Virtual Battery Parameter Identification Using Transfer Learning Based Stacked Autoencoder

Abstract

Recent studies have shown that the aggregated dynamic flexibility of an ensemble of thermostatic loads can be modeled in the form of a virtual battery. The existing methods for computing the virtual battery parameters require the knowledge of the first-principle models and parameter values of the loads in the ensemble. In real-world applications, however, it is likely that the only available information are end-use measurements such as power consumption, room temperature, device on/off status, etc., while very little about the individual load models and parameters are known. We propose a transfer learning based deep network framework for calculating virtual battery state of a given ensemble of flexible thermostatic loads, from the available end-use measurements. This proposed framework extracts first order virtual battery model parameters for the given ensemble. We illustrate the effectiveness of this novel framework on different ensembles of ACs and WHs.

Hierarchical, Grid-Aware, and Economically Optimal Coordination of Distributed Energy Resources in Realistic Distribution Systems

“…Each STL controller is tasked with: (a) updating an aggregated dynamic flexibility model for the DERs; and (b) real-time dispatch of the DERs to track certain power set-points. In particular, each STL controller constructs a dynamic representation of energy and power flexibility limits for the group of DERs at the service transformer, that together is denoted by a VB model [22][23][24][25][26]. The VB's power limits represent the (maximal) range of the control set-points that can be successfully tracked by the DERs at each service transformer; while VB's energy limits encode the end-user quality of service constraints and, along with the estimated state of charge (SoC), determine the duration of successful tracking performance.…”

“…The VB's power limits represent the (maximal) range of the control set-points that can be successfully tracked by the DERs at each service transformer; while VB's energy limits encode the end-user quality of service constraints and, along with the estimated state of charge (SoC), determine the duration of successful tracking performance. Different methods exist for characterizing the VB model of an aggregation of DERs, including closed-form expressions [22], optimization-based methods [23,25], as well as deep learning techniques [24,26]. Finally, the STL controller performs a real-time optimal control of the DERs (e.g., switching thermostatic loads on/off) to track the power set-points by explicitly accounting for service transformer and DER quality of service constraints, as necessary [25].…”

Identification and Validation of Virtual Battery Model for Heterogeneous Devices

“…The scope of this work extends to ACs, EWHs, and a heterogeneous ensemble of ACs and EWHs. We refer the readers to [8,15] for the first order hybrid models of ACs and EWHs. In what follows, we describe the method to identify the VB parameters, φ.…”

“…This results in a value of s i lying in the set [0, 1] which essentially implies that the power applied to either AC or EWH device is continuous which is not correct (refer Eqs. (8) and (9) in [15] or Eqs. (11) and (14) in [8]).…”

“…The temperature evolution of AC and EWH devices is given in Eqs. (8) and (9) [15] respectively. The parameters such as set-point temperature, outside air temperature, thermal resistance and thermal capacitance for the 100 AC devices is chosen such that no two devices have the same parameters.…”