KR20210055238A - System and method for analyzing failure of linkage type of MES of automated system - Google Patents

Abstract

A failure analysis system and method capable of analyzing causes of failure of an automation facility in connection with a manufacturing execution system shop floor (MES) of the automation facility are provided. An MES-linked failure analysis method according to an embodiment of the present invention comprises the steps of: collecting, by a data collection unit, data input to or output from each control system individually controlling each facility; generating, by a data learning unit, each of the collected data by machine learning to generate each piece of model data; predicting, by a data prediction unit, the next data individually on the basis of each piece of model data; determining, by a facility abnormality determination unit, the abnormality of contact points constituting each facility on the basis of each piece of model data and each piece of prediction data in units of cycles composed of the contact points; and inferring, by a cause diagnosis unit, the cause of failure of the cycles in an abnormal state by using the characteristics of the model data used to determine whether a specific cycle is abnormal. Accordingly, a failure analysis means of an automated facility automatically performs the analysis of the cause of failure of a production facility, and results can be linked with the existing MES. Thus, it is possible to reduce facility maintenance time and facility downtime, and troubleshooting time and cost can be saved. In addition, whether or not contact points constituting each facility is abnormal can be determined in the units of cycles composed of a plurality of contact points, and thus it is possible to improve the efficiency of facility abnormality determination.

Description

System and method for analyzing failure of linkage type of MES of automated system}

The present invention relates to a failure analysis system and method of an automated facility, and more particularly, in connection with the MES (Manufacturing Execution System Shop Floor) of an automated facility, to a failure analysis system and method capable of analyzing the cause of the failure of the automated facility. About.

In order to automatically control process equipment, conventionally, as shown in Fig. 1, when a failure occurs in a process automation facility using a PLC, a highly skilled technician who can maintain the facility directly goes to the automation facility in the field and uses a laptop computer. And by monitoring the PLC program, which is the controller of the automation facility, by using various test machines, the part where the failure occurred and analyzed the cause.

However, when a factory installed with an automation facility is located in a remote location such as overseas, it takes a lot of time and cost because a highly skilled technician must move for repair in order to repair the automation facility.

In addition, in the case of arranging high-level technicians in a number of factories, a problem arises that it takes a lot of time and cost for technician training.

In addition, in the event of a facility failure, in order to check the contents of the failure analysis, there is an inconvenience of having to go to the front of the failed facility and look at the monitor to confirm.

Therefore, there is a need to find a way to perform the failure cause analysis of the production facility in the failure analysis means of the automated facility and link the result with the existing MES.

The present invention was conceived to solve the above problems, and an object of the present invention is to automatically perform a failure cause analysis of a production facility in a failure analysis means of an automated facility, and link the result with the existing MES. It is to provide an MES-linked failure analysis system and method of an automated facility.

Another object of the present invention is to provide an MES-linked failure analysis system and method of an automated facility capable of determining whether or not a plurality of contact points constituting each facility is abnormal in a cycle unit composed of a plurality of contact points.

According to an embodiment of the present invention for achieving the above object, the MES-linked failure analysis method includes the steps of individually collecting data input or output from each control system that individually controls each facility by a data collection unit. ; Generating, by a data learning unit, each of the collected data by machine learning; The data prediction unit individually predicting next data based on each model data; Determining, by a facility abnormality determination unit, whether or not a plurality of contact points constituting each facility is abnormal based on each model data and each prediction data in a cycle unit composed of a plurality of contact points; And when determining that a specific cycle is an abnormal state, the cause diagnosis unit infers a cause of a failure of the abnormal state of the cycle by using the characteristics of the model data used to determine the abnormality.

In addition, in the collecting step, input or output data including tag information for identification of each control system may be collected.

In addition, each of the control systems may allow tag information to be automatically registered in the operation tag list of the data collection unit at the start of the operation.

In addition, in the collecting step, the data collection unit checks the operation status of the control systems automatically registered in the operation tag list according to a preset period, and separates tag information of the control systems whose operation is currently stopped from the operation tag list, In the step of moving to the interruption tag list, transferring the moved interruption tag list to the facility abnormality determination unit, and determining, the equipment abnormality determination unit preferentially determines whether or not the cycles controlled by the control system included in the interruption tag list are abnormal. can do.

In addition, the characteristic of the model data may be at least one of a contact order, a contact appearance frequency, a hit rate of predicted data, and a degree of association between contacts.

In addition, the determining step may determine whether the cycles are abnormal based on whether the prediction data is present or whether the prediction data is hit.

