What are the elements of the RFID asset management programme?
One count of the crash.
Late last year we visited a third-party logistics client. They were doing inventory that day: three workers had been scanning barcodes for four full days, but reconciliation showed a discrepancy of over 40 pieces of equipment and more than 2,000 reusable bins. The exhausted warehouse supervisor said, “We scan all day until our hands cramp, and the numbers still don't match.” A stack of unapplied asset labels lay on his desk, their edges curling slightly from sweat. This is typical of companies still using barcodes for asset management: slow manual scans, missed items, and discrepancies between records and actual stock.
And then, with RFID, the same size warehouse, two people can complete a full-volume inventory in one day, and the difference can be within a thousand points. The fundamental reason for this is that the logic of the count has been completely changed. A landing RFID asset management program, which contains what should be, which pits are easy to step on, and we're going to make clear the key points that we've drawn from the project.
A complete set of plans, four layers.
Many clients come up and ask "How much is a RFID system?" But the real thing about this project is, do you think that these four levels are going to be successful or not?
- Sensory Layer (hardware): Electronic labels, readers, antennas, handheld terminals, and print-coded integrated machines. Of these, the UHF tag (working frequency 860-960 MHz, suitable for remote, multilabel reading at the same time) is the main choice of the asset count.
- Network Layer (Transfer): How do you make a reading and writing device, whether it's wired or wireless (POE Ethernet, Wi-Fi, Industrial bus), how do you aggregate data to servers? The metal shelf and concrete wall in the storeroom will eat signals, and this layer cannot be surveyed.
- Platform level (software)The following are the following: Asset desk accounts, binding relationships, access returns, inventory missions, unusual alerts, and report output.
- Process layer (system): Who will mark, who will bind, how to store new assets and how to process them. The system is not always a technical problem, it is the process that is not right.
When Shanghai-based technology works with its clients, it usually folds these four layers together with the current status of its customers — whichever of the weak hardware, software, network, process is the one from which the programme is made instead of coming up and throwing a list of equipment.
Design phase, several of the easiest pits to step on.
The label selection is price-only, ignoring the asset material.
Metal cabinets, liquid containers, and plastic housings reflect and absorb RFID signals very differently. Put an ordinary paper tag on a metal server and the read range can drop from five or six meters to a few dozen centimeters. On one project we saw, more than 1,000 tags had been applied before field testing showed that nearly half were unreadable, forcing the team to redo them all. Test tags on actual assets when selecting a model. Cover typical materials, including metal, plastic, wood, and liquids, before deciding whether to use standard, on-metal, or flexible tags.
The reading and writing dots, of course. No site survey.
RFID's reading range is not a fixed circle, influenced by antenna direction, shelf height, metal shielding. Some customers follow the drawing and sheet points, and then they can't read through the top-line and back-channel intersections, and the signal over the corner of the wall overlaps. The standard process is to test the antenna angle and number at the candidate point with a reader, and adjust it to the actual cover map. For every 1,000 square meters or so, the warehouse needs to do one or two rounds of field surveys.
The tags and the assets are not tied.
Tags are labeled, but tags and asset numbers are not tied to the system, or are wrongly tied, missing, data coming out is a confused one. It's common for print coded as one machine printing on site, stickers, scans, three steps, and then paste one and activate it in the system instead of putting it back and re-recording it. Our project requires that the postpoint link be double-reviewed, so that the error rate of this step can be kept manageable.
System isolation, financial incompetence.
Asset management says it's for finance and auditing. If RFID ZZZQQZ data and ERP, the financial system's asset card is not working, and it's accurate and does not reflect value. The programme design phase identifies the data interface, whether it's one-way synchronized or two-way verification, assets added, transfers, and the end of life status is in line with which system. This is a very uncertain step, and then it falls into the muddy of "mixing systems."
Permission and security are not designed.
RFID readers can read all the labels within a few metres of the surrounding area, and the data is likely to involve asset amounts, where these sensitive information will be stored. Readers are required to go through a separate VLAN when they enter the network, which is equivalent to a quarantine channel in an internal network, and backstage operations are assigned role-specific privileges, and the count logs are kept for traceability. We write this in a single receipt list at delivery.
When you get online, you can really run the running mechanism.
The system is just starting, and it depends on the day-to-day mechanisms. We've seen some customer systems that are pretty good, and they won't be used in three months, and the problem is basically that the process isn't keeping up.
- Daily access and returnThe following is a list of the items that are being processed:
- Periodical inventoryThe whole recommendation is made quarterly, and the area count can be done on a weekly or monthly basis.
- Unusual treatment: Label damage, asset transfer, loss, end-of-life need to be separated by separate process branches, especially the end-of-life label, which is to be invalidated and avoid being misread on other assets.
- Statements and decision-making(b) Indicators of the actual discrepancy rate, idle rate and frequency of access should provide an automatic statement that provides a basis for procurement and asset management rather than "overtakes"
Back to the original warehouse.
The logistics client ended up using the hand-held terminal plus fixed-reader combination of our program, half a month on line, the first quarterly count three months later, two people working in one and a half days, and the difference fell from a few percent to two thousand. The biggest feeling on the financial side was not fast, but that data was finally used -- the count was straight to the ledger, and the audit didn't have to flip another stack of paper files.
Looking back, the RFID asset management program is not a list of equipment, but a whole of senses, networks, platforms, four layers of process. Design captures three main points: labels and hardware selection must be based on real assets on site testing; data stream to connect with existing systems; and routine processes to be accountable, to run. These three things do, and the plan is really down.

Ms. Lei
Mr. Pang