Master Thesis: Design and development of a PLC based vision system application for industrial robot in a hazardous environment

The aim of this project consist of developing modular upgrades for tube cleaners until fully automated heat exchanger cleaning machines. The advantages of automating the process are safety for the operator, who would be outside the operating area, and shorter process time. These machines could be used with smaller amount of water.

The thesis consist of develop a system to automatize a cleaning process in a dangerous environment, analysing the pipelines in order to apply the cleaning protocol. The hardware used was a Raspberry Pi, a Beckho's PLC and a camera IP. Codesys is the used PLC program with a license for Raspberry Pi, with OpenCV for the video recognition

Methodology

The proposed process for automating the cleaning and monitoring the parameters obtained consists of the following steps:

1. User interface: Scan of the hole pattern (Video PLC).

2. Sensor system. Collecting data for closed control to verify the result (Vision System).

3. Database: Selection of the correct method.

4. Cleaning process of the impurities and contamination (Autoclean PLC).

Development

Video PLC subsystem is in charge of the camera (streaming and recording) and image processing through a vision system. Autoclean PLC (cleaning machine) consists of the sequence programming software (Codesys) and the user interface (touch screen controller).

Video PLC is the part of the project in charge of analysing the video recorded from inside the pipeline to determine the degree of cleanliness. This system comply with:

Controls the Camera mover and records the Video
Connection to the Autoclean PLC

System

The endoscopic camera is connected via USB port to the Raspberry as a PLC unit (external device), where the video recognition and classification of contaminants is carried out using the OpenCV library.

The interface is included (accessible from a network-connected device) to complete the system to allow user action: entering parameters and reading output files. These files consist of:

Cleaning ratio (%): values.txt
Video recorded: video.avi
Video analysed: data:time.avi

Algorithm

The endoscopic camera is used to record the inside of the pipe and to obtain the impact of the heat treatment. Once the video has been recorded, it is loaded into a vision system (vision.py) to analyse its characteristics and the degree of contamination found (cleanliness score). the modules that make up the system and achieve its operation correspond to:

• camera.py: Camera streaming and grabbing

• vision.py: Analysis of Video Recorded

• server.py: User interface and web browser

Autoclean PLC

This section details the automated process of inspection and cleaning of the tubes, controlled by the operator from a safe place, by means of the graphic interface and the vision system, acting on the movement of the camera and the water nozzle. This section uses the Codesys software and its visualisation screen where the automation process will be represented. The variables controlled and supervised are:

• Length and position of tube

• Move the platform with camera, Timing the recording

• Degree of cleanliness of each pipe

Shown a diagram of the main blocks, together with their activation, that make up the programme. This diagram corresponds to the programming Grafcet, and is represented in SFC (Sequential Function Chart) language with the use of stages (actuators, outputs) and transitions (sensors, inputs)