Adriano Desideri

ThermoCycle: A Modelica library for the simulation of thermodynamic systems

This paper presents the results of an on-going project to develop ThermoCycle, an open Modelica library for the simulation of low-capacity thermodynamic cycles and thermal systems. Special attention is paid to robustness and simulation speed since dynamic simulations are often limited by numerical constraints and failures, either during initialization or during integration. Furthermore, the use of complex equations of state (EOS) to compute thermodynamic properties significantly decreases the simulation speed. In this paper, the approach adopted in the library to overcome these challenges is presented and discussed.

Dynamic Modeling and Control System Definition for a Micro-CSP Plant Coupled with Thermal Storage Unit

Organic Rankine cycle (ORC) systems are gaining ground as a means of effectively providing sustainable energy. Coupling small-scale ORCs powered by scroll expander-generators with solar thermal collectors and storage can provide combined heat and power to underserved rural communities. Simulation of such systems is instrumental in optimizing their control strategy. However, most models developed so far operate at steady-state or focus either on ORC or on storage dynamics. In this work, a model for the dynamics of the solar ORC system is developed to evaluate the impact of variable heat sources and sinks, thermal storage, and the variable loads associated with distributed generation. This model is then used to assess control schemes that adjust operating conditions for daily environmental variation.Copyright

Increasing the efficiency of Organic Rankine Cycle Technology by means of Multivariable Predictive Control

Abstract The Organic Rankine Cycle (ORC) technology has become very popular, as it is extremely suitable for waste heat recovery from low-grade heat sources. As the ORC system is a strongly coupled nonlinear multiple-input multiple-output (MIMO) process, conventional control strategies (e.g. PID) may not achieve satisfactory results. In this contribution our focus is on the accurate regulation of the superheating, in order to increase the efficiency of the cycle and to avoid the formation of liquid droplets that could damage the expander. To this end, a multivariable Model Predictive Control (MPC) strategy is proposed, its performance is compared to the one of PI controllers for the case of variable waste-heat source profiles.

Experimental study of Predictive Control strategies for optimal operation of Organic Rankine Cycle systems

In this paper the performance of Model Predictive Control (MPC) and PID based strategies to optimally recover waste heat using Organic Rankine Cycle (ORC) technology is investigated. First the relationship between the evaporating temperature and the output power is experimentally evaluated, concluding that for some given heat source conditions there exists an optimal evaporating temperature which maximizes the energy production. Three different control strategies MPC and PID based are developed in order not only to maximize energy production but to ensure safety conditions in the machine. For the case of the MPC, the Extended Prediction Self-Adaptive Control (EPSAC) algorithm is considered in this study as it uses input/output models for prediction, avoiding the need of state estimators, making of it a suitable tool for industrial applications. The experimental results obtained on a 11kWe pilot plant show that the constrained EPSAC-MPC outperforms PID based strategies, as it allows to accurately regulate the evaporating temperature with a lower control effort while keeping the superheating in a safer operating range.

Low-order models of a single-screw expander for organic Rankine cycle applications

Screw-type volumetric expanders have been demonstrated to be a suitable technology for organic Rankine cycle (ORC) systems because of higher overall effectiveness and good part-load behaviour over other positive displacement machines. An 11 kWe single-screw expander (SSE) adapted from an air compressor has been tested in an ORC test-rig operating with R245fa as working fluid. A total of 60 steady-steady points have been obtained at four different rotational speeds of the expander in the range between 2000 rpm and 3300 rpm. The maximum electrical power output and overall isentropic effectiveness measured were 7.3 kW and 51.9%, respectively. In this paper, a comparison between two low-order models is proposed in terms of accuracy of the predictions, the robustness of the model and the computational time. The first model is the Pacejka equation-based model and the second is a semi-empirical model derived from a well-known scroll expander model and modified to include the geometric aspects of a single screw expander. The models have been calibrated with the available steady-state measurement points by identifying the proper parameters.

Nonlinear identification and control of Organic Rankine Cycle systems using sparse polynomial models

Development of a first principles model of a system is not only a time- and cost-consuming task, but often leads to model structures which are not directly usable to design a controller using current available methodologies. In this paper we use a sparse identification procedure to obtain a nonlinear polynomial model. Since this is a NP-hard problem, a relaxed algorithm is employed to accelerate its convergence speed. The obtained model is further used inside the nonlinear Extended Prediction Self-Adaptive control (NEPSAC) approach to Nonlinear Model Predictive Control (NMPC), which replaces the complex nonlinear optimization problem by a simpler iterative quadratic programming procedure. An organic Rankine cycle system, characterized for presenting nonlinear time-varying dynamics, is used as benchmark to illustrate the effectiveness of the proposed combined strategies.