Sequential picking of vra4

From PlcWiki

So far, only a raw chat record, when Honza and I solved the problem with the sequence on sequential picking:

On vra4 there is a special workplace for sequential picking, which is outside the SAS building in Vrchlabí, about 500m away. The so-called "sequential kernel" runs there and one client station with a sequential client is connected to it. Everything runs on one computer - seqvr. On the screen is the screen of that sequ. client. It is a different application than a normal cage, so there is no sequence to see. The sequential client communicates with the sequential kernel, which provides it with information on what to pick.

The dataflow is as follows. The kernel on the line receives the data, writes it to the production sequence (as everywhere else) and also spits it (received original data) into a file, which is then sent by the plc_seq_sender module to the sequence kernel. It receives data and divides individual parts into pallets for defined workplaces. The pallet for each workplace has a defined how many pieces can fit on it. The kernel knows this and counts how many parts (cockpits) it has already put on the pallet. When he fills it, he creates a new pallet with a new number and starts filling it. The number of pieces on a pallet is different for each workplace.

The worker sees on the screen which pallets are in which condition. Because there are more than 6 pallets, as there are boxes on the screen, the screens change automatically, such as aircraft departures at the airport. What a window on the screen, the workplace for which to pick. E.g. KO - station wagon - there is a pallet 7416 - it is already on the line, etc. The employee decides for himself which workplace he will pick up. E.g. wants to pick KO for the workplace. It has its own barcodes for sequential picking, so it scans a code that tells the client that he needs to pick KO and he prints a so-called pallet guide with the pallet number and a list of parts to straighten on it (in the specified order). This marks the pallet as "in preparation". When he straightens it all, he takes the pallet guide and scans its barcode. this switches the pallet to the "ready" state and they can take it to the line. When they load it on the line, they are marked as "on the line", when they take the last part from it, they disappear completely.

All pallet data is provided by the sequential kernel. Both the sequential picking client and the line clients that notify the sequential kernel that they have removed a part from the palette. This data is in the files pallet.dat - header records of individual pallets and sequence.dat, where there is a list of parts and their distribution on the pallets. As data comes from the normal kernel, the sequential kernel writes it to these files. As clients remove them, the sequential kernel (in palettes) deletes them from these files.

For a sequence reset, the sequence kernel has no tools, everything must be edited manually.

[20:09:25] Petr Zalabák: what exactly needs to be done?

[20:11:23] jan_vejvoda: since seq number 616 there were bad cockpits, I deleted them, but I will probably need to do something with the pallet.dat but its structure doesn't tell me anything ...

[20:13:02] Petr Zalabák: the first line for this cockpit in the sequence.dat file is this:

'06483333 BV03 BV03-1J1614105AM PS PALPS3878 1 # 0616'

He says that for the cockpit 06483333 part BV03-1J1614105AM for the PS workplace is on the PALPS3878 pallet.

There is a line for this pallet in the pallet.dat file:

'PALPS3878 13:47:46 13:48:35 20:09:53 04 0032'

He says that the pallet is in the "04" state and there are still 32 unremoved parts on it. Pallet states can be as follows:

. 0 - PALLET_OPEN . 1 - PALLET_FULL . 2 - PALLET_BEEN_LOADED . 3 - PALLET_LOADED . 4 - PALLET_RECEIVED . 5 - PALLET_DELETED

[20:15:21] jan_vejvoda: so I'll lubricate him too ....?

[20:15:31] Petr Zalabák: beware [20:16:03] Petr Zalabák: if there is another part on the pallet (for another cockpit), then it must not be deleted [20:16:41] Petr Zalabák: ie. the pallet number can be extracted from the sequence.dat and this tells us which cockpits on the pallet have parts [20:17:25] Petr Zalabák: in those data it is very badly erased by hand. [20:18:01] Petr Zalabák: it would almost be best to find out which cockpit they are not picked from, then delete both files and send the data again. [20:19:05] Petr Zalabák: the problem is that for different workplaces the first unpicked cockpit is different, because the pallets do not have the same or unpaired numbers, how many can fit on them. [20:19:28] Petr Zalabák: but he can usually handle it. [20:23:48] Petr Zalabák: if we want to set it to the cockpit 616, for example, it is necessary to look in the sequence data file on which pallets it has parts, stop sending data, stop the sequence kernel, delete from pallet.dat these palettes and all palettes with higher numbers, delete this cockpit and all others from the sequence.dat file, unstop the sequence kernel and release the data again from the cockpit 616 [20:24:07] Petr Zalabák: it's quite complicated. [20:25:06] Petr Zalabák: Do you dare? [20:25:14] jan_vejvoda: in addition, I can repeat the problem that was solved at the beginning - send the old cockpit