Robotic arc welding gets smart in real time
Immediacy of information can benefit shop floors
Whether you're dealing with low part counts or wild welding variables or the challenges of just-in-time production, monitoring software can help smooth your operation.
Though robots probably are best-known for their ability to accomplish repetitive manufacturing tasks, it's no secret that, by themselves, they are unintelligent and must be told what to do. They are silent, unable to talk over problems with their controller, power sources, the robot in the next cell, the line foreman, or the manager in the front office.
When shops use robotic systems and equipment, that lack of communication often can prove to be problematic, creating a vacuum for needed information. Consider three typical scenarios. Each illustrates to quality- and productivity-conscious managers a different value of the real-time production data provided by specifically designed robotic software.
Scenario No. 1: Variable Welding Current
Imagine an operation in which production begins routinely in the early morning, apparently without a glitch. Before beginning its automated welding production for the day, the shop develops procedures for so many amps, volts, travel speed, torch angle, and so forth. Given this knowledge and preproduction routine, if the robotic cell begins operations within the proper parameters, the cell should keep on repeating the set procedures, right? Perhaps, but one factor, such as arc current, can get in the way.
With traditional technology, few welding engineers can ever tell when the arc current drifts slightly out of spec. If, for example, arc current changes from 120 to 140 amperes, nobody can visually see it. But, the deposited bead shape can change; penetration can become excessive (or inadequate if current decreases). If arc voltage changes moderately, the result is post-weld spatter, which requires labor-intensive cleanup.
One reason for arc current changes might be nonuniform dimensions of incoming parts that cause wire extension, popularly called stick-out, to decrease or increase. As a result, arc current may increase or decrease, thus changing bead shape and penetration or introducing excessive heat, which leads to distortion. Only when finished parts are checked later in the day do plant engineers learn that defective parts were produced.
"Exactly when did production of the defective parts begin and how many are there?" are questions easy to imagine hearing from a plant manager, welding engineer, or QA manager.
One solution is software that documents the actual parameters used by a group of robots that are linked by a standard Windows® NT network. Data is compiled periodically and sent to a server for storage and analysis. Once data is captured, action can be taken. If the magnitude of a preset welding parameter—such as arc current, voltage, and travel speed—exceeds an upper control limit or dips below a lower control limit, the system can alert a supervisor via e-mail or page. As a result, an immediate and focused corrective action can be taken.
Scenario No. 2: Low Part Count
As a second example, imagine that a plant is producing fewer parts than anticipated. The critical questions in this case are many: Which robots are underperforming? What resources are required? How should we redirect them?
The answer here again rests in how and whether managers receive timely, real-time information. If robots are monitored in real time to document parts per hour or specific welding procedure errors, the correct resources—whether maintenance or engineering—can be directed to those robots that require them.
In arc welding, for instance, software can generate customizable reports and analyses. Utilization reports document how much time each function requires, such as operating time, as well as teach, working, and error times during a two-day period the customer specifies for one individual robot or group of robots. Management can monitor actual utilization in real time via Web-based software that is accessible to authorized persons using a standard Web browser. Managers can capture actual welding parameters via an error history report confirming errors, or lack of errors, that occurred for an individual robot at a preset time and interval. Essentially, the report documents that the process has run within tolerance limits during a specific time period.
Scenario No. 3: Problems in JIT Manufacturing
In today's manufacturing environment, many subcontractors are pulled into the world of just-in-time (JIT) manufacturing. If a manufacturer can monitor the production of parts accurately in real time, it can send this information backward to raw materials suppliers or upstream to a contractor's customer. Each member of the supply chain can find value in a system that ensures incoming materials and parts will satisfy its own needs, as well as provide the same data to customers.
For example, if the fabricating company communicates to its suppliers that parts production is up, the upstream supplier can ship more parts to allow continuous flow of production (welding). Similarly, the fabricator can alert its downstream customers that production is on target to give them a little assurance.
It may be true that the ability to document such production data creates value beyond increased product quality and productivity. Many customers ask their contractors to verify that procedures are always to customer spec, so having records to show those customers helps a contractor's marketing and sales efforts. A fabricator that has hard documentation that every part was manufactured to agreed-upon specifications can acquire a high comfort level in its operations and protect itself from liability and possible litigation.
The ultimate impact of real-time monitoring software can indeed be significant. It can isolate problems with robotic arc welding equipment by identifying out-of-spec procedures, minimizing downtime, improving quality and productivity, and optimizing costly labor time. Shops that put monitoring software to proper use often find that they can assign their valuable resources cost-effectively and only where required. Groups of robots (as few as five)-instead of remaining silent-now can be programmed to keep everyone and everything in the loop.