Scribd
Upload
3 views

Data Filtering Methods

Uploaded by

Ogada
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
3 views

Data Filtering Methods

Uploaded by

Ogada
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 14

Brian ,

Ogada
INSTITUT FÜR
MASSIVBAU

Structural Health Monitoring of


Bridges
Data Filtering
Methods
A Brief Introduction to Data Normalization Statistical
Methods
Monday ,
5th ,2024
1.0 Introduction
Data filtering part of Data cleaning that aims to achieve quality data.
Data Cleaning is the process of implying hard limits in which the data is not usable and must be
discarded (noise/ outlier treatment) or missing data to be imputed.

Structural Health Monitoring Devices record and transmit data if form of


 Vibration Data For accelerometers , this data includes frequencies and amplitude of
Vibrations
 Strain Gauge Measurements Raw strain measurements for deformations
 Temperature data Thermocouples or infrared sensors measure temperature
changes
 Acoustic Emissions Capturing Soundwaves emitted by cracks or other damage
processes
1.0 Introduction – Objective

The Mentioned sensors collect information and transmits the information for Processing.
 Noise Reduction - Removing random noise and interference.
Relevance : Ensures clearer and more reliable data.
Application : Removal of ambient vibrations or electrical interference.
 Signal Enhancement - Enhancing key features of the data signals.
Relevance : Facilitates easier detection and analysis of structural responses
Application: We can Isolate for instance vibration modes or strain levels
 Data Consistency - Ensuring uniformity and reliability over time.
Relevance : Accurate monitoring of structural changes and trends
Application: Consistent readings across different sensors.
1.0 Introduction – Objective

The Mentioned sensors collect information and transmits the information for Processing.
 Removal of Outliers –Identifying and removing erroneous Data Points.
Relevance : Prevents distortion in analysis and conclusions.
Application : Unusual spikes due to sensor malfunctions.
 Frequency Isolation- Focusing on specific frequency bands of interest.
Relevance : Identifies structural resonances and specific responses
Application: Detecting natural frequencies associated with bridge vibration
 Data Consistency - Ensuring uniformity and reliability over time.
Relevance : Accurate monitoring of structural changes and trends
Application: Consistent readings across different sensors.
2.0– Methods of Data Filtering
In signal processing, particularly for linear time-invariant (LTI) systems used in filtering, the
relationship between the input and output signals can be represented using difference
equations:
2.0– Methods of Data Filtering
2.0– Methods of Data Filtering
2.0– Methods of Data Filtering
Low-Pass Filtering High-Pass Filtering
 Purpose: Allows frequencies below a certain  Purpose: Passes frequencies above a
cutoff to pass through, removing high-frequency certain threshold, removing low-frequency
properties. properties.
 Application: Useful for eliminating high-  Application: Helps in isolating high-
frequency sensor attributes such as noise while frequency components, such as those from
preserving the structural response signals. damage-induced vibrations.
2.0– Methods of Data Filtering

Low-Pass Filtering High-Pass Filtering


 Allows all frequencies
above the cut off
frequency to pass and
attenuates the level of the
2.0– Methods of Data Filtering
Band-Pass Filtering Notch Filtering
 Purpose: Allows only a specific range of  Purpose: Eliminates specific narrow
frequencies to pass, removing frequencies frequency bands, often to remove known
outside this range. interference or resonant frequencies.
 Application: Used to focus on specific frequency  Application: Effective in removing power
bands associated with certain structural line interference or specific vibrational
behaviors or modes. modes.
2.0– Methods of Data Filtering

Band-Pass Filtering Notch – Band Stop Filtering


 The coefficients are utilised to create  Attenuate signals within a

a peak in the frequency response at the specific frequency range while


desired band and attenuation allowing signals outside this
elsewhere
3.0– Criteria for selection and Conclusion
Selection Criteria
 Nature of Data :
-Consider whether the data is stationary or non-stationary, linear or non-linear.
-Identify the frequency range of interest and the presence of noise.
 Objective Of Filtering :
-This Considers filtering to remove noise, extract features such as damage induced
-Vibrations or smoothing for long term monitoring and reliability of data.
 Structural Characteristics :
-Different materials (steel, concrete ) have different dynamic properties influencing
choice.
-Known Resonances: The presence of known resonant frequencies can dictate the
use of notch filters for instance to eliminate specific frequency components.
3.0– Criteria for selection and Conclusion
Selection Criteria
 Sensor Specifications:
-The type and placement of sensors can affect the noise characteristics and the type of
filtering required. For example, accelerometers may require low-pass filtering to remove
high-frequency noise.
-High sensor Accuracy may necessitate less aggressive filtering.
 Environmental Factors
-Ambient Conditions: Environmental factors such as wind and temperature, can introduce
noise. Filters should be chosen based on the predominant environmental conditions.
-Operational Environment: Consideration of the operating environment, such as
high-vibration or high-noise areas, is crucial for filter selection.
3.0– Criteria for selection and Conclusion
Conclusion
 In conclusion, the selection and implementation of appropriate data filtering methods are
essential for maximizing the effectiveness of SHM systems in bridges.
 Different filters are applied for different structures or materials , objectives and ambient
conditions based on the sensor types also being utilized
 These techniques not only ensure data quality but also contribute to a deeper understanding of
structural behavior, ultimately leading to safer, more reliable, and longer-lasting infrastructure.

You might also like