KR102734466B1 - Smart factory control system based on ai technique - Google Patents

Abstract

The present invention is a system for efficiently controlling the operation of equipment using process operation record data of manufacturing equipment of a smart factory, comprising: a data acquisition unit provided in each process device of the manufacturing equipment to acquire process operation record data; and an AI analysis unit that analyzes the process operation record data of each device collected by the data acquisition unit to derive an improved process.

Description

{SMART FACTORY CONTROL SYSTEM BASED ON AI TECHNIQUE}

The present invention relates to a smart factory manufacturing process control system, and more particularly, to a system that effectively controls a manufacturing process performed in a factory using an IT device.

The four core technologies required to implement smart factories are “big data, cloud, artificial intelligence, and robots,” and the main features of implementing smart factories are connectivity, flexibility, and intelligence.

Connectivity means managing all manufacturing-related processes in real time, and flexibility means data-based decision-making for flexible response to manufacturing environment fluctuations. And intelligence means minimizing unverified human intervention to achieve uniformity of quality and optimization of operations.

Recently, in order to implement such smart factories, existing manufacturing operation management data generated at manufacturing sites are collected/managed/analyzed, and the results are integrated for data analysis, and a machine learning model is built based on the analyzed results.

As an example of achieving the value of a conventional smart factory, domestic patent No. 10-2016-0171549 provides a standard block device that enables combination of module units by modularizing process devices. However, in small and medium-sized manufacturing companies, in order to achieve the value of such a smart factory, cost and system risks arise as additional modules are installed in existing facilities or replaced with new facilities.

Therefore, in order to achieve the value of a smart factory without replacing existing facilities and systems, the development of a probability-based machine learning system using data obtained from existing facilities and systems is required.

Patent Document 1: Republic of Korea Publication Patent No. 10-2017-0064291 Publication Date June 9, 2017

The purpose of the present invention is to provide a system that effectively controls a manufacturing process performed in a factory using an IT device.

In order to achieve the above purpose, a smart factory manufacturing process control system according to one embodiment of the present invention is a system that efficiently controls the operation of equipment by using process operation record data of manufacturing equipment of a smart factory, and includes a data acquisition unit provided in each process device of the manufacturing equipment to acquire process operation record data; and an AI analysis unit that analyzes the process operation record data of each device collected by the data acquisition unit to derive an improved process.

At this time, the manufacturing facility is a facility for producing a casting, and may include a supply device for supplying steel and other raw materials, which are raw materials for the casting; and a melting device for melting the steel and other raw materials supplied through the supply device.

In addition, the data acquisition unit may include a raw material recognition sensor that recognizes steel and other raw materials supplied to the supply unit; a temperature detection sensor that recognizes the temperature of the melting vessel and the temperature of the molten raw material; and a strength measurement sensor that measures the strength of the mold.

In addition, the AI analysis unit may include a communication unit that receives measurement data from each sensor of the data acquisition unit; a process data DB that collects and records measurement data from each sensor; and a statistical-based learning derivation unit that analyzes data in the process data DB using an artificial intelligence method based on machine learning techniques to derive an improved process.

According to the smart factory manufacturing process control system according to the present invention, the most efficient resource utilization process conditions can be derived by analyzing and learning operation record data collected from existing manufacturing facilities, and also productivity can be improved through yield improvement and quality management by predicting and preventing malfunctions such as manufacturing facility load and facility failure.

FIG. 1 is a block diagram of a smart factory process control system according to one embodiment of the present invention.
Figure 2 is a block diagram illustrating the relationship between the manufacturing facility and the data acquisition unit.
Figure 3 is a detailed block diagram of the AI analysis unit.
FIG. 4 is a flowchart illustrating an improved process output process performed in the smart factory process control system of the present invention.
Figures 5 and 6 illustrate steps S400 and S500.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the attached drawings. At this time, it should be noted that the same components in the attached drawings are indicated by the same symbols as much as possible. In addition, detailed descriptions of well-known functions and configurations that may obscure the gist of the present invention will be omitted. For the same reason, some components in the attached drawings are exaggerated, omitted, or schematically illustrated.

FIG. 1 is a block diagram of a smart factory process control system according to one embodiment of the present invention, FIG. 2 is a block diagram explaining the relationship between a manufacturing facility and a data acquisition unit, and FIG. 3 is a detailed block diagram of an AI analysis unit.

Referring to FIGS. 1 to 3, the smart factory manufacturing process artificial intelligence control system (1000) according to the present invention is a system that efficiently controls the operation of the equipment by using process operation record data of the manufacturing equipment of the smart factory, and includes a data acquisition unit (200) and an AI analysis unit (300). In the present embodiment, the manufacturing equipment (100) may be equipment for producing a casting.

