Calculating the profits of an economic MPC applied to CSP plants with thermal storage system

Abstract

Electricity producers participating in a day-ahead energy market aim to maximize profits derived from electricity sales. The daily generation schedule has to be offered in advance, usually the previous day before a certain moment in time. The development of an economically-optimal generation schedule is the core of the generation scheduling problem. To solve this problem, renewable energy plant owners need, besides energy prices forecast, weather prediction. Among renewable energy sources, concentrated solar power (CSP) plants with thermal energy storage (TES) may find it easier to participate in electricity markets due to their semi-dispatchable generation. In any case, the limited accuracy of forecasting solar resource brings about the risk of penalties that may be imposed to CSP plants for deviation from the submitted schedule. This paper proposes a model-based predictive control (MPC) approach with an economic objective function to tackle the scheduling problem in CSP plants with TES. By this approach, the most recent forecast and the current status of plant can be used by the proposed economic MPC approach to reschedule the generation conveniently at regular time intervals. On the other hand, a more feasible generation schedule for the next day is performed at the appropriate time thanks to the use of short-term forecast. The proposed approach is applied, in a simulation context, to a 50 MW parabolic trough collector-based CSP plant with TES under the assumptions of perfect price forecasts and participation in the Spanish day-ahead energy market. A case study based on a half-year period to test several meteorological conditions is performed. In this study, an economic analysis is carried out using actual values of energy price, penalty cost, solar resource data and its day-ahead forecast. Results show an economic improvement in comparison with a traditional day-ahead scheduling strategy, especially in periods with a bad weather forecast. To overcome the lack of short-term weather forecast data for this study, a synthetic short-term predictor, whose accuracy level can be tuned by means of a parameter, is used. Sweeping this accuracy level between the situation with no forecast improvement and perfect shortterm forecast, the MPC strategy reaches an improvement in total profits during the six months period between 13.9% and 33.3% of the maximum room for improvement. This maximum ideal improvement is defined as the difference in profits between the MPC strategy with perfect forecasts and the dayahead scheduling strategy.

A binary‐regularization‐based model predictive control applied to generation scheduling in concentrating solar power plants

“…The value ω = 0.001 indicates an almost perfect short‐term DNI forecast during the first hour and no improvement over the day‐ahead forecast during the sixth hour. Moreover, the forecast worsening is lineal over time (see the work of Vasallo et al). It can be observed in Figure that profits with ω = 0 and 1 bound all results with ω = 0.001.…”

“…The MPC approach described in the works of Vasallo and Bravo and Vasallo et al is adopted for this work as starting point. Section 3.1 describes the sliding window on which the approach is based.…”

“…In order to take into account fresh data, short‐term DNI forecasts with a time horizon of 5 hours are also included in this study. Several methods to obtain this kind of forecast can be consulted in the work of Law et al A parameterized model is used to emulate the short‐term predictor (see the work of Vasallo et al) and to overcome the lack of short‐term DNI forecast data for this study. One of its parameters, ω , is a value between 0 and 1, and represents the accuracy of the forecasts (from ideally perfect short‐term DNI forecast to nonexisting short‐term DNI forecast, respectively).…”

Modelos empíricos del campo solar en plantas termosolares de concentración. Aplicación a la planificación

“…El desarrollo de modelos que estimen la potencia térmica proporcionada por el SF a partir de cierta cantidad de radiación tiene gran interés por su aplicabilidad en problemas de ingeniería. Se han aplicado modelos del SF a problemas de control [3], en herramientas de simulación [2] [5], en problemas de planificación de la producción [11,13] o de asignación de tareas [8,12] y en problemas sobre dimensionado de la planta [6,4]. Estos modelos están principalmente basados en primeros principios, y pueden ser tanto dinámicos [2,5] como estáticos (quasi-stationary) [7], aunque también pueden encontrase trabajos donde el campo solar es modelado simplemente con una relación lineal entre la radiación solar y la potencia térmica [10,9].…”

Transient Analysis and Execution-Level Power Tracking Control of the Concentrating Solar Thermal Power Plant

“…To fill the generation gaps in intermittent solar energy, the CSP plant is generally integrated with thermal energy storage (TES), which enables the CSP plant to control its power output flexibly in the presence of solar uncertainty [3]. Actually, the TES enables the CSP plant to be partly independent from constantly changing solar radiation [4], reducing the short-term load variation and extending or shifting the power supply period [5].…”