Optimizing O&M plans for flexible hydropower systems

Abstract

Nowadays, we consider electricity is guaranteed, and it is inconceivable to imagine a world without it. This electricity can be provided and produced in different energy source types, for instance, renewable and non-renewable energy production plants. Taking into account the unpredictable behavior of our planet, there is an extra political and social pressure to reduce the use of energy that come from non-renewable sources, which place a burden on renewable energy plants, as they must be able to respond to all the needs of the power grid. And this is where the greatest difficulties of clean energies are found. Some of them, depend directly on the climatic state, such as wind (wind), solar (sunlight), and even ondomotriz (waves and tides), failing to respond throughout the year to the demand defined by the power grid. Therefore, energies that are easier to control its production and predict its states have started to be studied with the hope to solve this seasonal problem. Hydropower and biomass energies are the most common energies that fit these requirements. These energies are able to compensate moments that energy demand is higher, however, a problem common to all renewable energies is the speed with which they respond to peaks in energy demand that can not be predicted. And this is where projects like XFLEX come in. XFLEX is a European project that institutions like universities, research centers, and turbine producing companies are part of, in which one of the main objectives is to combat the delay in the response of renewable energy, more specifically in power plants hydroelectric plants. That being said, the project is divided into 13 workpackages which each has their different targets, but with a common end goal. This dissertation is part of workpackage 3, in which the aim is the creation of an Intelligent Supervisor of the exchange, whereupon it integrates monitoring, modeling, and control of all the surrounding subsystems to be able to evaluate the best decision to be taken in each situation to increase the plant's productivity, reliability, and flexibility. The objective that was proposed in this dissertation was the development of a health index. This index represents the health status of the entire system to better manage the maintenance policies of the system while having the maximum performance, safety, and with the minimum possible costs. Therefore, in an initial phase of the dissertation, a reliability study is presented, motivated by the fact that the assets of a plant are subject to numerous factors of degradation. Consequently, the work developed is a model of fault detection and health status assessment. This entire analysis was developed only with access to sensory data. Finally, an agglomeration and combination algorithm of the various health indexes of the different subsystems was developed, but not tested, in order to achieve the complete health status of the hydroelectric power plant. Optimization algorithms were also used to improve the results obtained so that a health index represents the health status of the system as accurately as possible to assist in maintenance policy decision making.