The manufacturing facility (100) includes a feeder (110) and a melter (120). The feeder (110) is a device that supplies raw materials for castings, such as steel and other raw materials, and supplies raw materials of metal or non-metal materials used for steel ingots and alloys.

The melter (120) melts steel and other raw materials supplied through the supply unit (110). An electric furnace, etc. corresponds to the melter (120).

The data acquisition unit (200) is equipped with a device for each process of the manufacturing facility (100) to acquire process operation record data.

The AI analysis unit (300) analyzes the process operation record data for each device collected from the data acquisition unit (200) to derive an improved process.

Specifically, the data acquisition unit (200) includes a raw material recognition sensor (210), a temperature detection sensor (220), and a strength measurement sensor (230).

The raw material recognition sensor (210) recognizes steel and other raw materials supplied to the feeder (110).

If the feeder (110) is in the form of a silo, the raw material recognition sensor (210) may be provided in the form of a sensor that measures the weight and input amount of each raw material supplied. If the raw material is input manually, the raw material recognition sensor (120) may be replaced by a method in which a computer-processable tag such as NFC is used to recognize the input amount of steel ingots to a reader before the worker inputs the raw material into the melter (120).

If the feeder (110) is in the form of a conveyor belt, the raw material recognition sensor (210) may be provided as a metal detector type that is placed at a certain distance on the conveyor belt. If the steel is in the form of an ingot, it is efficient to transport it by a conveyor belt. The raw material recognition sensor (210) of the metal detector type is placed on the conveyor belt and detects the transported metal ingots one by one. That is, the quantity of ingots passing through is measured by measuring the magnetic reaction according to electric induction.

The temperature detection sensor (220) detects the temperature of the melting vessel (120) and the temperature of the molten raw material. Although the illustration shows that there is only one temperature detection sensor (220), this is only for simplicity, and multiple sensors are placed in the melting vessel (120) and the path through which the molten raw material is fed. In addition to the temperature of the molten raw material, sensors for measuring the temperature, humidity, etc. of the room where the manufacturing facility is placed may be further included.

The strength measurement sensor (230) measures the strength of the mold. The strength of the mold may change depending on the manufacturing process, temperature, and humidity conditions of the mold (M). The strength measurement sensor (230) converts the strength of the mold into a digital number and transmits it to the AI analysis unit (300).

The strength measuring sensor (230) is in the form of a probe, and a plurality of probes arranged at regular intervals are pressed against the mold (M) with a set pressure. The strength measuring sensor (230) applied to the probe can measure the strength of the mold (M) by measuring the amount of displacement. The strength measuring sensor (230) further includes an internal strength measuring device inserted into a mold frame inside the mold (M). The internal strength measuring device includes a balloon-shaped mold insertion tube whose volume increases according to air injection, a pump that injects air into the mold insertion tube at a set pressure, and a pressure sensor that measures the pressure change of the mold insertion tube.

The mold insertion tube is inserted into the mold frame inside the mold (M) in a deflated state and air is injected up to the set pressure. As the air is injected, the mold insertion tube applies force to the mold frame inside the mold (M). At this time, if the mold frame inside the mold (M) is deformed, the volume of the mold insertion tube increases and the pressure decreases. The pressure sensor can measure whether there is deformation inside the mold (M) by measuring whether the pressure decreases.

Meanwhile, although omitted in the city, the data acquisition unit (200) may further include a viscosity measurement sensor that measures the viscosity of the base dough (i.e., clay) for forming the mold (M).

In addition, a plurality of sensors may be further included to measure changes in the mold (M) during the production process of the casting. For example, a sensor for measuring the first injection amount and the second injection amount of the molten metal, and various types of sensors for measuring temperature changes and strength changes in the mold (M) according to the first injection amount and the second injection amount may be further included.

The AI analysis unit (300) includes a communication device (310), a DB (320), and a statistical learning derivation device (330).

The communication device (310) receives measurement data from each sensor of the data acquisition unit (200). That is, it receives measured data from the raw material recognition sensor (210), the temperature detection sensor (220), and the intensity measurement sensor (230). The raw material recognition sensor (210), the temperature detection sensor (220), and the intensity measurement sensor (230) may be connected to the communication device (310) in a wired manner, or may be connected in a wireless manner such as Wifi or Bluetooth.

The process data DB (320) collects and records measurement data from each sensor. At this time, it is desirable to collect and store data from the raw material recognition sensor (210), temperature detection sensor (220), and strength measurement sensor (230) at regular time intervals.

