Given the adverse impact and unnecessary costs associated with harmonic pollution in electrical power systems, such as increasing losses and decreasing life expectancy of equipment, development of new filtering techniques with higher quality of power, higher efficiency of transmission and distribution, and lower cost of filters is urgently required. This has become all the more important as society increasingly moves towards an environment filled with sophisticated electronic equipment and power-sensitive electrical machinery. The present work was conceived to provide a way forward towards the goal of improving Power Quality (PQ) by designing a new passive shunt filter based on minimizing the responsibility of the consumer for harmonic pollution, while simultaneously taking into account other relevant objective functions and related techno-economic considerations. Shunt passive filters are commonly found in power industries, where quality of power is paramount, in view of their simplicity of design, lower cost and their having possibly different range of frequency response characteristics that can satisfy a specific harmonic filtering needs. In addition, passive filters, in general, have two primary favourable features, they support reactive power to correct the load power factor and they mitigate harmonic distortion. This study favours passive filters while conceptualizing the research problem and formulating a proposed design. This thesis presents an optimal design of single-tuned passive filters that mm1m1zes consumer's responsibility for harmonic pollution. A solution for the allocation ofresponsibility between supplier and consumer has proved complex and elusive and it remains, therefore, absent. When, whoever takes whatever steps to contain the harmonic distortion, involves costs and expenditure, the situation inevitably leads to conflict between consumer and utility. In order to avoid this unwholesome state of affairs and introduce a degree of rationality and clarity, this study focuses on quantifying harmonic pollution responsibility by substantiating a new "NonLinearity Current Index" (NLCI) and presenting the mathematical modelling of the filter accordingly. The task of designing the filter, then, is posed as a quadratic optimization problem with nonlinear constraints, whose objective function includes the newly proposed NLCI, which will be minimized, as well as five other conventional objective functions that are most frequently used in the design of passive filters. The five objective functions are: i. maximization ofload power factor; ii. Minimization of the voltage total harmonic distortion; iii. minimization of the current total harmonic distortion; iv. minimization of transmission loss; and v. minimization of the filter investment cost. A suitable analytic framework is established, in which circuit analysis is made and relevant auxiliary mathematical expressions are derived for the system under study, followed by the filter design problem posed as an optimization question with the required objective functions and constraints. This thesis uses FORTRAN Feasible Sequential Quadratic Programming (FFSQP) to obtain the optimal values for the single-tuned shunt passive filter elements. The performance of the proposed index and the other five conventional objective functions is evaluated using six numerical 'benchmark' case studies to demonstrate the applicability, viability, validity and usefulness of the proposed harmonic passive filters. A rigorous and thorough comparison of the results of the design with and without compensation is included. As resonance is a critical design issue of passive filters, special care was taken, during the problem formulation, to avoid any risk of it in the compensated system. Care was also taken to meet capacitor loading specifications as given in IEEE Standard 18-2012 and practical constraints for the voltage total and individual harmonic distortion limits were followed in accordance with IEEE 519-2014 guidelines. The results of this study demonstrate the validity and effectiveness of the optimization method, for the given objective functions, since improvements were seen in both power factor and displacement power factor in all the case studies as well as notable reductions in rms values of the line current which led to reduced transmission losses, voltage drop and increased transmission efficiency, to various degrees and in a variety of combinations. Incorporation of the newly designed passive filter resulted in a considerable reduction in NLCI, indicating increased linearity because of minimization of the difference between supply current rms and linear current m1s at the point of common coupling. Critical impedance values also improved, because of the greatly reduced m1s line current. The NLCI minimised passive filter design differentiates the contributions of the supplier and consumer to harmonic pollution in a given system, and, therefore, provides a valuable solution where currently there is none in the literature, reflected in the absence of recognised standards and methodologies in this area. The small size of the proposed filter, in comparison to many other published designs, entails considerably lower financial investment costs, which, combined with its effectiveness at reducing harmonic pollution makes it a very valuable contribution to power system design and installation.