How does RFID model management work? Marking binding and position tracking.
RFID Knowledge

How does RFID model management work? Marking binding and position tracking.

PublishedShanghai Zejie Smart Technology Co., Ltd.

Last year at an automotive stamping plant in Zhejiang, we saw a typical problem. A new setup technician on the night shift got a rush order and needed a draw die for a left front fender. He searched three rows of shelves, asked two experienced colleagues, and finally found the dusty tooling plate in a corner of the spare parts area. It took 40 minutes. The line was down the whole time.

And then you realize that the problem is not human, and there's no reliable "identity" or "location" information at every point from the entry to the line. To solve this problem, it's basically up on the table and the eye; one step upward, with RFID to mark and track it in place, is a way of getting better balanced costs and results. Here we are going to open up the steps we have taken to implement these projects.

Let's get this out of the question: why does a mold have to be "unmanageable"?

The molds are difficult to control because they have several natural properties: a single set of high value (tens to hundreds of thousands of dollars), many models (a medium-size flush plant may have thousands), frequent flow (in storage, workshop, back-up, between production lines), and similar appearances (many of the moulds can't see the left or right).

The traditional practice is the spray code + paper table, which is a problem with the spray code wears, the paper loses, Excel updates are not timely, and it's common in the workshop to say "Who takes it, take it over the side" and come down for months.

The value of RFID is here: labels can be anti-oil, anti-metal, and heat-resistant (part of the scene), as long as the attached links are solid, every entry and exit post is automatically recorded, no molds are found, wrong tie-in is found, and the moulds are unmaintained for extended periods of time.

Step one: "E-ID" for every set of molds.

The key word here isMarking binding, that is, to connect an RFID label with a specific set of moulds, and to make the chip code in the labels the only number in the system.

The label selection depends on the location of the mould and the environment in which it is used. If you are attached to the side of the panel, the model hanger is relatively mild, and the normal ABS seal label is sufficient; if you are placed directly on the mold, you need to consider temperature resistance (many flushing the moulds have surface temperatures of 80 - 120 °C, higher heat treatment) and metal resistance, when you select high temperature tolerance, metal resistance labels, which usually work between -40 °C and 200 °C, pasting or swirling at the flat level of the mould.

The simplest step of binding this is the most problematic link in practice.Initial Data Entry. On initial linking, use a PDA (handheld reader/writer) to scan each mold set and enter its code, name, specifications, compatible machine models, storage area, last maintenance date, and current status (available, pending repair, or scrapped). Allow one to two minutes per set. Two or three people can finish 1,000 sets in about a week.

Once the entry is complete, one thing to do: to cross-check the old code and the newly bound electronic code in a paper or electronic billboard so as to avoid "two sizes for a mold" in the system.

Step 2: Set up sensing points at key locations to enableIn-place tracking.

The label is an "idid" and to get the system to know where the mold is now, it needs to be equipped with a reader at key nodes, i.e. RFID reading devices. These dots usually include: every shelf entrance in the model storage room, every workshop liner entry, back-room door, the mould distribution window.

There are two common ways to do this:

Antenna power needs to be calibrated on-site. We've had an example of a project where the reading and writing machine is too high, and the labels for the next post are read in, so the symmetry on the system presses number three and four is recorded in serial numbers. Then the power is just over the target area, plus the RSSI threshold (sign strength filter) and the problem is solved. This is a piece of detail that many plants can easily ignore for the first time, and it is suggested to allow time for the transfer of test revenues during the implementation phase.

Step three: feed the data into the production system and let the tracking really "use it."

Hardware is not as good as the system can work. The tag numbers read in the RFID system, if they stay in a local or an independent database, are "seeing a hot air" for the workshop. The real value of this is to tie the molds and their sheets to MES or ERP.

I'm going to take a few of the scenes that actually work:

Shanghai Select Tech will usually align data interfaces with client IT teams in similar projects, then deploy the equipment, and avoid the awkwardness of "reading data without access to systems." The interface is typically based on intermediate databases or WebService, depending on how open each other's system is.

Implementation of several less visible but critical recommendations

First, the location of the label is standardized. The labels for each panel are placed in a position, height and direction, with written standards and strict enforcement, otherwise the recognition rate will vary at the end-stage antenna debugging.

Second, a "back-up label" is left for each set of moulds. The in-situ use of the tags is inevitable, with pre-set placing of the tag replacement binding process in the system, which saves much of the communication costs later.

Third, piloting in a workshop or a production line, running through and then replicating in bulk. The pilot phase is often more exposed than expected, for example, to the decline of UHF signals in metal environments, the impact of cryogenic workshops on the work of readers, which need to be discovered under real conditions.

Fourth, value the feedback of first-line users. The system is used daily by three categories: model workers, dispatchers and storekeepers, who judge whether interfaces are appropriate, sound is appropriate, and scanned distance is sufficient to address these issues, more than any indicator.

Back to the scene.

If the stamping plant's dies had all been tagged with RFID and key stations had readers, the night-shift setup technician could have walked the shop floor with a PDA and found the exact shelf and level housing the left front fender draw die within 30 seconds, or seen which press it was on. That would have saved 40 minutes of line downtime.

The RFID model management is not "how much hardware" at the core, but it's a sort of asset that binds up correctly, keeps work-levels steady and keeps data smooth. And all three things are in place, and the mold is no longer in the workshop "no, not in time to fix, not with any "in the mind."