RRG Research Actuator Testbeds

Research

Actuator Performance Maps and Testbed

Objective

Approach

Operations of Actuator Test Bed

Performance Maps from Actuator Test Bed

Videos

Related Links

Objective

The goal of this research was to generalize actuator test data in the form of performance maps, which would document the characteristics of each family of motor types (SRM, BLDC, etc.), with a range of rated output power, under dynamic load operations. A collection of performance maps is our approach to representing the complexity of the actuator response (in both the model and the sensor data) and also provides metrics for measuring the actuator state. These actuator criteria maps can be used to build performance envelopes, and to provide a reference basis for CBM. This research also includes the development of a complete test system to emulate all possible real situation.

In order to complete the actuator test architecture considering all the possible situations, a dynamic and nonlinear test environment was required. The Nonlinear Test Bed for Actuators (NTBA) was created to measure and record an array of physical properties during nonlinear load experiments. Many different types of loads (such as a programmed nonlinear load by a load motor, and a constant load by hysteresis brake) were applied to emulate the most general situations during real operation of the actuator. The NTBA validated that the proposed prime mover criteria were accurately obtained through experimentation and their usefulness was proved in developing the operational performance envelopes.

Approach

I. Development of Actuator Performance Maps

A set of the experimental data developed in Nonlinear Test Bed is used to generate performance maps. Quantitative metrics thorough testing of the actuator provide necessary feedback information on the effectiveness of the actuator for a specific task. Performance maps and envelopes generated in this test bed will improve understanding of an actuator’s nonlinear characteristics.

Once all of the data sets are collected, they are combined based on each actuator criterion. The preferred maps are continuous, limited in magnitude, implicitly time dependent, graphically clear, and vary significantly over the actuator’s regime of operation. A full set of performance maps based on operating criteria can be generated using appropriate operational parameters monitored during the full rage of tests. The performance map is defined as

where X is the vector of the states, B is the unknown coefficients, and u is the input.

II. Nonlinear Test Bed

RRG’s initial motivation for building the Nonlinear Test Bed was to find the performance envelope plots for Permanent Magnet Synchronous Motors, but the Test-bed had to be flexible enough to be used for a variety of prime movers as well as future research efforts. The maximum power of the load emulator is approximately 10 HP with a maximum speed up to 5000 RPM and a maximum torque up to 40 lb-ft.

The mechanical component set-up for the test bed is shown in Figure 1. This test bed is for the dynamic test, so the load motor is used to generate the various load types. The special load type is selected to express the realistic situation for the given test motor. The clutch is used to engage to tranfer a load type from the load motor and disengage for emergency stop. The torque sensor is needed to measure the torque between the load motor and the test motor. The encoders estimate the position, velocity and the acceleration. In addition, three bellows couplings are used to connect each component to the rest of the system. The couplings were carefully chosen to endure more than 40 Lf-ft in the rotational direction. Figure 2 shows the electrical component set-up for the test bed.

Figure 1. Nonlinear Test Bed for Actuator

Figure 2. Electrical arrangement of NTBA

Operations of Actuator Test Bed

The National Instruments (NI) motion controller has eight input channels so a NI’s data acquisition board will be used to take the additional signals. In the test motor, encoder and hall signals are needed for feedback to the amplifier. These signals are needed to complete the sinusoidal commutation loop to run the test motor. The load motor does not need to feedback the position signals and hall signals to the amplifier because the load motor commutates the current signals in the motion controller and not in the amplifier. The test motor amplifier ground is not isolated from the power line, so the grounds for the amplifier and motion controller are connected to the earth. Moreover, the one to one ratio of isolated power transformer is used to generate the voltage output range of 0VDC to 149VDC. Finally, in order to control the engagement of the brake and clutch, one of the digital output channels is used. This signal will be sent to the brake and clutch in an emergency case or during operational. The NI motion controller provides over 100 vi's for specifying motion and machine parameters. Commands are included to initiate action, interrogate status and configure the controllers and filters. These NI’s commands help the user to generate, store and execute many complex application programs. Additionally, NI provides DLL files to develop the program by using C/C++ or Visual Basic. All of these features will help us to build real time operation system with fast sampling and feedback update speed.

Figure 3. Labview program to drive actuator control system

Performance Maps from Actuator Test Bed

The state values of interest from the sensors in the test bed are position, velocity, torque, temperature, voltage, current, and magnetic flux density. However, the parameters estimated from the embedded sensors in the test bed include signal noise, so mathematical tools (e.q. butterworth, bessel filters) are utilized to reduce this noise. The raw data from the test bed are conditioned by monitoring the stochastic process, developing dynamic regression models and tracking test protocol of the actuator. Additionally, the data obtained in real time will have associated upper and lower error bounds to indicate the uncertainties in the sample data.

The actuator test bed can develop the maximum performance envelope for a specific actuator (based on a specific set of performance criteria) from a complete set of performance maps. A particular map will have several different layers based on the different input loads with the same control parameters. The performance envelope is the line with the maximum performance capacity from the full set of maps which meets the desired criteria for the specific actuator. Figure 4. shows the sample plots of performance maps based on motor test protocols.



Figure 4. Experimental plots of performance maps

Videos

Monitoring signals in Labview : The frequency and magnitude of sinusoidal command inputs to the test motor were 1.1 Hz and 4.5 Amps. The load motor ran at 50 RPM. The current/voltage in three phases and magnetic flux density signals were monitored. Also, the temperature inside the test motor was measured.

Operational margin test with sinusoidal torque input : In the first stage, the movie showed the load motor was running at 150 RPM without engaging the clutch for loading. Then, the test motor was tested with the frequency of 0.5 Hz and the magnitude of 3 Amps in sinusoidal command inputs. At this time, the clutch was disengaged, either. Once each of the motors was tested respectively, the clutch was engaged and both of the motors were connected for emulating a real situation. They started running at the same time with a given load profile.

Torque ripple test using Hysteresis Brake : The test motor was tested with Hysteresis Brake to develop torque ripple and acceleration performance maps. Torque was produced through a magnetic air gap, without the use of friction or shear force using the Hysteresis Brake. Also, full torque was available at zero slip speed and torque remained smoothly at any rotor speed.