Civil Engineering Reference
In-Depth Information
FORCE
ELECTRODE
INDENTATION
SHEET
SE ARATION
WELD
NUGGET
TIP
DIAMETER
ELECTRODE
NUGGET
DIAMETER
FORCE
FIGURE 7.8
Resistance spot welding process.
where
N
is the number of samples, and
n
1
and
n
3
are the number of elements in the input and output
layer, respectively.
Kovacevic found that while none of the input parameters (width, length, length-to-width ratio, or rear
angles) alone can accurately assess the penetration, a combination of the length and rear angles of the
weld pool provided sufficient information to accurately model the relationship between the weld pool
and weld penetration. Furthermore, monitoring was possible in real-time at a rate of 5 Hz.
Dynamic Representation of Resistance Spot Welding
Process Input/Output Relationships
Used typically for joining one sheet-metal panel to another, spot welding, also known as resistance welding
or resistance spot welding (RSW), is one of the oldest production welding methods. The resistance
welding process differs from the arc welding processes in that filler wire is rarely used and fluxes are not
employed. As depicted in
Fig. 7.8
,
resistance welding uses pressure and electrical resistance through two
pieces of metal to form a molten nugget at the faying surface between the two sheets. Weld quality is
affected by the following factors: current applied, on-and-off cycle timing, electrode pressure, and the
shape and condition of the electrode.
Messler, et al. use a neural network and fuzzy logic controller to adjust the amount of heat input in
order to compensate for variations and errors in the RSW process [27]. Their intelligent control system
relies on the electrode displacement, caused by thermal expansion between the electrodes during nugget
formation, as the single parameter input to indicate weld nugget growth. Desired electrode displacement
vs. percent heat input and welding time is calculated by a neural network. A fuzzy logic controller
compares the desired electrode displacement with the actual electrode displacement and then appropri-
ately adjusts the power delivered in real-time.
Messler states that there are numerous sources of variations and errors which include: material com-
position and coatings, workpiece stack-up thickness, fit-up, surface roughness, and cleanliness; electrode
deformation, wear, or contamination; welding machine cooling efficiency, force consistency, and mechan-
ical compliance; and electrical dips or surges. Nevertheless, numerous mathematical models have been
developed to represent the ideal electrode displacement curve [70]. Unfortunately, these mathematical
models have limitations in a practical application due to the previously mentioned sources of variations
and errors. Furthermore, a comprehensive analysis of the RSW process is extremely complex due to the
interactions between electrical, thermal, mechanical, and metallurgical phenomenon of the process.
