Electrical power systems play a key role in production and services in both the industrial and commercial sectors and significantly affect the private lives of citizens. A major asset of any power delivery system is the transformer. Transformers represent extensive investment in any power delivery system, and because of the notable effect of a transformer outage on system reliability, careful management of this type of asset is critical. In North America, a large proportion of transformers is approaching the end of their life and should be replaced. In many cases, unexpected transformer outages can be catastrophic and cause both direct and indirect costs to be incurred by industrial, commercial, and residential sectors. Direct costs include but are not limited to loss of production, idle facilities and labour, damaged or spoiled product, and damage to equipment. For commercial customers, the effects may include damage to electrical and electronic equipment, and in some cases damage to goods. For residential customers, outages may cause food spoilage or damage to electrical equipment. In addition to direct costs, there are several types of indirect costs may also result, such as accidental injuries, looting, vandalism, legal costs, and increases in insurance rates. The main goal of this research was to assess the health and remaining lifetime of a working transformer. This information plays a very important role in the planning strategies of power delivery systems and in the avoidance of the potentially appalling effects of unexpected transformer outages. This thesis presents two different methods of assessing transformer end of life and three distinct methods of determining the health index and health condition of any working transformer. The first method of assessing transformer end of life is based on the use of Monte Carlo technique to simulate the thermal life of the solid insulation in a transformer, the failure of which is the main reason for transformer breakdown. The method developed uses the monthly average ambient temperature and the monthly solar clearness index along with their associated uncertainties in order to estimate the hourly ambient temperature. The average daily load curve and the associated uncertainties in each hourly load are then used to model the transformer load. The inherent uncertainties in the transformer loading and the ambient temperature are used to generate an artificial history of the life of the transformer, which becomes the basis for appraising its remaining lifetime. The second method of assessing transformer end of life is essentially an economic evaluation of the remaining time to the replacement of the transformer, taking into consideration its technical aspects. This method relies on the fact that a transformer fails more frequently during the wear-out period, thus incurring additional maintenance and repair costs. As well, frequent failures increase during this period also costs related to transformer interruptions. Replacing a transformer before it is physically damaged is therefore a wise decision. The bathtub failure model is used to represent the technical aspects of the transformer for the purposes of making the replacement decision. The uncertainties related to the time-to-failure, time-to-repair, time-to-switch, and scheduled maintenance time are modeled using a Monte Carlo simulation technique, which enables the calculation of the repair costs and the cost of interruptions. The repair, operation, and interruption costs are then used to generate equivalent uniform annual costs (EUACs) for the existing transformer and for a new transformer, a comparison of which enables the determination of the most economical replacement year. The case studies conducted using both methods demonstrate their reliability for determining transformer end of life for assessing the appropriate time for replacement. Diagnostic test data for 90 working transformers were used to develop three methods of estimating the health condition of a transformer, which utilities and industries can use in order to assess the health of their transformer fleet. The first method is based on building a linear relation between all parameters of diagnostic data in order to determine a transformer health index, from which the health condition of the transformer can be evaluated. The second method depends on the use of artificial neural networks (ANN) in order to find the health condition of any individual transformer. The diagnostic data for the 90 working transformers together with the health indices calculated for them by means of a specialized transformer asset management and health assessment lab, were used to train an ANN. After the training, the ANN can estimate a health index for any transformer, which can be used in order to determine the health condition of the transformer. The third method is based on finding a relation between the input data and the given health indices (calculated by the specialized transformer asset management and health assessment lab) using the least squares method. This relation then can be used to find the health index and health condition of any working transformer. The health condition determined based on these methods shows excellent correlation with the given health condition calculated by the specialized transformer asset management and health assessment lab.

Electrical power systems play a key role in production and services in both the industrial and commercial sectors and significantly affect the private lives of citizens. A major asset of any power delivery system is the transformer. Transformers represent extensive investment in any power delivery system, and because of the notable effect of a transformer outage on system reliability, careful management of this type of asset is critical. In North America, a large proportion of transformers is approaching the end of their life and should be replaced. In many cases, unexpected transformer outages can be catastrophic and cause both direct and indirect costs to be incurred by industrial, commercial, and residential sectors. Direct costs include but are not limited to loss of production, idle facilities and labour, damaged or spoiled product, and damage to equipment. For commercial customers, the effects may include damage to electrical and electronic equipment, and in some cases damage to goods. For residential customers, outages may cause food spoilage or damage to electrical equipment. In addition to direct costs, there are several types of indirect costs may also result, such as accidental injuries, looting, vandalism, legal costs, and increases in insurance rates. The main goal of this research was to assess the health and remaining lifetime of a working transformer. This information plays a very important role in the planning strategies of power delivery systems and in the avoidance of the potentially appalling effects of unexpected transformer outages. This thesis presents two different methods of assessing transformer end of life and three distinct methods of determining the health index and health condition of any working transformer. The first method of assessing transformer end of life is based on the use of Monte Carlo technique to simulate the thermal life of the solid insulation in a transformer, the failure of which is the main reason for transformer breakdown. The method developed uses the monthly average ambient temperature and the monthly solar clearness index along with their associated uncertainties in order to estimate the hourly ambient temperature. The average daily load curve and the associated uncertainties in each hourly load are then used to model the transformer load. The inherent uncertainties in the transformer loading and the ambient temperature are used to generate an artificial history of the life of the transformer, which becomes the basis for appraising its remaining lifetime. The second method of assessing transformer end of life is essentially an economic evaluation of the remaining time to the replacement of the transformer, taking into consideration its technical aspects. This method relies on the fact that a transformer fails more frequently during the wear-out period, thus incurring additional maintenance and repair costs. As well, frequent failures increase during this period also costs related to transformer interruptions. Replacing a transformer before it is physically damaged is therefore a wise decision. The bathtub failure model is used to represent the technical aspects of the transformer for the purposes of making the replacement decision. The uncertainties related to the time-to-failure, time-to-repair, time-to-switch, and scheduled maintenance time are modeled using a Monte Carlo simulation technique, which enables the calculation of the repair costs and the cost of interruptions. The repair, operation, and interruption costs are then used to generate equivalent uniform annual costs (EUACs) for the existing transformer and for a new transformer, a comparison of which enables the determination of the most economical replacement year. The case studies conducted using both methods demonstrate their reliability for determining transformer end of life for assessing the appropriate time for replacement. Diagnostic test data for 90 working transformers were used to develop three methods of estimating the health condition of a transformer, which utilities and industries can use in order to assess the health of their transformer fleet. The first method is based on building a linear relation between all parameters of diagnostic data in order to determine a transformer health index, from which the health condition of the transformer can be evaluated. The second method depends on the use of artificial neural networks (ANN) in order to find the health condition of any individual transformer. The diagnostic data for the 90 working transformers together with the health indices calculated for them by means of a specialized transformer asset management and health assessment lab, were used to train an ANN. After the training, the ANN can estimate a health index for any transformer, which can be used in order to determine the health condition of the transformer. The third method is based on finding a relation between the input data and the given health indices (calculated by the specialized transformer asset management and health assessment lab) using the least squares method. This relation then can be used to find the health index and health condition of any working transformer. The health condition determined based on these methods shows excellent correlation with the given health condition calculated by the specialized transformer asset management and health assessment lab.