be seen that the mechanical learning action (yellow line part) and the real-time sensing signal (white line part) have a large difference (such as the red circle), and the relevant sensing data can be further viewed, and relatively, it can be seen that the health degree also correspondingly decreases.

[0079] Referring to FIGS. 3G-3H, which are evaluation result patterns of an embodiment of the present application, taking tracking of each action of the automated mechanical arm (mechanical micro-impact Pass) as an example, it can be seen from the figures that when each action includes action one and action two, from the overall dynamic similarity (Dynamic Similarity), it can be seen that the real-time sensing signal (white line part) at the red circle has a surge, but the health state does not have too much decrease, but the stability degree decreases a little.

[0080] Referring to FIG. 3I to FIG. 3J, which are evaluation result patterns of an embodiment of the present application, taking each action of the automated mechanical arm as an example, it can be seen from the figures that when each action contains action one and action two, from the overall dynamic similarity, it can be seen that the real-time sensing signal (white line part) has a continuous surge compared with the mechanical learning action (yellow line part), but the health status does not have a large decline.

[0081] In summary, the present application provides an automated equipment evaluation optimization method and system, which operates each periodic signal in the sensing signal and generates an evaluation result according to the weight, thereby obtaining the real-time health status of the machine and whether there is an abnormal situation, which can be applied to all periodic equipment monitoring, has high universality, greatly reduces the construction cost, and achieves the purpose of the present application.

[0082] The above-mentioned is only a preferred embodiment of the present application, but cannot limit the patent protection scope of the present application; therefore, any simple equivalent change and modification made according to the content of the claims and the specification of the present application still falls within the patent protection scope of the present application.

[0083]

Symbol Description

[0084] 1 Operation processing module

[0085] 2 Monitoring module

[0086] 3 Display interface

[0087] 31 Equipment overall score area

[0088] 32 Equipment state management area

[0089] 33 Equipment operation trend management area

[0090] 4 Monitoring device

[0091] D Equipment

Claims

1. An automated equipment evaluation optimization method, characterized by, The steps comprise: obtaining judgment information of the equipment and corresponding initial sensing signals; extracting periodic signals from the initial sensing signals; operating the periodic signals to obtain at least one initial characteristic parameter; operating the judgment information and the at least one initial characteristic parameter to obtain corresponding initial weight information; and establishing an evaluation model with the at least one initial characteristic parameter and the initial weight information.

2. The automated equipment evaluation optimization method of claim 1, wherein, The steps comprise: extracting sensing signals; operating the sensing signals to obtain corresponding at least one characteristic parameter; and inputting the at least one characteristic parameter into the evaluation model to generate corresponding evaluation results.

3. The automated equipment evaluation optimization method of claim 1, wherein, The sensing signals and the initial sensing signals are selected from vibration signals, current signals, pressure signals, torque signals, or combinations thereof.

4. The automated equipment evaluation optimization method of claim 1, wherein, The at least one characteristic parameter and the at least one initial characteristic parameter are selected from amplitudes, frequencies, phases, wavelets, standard deviations, stabilities, or combinations thereof.

5. The automated equipment evaluation optimization method of claim 1, wherein, The evaluation results comprise operation analysis information, abnormality detection information, and operation trend information of the equipment.

6. An automated equipment evaluation optimization system, characterized by, The steps comprise: an operation processing module that obtains sensing signals of the equipment and operates according to the sensing signals to obtain at least one characteristic parameter; and a monitoring module that is in signal connection with the operation processing module, receives and evaluates according to the at least one characteristic parameter to obtain corresponding evaluation results.

7. The automated equipment evaluation optimization system of claim 6, wherein, A display interface that is in signal connection with the monitoring module to display the evaluation results, the display interface comprising an equipment overall score area, an equipment state management area, and an equipment operation trend management area, the equipment overall score area displaying operation analysis information of the equipment, the equipment state management area displaying abnormality detection information, and the equipment operation trend management area displaying operation trend information.

8. The automated equipment evaluation optimization system of claim 6, wherein, A monitoring device that is in signal connection with the operation processing module to detect the equipment to obtain corresponding sensing signals, wherein the sensing signals are selected from vibration signals, current signals, pressure signals, torque signals, or combinations thereof.

9. The automated equipment evaluation optimization system of claim 6, wherein, The at least one characteristic parameter is selected from amplitudes, frequencies, phases, wavelets, standard deviations, stabilities, or combinations thereof or combinations thereof.

10. The automated equipment evaluation optimization system of claim 6, wherein, The operation processing module obtains judgment information of the equipment and corresponding initial sensing signals, operates from the initial sensing signals to extract periodic signals, obtains the at least one initial characteristic parameter and corresponding initial weight information from the judgment information and the periodic signals, establishes an evaluation model according to the at least one initial characteristic parameter and the initial weight information, and transmits the evaluation model to the monitoring module, so that the monitoring module evaluates the sensing signals through the evaluation model to obtain corresponding evaluation results.

Images