This paper describes a turbine engine thrust estimator that computes “virtual measurements” of dynamic engine thrust and other parameters of interest from test cell data in a very short amount of time. The system ‘tunes’ a user’s engine model, as developed in the commonly used Numerical Propulsion System Simulation (NPSS), by optimizing system biases and health parameters to match the sensor outputs of a set of steady state data points across the operating range. The tuned model is then used to create a constant gain extended Kalman filter that is added directly within the NPSS model code. The results, including thrust, from this NPSS model with Kalman filter are then presented as the ‘actual’ corrected data.

Key aspects of the system include:

• Utilization of and tight integration with NPSS. This ensures that the results always preserve mass and energy, and are realistic from an engine performance point of view.

• Flexibility; any NPSS model for any gas turbine engine can be used.

• The ability for the whole tuning process to ‘correct’ not just noisy test data, but also performance parameters within the user’s actual NPSS model.

• A GUI that leads the user through each step of the process, such as matching NPSS variable names to signals in the actual test cell data files, and selection of ‘tuners’ (performance parameters, sensor errors) and Kalman filter variables.

The system has been tested both on a simulated two spool turbojet engine, plus an actual two spool turbojet engine, with recorded test cell data.

This content is only available via PDF.
You do not currently have access to this content.