自动换刀在可重构制造系统英文文献和中文翻译(6)

菜单

b) Measurement - The end effecter positions are then measured and compared with the predicted positions of the mathematical model

c) Identification – This is the process of identifying how the degrees of freedom or joint angles affect the position of the end effecter

d) Compensation – The software commands are then reprogrammed to allow for accurate correlation between the user’s input and the resulting output.

5.2 Automatic calibration of the tool changer

The tool changer will need to be able to adjust to a new reconfiguration of equipment, or it may even be required to service a different machine altogether. As mentioned earlier, its integrability will be affected by its ability to be re-calibrated to the new environment. Recalibration must not only be possible, but also efficient. The more efficiently the unit is able to be calibrated, the more integrable it becomes.

Accurate sensors are the key to calibrating the unit. Often calibration is performed with the use of additional sensors or specialised equipment. As this equipment is not usually required for a machine’s day-to-day running, it is a suitable solution for machines that do not need to be calibrated very often. In a reconfigurable system, however, it may be difficult to predict how often calibration may be needed. The solution is therefore to have the sensors built into the machine. The number and accuracy of the sensors will mainly be limited by financial constraints [14].

There is an additional benefit to having an automatically-calibrated machine: it could be programmed offline [16]. This would significantly reduce the down-time of equipment during a reconfiguration. It would also have the knock-on effect of reducing ramp-up time;and a new product might gain a competitive edge by being brought to the market that much sooner.

In industry, a majority of robots are programmed using the ‘teach’ method. The end effecter is placed in position for each of its required motions. These positions are then recorded and used to produce the program for the desired function of the machine. While the benefit of this procedure is that an inverse kinematic model is not required to produce a program, teaching the robot is very time-consuming. This method may be advantageous in some circumstances, but for reconfigurable systems it may lead to excessive amounts of equipment down-time.

In terms of the tool-changing unit under discussion, one of the limitations mentioned previously has a significant effect on its ease of calibration and hence its integrability. The machine had to be taught the required positions in order to effect a tool change. This would mean either that the unit would have to be positioned in exactly the same position relative to each machine tool, or reconfiguration. That might not be possible in every circumstance, in which case the unit would have to have a new position table configured.This would drastically increase reconfiguration time.

High integrability of the unit would allow it to be positioned within a reasonable working range of the machine it would be interacting with, and for it to detect its necessary working positions automatically. Further work on this project will explore ways to make this a reality.

Due to the constant reduction in the price of electronics, and the common availability of sophisticated equipment, location-sensing technologies are easily accessible [17]. These technologies will be researched and an appropriate solution developed to increase the tool changer’s ability to calibrate itself automatically. Further research will integrate the toolchanging unit with a 5-axis modular reconfigurable machine that has been developed by the research group.

5.3 Improving the unit’s diagnosability

Diagnosability has two facets. The first is defined by a system’s ability to ascertain the causes of poor part quality, and then to be able to adjust in order to produce the acceptable parts. The second facet is the system’s ability to detect failure or damage in the machine itself [1], [2].