Experimental Procedure

The experimentation was split into three campaigns. The first two ones involved HJs, with four different ERs. Two different interference levels were considered: in particular a higher level in the first campaign and a lower level in the second one. The third campaign involved SFJs with two different ERs (the extreme levels only were considered).

At a preliminary stage all the shaft and hub specimens to be connected by interference and adhesive were carefully measured by micrometers (for internal and external surfaces) and by a profilometer, in order to determine the
actual values of the external diameter of the shaft, of the internal diameter of the hub and of the roughness of the mating surfaces. The samples were arranged in pairs so that the interference values could range in a not too large interval.. This procedure led to a theoretical interference U between 12|am and 26|am, with a mean value of 18|am and a standard deviation of 3|am. The actual interference was estimated by Eq. (8), introducing the measured values for Ra on the mating surfaces. Z stood in the range 8|am-22|am, with an average value of 14 ^m and a standard deviation of 3|am.

Specimens with different ERs were coupled and released, without adhesive, for the estimation of the coupling pressure and of the friction coefficient, as exposed in Section 2. These specimens are later indicated as dry press-fitted joints (PFJs)

Before starting the coupling procedure, all the specimens (both shafts and hubs) were accurately cleaned by LOCTITE7063® a multi-purpose cleaner. Then, the adhesive was spread both on the male and on the female parts. Shafts and hubs were connected under a hydraulic press, equipped by a 100kN load cell, operating in the displacement control mode at the fixed actuator velocity of 0.5mm/s. Since misalignments between the mating parts may seriously affect results [21], the coupling procedure was assisted by a specifically developed self-aligning device (partly visible in Figure 2). The coupling force, Fcouph was monitored and sampled throughout each test. Moreover, in order to prevent results from being affected by the test order, this order was fully randomized, involving the entire specimen population, with different ERs.

Afterwards, the specimens cured more than 24 hours at room temperature (about 25°C), in order to achieve a complete polymerization [16].

The decoupling procedure was performed by the same standing press and the same fixtures, again with a (different) randomized order of the trials and with on-line monitoring and recording of the exerted release force.

The same specimens were re-used for the second experimental campaign under interference and adhesive. This procedure is justified by the outcomes of previous researches [3, 12], where it was shown that, if only steel components are involved, the specimens can be pushed out and re-used several times without influencing the strength of the joint, provided that they are cleaned accurately after pushing out. For this purpose LOCTITE7063® multi-purpose cleaner was used again.

The male and female parts were measured again for the estimation of coupling dimensions and roughness and then rearranged in (different) pairs. The theoretical interference U was estimated between 2|am and 14|am, with a mean value of 5 pm and a standard deviation of 2 pm. By subtracting the roughness values, as in Eq. (8), a mean actual interference (Z) of 2pm (with a standard deviation of 2pm) was determined. The lower values are due to the general increase of the hub internal diameter, after previous coupling and pushing out.

The same procedure, regarding sample cleaning, order randomizing, applied force recording and specimen curing was carefully followed for both the coupling and the decoupling tests. Two specimens were tested without adhesive for a further estimation of the friction coefficient.

The last stage of the experimental part was devoted to the tests on Pin — and-Collar specimens (SFJs). They also were measured, determining the coupling diameters and roughness values. Afterwards they were arranged in pairs, in order to make clearance as uniform as possible over the entire population. The specimens exhibited a clearance between 27pm and 65 pm (mean value=48pm; standard deviation=10pm), within the interval 25-75pm recommended by Standard [13], and consistent with the application field of the studied adhesive [16].

Before starting the mating process, all the samples, both shafts and hubs, were cleaned by the aforementioned cleaner. The samples were coupled, applying the adhesive to the external surface of the pin and to the interior of the collar and slipping the collar over the pin, with a helicoidal back-and-forth rapid movement. All the specimens cured for 24 hours at room temperature. The pushing out operation was run with the same procedure followed for HJs. The entire specimen population was assembled and disassembled following a randomized order.

Комментирование и размещение ссылок запрещено.

Комментарии закрыты.