Production line equipment abnormal behavior judgment method based on correlation analysis

Abstract

The invention relates to the field of production line protection and maintenance, and discloses a method for judging abnormal behaviors of production line equipment based on correlation analysis. A traditional fault diagnosis method has certain limitations, is mainly based on artificial experience and fault codes, and lacks real-time performance and accuracy. The method is mainly carried out from two stages of fault detection and fault diagnosis, and comprises the following steps: switching from a power supply end, monitoring the correlation of electric energy output signals in real time, identifying faults by calculating correlation coefficients and setting a fault threshold value, quickly and accurately positioning fault equipment, and effectively capturing non-transient faults occurring in the change process of the electrical equipment. The method is characterized in that the method is high in accuracy and timeliness, is not interfered by equipment, is suitable for production environments under various rhythm work, and provides reliable support for troubleshooting and maintenance of a production line.

Description

Technical Field:

[0001] The present invention relates to the field of production line protection and maintenance, and particularly focuses on various types of faults that may occur during the rhythm operation of equipment. Specifically, it relates to a method for judging abnormal behaviors of production line equipment based on correlation analysis. Background Art:

[0002] Fault diagnosis of electrical equipment in industrial production lines is a core link to ensure the smooth operation of production lines and improve production efficiency. Given the continuous improvement of industrial automation levels, production line equipment has become increasingly complex and diverse. Therefore, fault diagnosis work has become increasingly important. The main goal of fault diagnosis is to quickly and accurately identify and eliminate potential fault hazards in production line equipment or systems. Generally speaking, fault diagnosis of industrial production lines plays an indispensable role in maintaining the stable operation of production lines, improving production efficiency, and reducing maintenance costs. With the help of scientific and effective fault diagnosis methods and technical means, enterprises can timely discover and solve various problems, thereby enhancing the overall competitiveness and production efficiency of production lines.

[0003] Traditionally, the operating state of equipment has often been judged based on manual experience. This method mainly relies on the historical maintenance information and current operating state of the equipment, has strong subjectivity, and is easily affected by equipment operating intensity and environmental factors. Therefore, the fault diagnosis efficiency is relatively low, and there is a risk of triggering serious production safety accidents. Currently, production line fault diagnosis mainly uses fault codes as the basis. However, this method lacks real-time performance and cannot give early warnings of the occurrence of faults. In addition, for electrical equipment faults on production lines, in addition to sudden hard faults (such as open circuits and short circuits), more are formed by the gradual accumulation of abnormal phenomena over a long period of time. Existing diagnosis methods are difficult to effectively reflect the situation of non-instantaneous faults during the change process of electrical equipment, and there are great limitations. Summary of the Invention:

[0004] The purpose of the present invention is to overcome the limitations of the prior art and particularly propose a method for judging abnormal behaviors of production line equipment based on correlation analysis. The entire diagnosis process of this method starts from the power supply end, and its characteristic is that it is not affected by the mutual interference between devices, reduces the cost of fault diagnosis, and shows strong environmental adaptability. It aims to solve the problem of poor real-time performance in traditional quasi-fault detection and challenges such as difficulty in effectively capturing non-instantaneous faults during the change process of electrical equipment. By introducing the correlation analysis technology of power output signals under rhythm work, the accuracy and timeliness of fault diagnosis can be improved, providing more reliable technical support for production line fault troubleshooting and maintenance.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] The method for judging abnormal behavior of production line equipment based on correlation analysis includes the following steps:

[0007] Step 1. When the production line is running stably and the equipment is in normal working condition, the time required to complete the entire set production process is defined as a working beat. Each device on the production line performs collaborative operations according to this fixed beat. The detection device is responsible for capturing and monitoring the electrical energy output signals supplied to each device in real time. During a normal working beat, the sampled electrical energy output signal of this working beat is selected as the initial reference signal and then stored inside the detection device;

[0008] Step 2. Using the initial reference signal and the real-time collected electrical energy output signals in the detection device, calculate the correlation coefficient within a fixed time window to obtain the real-time correlation coefficient of each working beat;

[0009] Step 3. According to the 3σ threshold criterion, conduct in-depth analysis on the real-time correlation coefficient and set a reasonable fault threshold accordingly. Once the real-time monitored correlation coefficient is lower than the preset fault threshold, it means that a fault has occurred in the production line;

