How does RFID production management work? Workspace recognition and ongoing product tracking
RFID (radio frequency recognition) is used in production management, with the core of which is to automatically connect "situation-material-time-operation" four dimensions with electronic labels. The site identification reader reads the tag into the reading area, and the product tracking is the deployment of antenna+ backstage systems at key nodes of the process route, where the location and status of each semi-finished item is written into the database in real time. Hardware itself is not difficult to layout in the antenna, anti-disturbed design and aligning data with MES/ERP.
How can the location identification be achieved? Where are the differences in the various deployment modalities?
Workspace recognition usually has three paths of realization in the project.
The first is a fixed-reader + antenna array. One to four URHF antennas (UHF means UHF frequency bands, frequency of 860-960 MHz) are installed on the top or side of each work slot, forming a reading area approximately 1-3 metres in diameter, with the package being automatically identified when the conveyor belt or material car enters. This is suitable for the convulsion of stable water lines, and single-place hardware input (with readers and a side antenna) is roughly at the thousands of dollar level.
The second is to hold the PDA (a handheld data terminal for industry). Workers use the PDA to scan labels of materials during process switching, flexible but discipline-dependent, with a higher rate of failure than fixed, more suitable for rhythm, and a large variety of small batch production.
The third is a closed-door reader, which is identified in bulk when the material vehicle or swing tank passes through and is used for the trans-shipment between the workshop and the warehouse.
When using this type of programme, the main waterlines are used to capture process and operator data with a fixed reader; metal-covered assembly places are replaced by handheld PDAs; and the turnbox is read in a batch of traffic through the workshop entrance. The recognition and acceptance of workers are better balanced in one section of each tube.
How do you get every single one of them in the product tracking?
The product tracking is extended on the basis of a workspace recognition. The key is not to have a reader at each site, but to deploy several core nodes along the process route, such as feed, processing, critical processes, reordering, mass checking, entering the library in six locations. Each node is accompanied by a reader, recording the time that the material passes, operating the employee number, which is the time of the process.
The logic behind the data is that each product is given a unique global RFID code (e.g., e-product code, usually 96 or 128 bit in industry), which is a production order in MES systems, and the material BOM (Bil of Material, List) and process routed. Readers read labels, EPC, corresponding antenna numbers, read time packaged to MES. MES determines at what step "it should be" and compares "real" position with "planned" position, and triggers an early warning before falling behind or going beyond.
The mechanism addresses three specific issues: first, the rapid location of a particular item at which post is currently being worked as several hundred items are flowing from the workshop; secondly, the measurement of the true time-frame and processing time for each job to provide data support for the assessment of capacity; and thirdly, the rechecking of complete curricula vitae by the EPC in which, when quality arises, including who operates, which equipment is used, and what environmental parameters were used.
What are the issues to be noted on labels and antenna selection?
Tag selection directly affects read rates. Metal parts and environments with liquids pose particular problems for UHF RFID because the signal is reflected or absorbed.
The industry generally uses on-metal tags for metal workpieces. A layer of magnetic RF-absorbing material on the back confines RF energy near the tag antenna. On-metal tags cost three to five times as much as standard tags, but they are almost indispensable for metal parts because of the substantial difference in read performance.
Liquids are more difficult, and high water-bearing materials can absorb a large amount of radio frequency and require special liquid labels or repositioning to the outside of the swinger to avoid being attached directly to the tank.
The general principle is that the reading and writing device antenna and label distance are at 3-5 meters, and the one to which the label is right is the best entry; two antennas on both sides of the conveyor belt can reduce the readable blind zone by installing a wrong-side antenna; and a metal frame is required for heavy workspaces.
The three types of interference that are common to field debugging are: a modulator, a server, a power source, and a wide-frequency noise influence the reader ' s sensitivity when working. The method is to ground the outer shell, with an antenna feed coded for as long as 10 metres in length, and if necessary, to physically isolate the antenna from the jamming source by adding metal panels.
How do we get through to the MS, the ERP?
The RFID hardware is not complex in itself, and it's difficult to integrate into the system. Most MES and ERPs do not receive the original RFID data directly, and one of them needs a data capture and conversion layer.
The typical structure is three-tiered: the bottom is a reader and label, which collects physical signals; the middle is an RFID medium, also called a data acquisition server, which is responsible for tags to be reweighted (the same label is read dozens of times in the reading area, the intermediate is only valid records according to the rules of practice), multireader data are merged and interfaced with business systems; the upper layer is an operational system such as MES, ERP, WMS (storage management system) that consumes these data.
The middle layer is the most problematic. Many business charts are directly connected to a reader at MES, and the result is that the reader is wrong when it drops the line or repeats data to boost the database. The standard approach is for intermediates to take full exception: Cache data when the reader breaks off, reconnecting subsequent batch back-up; filtering to MES when the label is read over and over again.
The mainstream RFID intermediate generally supports the RESTful interface (a standardized data interface based on HTTP protocols), and the MES end simply analyzes JSON by standard, without concern for which brand of reader is listed below. This decomposition design has a clear advantage at later stages when switching equipment or increasing the product line, and does not involve a whole body.
What pits are the easiest to step on when new men deploy?
Talk about recurring issues in several projects.
First, blindness requires 100% recognition. The UHF RFID is already very good at over 98% of industrial sites, and the remaining edges are usually based on process error prevention and business processes (such as automatic alarms for bugs). Taking a consumption-level standard requires industry scenes, which often end up with repeated hardware adjustments.
Second, the label position is not standardized. Some workers have a front, back and back, some tilt, which leads to significant differences in reading and writing stability.
Third, cycle time does not match hardware capability. A fixed reader usually takes 100 ms to several hundred milliseconds per read. If the conveyor is too fast or cycles are under one second, use the reader's multi-tag anti-collision algorithm to read several tags at once or slow the conveyor. Calculate the cycle time before choosing hardware to avoid rework.
And fourth, use RFID as a panacea. RFID is good at "non-contact, automatic, mass recognition," but it's not appropriate to write in a lot of data, to encrypt complex information, to compare the visualization of objects with the scene. Stronger RFID costs more than sweeps, and more complicated to maintain.
The value of RFID in production management is not the label itself, but rather the task of setting a place, person, material, and operating these data points into a complete resume. It is recommended that a small pilot line or workshop be set up to run the reader-media-MES link, confirming recognition and data accuracy, before expanding to other regions. The hardware selection phase leaves a balance of 10% to 20% for field retracing, which is more economical than the budget compression at the outset, and more manageable.

Ms. Lei
Mr. Pang