And the determining step is, if the model data includes characteristics that proceed in a specific contact order with respect to a plurality of contact points constituting one cycle, if the predicted data does not exist, it is registered in the unpredictable data list, and the specific contact point When a contact point different from the order is performed and the prediction data does not hit, it can be registered in the prediction failure data list.

In addition, in the predicting step, the priority of the data registered in the unpredictable data list and the predicted failure data list is determined, but the data registered in the unpredictable data list and the predicted failure data list are rearranged in consideration of the characteristics of the model data. In this case, high priority can be given to data with a high probability of a cause of failure.

In addition, in the MES-linked failure analysis method according to an embodiment of the present invention, when the monitoring unit determines a specific cycle as a normal state according to whether or not a specific cycle is abnormal, operation information of the control system for controlling the normal cycle And outputting failure cause information of the cycle in the abnormal state when it is determined that the specific cycle is in an abnormal state.

Further, in the outputting step, at the request of the administrator, the operation time and the number of operations operated during a preset period may be output for each control system.

In the outputting step, at the request of the administrator, the real-time operation state of each control system may be individually output.

In addition, in the outputting step, when it is determined that a specific cycle is in an abnormal state, information on a cause of failure of a cycle in an abnormal state and information about an abnormality in another cycle of a control system controlling the cycle in an abnormal state may be output together.

In the outputting step, at the request of an administrator, history information for a cycle determined as an abnormal state among a plurality of cycles controlled by each control system may be output.

Meanwhile, it goes without saying that the technical idea of the present invention can be applied to a computer-readable recording medium containing a computer program for performing functions of the apparatus and method according to the present embodiment. Further, the technical idea according to various embodiments of the present invention may be implemented in the form of a computer-readable code recorded on a computer-readable recording medium. The computer-readable recording medium can be any data storage device that can be read by a computer and can store data. For example, a computer-readable recording medium may be a ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical disk, hard disk drive, or the like. In addition, a computer-readable code or program stored in a computer-readable recording medium may be transmitted through a network connected between computers.

In addition, according to another embodiment of the present invention, the MES-linked failure analysis system includes: a data collection unit that individually collects data input or output from each control system that individually controls each facility; A data learning unit that generates each model data by machine learning the collected data; A data prediction unit that individually predicts next data based on each model data; A facility abnormality determination unit that determines whether or not a plurality of contact points constituting each facility is abnormal based on each model data and each predicted data in a cycle unit composed of a plurality of contact points; And a cause diagnosis unit for inferring a cause of a failure of the cycle in the abnormal state by using the characteristics of the model data used to determine whether the specific cycle is abnormal.

As described above, according to the embodiments of the present invention, the failure analysis means of the automated facility automatically performs the failure cause analysis of the production facility, and the result can be linked with the existing MES, so that the facility maintenance time and the facility cost are reduced. The same time can be shortened, and trouble-shooting time and cost can be saved.

According to embodiments of the present invention, it is possible to determine whether or not a plurality of contact points constituting each facility is abnormal in a cycle unit composed of a plurality of contact points, thereby improving the efficiency of facility abnormality determination.

1 is a diagram provided for explanation of a response method in case of failure of an existing automation facility,
2 to 3 are views provided to explain the MES-linked failure analysis system and method according to an embodiment of the present invention;
4 is a view provided to explain the configuration of the MES-linked failure analysis system according to an embodiment of the present invention;
5 is a flow chart provided for explaining the MES-linked failure analysis method according to an embodiment of the present invention, and
6 to 7 are diagrams illustrating a state in which a monitoring result of a facility is output using the MES-linked failure analysis method according to an embodiment of the present invention.

Hereinafter, the present invention will be described in more detail with reference to the drawings.

2 to 3 are views provided for explanation of the MES-linked failure analysis system 100 and method according to an embodiment of the present invention, and FIG. 4 is a MES-linked failure analysis system according to an embodiment of the present invention ( 100) is a diagram provided in the description of the configuration.

The MES-linked failure analysis system 100 according to an embodiment of the present invention can automatically perform a failure cause analysis of a production facility and link the result with the existing MES 10, and configure each facility. Whether or not the plurality of contact points are abnormal may be determined in a cycle unit composed of a plurality of contact points.

To this end, the MES-linked failure analysis system 100 includes a control system 110, a data collection unit 120, a data learning unit 130, a data prediction unit 140, a facility abnormality determination unit 150, and a cause. A diagnosis unit 160, a monitoring unit 170, and a storage unit 180 may be included.

