Ensuring Replicability of Power HIL Simulations: A Proposal towards comparable PHIL tests (PHIL-Rep)

Power Hardware-in-the-loop (PHIL) testing has emerged as an appropriate validation ap proach for electrical systems. PHIL set-ups provide an environment for repeatable, economi cal, flexible, and scalable hardware and software verification under realistic conditions. How ever, PHIL set-ups are complex systems which design, and hardware components significantly influence its results. Therefore, a potential concern is the replicability of PHIL results over dif ferent lab infrastructures. In that context, this work aims to investigate the aspects related to the PHIL results replicability through analysis of the behavior of a same case of study in two different PHIL set-ups.

Lab Access (LA) User Group (UG) works at CEPEL’s Smart Grid laboratory (Lab SG), a recent inaugurated infrastructure located in Brazil, that has, within its resources, a PHIL test bench composed by one digital real-time simulator (OPAL-RT), two linear amplifiers (Spitzenberger & Spies) and three programable inverters (Triphase). Part of these devices were used, before LA period, to stablish a PHIL set-up for distributed PV generation integration analysis in which an IEEE benchmark feeder interacts with a real inverter that represents PV generation (HUT – Hardware under test). The tests performed at CEPEL Lab focused on voltage profile behav ior, using the Brazilian standard for power quality in distribution grids as the main standard reference.

During the 4 weeks of LA period detailed in this report, a replicated PHIL set-up was estab lished using Host Laboratory (ESP Lab) infrastructure, in which experimental simulations of two test cases were carried out. A set of real electrical measurements and virtual variables were obtained as results of the performed test. This report presents a broad description of the main hardware and software configurations as well as the differences between both set-ups. LA period was also used to implement a simplifier PF closed loop control in which active and reactive power references are sent to HUT, to implement an ancillary service for voltage profile control. This control was replicated at CEPEL Lab after LA period.

Preliminary findings

• Even though ESP Lab’s PHIL set-up significatively differs from the one available at CEPEL Lab, both set of results lead to the same conclusion regarding PV hosting ca pability levels due to overvoltage issues.
• The implemented PF control loop proved to be appropriate, for the given modeled grid, to avoid overvoltage deviations.
• It was not possible verify the reproducibility of CEPEL’s and ESP lab’s results at wave form-comparison level because of the relevant differences between both PHIL set-ups. As: i) four-wire equipment (CEPEL) vs three-wire equipment (ESP Lab), ii) interface algorithm based on instantaneous variables (CEPEL) vs interface algorithm with rms quantities (ESP Lab) and iii) non-real time MODBUS communication for PF control set point establishment (CEPEL) vs Aurora real time communication (ESP Lab).

Open threads:

• Change of the electrical model represented at the RTS for one that can be fully repre sented by ESP Lab, for instance, with no unbalanced voltages and currents.
• Modification of internal Triphase converter controllers to have the same set of adjust ments in both PHIL set-ups.