Today, power electronic-based loads, which draw harmonically distorted currents, are widely used in the industrial distribution systems. Thus, line currents and bus voltages have high harmonic distortion levels. It is known in the literature that harmonics cause adverse effects such as loss increase in power system equipment, decrease of power factor, torque pulsations in rotating electric machines, malfunction of measurement and protection devices. Accordingly, to reduce these adverse effects, harmonic mitigation devices as passive and active filters are widely employed by consumers and utilities. For designing passive filters, it is conventionally aimed to achieve different goals as minimization of current and / or voltage total harmonic distortion, power factor maximization and minimization of filter cost and losses while keeping individual / total harmonic distortion levels and power factor in the recommended intervals defined by international standards. On the other hand, recent studies in the literature have designed optimal passive filters to maximize transformer's loading capacity under harmonic distortion conditions while considering conventional optimal passive filter design constraints. In this study, firstly, for the solution of the optimal C-type passive filter design problem, which aims to maximize the loading capacity of the transformer in a typical industrial power system with harmonic distortion, the big bang big crunch, ant lion optimization and dragonfly optimization algorithms was comparatively evaluated in terms of their results and speed. And then, the performance of the designed optimal passive filter was analysed for the cases where the total harmonic distortion of the source, line impedance's magnitude, and line impedance's X/R ratio change.