[0010] Step 4. During the normal operation of the production line, once a fault is detected, the detection device will immediately activate the fault source identification program and quickly lock the working beat that matches the abnormal signal. Since the running time sequence and electrical energy requirements of each device in the working beat are different, within the set time window, the waveforms of the electrical energy output signals corresponding to each device will show significant differences and different generation sequences. Therefore, the detection device will use an accurate correlation analysis method to carefully distinguish the waveforms of the electrical energy output signals within the time window to ensure that the specific waveform of the electrical energy output signal that has changed can be accurately identified. When the detection device finds that the waveform corresponding to the device is abnormal within the time window, the waveform will be marked with a dotted line. Based on the identified changed waveform, the detection device will accurately locate the specific device with the fault. Subsequently, the detection device will follow the preset program logic and emit a flashing alarm signal to prompt relevant personnel to handle it in a timely manner.

[0011] Further, in the above Step 1, the normal working beat refers to the time required for each device on the production line to complete the entire set production process without any faults.

[0012] Further, without human intervention, the initial reference signal will be continuously used by the system as a reference standard. When the production line undergoes necessary maintenance operations, in order to ensure the accuracy and stability of the production process, it will be necessary to re-sample and update the initial reference signal accordingly.

[0013] Further, the method and device for judging abnormal behavior of production line equipment based on correlation analysis are applicable to various beat working environments. On the premise that the production line runs stably and the equipment is in normal working condition, there are many sub-processes in its overall process. The present invention can be applied not only to the entire production line process, but also to each corresponding sub-process link inside it.

[0014] Further, the electric energy output signal refers to the signal of the electric energy output from a power source or a power system. In a system, the electric energy output signal indicates the operating state of the power source or the power system, such as the changes in parameters such as current, voltage, and power.

[0015] Further, in the second step, when the time window collects the electric energy output signal in real time under the same working beat, its corresponding length is fixed.

[0016] Further, in the second step, the calculation formula of the correlation coefficient is as follows:

[0017] R = ρ ab ×p (1)

[0018]

[0019] where a, b = 1, 2,......T, P a,n and P b,n correspond to the nth data point of the ath and bth time domain sample quantities respectively, N is the number of data points of each time domain sample quantity, and are the mean values of the data points of the ath and bth time domain sample quantities respectively; R is the real-time correlation coefficient; p is the correlation coefficient calculation adjustment factor.

[0020] Further, in the third step, the 3σ threshold criterion formula is as follows:

[0021]

[0022] Compared with the prior art, the beneficial effects of the present invention are:

[0023] The entire diagnostic process of the present invention starts from the power supply end, and proposes a method for in-depth correlation analysis of the electrical energy output signal under rhythmic operation, realizing the fault diagnosis of various electrical equipment on the production line. It is characterized by being unaffected by the mutual interference between devices, reducing the cost of fault diagnosis, and showing strong environmental adaptability. The present invention is applicable to various rhythmic working environments. On the premise that the production line operates stably and the equipment is in a normal working state, there are many sub-processes in its overall process. The present invention can be applied not only to the entire production line process, but also to each corresponding sub-process link inside it. In the process of deriving the correlation coefficient, the present invention innovatively proposes a Pearson improvement factor to improve the accuracy of the detection process. Compared with traditional fault detection means and the current mainstream fault code method on the production line, the present invention can monitor the fault condition of the production line in real time and effectively capture non-instantaneous faults that occur during the change process of electrical equipment. BRIEF DESCRIPTION OF THE DRAWINGS:

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments:

[0025] Figure 1 . Flowchart of the method for judging abnormal behavior of production line equipment based on correlation analysis;

[0026] Figure 2 . Flowchart of the application method in the fault detection stage;

[0027] Figure 3 . Flowchart of the application method in the fault location stage;

[0028] Figure 4 . Production line simulation diagram;

[0029] Figure 5 . Initial reference signal diagram;

[0030] Figure 6 . Waveform diagram of the electrical energy output signal corresponding to each device within the time window;

[0031] Figure 7 . Analysis diagram of the real-time correlation coefficient of production line faults;

[0032] Figure 8 . Analysis diagram of the electrical energy output signal at the fault moment. DETAILED DESCRIPTION OF THE EMBODIMENTS:

[0033] The following will further illustrate the present invention in conjunction with the drawings and examples:

[0034] See Figure 1 The specific implementation manner of the present invention is described in detail as follows:

[0035] Step 1. Enter the production line fault detection phase. When the production line is running stably and the equipment is in a normal working state, the time required to complete the entire production process is defined as a working cycle. Each device on the production line performs collaborative operations according to this fixed cycle. The detection device is responsible for capturing and monitoring the electrical energy output signals supplied to each device in real time. During a normal working cycle, the sampled electrical energy output signal of this working cycle is selected as the initial reference signal and then stored inside the detection device.

