Benchmark testing methodology for continuous modal-estimation Algorithms

A practical methodology for defining the required characteristics of sets of stationary and non-stationary test-signals for benchmarking the performance of modal-estimators is described along with the assessment criteria. The algorithm for the modal-estimator assessed in this paper is used in the Australian National Grid for system damping performance and model calibration. The efficient computational methods used to generate the test-signals are described and are based on detailed models of the operator's own system. The benchmarking methodology has been applied in practice to rigorously and independently assess the relative performance of a number of practical modal-estimators. As a result the system operator is able to identify the key strengths and limitations of the various modal-estimators for their own system and apply the appropriate estimator to the desired application, i.e. short-term or long-term estimation. The set of generated test-signals has proved to be of benefit to the developers of modal-estimators as they endeavour to improve their estimation algorithms.