PSI - Issue 77

Carolina Francisco et al. / Procedia Structural Integrity 77 (2026) 567–574 C. Francisco et al. / Structural Integrity Procedia 00 (2026) 000–000

571

5

monitoring, the acquired data is compared with these predefined limits to assess the press condition; if the measured values exceed the established thresholds, an anomaly is detected.

3. Experimental analysis

To develop the condition monitoring system and evaluate the proposed processing methods, acceleration data were collected from a stamping press. Although this press di ff ered from the one under development in loading conditions and size, it shared the same main components—the slide, frame, and shaft (Figure 3a). Three operating scenarios were tested: continuous operation without load, with load, and under overload. Since the press was newly manufactured and not yet equipped with a stamping tool, a metal block was used to simulate the process, with the slide applying the load upon contact. The slide stroke was set to 52 cm, and two speeds were tested: 50 and 75 strokes per minute. In the no-load case, the slide moved freely; in the load case, contact occurred at bottom dead center. For overload, the slide traveled slightly below the block surface, producing an increased load that triggered an automatic stop

(a) Press parts.

(b) Acquisition equipment.

Fig. 3: Accelerometer data acquisition from a real press.

Triaxial accelerometers were installed on key components to measure acceleration. Each sensor was connected to an NI 9232 module and an NI cDAQ-9189 chassis (as shown in Figure 3b), with data transmitted to a local PC and stored in CSV format. Measurements were taken from the slide, the table representing the frame, and the acces sible housing of the shaft bearing, simulating vibration monitoring of crankshaft bearings, places where the use of accelerometers was recommended on the sensing plan (see Figure 4). Figure 5 to 7 present the acceleration signals measured on the slide, frame, and shaft under no load, load, and overload conditions. Analyzing the acceleration signals measured on the slide with no load applied (Figure 5a), the signal in the direction of the main movement of the slide (z axis) shows clearly the descending and ascending motion. The measurements under load (Figure 5c) confirm that the crests and valleys correspond to the bottom dead center (BDC) and top dead center (TDC) since the slide-block is noticeable at the crests of the signal. The signals of the perpendicular directions capture the excitation from these impacts, followed by a damping e ff ect, while under no load, these signals are close to zero. The signals from the overload (Figure 5b) condition are similar to the load condition, as there is an impact and a response to that impact, but at amplitudes nearly four times greater, underlining the strong influence of load on slide dynamics. The frame signals, by contrast, show very low amplitudes under no-load conditions (Figure 6a), though a slight increase can be observed as the slide velocity rises. These signals reflect the overall vibrations from the combined action of the slide, motor, and other components; that’s why the amplitudes change with velocity. With load applied (Figure 6b and 6c), the press cycles become distinguishable, and the moment of slide–block contact produces a clear response across all three axes. In the shaft acceleration under no load (Figure 7a), when comparing with the acceleration on the slide in the motion direction, some small vibrations appear when the slide begins its ascending and descending movement. These