Speaker
Description
This study presents optimizations to the Massive Temperature Readout System (MTRS), a low-cost alternative to traditional PLC-based systems for large-scale temperature monitoring. Our improved MTRS design streamlines hardware by eliminating intermediate microcontrollers and communication modules, reducing complexity and potential failure points. A unified multi-threaded C++ application replaces the original multi-process software, enhancing performance and managing sensor data acquisition and OPC UA server functions within a single process. Furthermore, a novel sensor reading strategy that achieves higher system performance and provides increased channel capacity is implemented. Our optimizations yield a feasible, cost-effective, high-performance MTRS solution for industrial infrastructures.
Summary (500 words)
This paper introduces an improved Massive Temperature Readout System (MTRS), designed as a cost-effective alternative to traditional Programmable Logic Controllers (PLCs) for large-scale temperature monitoring in industrial applications. The need for reliable and precise temperature monitoring is crucial in modern industrial systems, particularly in demanding environments like those found at CERN, to ensure stability and safety. While PLCs and Supervisory Control and Data Acquisition (SCADA) systems are traditionally used for this purpose, they can become costly and complex when dealing with a large number of sensor channels.
The original MTRS was designed to address these limitations by utilizing Raspberry Pi (RPi) embedded platforms for data acquisition and Open Platform Communications Unified Architecture (OPC UA) for communication with Detector Control Systems (DCS), critical for ensuring safe operation at facilities like CERN. However, the original design faced challenges with high-frequency noise affecting the reliability and precision of direct sensor readings with the RPi. This led to the introduction of intermediate hardware components like ATMEGA328P microcontrollers and RS485 communication modules, thereby increasing system complexity.
In contrast, the proposed system in this paper introduces a more streamlined approach. It eliminates the intermediate microcontrollers and communication modules, enabling direct sensor reading operations via the RPi. The improved MTRS leverages the LTC2984 chip for analog-to-digital conversion of temperature sensor data. Communication between the RPi and the LTC2984 chips is handled via the SPI bus, with a demultiplexer used to manage chip-select signals. This hardware simplification reduces system complexity and potential failure points. Furthermore, the original multi-process software architecture is replaced with a single, high-performance multi-threaded C++ application, improving overall system performance. The software utilizes the Linux spidev library for SPI communication and the WiringPi library for GPIO control. OPC UA communication is implemented using the FreeOpcUa library, enabling seamless integration with SCADA systems.
A novel sensor reading strategy is also implemented to enhance performance and channel capacity, achieving up to 25.8 sensor readings per second with four LTC2984 devices, compared to 6.38 readings per second with the sequential approach. This new strategy uses an overlaying scheme, initiating new conversions on other LTC2984 devices instead of waiting for each conversion to complete, significantly increasing the number of channels that can be read.
The paper details the experimental setup used to evaluate the proposed system, including the hardware and software components. The results of the experiments demonstrate the system's stability, performance, and scalability. Notably, the improved sensor reading strategy significantly enhances the system's performance compared to the original sequential approach, allowing the system to scale up to 1600 LTC2984 devices and a channel capacity of 9600.
In conclusion, the study demonstrates the successful optimization of the MTRS system, offering a more efficient, reliable, and scalable solution for large-scale temperature monitoring. The improvements in hardware and software architecture, along with the novel sensor reading strategy, provide enhanced performance and scalability.