The control system 110 may individually control each facility, collect data for each facility, and transmit it to the data collection unit 120.

In addition, the control system 110, by adding unique tag information to the data to be transmitted to the data collection unit 120, the data collection unit 120, to identify the data received from which control system (110). You can do it.

To this end, the control system 110 includes a sensor 111 that collects data on equipment connected for control, a tag unit 112 in which unique tag information is stored, and a processor 113 that adds tag information to the collected data. ), and a communication unit 114 that transmits the data to which the tag information is added to the data collection unit 120.

The data collection unit 120 may individually collect data input or output from each control system 110 that individually controls each facility.

At this time, each control system 110 adds tag information to data notifying that the operation is started, and transmits the tag information to the data collection unit 120, so that the tag information of the control system 110 is transferred to the data collection unit ( 120) can be automatically registered in the operation tag list.

In addition, the data collection unit 120 checks the operation status of the control systems 110 automatically registered in the operation tag list according to a preset period, and stores the tag information of the control systems 110 in which the operation is currently stopped. Separated from, it may be moved to the stop tag list, and the moved stop tag list may be transmitted to the facility abnormality determination unit 150.

The data learning unit 130 may generate each model data by machine learning the collected data.

The data prediction unit 140 may individually predict next data based on each model data.

The control system 110 is configured to control a plurality of contact points in order to automate the facility. Specifically, for example, if a task performed by a specific facility is composed of four individual contact points of A, B, C, D , A, B, C, D, each contact point can be controlled so that the task is performed in the order of ABCD.

In this case, the data prediction unit 140 may predict next data based on the model data generated by machine learning by the data learning unit 130.

That is, the model data includes the characteristic that the contact points proceed in the order of ABCD, and the data prediction unit 140 monitors each control system 110 in real time based on each model data, and individually predicts the next data. You can proceed with

For example, when it is confirmed that the contact A is performed, the data prediction unit 140 predicts that the next contact B will proceed, and predicts input/output data corresponding to the contact B.

In this way, the data prediction unit 140 may generate prediction data by predicting a contact point to be performed next based on the model data generated by the data learning unit 130.

Here, each facility may include a plurality of automated process cycles, and each cycle may be configured to include a plurality of contact points.

The facility abnormality determination unit 150 may determine whether each facility is abnormal based on each model data and each prediction data.

For example, the facility abnormality determination unit 150 may determine whether cycles are abnormal based on whether the prediction data is present or whether the prediction data is hit.

In addition, the facility abnormality determination unit 150 determines whether or not a plurality of contact points constituting each facility is abnormal based on each model data and each prediction data, in a cycle unit consisting of a plurality of contact points, It is possible to improve the efficiency of abnormality judgment.

This means that when the number of unit contact points is composed of thousands, it is data from the facility abnormality determination unit 150 to determine the abnormality in a cycle unit including a plurality of contact points rather than the equipment abnormality determination unit to determine the abnormality in a contact unit. This is because the throughput is much reduced.

In addition, the facility abnormality determination unit 150 registers in the unpredictable data list when the model data includes characteristics that proceed in a specific contact order with respect to a plurality of contact points constituting one cycle, and if predicted data does not exist, , When a contact point different from the specific contact order is performed and the prediction data does not hit, it may be registered in the prediction failure data list.

In this case, the cause diagnosis unit 160 determines the priority of the data registered in the unpredictable data list and the prediction failure data list, but determines the characteristics of the model data by using the data registered in the unpredictable data list and the prediction failure data list. By taking into account and rearranging, high priority can be given to data with a high probability of failure.

In addition, the facility abnormality determination unit 150, when the interruption tag list is received, firstly determines whether or not the cycles controlled by the control system 110 included in the received interruption tag list are abnormal, so that the efficiency of equipment abnormality determination Can improve.

When the cause diagnosis unit 160 determines a specific cycle as an abnormal state, the cause diagnosis unit 160 can infer the cause of the failure of the abnormal cycle by using the characteristics of the model data used to determine the abnormality. have.

The characteristic of the model data may be at least one of a contact order, a contact appearance frequency, a hit rate of predicted data, and a degree of association between contact points.

That is, in the model data, reference information on the contact order in which the contact points proceed in the same order as ABCD, the frequency of appearance of each contact point such as A, B, C, B, and data predicted by the data prediction unit 140 (or contact point ) May include characteristics such as the hit rate that is actually performed, the degree of association between the contact A and the contact B or the contact C.

