Two-layer predictive control of a micro-grid including stochastic energy sources

Abstract

A two-layer control scheme based on Model Predictive Control (MPC) operating at two different timescales is proposed for the energy management of a micro-grid (MG), including a battery, a gas-turbine generator, a photovoltaic (PV) generator and the input from the electrical network. The highlevel optimizer, which acts at a slow timescale and relies on a simplified model of the system, is in charge of computing the nominal operating conditions for each MG component so as to optimize an economic performance index on the basis of available predictions for the PV generation and load request. The low-level controller acts at higher frequency, adjusts the MG operation, and relies on a stochastic MPC algorithm that ensures probabilistic constraints satisfaction. Detailed models and simulations of the overall control system are presented.

Stochastic linear Model Predictive Control with chance constraints – A review

“…Specifically, some works have been recently devoted to this issue, with specific reference to distributed control of independent systems [107,108], while currently results on the application to interconnected systems are not available, with the exception of [109]. • SMPC methods could be efficiently used in multilevel control structures, where at the higher layer low frequency and deterministic methods are in charge of long-term predictions and control, while at the lower layer SMPC can be used at a higher frequency to compensate for the effect of stochastic noise as, e.g., in [31]. • Since classical methods for online estimation of the noise characteristics (e.g., their covariances) from the measurements are available, possible extensions of the SMPC schemes can be envisaged to make them adaptive and more reliable with respect to time-varying noise characteristics in a real-time framework.…”

“…A remark is due at this point. From the comparison of the constraint (28) with (29), as well as of ( 30) with (31), it is apparent that the use of a feedback control strategy is desirable for reducing the conservativity of the robust constraints, both in the deterministic and in the stochastic framework, since 1−a k 1−a > 1 and…”

“…The use of chance constrained optimization and stochastic MPC has been considered as a promising solution in a number of application domains, such as water reservoir management [5][6][7], temperature and HVAC control in buildings [8][9][10][11][12][13][14][15][16][17], process control [18][19][20][21][22], power production, management, and supply in systems with renewable energy sources [23][24][25][26][27][28][29][30][31][32][33][34], cellular networks management [35], driver steering, scheduling, and energy management in vehicles [36][37][38][39][40][41][42][43][44], path planning and formation control [45][46][47][48], air traffic control [49], inventory control and supply chain management [50][51]…”

A receding horizon control approach for re-dispatching stochastic heterogeneous resources accounting for grid and battery losses

“…In [20], RHC and scenario-based optimization are jointly applied for the operational control of islanded microgrids using the DC approximation of the power flow equations and assuming no energy storage losses. [21] develops a twolayer RHC-based control scheme for a microgrid in two different time-scales. The power grid and the associated constraints are not modeled, whereas uncertainties are considered only in the fast time-scale control via chance constraints.…”

Data-driven non-parametric chance-constrained model predictive control for microgrids energy management using small data batches

“…Moreover, the increasing availability of local energy storage devices, like batteries, requires proper management to extract their full potential, and it provides additional degrees of freedom for optimal energy management strategies. The topic of energy management in microgrids has been widely investigated in recent years, both for the single-energy and for multi-energy scenarios (Dörfler et al, 2016;Raimondi Cominesi et al, 2018;Khayat et al, 2020;Ceusters et al, 2021;Wang et al, 2021).…”