Vibrationdata Unit 21 Integration and Differentiation of Time
Vibrationdata Unit 21 Integration and Differentiation of Time Histories 1
Accelerometer Vibrationdata • Mechanical vibration is usually characterized in terms of acceleration • The main reason is that acceleration is easier to measure than velocity or displacement • Acceleration can be measured with a piezoelectric, piezoresistive or variable capacitance accelerometer 2
Velocity Criteria Vibrationdata • Hunt, Gaberson, Bateman, et al, have published papers showing that dynamic stress is directly proportional to modal velocity (future webinar) • A peak velocity of 50 in/sec is sometimes considered as the shock severity threshold for military components • Allowable building floor vibration limits are typically < 2. 0 in/sec • Colin Gordon has established a generic vibration criteria for building floor vibration in terms of velocity (see ISO Generic Vibration Criteria for Vibration. Sensitive Equipment) 3
Velocity Sensor Vibrationdata • Velocity measurements require a Doppler laser or a geophone • The laser is expensive and requires a direct line of sight • The geophone is bulky and is intended for seismology measurements 4
Geophone Vibrationdata 5
Laser Vibrometer Vibrationdata Advantage No mass loading effect from laser on object. Disadvantage Laser system actually measures relative velocity between laser source and object, so laser source must be kept still. A single point laser vibrometer is used to compare the vibration of two similar guitars 6
Scanning Laser Vibrometer Vibrationdata • A Scanning Laser Vibrometer measurement shows the velocity profile of a vibrating turbine blade • The measurement grid has been tailored to match the specific shape of the blade 7
Displacement Sensor Vibrationdata • Dynamic displacement can be measured by a linear variable displacement transducer (LVDT) • The frequency response is only suited for lowfrequency measurements LVDTs used to measure traffic-induced vibration on underside of bridge 8
Old School Analog Method for Measuring Velocity & Displacement • Measure vibration with charge mode piezoelectric accelerometer • Analog signal goes through Bruel & Kjaer 2635 signal conditioner • Select acceleration, velocity or displacement output with this knob • Analog integration & double integration applied for velocity & displacement, respectively • Highpass filtering needed to prevent spurious offsets, drifts, etc. • Minimum highpass filtering frequencies: 0. 2 Hz for acceleration 1 Hz for velocity & displacement Vibrationdata 9
Typical Building Vibration Limits Vibrationdata Transportation Research Board Building Maximum Structure Vibration Criteria Limiting Peak Particle Velocity Structure and Condition (in/sec) (cm/sec) Historic buildings, Certain other old buildings 0. 5 ~1. 3 Residential structures 0. 5 ~1. 3 New residential structures 1. 0 ~2. 5 Industrial buildings 2. 0 ~5. 1 Bridges 2. 0 ~5. 1 10
Hyatt Regency Hotel, Phoenix, Arizona Vibrationdata Typical Elevator Recommended Limits Parameter Limit acceleration/ deceleration < 1. 0 - 1. 5 m/sec^2 Speed < 7. 0 m/sec Jerk rates < 2. 5 m/sec^3 Sound < 50 d. Ba Ear-pressure change < 2000 Pa Fast elevator ride from ground floor to top restaurant! 11
Vibrationdata Accelerometer Measurement Integrated Velocity 12
Hyatt Regency Elevator Vibrationdata Accelerometer Measurement Differentiated Jerk 13
Integration, Trapezoidal Rule Vibrationdata The integration of a time history is carried out on a “running sum” basis. Let the acceleration time history be represented by a 1, a 2, a 3, . . . , an. The velocity time history is calculated as follows. 14
Differentiate, Matlab Function Vibrationdata function[v]=differentiate_function(y, dt) % ddt=12. *dt; % y = input amplitude v = output amplitude dt = time step n=length(y); % v(1)=( -y(3)+4. *y(2)-3. *y(1) )/(2. *dt); v(2)=( -y(4)+4. *y(3)-3. *y(2) )/(2. *dt); v(3: (n-2))=(-y(5: n)+8*y(4: (n-1))-8*y(2: (n-3))+y(1: (n-4)))/ddt; v(n-1)=( +y(n-1)-y(n-3) )/(2. *dt); v(n) =( +y(n-1)-y(n-2) )/dt; 15
Sine Example Vibrationdata Generate sine function: Amp = 1 Dur = 10 sec Freq = 1 Hz Sample Rate = 40 Hz (assume amp unit: G ) Save as: sine_accel 16
Integrate from Acceleration to Velocity Vibrationdata Baseline shift Mean 61 in/sec Vibrationdata > Time History > Integrate Input File: sine_accel Trend Removal = None (prior & after) Output File: sine_vel 17
Integrate from Velocity to Displacement Vibrationdata Ski Slope Effect! Vibrationdata > Time History > Integrate Input File: sine_vel Trend Removal = None (prior & after) Output File: sine_disp 18
Integrate from Acceleration to Velocity with Mean Removal Vibrationdata Symmetric Oscillation about zero baseline Vibrationdata > Time History > Integrate Input File: sine_accel Trend Removal Prior = None After = Mean Output File: sine_vel 19
Integrate from Velocity to Displacement with Mean Removal Vibrationdata Stable oscillation about zero baseline But with some distortion Vibrationdata > Time History > Integrate Input File: sine_vel Trend Removal Prior = None After = Mean Output File: sine_disp 20
Differentiate from Displacement to Velocity Vibrationdata > Time History > Differentiate Input File: sine_disp Output File: sine_vel 21
Review Exercise, Sine Amplitude Vibrationdata Agrees with integration & differentiation results on previous slides Vibrationdata > Miscellaneous Functions > Steady-state Sine Amplitude 22
Launch Vehicle Separation Test Vibrationdata Filename: pyro_test. txt 23
Integrate from Acceleration to Velocity Vibrationdata > Time History > Integrate Input File: pyro_test. txt Trend Removal = None (prior & after) 24
Integrate from Acceleration to Velocity with HP Filtering Vibrationdata > Time History > Integrate Input File: pyro_test. txt Trend Removal Prior: Highpass filter at 30 Hz After: none 25
Recall PSD Synthesis Vibrationdata 26
PSD Synthesis Review Vibrationdata 1. Generate acceleration white noise 2. Manipulate the time history via FFTs and inverse FFTs so that its satisfies the PSD specification 3. Integrate resulting acceleration time history to velocity 4. Integrate resulting velocity time history to displacement 5. Remove third-order polynomial trend from displacement 6. Apply tapering using half-cosine function to beginning and end of displacement 7. Differentiate displacement to velocity and again to acceleration Steps 3 through 7 allow the set of acceleration, velocity and displacement time histories to each have zero mean values. 27
- Slides: 27