That is, the cause diagnosis unit 160 determines that the abnormal state is abnormal and rearranges the list-up abnormal data in consideration of the characteristics of the model data, and gives high priority to data with a high probability of a failure.

Specifically, the cause diagnosis unit 160 determines the priority of the data registered in the unpredictable data list and the prediction failure data list, but determines the characteristics of the model data by using the data registered in the unpredictable data list and the prediction failure data list. By taking into account and rearranging, high priority can be given to data with a high probability of failure.

The monitoring unit 170 outputs operation information of the control system 110 for controlling the cycle in the normal state, and determines the specific cycle as an abnormal state when determining a specific cycle as a normal state according to whether or not a specific cycle is abnormal. In this case, information on the cause of failure of the cycle in an abnormal state can be output.

For example, the monitoring unit 170 may display the fault region on a three-dimensional image based on the analyzed cause of the fault, and mark the fault region with text so that the operator can more intuitively and quickly identify the fault region.

In addition, the monitoring unit 170 may output the operation time and the number of operations operated during a preset period for each control system 110 at the request of the administrator.

In addition, the monitoring unit 170 may individually output a real-time operation state of each control system 110 upon the request of the administrator.

In addition, when determining that a specific cycle is in an abnormal state, the monitoring unit 170 may output information on the cause of failure of the abnormal cycle and whether other cycles of the control system 110 controlling the cycle in the abnormal state are abnormal. have.

In addition, the monitoring unit 170 may output history information on a cycle determined to be an abnormal state among a plurality of cycles controlled by each control system 110 at the request of the manager.

The storage unit 180 is a storage medium capable of storing programs and data necessary for the operation of the MES-linked failure analysis system 100.

For example, the storage unit 180 may store machine learning data, model data and prediction data, and data for each control system 110.

5 is a flowchart provided to explain the MES-linked failure analysis method according to an embodiment of the present invention, and FIGS. 6 to 7 are diagrams of a facility using the MES-linked failure analysis method according to an embodiment of the present invention. This is a diagram illustrating the output of the monitoring result.

5 to 7, in the MES-linked failure analysis method according to the present embodiment, the data collection unit 120 is input or output from each control system 110 that individually controls each facility. When data is individually collected (S510), the data learning unit 130 may generate each model data by machine learning the collected data (S520).

And when the data prediction unit 140 individually predicts the next data based on each model data and generates each prediction data (S530), the facility abnormality determination unit 150, each model data and each Based on the predicted data of, it may be determined whether or not a plurality of contact points constituting each facility are abnormal in a cycle unit composed of a plurality of contact points (S540).

In addition, when a specific cycle is determined to be an abnormal state (S550-N), the cause diagnosis unit 160 may infer the cause of the failure of the abnormal cycle by using the characteristics of the model data used to determine whether the abnormality is present. (S570).

That is, when determining a specific cycle as a normal state according to whether or not a specific cycle is abnormal (S550-Y), the monitoring unit 170 outputs operation information of the control system 110 for controlling the normal cycle ( S560), when it is determined that the specific cycle is in an abnormal state (S550-N), the monitoring unit 170 may output information on the cause of the failure of the cycle in the abnormal state (S580).

At this time, the monitoring unit 170, as illustrated in FIG. 6, may output whether or not a specific facility (ex. a fusion facility of a vehicle door trim) is abnormal, for each automation process cycle, and as illustrated in FIG. 7 The past failure history of (ex. car door trim fusion facility) can be output.

Through this, in order to save trouble resolution time and cost, the failure analysis means of the automated facility automatically performs the failure cause analysis of the production facility, and the result can be linked with the existing MES, It is possible to determine whether or not the contact points of are abnormal in a cycle unit composed of a plurality of contact points, thereby improving the efficiency of equipment abnormality determination.

Meanwhile, it goes without saying that the technical idea of the present invention can be applied to a computer-readable recording medium containing a computer program for performing functions of the apparatus and method according to the present embodiment. Further, the technical idea according to various embodiments of the present invention may be implemented in the form of a computer-readable code recorded on a computer-readable recording medium. The computer-readable recording medium can be any data storage device that can be read by a computer and can store data. For example, a computer-readable recording medium may be a ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical disk, hard disk drive, or the like. In addition, a computer-readable code or program stored in a computer-readable recording medium may be transmitted through a network connected between computers.

In addition, although preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above, and the technical field to which the present invention belongs without departing from the gist of the present invention claimed in the claims In addition, various modifications are possible by those of ordinary skill in the art, and these modifications should not be understood individually from the technical spirit or prospect of the present invention.

