Abstract
This thesis deals with the development of a Risk Based Inspection Programme for the Eskom Fossil Fired Power Stations. This study comes as a result of external and internal factors that affect Eskom from producing electricity to ensure security of supply. Eskom current supplies 95% of South Africa’s electricity needs. AS part of Eskom’s drive to remain financially sustainable Eskom has move to cost reflective tariff model. This model considers all the input costs required to generate electricity and this cost is then passed on the consumer. As part of Eskom’s multi-year price determination Eskom requested a 16% increase for the next five years (2014-2019). This was turned down by the National Energy Regulator (NERSA) and instead Eskom was awarded an 8% tariff increase. For Eskom this means that current funds must be spent wisely. With more than 50% of Eskom’s current running fleet being past the mid-life of their design it means that more maintenance activities need to be carried out on these plants. With a decrease in the tariff Eskom now has to decide on which plants take preference over maintenance activities. The Pressure Equipment Regulations (PER) came into effect in October 2009 and compliance to these regulations are mandatory. The PER requires that Eskom pressure test all equipment regarded as pressure equipment according to the PER on a 3 yearly basis. Currently these pressure tests are carried out on a 6 yearly basis. Should Eskom move to the 3 yearly frequencies it would mean that Eskom would experience some difficulties in ensuring security of supply of electricity to South Africa. The PER however allows the users of pressure equipment to apply a Risk Based Inspection approached to the pressure equipment as opposed to the fixed term pressure testing. This RBI would give the user the latitude to make fitness for service decisions and determine the frequency of test and inspection 3 activities. Eskom has opted to implement a risk based approached to all pressure equipment on within the generating fleet. This thesis looks at the development of the Risk Based Inspection programme. The objective of the thesis is to determine whether the CWA15740 process can be applied within the Eskom environment to ensure compliance to the OHSACT. The thesis also investigates the use of fuzzy logic to determine risk plots for boiler valves. This thesis consists of 9 chapters. Chapter one looks at the current situation that Eskom is faced with both from a funding and statutory point of view. Chapter two is a literature survey that looks the development of RBI and is applications in other industries. Chapter 3 discusses the research approach adopted by the author and it defines the research question and the research objectives. This research follows the case study approach. Chapter fours describes the CWA15740 approach looking at the possible advantages and disadvantages of the approach. Chapter five describes the data gathering process for the data required for the RBI assessment. Chapter six looks at the development of the RBI model and delves into the multi-level risk assessment approach. Chapter seven discusses the outcomes of the RBI pilot that was carried out at Lethabo Power Station. Chapter eight discusses the development of the Fuzzy C Mean (FCM) algorithm to predict valve risk based on the clustering concept. Finally chapter nine discusses the learnings from the research as well as the conclusions that can be drawn and finally, future work based on the findings of the research. The findings of the research indicate that the CWA15740 process is an effective process for the power generation industry given its non-prescriptive approach. This was proven by the successful certification of Lethabo Power Station by an independent Certification Body. Further the FCM algorithm proved that it is possible to predict boiler valve risk using the clustering approach. 4 Finally the research found that by implementing a risk based strategy saving of a minimum of 84% could be achieved. The contributions of this work are twofold. Firstly the development of a Risk Based Inspection process for the fossil Fired power generation industry. To the best of the author’s knowledge the process using the CWA15740 process has not been rolled out at any fossil power generating utility. Typical maintenance processes followed by fossil fired power generating utilities do not include the risk component. This statement is supported by the benchmarking exercise as well as the literature survey. Secondly the use of the FCM algorithm shows that the FCM clustering process can be effectively used to predict component risk. Keywords: Probability of Failure; Likelihood of Failure; Component Damage Mechanism; Fuzzy C Mean; Pressure Equipment Regulation; Conformity Assessment; Multi Level Risk Assessment.
D.Ing. (Engineering Management)