Strain Virtual Sensing Applied to Industrial Presses for Fatigue Monitoring

Area of Science:

• Mechanical Engineering
• Structural Health Monitoring
• Signal Processing

Background:

• Industrial equipment, especially under high cyclic loads, is prone to fatigue damage.
• Condition monitoring systems are crucial for maintaining industrial machinery.
• Virtual sensing offers a method to estimate measurements at unmonituted locations.

Purpose of the Study:

• To evaluate the effectiveness of different virtual sensing algorithms for strain estimation.
• To assess the performance of these algorithms with a reduced sensor count.
• To determine the feasibility of using virtual strain sensors for fatigue crack detection.

Main Methods:

• Tested three virtual sensing algorithms: Direct Strain Observer, Least-Squares Strain Estimation, and Static Strain Kalman Filter.
• Utilized real measurement data from a scaled bed press prototype under cyclic loading.
• Employed strain gauges for actual strain measurements on the prototype.

Main Results:

• Virtual sensing algorithms achieved sufficiently accurate strain estimations.
• Reduced numbers of strain gauges proved effective for static and quasi-static load conditions.
• Virtual strain sensors successfully detected the initiation of fatigue cracks.

Conclusions:

• Virtual sensing is a viable technique for condition monitoring of industrial equipment.
• It allows for accurate strain estimation with minimal sensor deployment.
• This approach aids in the early detection of structural damage, such as fatigue cracks.