23
Problems in Validating Control Feel in Simulators By T. Scott Davis Aerospace Engineer, Member AIAA [email protected] Bruce Hildreth [email protected]

Problems in Validating Control Feel in Simulators By T. Scott Davis Aerospace Engineer, Member AIAA [email protected] Bruce Hildreth [email protected]

Embed Size (px)

Citation preview

Problems in Validating Control Feel in Simulators

ByT. Scott Davis

Aerospace Engineer, Member [email protected]

Bruce [email protected]

Overview

Introduction

Control system– Model – Control loading system– Flexure

Testing– External Instrumentation– Internal Instrumentation– Test results

Conclusions

Introduction

US Marine Corp KC-130T APT– Under subcontract to JF Taylor Inc / Boeing.– SAIC role

• Re-host KC-130T software for modeling flight dynamics

• Procure control loading system hardware• Acceptance testing of control loading system

Introduction (continued)

Fixed based Aircrew Procedures Trainer– Full fidelity with malfunctions– Pilot, Co-pilot, Flight Engineer, Navigator– Wide field of view visuals– Fully functional cockpit – Transportable enclosure

Introduction (continued)

SCT Control Loading system– 4-axes

• Column• Wheel• Rudder• Tiller (Nose wheel steering)

– Pilot and Co-pilot controls physically connected

– Model run on a DOS based PC

– Ethernet communication to the host

Overview

Introduction Control system

– Model – Control loading system– Flexure

Testing– External Instrumentation– Internal Instrumentation– Test results

Conclusions

Control System Model

Complex system of flexible linkages/cables Position of control versus position of surface

dependent on loading

Mass and Flexibility Distributed throughout system

Linkage or cables from cockpit to control surface

Column

Mechanical components in or under cockpit

Mechanical components, actuators, linkages and control surface

Elevator

Simulation Control Loading Model

Reduce system to two masses with single flexible linkage/cable

Forward Mass Aft Mass

Column

Flexible Cable or Linkage

Elevator

All components lumped into two masses and a

connecting single spring

Simulation Control Loading Trainer Layout

Actuator used to position control based on applied force

Column

Control Loader Actuator

Position Sensor

Force Sensor

Cockpit Floor

Effect of Flexibility

Physical control loading system will flex

This may be flexibility already in model and thus be double accounted

Flex in control loading system may not be similar to forward system flex in aircraft

Linkage locked

Column

Applied ForceApplied Force

Overview

Introduction

Control system– Model – Control loading system– Flexure

Testing– External Instrumentation– Internal Instrumentation– Test results

Conclusions

Control Loader Testing

External testing (SIMES)– Force sensor at Yoke (measures applied force)– Position sensor at column (measures control

position)– Effect of flexibility

• External position maybe different than control loader measured positionColumn

Control Loader Actuator

Position Sensor

Force Sensor

External Force Sensor

External Position Sensor

Cockpit Floor

Control Loader Testing

Internal (Control loader self test)– Loader applies a fictitious force to the model– Control loader model moves control to

appropriate position– Tests control loader model only, not control

loaders measurement of force or position– Effect of flexibility

• No flex due to no actual force applied• No ability to measure flex if it was there

KC-130T APT Test Results

Only the Cruise condition results for the column force versus displacement are shown

Testing method– Simulation was initialized to test condition and

frozen– Control loader remained active– SIMES used to record force and position as

control was slowly moved though its range of motion or to maximum measurable force

– Testing also done with internal control loader software

KC-130 APT Results

Post processing– Force translated to reference location (pilot

grip)– Position translated to angular measure from

linear transducer (from calibration)– Estimation of surface position from measured

control position• Initially, poor assumptions made when

analyzing the data– Assumed rigid cable– Assumed rigid control loader

KC-130 APT Results

Needed to explain differences between SIMES results and internal results

KC-130 APT Results

Initially removed assumption of rigid cable

KC-130 APT Results

Additional testing of flex in simulator controls-Simple linear spring determined to model control flex

KC-130 APT Results

Force vs. deflection plot clearly illustrates relative effects

RIGID SYSTEM

ACTUAL FLEXIBLE SYSTEM

Just Cable Stretch

Overview

Introduction

Control system– Model – Control loading system– Flexure

Testing– External Instrumentation– Internal Instrumentation– Test results

Conclusions

Conclusions

Simulator control feel liked by aircrew

Additional testing required to explain why the external test system test results were reasonable when compared to the criteria data

The above infers:– The criteria data did not include flexure of the

control system in the aircraft

Conclusions

Expectations– Did not expect flex in

simulator controls to be more significant than cable flex

Test Results– Control Loader Internal

testing not able to reflect true control position to surface position

– Control loader testing must measure

• Control force• Control position• Surface position

– Criteria should include relationship between these three parameters.

RIGID SYSTEM

ACTUAL FLEXIBLE SYSTEM

Just Cable Stretch

Conclusions

A/C Criteria data must:– Specify where the force and position sensors are

for the test– Test data should be taken with an external test

system at the point of pilot force application– For a two pilot a/c, both sets of controls should be

measured– Specific tests should be made to measure flexure

between the pilot and the control surface

Simulator control loading must: – Include aircraft control system flexure in the model

• Linkages and controls

Beware of double accounting