Primary refrigeration system commissioning based on a transcritical 2- stage R744 cycle

Teixeira, Daniella

Primary refrigeration system commissioning based on a transcritical 2- stage R744 cycle

Thesis / Dissertation

2024

Abstract

This report gives a brief background into the use of carbon dioxide as a refrigerant (R744) and describes the development of a two-stage trans-critical cooling system that is intended to be used as a chiller for the detectors at CERN's Large Hadron Collider (LHC). It then goes on to describe the steps taken to prepare the system for start-up. These steps include the process of defining how the system should operate and translating this into actuator and PLC logic; identifying the safety limits and implementing alarms to prevent accidents; testing the PLC redundancy to understand its failure modes; testing the programmed logic and wiring; and testing the alarms before clearing the system for start-up. Once the system is started, the controllers are manually tuned by an operator to achieve stable and reliable performance. However, this project aims to determine whether a better performance can be achieved by first modelling the system, determining the transfer function of each control loop and designing the controllers mathematically. To do this, the system is modelled in Simulink, and the performance of the model is verified by comparing the outputs of the model to that of the physical system while running with the same operating conditions. With the verified model, the transfer function of each control loop can be determined, and various control methods can be used to design the PI controllers. Due to the complexity of the control problem, and the interaction between the multiple control loops, care is taken when defining the desired performance of the controllers to maximise disturbance rejection and ensure that the controllers can operate independently without causing instability in other control loops. The designed controllers are implemented in the simulated model of the plant to verify the performance of the control loops under different operating conditions and with realistic disturbances. This is compared to the performance of the physical system with its manually tuned controllers. The comparison finds that the designed controllers perform better, with less oscillation and better disturbance rejection than the manually tuned controllers. From this it can be concluded that the process of simulating the system and designing the controllers mathematically provides more stable performance than the manual operator tuning. However, this process is much more time-consuming and requires a deep understanding of the instabilities, disturbances, and possible failures of the system. This may not be practical for the commissioning of multiple, large, complex systems with restrictive deadlines but may be worthwhile for systems that will be multiplied several times as the Primary R744 chiller at CERN will be.