Electric heating – Metal heating – For bonding with pressure
Reexamination Certificate
2001-07-06
2003-01-14
Shaw, Clifford C. (Department: 1725)
Electric heating
Metal heating
For bonding with pressure
C219S109000
Reexamination Certificate
active
06506997
ABSTRACT:
FIELD OF THE INVENTION
This invention relates generally to resistance spot welding and more particularly to a method and apparatus to sense welding conditions, calculate energy balance, predict weld diameter, and eliminate splash.
BACKGROUND OF THE INVENTION
Resistance spot welding is a process for joining sheet metal components that involves clamping two or more parts together between two electrodes and passing a series of low voltage, high current pulses through the parts. The parts are squeezed together in an area between the electrodes and heated by the high current pulses. The material between the electrodes melts forming a molten area known as a nugget. When the current is switched off, the nugget of molten material solidifies forming a joint which is referred to as a weld. Frequently during resistance spot welding, an expulsion of the molten material occurs when the parts to be welded, generally referred to as weldments, plates or workpieces, are momentarily overheated. The expulsion of molten material, conventionally referred to as splash, is detrimental because it creates a hazardous working environment and it also creates a weaker weld. Splash caused by overheating has been attributed to sudden melting of a portion of the corona bond zone at the faying surface interface (i.e. the contacting surface of parts to be joined). Continuous monitoring and prediction of nugget temperature including the dynamic behavior of the rise in temperature, is important for predicting splash caused by overheating.
For example, the temperature pattern of a weld can be continuously simulated by a system implementing a monitoring procedure based on a numerical simulation of the workpiece temperature. Prior art simulations have been used to estimate certain welding phenomena including current distribution, temperature spatial distribution, and temperature profile as a function of time. Such systems require welding current data, welding voltage data, and information concerning physical properties of the workpiece as inputs to the simulation.
However, these prior art simulations are not used as real time prediction tools, since the calculation time required to predict temperature profiles is several orders of magnitude longer than the actual welding time. Thus, although such systems are useful for determining weld diameter and temperature profiles of welds, these systems have not been used in feedback control systems for the elimination of splash.
A heat conductance distribution is used to estimate temperature profiles. Some prior art systems compute the heat conductance distribution using differential equations and finite element analysis methods that require time consuming calculations. The prior art methods require a large number of mesh points which are mathematical point approximations of a continuous volume used to calculate accurate temperature patterns and temperature profiles. Finite difference methods and finite boundary methods are similarly computationally intensive.
Some prior art resistance spot welding systems use displacement of the electrode tips to monitor the welding process. These techniques require further instrumentation of the welding systems to accurately measure electrode tip displacement and complex statistical models and regression analysis to estimate nugget diameter. U.S. Pat. No. 6,043,449 discloses a method to estimate nugget diameter based on inter-electrode distance. A system using electrode displacement measurements cannot correctly determine weld diameter when the weld is formed near a workpiece edge due to the typical deformation of the edge of the workpiece.
Other prior art resistance spot welding systems use a finite element temperature analysis to predict temperature distribution and nugget growth. These prior art systems typically do not run in real time and require a processing capability generally not found in spot welding systems. For example, U.S. Pat. No. 5,892,197 discloses a nugget growth predicting means based on energy distribution calculations using a 3-D micro-lattice model to estimate energy distribution. A very high performance work station, special purpose digital signal processor (DSP) hardware, or small supercomputer would be required to solve the 3-D micro-lattice model in real time whereas the typical resistance spot welding controller has the processing power of a slow personal computer.
Other techniques have been developed for monitoring weld quality based on a one-dimensional thermal equation and analog dynamic model of neural networks. However, these techniques use a thermal equation without considering the importance of plate or part thickness. Some of these techniques are described in “
Nugget Size Sensing of Spot Weld based on Neural Network Learning
,” Kin-ichi Matsuyama, 7
th
International Conference on Computer technology in Welding, Proceedings of the IIW Doc. III-1081-97, July, 1997, and “
Prediction of Spot Welding Diameter Using Neural Networks
, Monari et al., Proceedings of the IIW Doc. III-1108-98, 1998. Several systems have also been developed to detect splash, but these systems cannot predict the future occurrence of splash.
As is known in the art, a thermal similarity rule developed by Okuda, as described in “
Resistance spot welding of thick steels
,” (Report 1), Welding Technology, 19 (1971), pp 104-107, can explain the governing parameters of nugget formation in resistance spot welding. This rule is derived from the non-dimensional expression of the thermal conduction equation. The rule, however, cannot explain the relationships between ordinary welding conditions and welding results since the rule holds only in special cases, in which the electrode tip diameter is exactly proportional to the plate thickness. Under actual welding conditions, the electrode tip diameter is usually not exactly proportional to the plate thickness as recommended for resistance spot welding. Therefore, the thermal similarity rule described above cannot be used to predict the temperature profile under typical operating conditions.
The Resistance Welders Manufacturer's Association (RWMA) recommends that an electrode tip diameter should be five times the square root of the plate thickness. This suggested electrode tip diameter results from empirical data.
Other techniques have been developed for monitoring weld quality based on a one dimensional thermal equation and feed forward model of neural networks. The procedure based on the one dimensional thermal equation has been successfully applied to monitor weld quality, but the system requires a high speed DSP to provide real time control. The neural network solutions require a large amount of data for training, and the solutions are not applicable for any welding condition out of a trained zone because the underlying model is determined experimentally rather than theoretically. As a result neural networks are not widely used in the installed base of resistance spot welding systems.
There are several hundred thousand resistance spot welding systems in operation with a limited computational capacity. The total number of spot welding machines may be more than several million worldwide. It would be very expensive to replace even a relatively small number of resistance spot welding machines installed in automotive company manufacturing plants with controllers having real time monitoring because of the capital expense, installation costs, labor to reprogram the new controllers, and the factory downtime.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide methods to predict weld diameter, to eliminate splash and to assure a sound weld. It is a further object of the present invention to implement the methods on an installed base of controllers without significant or even any welding machine hardware modifications or the addition of processing capability.
It is a still further object of the present invention to provide such a method for real time use during the welding process and in applications in which the electrode tip diameter is not exactly proportional to the p
Daly, Crowley & Mofford LLP
Massachusetts Institute of Technology
Shaw Clifford C.
LandOfFree
Spot welding system and method for sensing welding... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Spot welding system and method for sensing welding..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Spot welding system and method for sensing welding... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3032909