[0036] Step 2. Use the initial reference signal and the real-time collected electrical energy output signals in the detection device to calculate the Pearson correlation coefficient within a fixed time window, and accurately calculate the real-time correlation coefficient of each cycle. The specific formula for the Pearson correlation coefficient is as follows:

[0037] R = ρ ab ×p(1)

[0038]

[0039] where s, t = 1, 2,......T, P s,n and P t,n correspond to the nth data point of the sth and tth time domain sample sizes respectively, N is the number of data points in each time domain sample size, and are the mean values of the data points of the sth and tth time domain sample sizes respectively; R is the real-time correlation coefficient; p is the Pearson correlation coefficient calculation adjustment factor.

[0040] Step 3. According to the 3σ threshold criterion, conduct an in-depth analysis of the real-time correlation coefficient and set a reasonable fault threshold accordingly. Once the real-time monitored correlation coefficient is lower than the preset fault threshold, it means that a fault has occurred in the production line. The specific formula for the 3σ threshold criterion is as follows:

[0041]

[0042] Step 4. Conduct the production line fault diagnosis part. During the normal operation of the production line, once a fault is detected, the detection device will immediately activate the fault source identification program and quickly lock the working rhythm that matches the abnormal signal. Given that the operating time sequence and power demand of each device in the working rhythm are different, within the set time window, the waveform of the power output signal corresponding to each device will show significant differences and different generation sequences. Therefore, the detection device will use an accurate correlation analysis method to carefully distinguish the waveform of the power output signal within the time window to ensure that the specific changed waveform of the power output signal can be accurately identified. When the detection device finds that the waveform corresponding to a certain device is abnormal within the time window, it will mark the waveform with a dotted line. Based on the identified changed waveform, the detection device will accurately locate the specific device with a fault. Subsequently, the detection device will follow the preset program logic and emit a flashing alarm signal to prompt relevant personnel to handle it in a timely manner.

[0043] Scheme verification: The following is an example process of the production line fault diagnosis by this method. First, we built a simulated production line (as Figure 4 shown). Suppose when the production line is running stably and the equipment is in a normal working state, the equipment configuration on the production line includes Equipment 1, Equipment 2, and Equipment 3. Each equipment collaborates according to a fixed rhythm to ensure the smooth progress of the production process. Taking the continuous operation of these equipments for fifteen rhythms as an example, it is artificially set that Equipment 3 has a fault during the tenth to the fifteenth rhythm. Based on the above settings, the following steps are an example of the fault diagnosis process of this method for the production line. First, when the production line is running stably and the equipment is in a normal working state, the detection device is responsible for capturing and monitoring the power output signals supplied to each equipment in real time, and in a normal working rhythm, select the sampled power output signal of this working rhythm as the initial reference signal (as Figure 5 shown), and then store it inside the detection device. The detection device uses the initial reference signal and the real-time collected power output signals to calculate the Pearson correlation coefficient within a fixed time window and accurately calculates the real-time correlation coefficient of each rhythm. According to the 3σ threshold criterion, conduct an in-depth analysis of the real-time correlation coefficient and set a reasonable fault threshold accordingly. As Figure 7 shown, when the real-time monitored correlation coefficient is lower than the preset fault threshold, it means that a fault has occurred in the production line. The detection device will immediately activate the fault source identification mechanism and quickly lock the working rhythm corresponding to the abnormal signal. As Figure 6 shown, within the set time window, the waveforms of the three power output signals respectively correspond to Equipment 1, Equipment 2, and Equipment 3. Therefore, the detection device will use an accurate correlation analysis method to carefully distinguish the waveforms of the power output signals within the time window to ensure that the specific changed waveforms of the power output signals can be accurately identified. As Figure 8As shown, when the detection device finds that the waveform corresponding to device 3 is abnormal within the time window, the waveform of the power output signal will be marked with a dotted line. Based on the identified changed waveform, the detection device will accurately locate the fault of device 3. Subsequently, the detection device will follow the preset program logic and emit a flashing alarm signal to prompt the relevant personnel to handle it in a timely manner. Through the above verification, this method can accurately detect the faults of the production line and accurately locate the specific faulty device.

