In-Process Variation Monitoring - Hose Crimping Applications 

Tubing is used to transport fluid and gas in pneumatic, hydraulic, or process applications and is combined with flexible hose into fluid handling assemblies.  The connections between tubing and hose are frequently crimped in a press in order to make a leak proof seal.

Errors in the crimp process lead to defects in the product.  The hardest process error to correct is the one you can’t see.

The most obvious unseen process variance or error is one that is hidden from view inside a component or finished part.  Inside a hose crimp there may be foreign material embedded, a component missing or out of alignment, the hose may be folded or under-inserted, or you can have a split of the hose inside the ferrule or partial collapse of the tubing inside the crimp.

For catching under-insertion of the hose, many systems rely on a poka-yoke check hole at the base of the ferrule.  This allows the operator to see the hose protruding from the hole if it has been inserted correctly for the crimp.  Towards the end of a shift that bit of black rubber may be indistinguishable from a shadow, or due to fatigue the operator just doesn’t see it – it’s become invisible.

Another way to catch an under-insertion error is to mark the tubing with some visual cue that indicates the tubing is correctly inserted for the crimping process.  This poka-yoke procedure assumes that the upstream process placed the mark in the right spot or that the tube end is cut “square” so it seats properly at the base of the ferrule. 

Consistency of component parts is critical to output quality - "Garbage in, Garbage out".  Variance in ferrule hardness, wall thickness, weak spots or seams, and deburring or trim preparation of the hose components can all lead to product defects.  Component parts may all be within spec limits, but a combination of one part at the top end of its spec matched with a component on the low end of its spec can lead to a weak part or an over crimp that can pass a quality check only to fail after limited use.

The process error that no one knows is occurring goes unseen because no one is looking for it.  Once it creates a visible defect then action is taken, but by then the costs have escalated and the confidence in the quality of product output eroded.  An example is a nick in the back bead on a crimp because of a mispresentation in the die.  Here the operators were only checking dimensions on the crimp.  Very likely the defect has no impact on the integrity of the crimp, but what about the consequences if a Supplier Quality Engineer quarantines a shipment of parts at final assembly until that can be established?

All of the errors outlined are variations in process that can affect the quality of the output but still meet the QA “metrics”.  Automated dimensional or physical inspection devices may catch the gross failures, but will not keep the other, less obvious but equally significant, defects from getting to the customer.  Some form of objective monitoring is required that can see the invisible process errors before they become a quality defect. 

The concept of process variation monitoring is simple: if a process is known to be capable of only producing good parts, then as long as the process is confirmed to be stable and repeatable, good parts can be assumed to be produced. The key is to be able to accurately determine, in real-time, whether or not the process is stable and therefore quality-capable.

For many processes where parts are formed, shaped, or fastened by the application of force, a simple “force-resistance-strain” relationship exists. If the force is known to be constant, the strain experienced by the machine will be directly related to resistance to that force – i.e. the work being done to the part. The resistance to the force includes all the other process variables such as raw material (thickness, hardness, etc), lubrication, tooling condition, and even machine condition (bearings, clutches, set-up, etc).  Accurately sensing the strain on the frame of the machine while it is forming, shaping, or bending a part, and comparing it to a known strain signature for good examples of that process, can determine whether or not the process has varied to the point where it has also become capable of producing bad parts.

New piezo electric strain sensors coupled with intelligent signal monitors employing application specific algorithms, make it possible to use this method of “process variation monitoring” as a very effective, and reliable, quality assurance tool for processes such as end-forming, riveting, hose-crimping, wire crimping, cold heading and thread rolling.

Every production cycle will create a specific strain signature for every part produced. In many cases, one properly positioned high sensitivity piezo-electric strain sensor is the only sensor required.

 

Process variation monitoring, using piezo electric strain sensors and simple, but intelligent monitors can offer the manufacturer of low cost, high volume parts many advantages:
• 100% monitoring of production for more than just dimensional defects, without adding time or processing to the part.
• Early detection of a defective part prevents any further value from being added, saving production time and reducing scrap costs.
• Maintenance and down-time can now be performed when the machine actually needs attention, and not according to a worst-case schedule.
• Tooling can be replaced only when it needs to be replaced - not too early when there is still life left in it, and not too late after it starts to affect quality or efficiency.

Most importantly, process variation monitoring puts the focus of attention (from maintenance, production and management) squarely on the manufacturing process - which is the only place where sustainable gains in efficiency, scrap reduction, and quality can really be made.