When DIC technology is used for vibration modal analysis, its reliability can be verified through several methods. These include comparing the results with theoretical models or simulations to calculate theoretical natural frequencies and mode shapes. Alternatively, DIC technology can be used to measure the vibration modes of the actual structure, and the measurement results can be compared with theoretical values to calculate the relative error of the natural frequencies and the similarity of the mode shapes (e.g., modal confidence factors). Errors within a reasonable range (e.g., ±5%) indicate that the DIC measurement is reliable.
Compared with data from contact sensors, contact sensors (such as accelerometers and displacement gauges) are placed at key locations on the structure to collect vibration data simultaneously.
Compare the DIC measurement results with the sensor data, and analyze the consistency of displacement time history curves, peak values, phases, etc. If the two are highly consistent, it indicates that the DIC measurement is accurate.
Repeatability testing: Perform multiple DIC measurements on the same structure, keeping the measurement conditions (such as excitation method and environmental parameters) consistent each time.
Calculate the statistical indicators (such as mean and standard deviation) of multiple measurement results. If the results fluctuate little, it indicates that the DIC technique has good repeatability and high reliability.
Verification and Uncertainty Assessment of DIC Technology Modal Analysis Results
Modal indication function: helps determine the location of modal frequencies in the frequency domain method.
Stability plots are crucial in time-domain methods, especially ERA and SSI. By changing the model order, stable (physical) modal poles (frequency, damping ratio) are identified, while unstable (computational) poles are eliminated. This is a key tool for improving the reliability of the identification process.
Modal Confidence Criterion (MAC): Quantitatively assesses the correlation between identified mode shape vectors or with the mode shapes of a reference model (such as FEM). A high MAC value indicates a high correlation.
Comparison with reference sensor: The main frequencies, damping ratios, and mode shape components at the reference point identified by DIC are compared with the results of traditional sensors to verify consistency.
Physical rationality check: Check whether the mode shape conforms to the principles of structural dynamics (such as the position of nodal lines and symmetry).
Repeatability test: Repeat the experiment under the same conditions to evaluate the reproducibility of the identification results.
Uncertainty quantification: This involves analyzing the impact of factors such as noise, parameter settings, and algorithm selection on the recognition results. Uncertainty sources specific to DIC (such as speckle quality and calibration error) also need to be considered.
Key points to improve the accuracy of DIC modal analysis
Initial investment: High-quality speckle patterns, high frame rate and high resolution cameras, stable and uniform lighting, solid installation, and accurate calibration are the basic guarantees.
Frame rate priority: Ensure the sampling rate is much higher than the highest target modal frequency.
Precise synchronization: The hammer/vibrator signal, camera trigger, and image timestamp must be precisely synchronized.
Noise control: High importance is attached to noise reduction and filtering of displacement data (time domain + spatial domain).
Data Management: Utilize data reduction techniques such as PCA/SVD to efficiently process massive amounts of data.
Algorithm selection: Select a suitable modal parameter identification algorithm (frequency domain curve fitting, SSI, ERA, etc.) based on the excitation type (known input/OMA) and data characteristics.
Make good use of tools: Modal indicator functions, stability plots (time domain method), and MAC values are indispensable verification tools.
Cross-validation: Validation must be performed using a small amount of data from traditional reference sensors.
Experience and iteration: Modal parameter identification often requires empirical judgment (such as stability plot interpretation and model order selection) and parameter adjustment iteration.