Claims

1. A method for judging abnormal behaviors of production line equipment based on correlation analysis, characterized in that, It includes the following steps: Step 1. When the production line is running stably and the equipment is in normal working condition, the time required to complete the entire set production process is defined as a working beat. Each device on the production line collaborates according to this fixed beat. The detection device is responsible for capturing and monitoring the electrical energy output signals supplied to each device in real time, and within a normal working beat, selects the sampled electrical energy output signal of this working beat as the initial reference signal, and then stores it inside the detection device; Step 2. In the detection device, using the initial reference signal and the real-time collected electrical energy output signals, calculate the correlation coefficient within a fixed time window to obtain the real-time correlation coefficient of each working beat; Step 3. Based on the 3σ threshold criterion, conduct in-depth analysis on the real-time correlation coefficient, and accordingly set a reasonable fault threshold. Once the real-time monitored correlation coefficient is lower than the preset fault threshold, it means that a fault has occurred in the production line; Step 4. During the normal operation of the production line, once a fault is detected, the detection device will immediately activate the fault source identification program, quickly lock the working beat that matches the abnormal signal. Since the running time sequence and electrical energy requirements of each device in the working beat are different, within the set time window, the waveforms of the electrical energy output signals corresponding to each device will show significant differences and different generation sequences. Therefore, the detection device will use an accurate correlation analysis method to carefully distinguish the waveforms of the electrical energy output signals within the time window to ensure that it can accurately identify the specific waveform of the electrical energy output signal that has changed. When the detection device finds that the waveform corresponding to this device is abnormal within the time window, it will mark the waveform of this electrical energy output signal with a dotted line. Based on the identified changed waveform, the detection device will accurately locate the specific device where the fault occurred. Subsequently, the detection device will follow the preset program logic and issue a flashing alarm signal to prompt relevant personnel to handle it in a timely manner.

2. The method for judging abnormal behavior of production line equipment based on correlation analysis according to claim 1, wherein, In the above Step 1, without human intervention, the initial reference signal will be continuously used by the system as a reference standard. When the production line undergoes necessary repair operations, to ensure the accuracy and stability of the production process, it will be necessary to re-sample and update the initial reference signal accordingly.

3. The method for judging abnormal behavior of production line equipment based on correlation analysis according to claim 1, characterized in that, The described invention is applicable to various beat working environments. On the premise that the production line is running stably and the equipment is in normal working condition, its overall process contains many sub-processes. The present invention can be applied not only to the entire production line process, but also to each corresponding sub-process link inside it.

4. The method for judging abnormal behavior of production line equipment based on correlation analysis according to claim 1, characterized in that, The described electrical energy output signal refers to the signal of the electrical energy output from a power source or a power system. In a system, the electrical energy output signal indicates the operating state of the power source or the power system, such as the changes in parameters such as current, voltage, and power.

5. The method for judging abnormal behavior of production line equipment based on correlation analysis according to claim 1, characterized in that In the above Step 2, when the time window collects the electrical energy output signals in real time under the same working beat, its corresponding length is fixed.

6. The method for judging abnormal behavior of production line equipment based on correlation analysis according to claim 1, characterized in that, In the above Step 2, the correlation coefficient calculation formula is as follows: R = ρ ab × p (1) where a, b = 1, 2,......T, P a,n and P b,n correspond to the n-th data point of the a-th and b-th time-domain sample quantities respectively, N is the number of data points of each time-domain sample quantity, and are the mean values of the data points of the a-th and b-th time-domain sample quantities respectively; R is the real-time correlation coefficient; p is the correlation coefficient calculation adjustment factor.

7. The method for judging abnormal behavior of production line equipment based on correlation analysis according to claim 1, characterized in that, In the above step, the 3σ threshold criterion formula is as follows:

8. The method for judging abnormal behavior of production line equipment based on correlation analysis according to claim 1, characterized in that, Perform the real-time correlation operation and determine whether the real-time correlation coefficient has a mutation. If the real-time correlation coefficient has a mutation and exceeds the set threshold, a fault occurs.

Images