In order to improve solar ground-mounted photovoltaic power plant's operations and maintenance, this article aims at minimizing the operating losses generated by the equipment shutdown for their preventive maintenance. The model we propose relies on the representation of a plant as a graph having interdependant nodes. Given a number of workers, this enables to create a metric counting the operating losses in day-equivalent of plant shutdown, which can then be converted to kWh. A scheduling Mixed-Integer Linear Program (MILP) is presented to optimize this new criteria on a single photovoltaic plant. A bi-objective optimization is then carried out to help the decision-maker determine the best-compromise number of technicians to perform preventive maintenance on a set of power plants. This study utilizes test instances designed to reflect real-world scenarios. The results exhibit the Pareto-optimal solutions and allow to quantify the gain on the operating losses for one more worker hired, confirming that the model is a relevant decision-aid tool for operations and maintenance decision-makers.