PSI - Issue 44

Mauro Mazzei et al. / Procedia Structural Integrity 44 (2023) 1212–1219 Mauro Mazzei et al. / Structural Integrity Procedia 00 (2022) 000 – 000

1218

7

Where P is the induced force, l = 0,01 m, distance from the frame of the suspended mass, E=200GPa the Young module, I=[0,03*(0,0004)3]/12 the moment of inertia. For P = 0,200 Kg it provides δ d = (0,2x10 -6 ) x 12 / (3 x 2 x 10 11 x 1.9 x 10 -12 ) = 2.1 um So far an acceleration of 1g would provide on the proof mass a d displacement larger than 20 um. If even one quarter of the area 8 cm x 8 cm proof area is parallelly faced to an equivalent plate to implement one of the Cx capacitors at a rest distance of 500 um we would observe on that a variation from 28,35 pF to 29,56 pf or 27,21 pf according the considered face or a to tal ΔC = 2.35 pF With this design parameter we would be ina maximum 1 g case to less that the 5% displacement and well matching the overall range of the of the CAP sampling part. This has a resolution better than 25 aF and an on chip temperature measurement of better than 0.1°C. 5. Data analysis Identification of anomalies beyond the signal-to-noise limits of the instrument is achieved by normalization of distributed strains measured during different positions of a passing vehicle relative to the theoretical line of influence of the bridge.

Fig. 5. Line of influence of the load point moment

Figure 5 shows a vehicle with three axles in five different load positions. Considering the relationship between bending stress and moment. Conversely, the strain profile can be determined from the distributed strain measured during load tests of the bridge. An example of this graph for a five-span bridge is shown in Figure 6. The theoretical strain profile is curvilinear because it represents the strain history of a specific point under a continuously moving load. The strain profile shown in Figure 6 is linear because it is constructed by measuring the strain for five distinct carriage positions.

Fig. 6. Deformation profile measured for Pi point due to vehicle loading.

The analysis of data through these devices can be of great interest in the study and optimization of infrastructure monitoring, then the problem of reconstructing the motion of a body in space and analyzing it precisely to monitor causes and effects becomes essential, starting from the knowledge of previous recurrences, the phenomenology of movements can be framed deterministically, or it can be characterized by having only some properties that follow statistical laws, we speak in that case of stochastic motion or random vibrations. Stochastic measurement errors are not modelable and generally do not have a fixed distribution but are dependent on the momentary physical context in which the sensor is located; this actually makes it very difficult to make