10: Existing MES system
100: MES linkage failure analysis system
110: control system
111: sensor
112: tag part
113: processor
114: communication department
120: data collection unit
130: data learning department
140: data prediction unit
150: equipment abnormality determination unit
160: cause diagnosis unit
170: monitoring unit
180: storage unit

Claims (14)

A step of individually collecting, by a data collection unit, data input or output from each control system that individually controls each facility;
Generating, by a data learning unit, each of the collected data by machine learning;
The data prediction unit individually predicting next data based on each model data;
Determining, by a facility abnormality determination unit, whether or not a plurality of contact points constituting each facility is abnormal based on each model data and each prediction data in a cycle unit composed of a plurality of contact points; And
When determining a specific cycle as an abnormal state, the cause diagnosis unit, using the characteristics of the model data used to determine the abnormality, inferring the cause of the failure of the cycle in the abnormal state; MES-linked failure analysis method comprising a.
The method according to claim 1,
The steps to collect are:
MES-linked failure analysis method, characterized in that collecting input or output data including tag information for identification of each control system.
The method according to claim 2,
Each control system,
MES-linked failure analysis method, characterized in that at the start of the operation, the tag information is automatically registered in the operation tag list of the data collection unit.
The method of claim 3,
The steps to collect are:
The data collection unit checks the operation status of the control systems automatically registered in the operation tag list according to a preset period, separates the tag information of the control systems whose operation is currently stopped from the operation tag list, and moves it to the stop tag list, Transfers the list of moved hang tags to the facility abnormality determination unit,
The step of judging is,
An MES-linked failure analysis system, characterized in that the facility abnormality determination unit first determines whether or not cycles controlled by the control system included in the interruption tag list are abnormal.
The method of claim 4,
The characteristics of model data are:
MES-linked failure analysis method, characterized in that at least one of a contact order, a contact appearance frequency, a hit rate of predicted data, and a degree of association between contacts.
The method of claim 5,
The step of judging is,
An MES-linked failure analysis method, characterized in that determining whether cycles are abnormal based on the presence of predicted data or whether the predicted data is hit.
The method of claim 6,
The step of judging is,
If the model data includes characteristics that proceed in a specific contact order with respect to a plurality of contact points constituting one cycle, if the predicted data does not exist, it is registered in the unpredictable data list, and a contact different from the specific contact order is performed. Thus, when the prediction data does not hit, the MES-linked failure analysis method, characterized in that registering in the prediction failure data list.
The method of claim 7,
The predicting step is,
Data with high probability of failure by determining the priority of data registered in the unpredictable data list and the predicted failure data list, and rearranging the data registered in the unpredictable data list and the predicted failure data list in consideration of the characteristics of the model data. MES-linked failure analysis method, characterized in that giving a high priority to the.
The method according to claim 2,
The monitoring unit outputs operation information of the control system controlling the normal cycle when a specific cycle is determined as a normal state according to whether a specific cycle is abnormal, and when a specific cycle is determined as an abnormal state, the abnormal cycle Outputting the cause of failure information; MES-linked failure analysis method further comprising a.
The method of claim 9,
The output step is,
At the request of a manager, the MES-linked failure analysis method, characterized in that outputting the operation time and the number of operations operated during a preset period for each control system.
The method of claim 9,
The output step is,
MES-linked failure analysis method, characterized in that at the request of the manager, the real-time operation status of each control system individually output.
The method of claim 9,
The output step is,
When it is determined that a specific cycle is in an abnormal state, the failure cause information of the abnormal state of the cycle and the abnormality of other cycles of a control system controlling the abnormal state of the cycle are output together.
The method of claim 9,
The output step is,
At the request of a manager, MES-linked failure analysis method, characterized in that outputting history information on cycles determined to be abnormal among a plurality of cycles controlled by each control system.
A data collection unit that individually collects data input or output from each control system that individually controls each facility;
A data learning unit that generates each model data by machine learning the collected data;
A data prediction unit that individually predicts next data based on each model data;
A facility abnormality determination unit that determines whether or not a plurality of contact points constituting each facility is abnormal based on each model data and each predicted data in a cycle unit composed of a plurality of contact points; And
MES-linked failure analysis system comprising; when a specific cycle is determined to be an abnormal state, a cause diagnosis unit that infers the cause of the failure of the abnormal state by using the characteristics of the model data used to determine the abnormality.

Images