Current practices of power quality mitigation in the industry are characterized by sub-optimal investment decisions where over compensation is often the norm such causing huge wastage in financial resources. Providing power quality management services to industrial customers in the form of power quality contracts could yield substantial return for the network operator. With better understanding of network parameters, and the option of installing network level mitigation devices, network operators could employ wider range of cost effective mitigation solutions. Tapping into the market however, entails bearing the risks for the customers which network operators are not always willing or encouraged to do. With potentially millions at stake, extensive risk assessments are crucial for any proposed power quality management scheme. This thesis investigates the voltage sag aspect of the problem as part of a larger power quality management scheme. The aim is to develop general framework for technical and financial assessments of voltage sags prior to the introduction of power quality management service. The thesis focuses on five major aspects of voltage sag assessment: identification of customer requirement, financial loss assessment, network sag performance estimation, sag mitigation, and financial appraisal of mitigating solutions. The first part of the thesis gives a comprehensive overview of current power quality problems faced by industrial customers and provides ranges of typical financial losses incurred by different types of industries around the world. It then proposes robust methodology for assessment of typical financial loss, i.e., customized customer damage function (CCDF), for a given industry based on available survey data and taking into account characteristics of the assessed customer plant. For failure and financial risk assessments, the thesis introduces new customer models employing probabilistic methods to quantify risks induced by voltage sags and proposes generic models that incorporate full flexibility in failure risk assessment, taking into account the effect of unbalanced sags on equipment behavior. It further quantifies the error introduced by sag performance estimation using limited monitoring data with a case study on actual sag profile. It demonstrates how different estimation methods and different durations of monitoring period affect accuracy of estimation of voltage sag profile and associated risk of industrial process failure. Following this, the thesis presents new models for plant and network level sag mitigation devices. They include power injecting mitigation devices, devices that reduce number of faults in the network and devices that reduce the severity of faults. Developed models are then used to investigate the cost-effectiveness of sag mitigation at different levels. Finally, the thesis presents Genetic Algorithm based methodology for deciding on optimal investment scheme in voltage sag mitigation in the network. The sensitivity of the solution to various influential parameters, including plant type and size, sensitive equipment type, process characteristics, financial loss resulting from process interruption, cost and effectiveness of mitigating solution and network fault rates is also established.