The statistical learning derivation unit (330) analyzes data collected in the process data DB (320) using an artificial intelligence method based on machine learning techniques to produce an improved process.

The statistical learning derivation device (330) can be implemented in a form including a calculation module (331) and a control module (332). However, the calculation module (331) and the control module (332) are implemented as software in a computing device such as a PC and are not hardware-separated configurations.

According to one embodiment, the operation module (331) analyzes the correlation between the amount of change in the process and statistical data collected in the process data DB (320) to derive process conditions that affect changes in the process results (i.e., the quality of the casting product).

FIG. 4 is a flowchart explaining an improved process calculation process performed in the smart factory process control system of the present invention, and FIGS. 5 and 6 explain steps S400 and S500. Referring to FIGS. 4 and 5, the calculation module (331) calculates a deformation amount from statistical data of the immediately previous time point (t-1) and statistical data of the present time (t), and when a deformation amount of its own occurs, it determines the changed process factor as an important process condition and transmits the important process condition to the control module (332), and the control module (332) adjusts the important process condition and performs a process simulation. (Corresponding to step S400)

The operation module (331) compares the statistical data of the important process conditions adjusted by the control module (3312) with the statistical data of the process conditions before adjustment, calculates the average variation of the process results, and deletes the process conditions according to the average variation of the process results smaller than the threshold value from the important process conditions, thereby deriving the most efficient resource utilization process conditions.

These calculations can be used to optimize the temperature and pouring speed of the molten metal.

When an event occurs in the process, such as a crack in a mold (M) due to a load or failure of the manufacturing equipment, the operation module (331) calculates the time difference between the time point (Tx) at which the event occurs and the first time point (Tx-1) before the event occurs from statistical data, and derives an average value between multiple time difference values to predict the time point at which the event occurs.

In addition, the operation module (331) adjusts the process conditions affecting the event at regular time intervals by applying a time deviation and transmits them to the control module (3322) so that it can respond to an event related to the process (corresponding to step S500). In other words, it can calculate a condition that prevents the event from occurring.

The embodiments of the present invention disclosed in this specification and drawings are merely specific examples presented to easily explain the technical content of the present invention and to help understand the present invention, and are not intended to limit the scope of the present invention. It will be apparent to those skilled in the art to which the present invention pertains that other modified examples based on the technical idea of the present invention can be implemented in addition to the embodiments disclosed herein.

1000: Smart Factory Manufacturing Process Artificial Intelligence Control System
100 : Manufacturing facilities
200 : Data Acquisition Department
300: AI Analysis Department

Claims (5)

A system that efficiently controls the operation of equipment by using process operation record data of manufacturing equipment in a smart factory.
A data acquisition unit equipped in each process device of the above manufacturing facility to acquire process operation record data; and
Includes an AI analysis unit that analyzes process operation record data for each device collected from the above data acquisition unit to derive an improved process.
The above manufacturing facility is a facility for producing castings.
A feeder for supplying steel and other raw materials, which are raw materials for the casting; and
Includes a melting furnace for melting steel and other raw materials supplied through the above-mentioned supply furnace,
The above data acquisition unit,
A raw material recognition sensor that recognizes steel and other raw materials supplied to the above feeder;
A temperature detection sensor that detects the temperature of the melting vessel and the temperature of the melted raw material; and
Includes a strength measuring sensor for measuring the strength of the mold,
The above AI analysis unit,
A communication device that receives measurement data from each sensor of the above data acquisition unit;
Process data DB that collects and records measurement data from each of the above sensors; and
It includes a statistical learning derivation engine that analyzes data in the process data DB using an artificial intelligence method based on machine learning techniques to produce an improved process.
The above feeder is in the form of a conveyor belt that transports steel in the form of ingots,
The above raw material recognition sensor,
A metal detector in the form of a metal detector placed on the conveyor belt detects the conveyed ingots one by one,
The above strength measuring sensor,
As a probe type, a plurality of probes arranged at regular intervals are pressed against the mold with a set pressure, and the strength of the mold is measured by measuring the displacement applied to the probes.
The above strength measuring sensor,
Further comprising an internal strength measuring device inserted into the mold frame within the mold,
The above internal strength meter is,
A balloon-shaped mold insert tube that increases in volume as air is injected;
A pump for injecting air into the mold insertion tube at a set pressure; and
A smart factory manufacturing process artificial intelligence control system characterized by including a pressure sensor that measures pressure changes in the mold insertion tube. delete delete delete delete

Images