As automatic sensing and information and communication technology get cheaper, building monitoring data becomes easier to obtain. The availability of data leads to new opportunities in the context of energy efficiency in buildings. This paper describes the development and validation of a data-driven grey-box modelling toolbox for buildings. The Python toolbox is based on a Modelica library with thermal building and Heating, Ventilation and Air-Conditioning models and the optimization framework in JModelica.org. The toolchain facilitates and automates the different steps in the system identification procedure, like data handling, model selection, parameter estimation and validation. To validate the methodology, different grey-box models are identified for a single-family dwelling with detailed monitoring data from two experiments. Validated models for forecasting and control can be identified. However, in one experiment the model performance is reduced, likely due to a poor information content in the identification data set.

No abstract

Abstract:We present the open-source software framework in JModelica.org for numerically solving large-scale dynamic optimization problems. The framework solves problems whose dynamic systems are described in Modelica, an open modeling language supported by several different tools. The framework implements a numerical method based on direct local collocation, of which the details are presented. The implementation uses the open-source third-party software package CasADi to construct the nonlinear program in order to efficiently obtain derivative information using algorithmic differentiation. The framework is interfaced with the numerical optimizers IPOPT and WORHP for finding local optima of the optimization problem after discretization. We provide an illustrative example based on the Van der Pol oscillator of how the framework is used. We also present results for an industrially relevant problem regarding optimal control of a distillation column.

No abstract

Abstract-This paper presents a hierarchical approach to feedback-based trajectory generation for improved vehicle autonomy. Hierarchical vehicle-control structures have been used before-for example, in electronic stability control systems, where a low-level control loop tracks high-level references. Here, the control structure includes a nonlinear vehicle model already at the high level to generate optimization-based references. A nonlinear model-predictive control (MPC) formulation, combined with a linearized MPC acting as a backup controller, tracks these references by allocating torque and steer commands. With this structure the two control layers have a physical coupling, which makes it easier for the low-level loop to track the references. Simulation results show improved performance over an approach based on linearized MPC, as well as feasibility for online implementations.

No abstract

No abstract

This paper describes the development of an optimization-friendly thermodynamic property model of water and steam that covers liquid, vapor, 2-phase as well as the super-critical region. All equations are at least twice continuously differentiable with respect to all model variables and can be used in dynamic optimization problems solved by efficient derivativebased algorithms. The accuracy has been verified against the industry standard IAPWS IF97 and performance and robustness have been tested by solving a trajectory optimization problem where the start-up time of a gas power plant has been minimized while satisfying constraints on temperature gradients, pressure and flows. Simulations of various plant models have also been performed to verify and benchmark the implementation. The results show that the new media can be used in both solving dynamic optimization and simulation problems yielding reliable results. The new media has been integrated into Modelon's Thermal Power library 1.13. This article is built upon the work in (Åberg, 2016).