Rapid Energy Optimization of Vapor Compression Systems Using Probabilistic Machine Learning and Extremum Seeking Control

Extremum seeking control (ESC) is a popular datadriven approach for optimizing the energy consumption of vapor compression systems (VCS). Tuning ESC control parameters can present a challenge to implementation, especially in advanced variants of ESC, because time-consuming and problemspecific manual tuning is often required to eliminate numerical and dynamical instabilities. In this paper, we propose an automatic ESC tuning mechanism based on a Bayesian optimization framework that systematically leverages closedloop ESC experiments to compute highperforming ESC parameters. We validate the proposed Bayesian-optimized ESC on a physicsbased Modelica model of a VCS. This new approach is six times faster and yields a 9% higher coefficient of performance than a stateoftheart timevarying ESC method under identical experimental conditions

[1]  Christopher R. Laughman,et al.  Simulation Failure Robust Bayesian Optimization for Estimating Black-Box Model Parameters , 2021, 2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC).

[2]  Claus Danielson,et al.  Accelerating self-optimization control of refrigerant cycles with Bayesian optimization and adaptive moment estimation , 2021 .

[3]  Victor M. Zavala,et al.  MPC Controller Tuning using Bayesian Optimization Techniques , 2020, ArXiv.

[4]  John Lygeros,et al.  Performance-Driven Cascade Controller Tuning With Bayesian Optimization , 2020, IEEE Transactions on Industrial Electronics.

[5]  Martin Guay,et al.  Proportional–Integral Extremum Seeking for Vapor Compression Systems , 2020, IEEE Transactions on Control Systems Technology.

[6]  Annika Eichler,et al.  Controller Tuning by Bayesian Optimization An Application to a Heat Pump , 2019, 2019 18th European Control Conference (ECC).

[7]  Yaoyu Li,et al.  Intermediate pressure optimization for two-stage air-source heat pump with flash tank cycle vapor injection via extremum seeking , 2019, Applied Energy.

[8]  Dieter Schwarzmann,et al.  Data-Efficient Autotuning With Bayesian Optimization: An Industrial Control Study , 2018, IEEE Transactions on Control Systems Technology.

[9]  Tao Wen,et al.  Energy-efficient and -economic technologies for air conditioning with vapor compression refrigeration: A comprehensive review , 2018, Applied Energy.

[10]  Martin Guay,et al.  A proportional integral extremum-seeking control approach for discrete-time nonlinear systems , 2015, 2015 54th IEEE Conference on Decision and Control (CDC).

[11]  James Hensman,et al.  Scalable Variational Gaussian Process Classification , 2014, AISTATS.

[12]  Martin Guay,et al.  A time-varying extremum-seeking control approach for discrete-time systems , 2014 .

[13]  Jasper Snoek,et al.  Practical Bayesian Optimization of Machine Learning Algorithms , 2012, NIPS.

[14]  Sigurd Skogestad,et al.  NCO tracking and self-optimizing control in the context of real-time optimization , 2011 .

[15]  Wenming Yang,et al.  Advances in heat pump systems: A review , 2010 .

[16]  Perry Y. Li,et al.  Efficient Operation of Air-Side Economizer Using Extremum Seeking Control , 2010 .

[17]  Christopher R. Laughman,et al.  Extremum Seeking Control for Energy Optimization of Vapor Compression Systems , 2012 .