Avoiding Surprises with Model Based Twin for Protective Relays

Simulation studies will be of value only when the models closely mimic the response of the equipment of interest for that specific study. In the case of studies with IBRs, this is a big challenge as the control algorithms are proprietary. However, there are good advancements in the modelling of the modern electronics dominated power systems, e.g. the evolving IEEE 2800.2 standard, which defines the recommended practices for test and verification procedures for IBRs interconnecting with bulk power systems. The protection algorithms on the other hand remain proprietary and these studies lack confidence when the relay models used in simulation studies do not actually mimic the true response of the relay. 1 The relay under evaluation can be represented/modelled in non-real time or real time simulation studies/testing in different ways as listed below,

•

•

•

•

•

• Steady state model.

Transient model

Generic models available in simulation software

Models which are recreated from actual relay code to mimic the performance

Research/Proof of concept model

Real-code model o Use of executable versions of code e.g. IEEE/CIGRE B4.82

▪ based on the high-level language e.g. C/C++, used in the relay firmware,

▪ based on the model-based design which is used in relay firmware. Each of the above-mentioned approaches has its own limitations, e.g., executable versions of code using high level language could be a promising option, but it cannot provide better insights, and it is relatively time-consuming. The paper also discusses briefly different levels of testing, i.e.,

• Model in loop,

• Software in loop,

• Processor in loop and

• Hardware in loop and their limitations. Also, how can these limitations help expose hidden failures with some real-world examples.

Over a century, the relay industry has seen many technological evolutions to realize protection i.e., electro-mechanical, static, digital and numerical relays. The built-in flexibility provided by the numerical relays helped to realize complex functions as the protection industry evolved to meet new challenges. However, there has been no significant change or improvement in the design, development and post-fault investigation phase over a couple of decades.

This paper showcases how significant improvement can be made in different stages of the protective relay development, as well as in simulation studies using model-based relay design.

It also shows how model-based relay design can provide valuable insights even at preliminary design stages and during post-fault investigation, especially of interest during interconnection studies with inverter-based resources. With over a decade of model-based relay design experience for complex algorithms which are successfully used in real-world applications, we share our lessons learned, benefits and their limitations in this paper.