Design of a microprocessor-based embedded fault diagnostic system and an FPGA-based improvement proposal

Abstract

In the past, the use of proprietary hardware, software and protocols in systems and the lack of continuous interaction between the systems /subsystems were a natural obstacle to prevent critical errors from occurring for systems. Therefore, troubleshooting, which has been important in engineering from the past to the present, was not as challenging as in current applications, and the detection of errors was not seen as important part of as it is at present in systems' lifecycle. However, today's applications are complex systems consisting of embedded systems, units, modules etc. that continuously interact with each other, with interfaced, chip to chip, wired and/or wireless communication between. In addition, the widespread use of new technologies, protocols, commercial off-the-shelf (CoTS) products and operating systems has made a remarkable effect to increasing the error's frequency. The increasing complexity of engineering applications has brought challenges associated with the exposure to multifarious failures or errors, affecting the systems' reliability, safety, availability and performability. In this sense, new concepts such as error detection, debugging, error recovery, correction, maintenance and repair over have become indispensable elements of engineering applications. In particular, the detection of errors has become even more important in order to correct the errors, ensure maximum efficiency, and provide optimized and sustainable systems. In this context, the systems, named as Fault Diagnostic Systems; designed specifically for the purpose of detecting and recording the error conditions on the software, hardware or equipment etc. have become widespread in different fields such as, defence industry, industrial control and automation systems, aerospace industry, railway transportation sector, smart factories, electric and hybrid vehicles, power plants, network and information technologies, automotive, medical/healthcare systems and have become an important part of the system design phase. In this thesis, a microprocessor-based embedded fault diagnostic system, which will operate in conformity with or integrated into main units such as control, automation, power units to be used in numerous fields is designed. This diagnostic system is fundamentally a logger system that continuously records sensor data to be sent from the main system and creates the "Error Messages" formed from the waveforms measured by the sensors. The related system is responsible for and able to record the measurements real-timely, operate in different modes according to the present scenarios, detect the data for a certain second before the error occurs in case of a situation that is considered as an "Error" by the main system, and store the error messages permanently in protected memory structures. In the the first chapter of the study, a comprehensive literature review on fault diagnostic systems and embedded systems is made. In relevant parts of the chapter, the related issues are researched in details such as their history, purposes, structures, operating logics, architectures, hardware and software components etc. In the second chapter, a microprocessor-based diagnostic system is designed by starting from the requirements and conceptual system design. Then, it is advanced to Detailed System Design Phase which the system architecture and operating scenario for the system are built and needed specifications such as sampling frequency, data sizes, required memory etc. are calculated. Afterwards, the hardware and software development processes are started respectively by taking the system level designs as reference. In hardware design studies, components used in the Diagnostic PCB such as memory units are selected according to the results of analysis on memory types. In the final part of the hardware design phase, a printed circuit board named Diagnostic PCB is designed in Altium PCB Design Software & Tools by using the recommended information in datasheets, PCB design rules and lessons learnt from past experiences. Then, the PCB is manufactured and verifications tests are performed. Thus, the hardware is verified and software development processes in C programming language are started. In software development process, software architecture is built and the architectural design is made based on the scenario determined in the system design section. The Diagnostic software is designed and verified one by one. In the third chapter, accuracy and performance tests are conducted for the Diagnostic System in the microprocessor-based inverter control unit (ICU) setup, and the results are given. Within the scope of the tests, a dummy traffic is generated and the external systems that Diagnostic System will work with or operate into are simulated. By doing so, real operation scenarios run, and essential optimizations for software are made. Then, the accuracy and performance of the system are retested again and again by simulating the real environment, and the system is verified and validated. The last development within the scope of the thesis is an FPGA-based improvement which is named Multitasking Diagnostic System proposed in the fourth chapter. The related system is developed starting from zero by using VHDL in the VIVADO Design Suite environment as high-level. However, it should be emphasized that no hardware design or tests on physical setup have been made, as in the system developed based on microprocessor. The new system model providing uninterrupted operation is created by considering the advantages of using FPGA for embedded applications such as parallel processing, speed, flexibility, extendibility. In the fifth chapter, accuracy and performance tests are performed also for the Multitasking Diagnostic System through behavioural simulation on the VIVADO Design Suite environment. For all tests, a dummy sensor traffic is generated and a similar scenario is built by taking the scenario followed in the microprocessor-based design processes as reference. Scenarios are run repeatedly by simulating the real environment, and essential optimizations for firmware are made also for. And consequently, the system is also verified and validated. In the last chapter, the microprocessor-based and FPGA-based system designs with their results are summarized, especially the worth-emphasizing parts are given by restating the dissertation main motivation which is the need of an Embedded Diagnostic System. The systems are examined, and a detailed comparison containing their advantages and disadvantages according to the requirements and circumstances is given. Lastly, the contributions to the literature, and the future considerations related to the dissertation is introduced. In conclude, both of the microprocessor-based and FPGA-based systems are successfully designed, verified and validated.