| Titre : |
Design & Implementation of a Cost-Effective Embedded Control System for Vaccine Research Laboratories |
| Type de document : |
Travail de fin d'études |
| Auteurs : |
Hugo DUBOIS, Auteur ; Brice NJINWOUA, ; Nicolas Marchand, |
| Editeur : |
ECAM |
| Année de publication : |
2024 |
| Langues : |
Anglais (eng) |
| Mots-clés : |
IOT |
| Index. décimale : |
TFE - Informatique (ECAM) |
| Résumé : |
In recent years, the rise in vaccine research has called for innovative technologies for laboratory experiments. To address this need, GSK's Automation & Robotics Team was formed within the Research & Development department. Their goal is to provide automated solutions that improve efficiency, produce high-quality data, and enable innovative data reuse. This master's thesis aims to solve the challenges faced by scientists in training, designing, configuring, and using control systems. It presents an affordable embedded control system for a new experiment, outlining its design, hardware configuration, implementation, and deployment. A micro-controller (MCU) and a micro-processor (MPU) are utilized to connect to the laboratory devices but also to provide and host the user interface. The selection and implementation of the Control Area Network (CAN) protocol, used for the communication between these two devices and the Open Platform Communications Unified Architecture (OPC UA) to establish the connection with the external world, are discussed. The initial challenge of configuring a Proportional-Integral-Derivative (PID) controller for a laboratory experiment is tackled. The project is adapted to a new dedicated experiment, involving a peristaltic pump and pressure transmitters, which is thoroughly explained. Considering the features of this experiment, the MCU, the MPU, and the chosen communications protocols, a comprehensive bill of materials is created based on a pre-designed electronic circuit. The chosen embedded architecture, featuring a Teensy 4.1 (which hosts the PID controller) connected to a Raspberry Pi 5 (which hosts the server for the user interface), is justified for several reasons. This architecture efficiently integrates essential components such as sensors, analog-to-digital converters, memory, input/output interfaces, digital-to-analog converters, actuators, and DC/DC converters. The Teensy 4.1's capabilities ensure precise control and real-time processing, while the Raspberry Pi 5 provides robust computational power and a versatile platform for user interface management. The connections between the MCU and the pressure monitor, that can be connected to 3 pressure transmitters, are established to read the pressure values using 4-20[mA] signals. This protocol is used to transmit sensor measurements from a sensing device to a controller in industrial settings. An analysis of the applicable filter to minimize noise and interference on the read signal is conducted. Based on the peristaltic pump's datasheet, the connections are set up and two methods for producing an external voltage control signal are implemented and compared to determine the most accurate one. The first one using a PWM signal and the second one using a DAC combined with an OP AMP. An initial PID controller is implemented and tuned on the MCU for pump control. The user interface is developed in collaboration with the scientists to meet their needs and requirements. A general discussion about the system's efficiency, strengths and weaknesses, and potential improvements is provided. The integration of this control system as a standalone solution concludes this work. |
Design & Implementation of a Cost-Effective Embedded Control System for Vaccine Research Laboratories [Travail de fin d'études] / Hugo DUBOIS, Auteur ; Brice NJINWOUA, ; Nicolas Marchand, . - ECAM, 2024. Langues : Anglais ( eng)
| Mots-clés : |
IOT |
| Index. décimale : |
TFE - Informatique (ECAM) |
| Résumé : |
In recent years, the rise in vaccine research has called for innovative technologies for laboratory experiments. To address this need, GSK's Automation & Robotics Team was formed within the Research & Development department. Their goal is to provide automated solutions that improve efficiency, produce high-quality data, and enable innovative data reuse. This master's thesis aims to solve the challenges faced by scientists in training, designing, configuring, and using control systems. It presents an affordable embedded control system for a new experiment, outlining its design, hardware configuration, implementation, and deployment. A micro-controller (MCU) and a micro-processor (MPU) are utilized to connect to the laboratory devices but also to provide and host the user interface. The selection and implementation of the Control Area Network (CAN) protocol, used for the communication between these two devices and the Open Platform Communications Unified Architecture (OPC UA) to establish the connection with the external world, are discussed. The initial challenge of configuring a Proportional-Integral-Derivative (PID) controller for a laboratory experiment is tackled. The project is adapted to a new dedicated experiment, involving a peristaltic pump and pressure transmitters, which is thoroughly explained. Considering the features of this experiment, the MCU, the MPU, and the chosen communications protocols, a comprehensive bill of materials is created based on a pre-designed electronic circuit. The chosen embedded architecture, featuring a Teensy 4.1 (which hosts the PID controller) connected to a Raspberry Pi 5 (which hosts the server for the user interface), is justified for several reasons. This architecture efficiently integrates essential components such as sensors, analog-to-digital converters, memory, input/output interfaces, digital-to-analog converters, actuators, and DC/DC converters. The Teensy 4.1's capabilities ensure precise control and real-time processing, while the Raspberry Pi 5 provides robust computational power and a versatile platform for user interface management. The connections between the MCU and the pressure monitor, that can be connected to 3 pressure transmitters, are established to read the pressure values using 4-20[mA] signals. This protocol is used to transmit sensor measurements from a sensing device to a controller in industrial settings. An analysis of the applicable filter to minimize noise and interference on the read signal is conducted. Based on the peristaltic pump's datasheet, the connections are set up and two methods for producing an external voltage control signal are implemented and compared to determine the most accurate one. The first one using a PWM signal and the second one using a DAC combined with an OP AMP. An initial PID controller is implemented and tuned on the MCU for pump control. The user interface is developed in collaboration with the scientists to meet their needs and requirements. A general discussion about the system's efficiency, strengths and weaknesses, and potential improvements is provided. The integration of this control system as a standalone solution concludes this work. |
|
Visionner les documents numériques
Affiner la rechercheDesign & Implementation of a Cost-Effective Embedded Control System for Vaccine Research Laboratories / Hugo DUBOIS
Ce document n'est visible qu'après identification
