Human machine interface crash phenomenon and treatment - Database & Sql Blog Articles

Human-machine interface (mainly operation station) crashes All dcs exist to varying degrees, but the frequency is different. Also related to the length of use of DCS. The display is no longer recoverable due to hardware crashes. If the operating station power supply voltage is low, it is easy to cause the operating station to crash. The crash caused by improper selection of the rate of software and communication connections can be recovered over time. If the operating system operating system and the monitoring software are not well-matched, it can also cause a crash. There is no way for the maintenance personnel to call it a primary crash. Due to some unreasonable configuration, the network is blocked, the memory is too small, and so on. It can recover after a few minutes.

First, the crash of DCS man-machine interface

Recently, since the MIS system needs to read production real-time data from the DCS. In the network to access the dynamic data server, and some use the DDE method, in the case of reading more data (such as 3000 points), because DDE consumes more resources, the network congestion phenomenon becomes very frequent, and makes various people The node crash of the machine interface becomes serious.

The severity of network congestion is related to a variety of factors. The data is taken from the operation station, and the DDE method is adopted. If the data is below 500 points, the problem is not too large. If you take a few thousand points of data, the operator station crash is almost inevitable. If you connect the node to the DCS communication network and take data from the network interface, the situation is better.

The controller sends the data to the interface, and the human machine interface reads the data from the network interface in compliance with the network communication protocol. The physical structure of the network is both ring type and bus type, and the bus network is also logically ring type. The star network is only used for small systems (within 100 I/O points). The common communication protocol is broadcast. When a node on the network sends data to the network, it must continuously broadcast data, and the node that needs the data receives the data. In addition to this method, a broadcast protocol network has a way in which a node queries data from other nodes on the network. If other nodes do not, it repeats the inquiry until the data is read. If there is no such data on the network, it will cause network congestion.

In order to make the operators proficient in the operation of the DCS operation station, the simulation system can be used to understand the keyboard of the DCS to reduce the crash caused by misoperation; also to understand the dead situation of the operation station.

In the case of a long DCS running time, the configuration is only added or not, and some of the configurations are actually not connected to the real I/O points, regardless of the control. There are such I/O points in the user application of the engineering station. When the dynamic data server is connected, all the data points on the DCS are to be read. Because a large number of data points are invalid, the network is blocked, so the human-machine interface crashes. At this time, the reverse engineer station can be used to read the configuration in the controller, compare it with the content of the forward engineer station, and delete the invalid point to avoid network congestion.

Another problem to be aware of is that when accessing a dynamic data server, you should check whether the software versions of all interfaces are consistent, otherwise it will affect data transfer.

A third way to address network congestion is to increase the exception reporting area to reduce network traffic if exception reporting is used. An exception report is when a site changes data to the network only when it changes. In order to prevent the situation where the point is broken and the man-machine interface is unknown, even if the point does not change, it will be reported after a certain period of time. In order to reduce the amount of traffic on the network, the two parameters of the exception report are appropriately increased, and the purpose of reducing the amount of data can also be achieved.

Recently introduced general operating stations mostly use WINOOWsNT/2000 operating system, monitoring software is also common, such as FIX, InTouch and so on.

Due to the large amount of software sales, there are fewer problems, the general operation station is open, and the crash phenomenon is greatly reduced. If the driver software is not very good, it will also cause a crash. Importantly, it is not limited by DCS manufacturers regardless of maintenance costs and spare parts purchases. But in order to prevent hackers and viruses from invading, firewalls are indispensable.

Second, the crash of the human-machine interface of the C/S structure

1.C/S structure

After the DCS man-machine interface adopts the ordinary PC and WINOOWs operating system, in order to increase the number of human-machine interfaces, the C/S structure is adopted. As long as the controller is connected to the server through the interface, the server and the client use the network interface adapter to connect with the cable, which becomes C/S structure.

The client shares the resources of the server. The server is usually installed in a secure location to hold the most valuable on-site production data from the DCS controller.

A computer acting as a server may serve several or more than a dozen clients at the same time, so the server needs to have a faster processor than the client, larger memory, and more storage space.

The client is a standard PC running the WINOOWs operating system, communicating only with the server, not communicating with other clients, and having its own software package. When working as a DCS operator station, the server and client can run the same monitoring software to share server resources. The difference between the two installed software is that the server should have DCS controller driver software and the client does not need it.

In another case, the OPC server is installed on the server monitoring software, and the OPCClient is installed on the client monitoring software, and the client can access the server. For example, if you do not need to graphically observe the operation of the field device on the server, you can also monitor the software on the server, but only the OPC server, only when you are looking for a fault, you need the monitor.

The C/S structure saves DCS' expensive network-specific interfaces. For the reliable operation of the system, dual-server redundant operation is used, which is a dual-machine dual network. A system uses up to two dedicated interfaces.

The end user reported that the multi-master architecture with multiple interfaces is better because the server and client are not properly arranged to crash. ConductorNT such as INFI90 and GUS from Honeywell are prone to crash in the lower version.

2.C/S structure crash phenomenon

Two types of hardware for a C/S fabric network, often referred to as a client PC and server. The client PC is in the central control room and the server is in a safe place. The reasons for the DCS operator station crash are complicated. This article is limited to analyzing the cause of the crash caused by the C/S structure. There are two cases:

There are two cases of connection between the controller, the server, and the client: First, the DCS controller has an Ethernet interface, and the three are connected by an Ethernet switch. The switch has multiple ports, and the data transfer rate of each port can vary. The number of ports is determined by the number of connected computers. The main indicators of the switch are the backplane width and the memory size. If an Ethernet cable such as 10BASE, 100BASE, etc. can be used for transmission, if the Category 5 cable is used, it can only be 10BASE. The Ethernet connection in Figure 2 is a star. Using discrete

The cable connects each computer to a central connection point, commonly referred to as a network hub. Each computer uses a separate cable, and the connection failure only affects the single computer involved, and other computers can continue to operate. If the speed of each machine adapter is the same, the Ethernet star connection usually uses a 10BaseT cable.

The controller sends the information to the server, and the client reads and writes data from the server. Due to the large number of clients, if they use the same connection between them, they may crash.

The data read by each client from the server can be the same or different. It can also be that all clients display the same content, and multiple operating stations operate redundantly. Each operating station exchanges more data with the server. The 100BASE port can be used. The server and controller exchange data less. The 10BASE port is used. The client does not exchange data between the clients. The 10BASE port can also be used. If this is not the case, it is likely that the client will crash.

Inappropriate configuration of client PC and server

The configuration of the client PC depends on the operating system you are going to use. Such as DOS and WINOOWs3.1 as long as 8MB of memory can run, and WINOOWs9x at least 16 ~ 32MB, WINOOWsNT at least 32MB, preferably 64MB, especially WINOOWs2000, more need 64MB of memory. In addition to memory, WINOOWs2000 should use a faster processor and a larger space.

When using a client PC, although it can be worse than the server machine, the memory must be higher than the above requirements. Because the machine has to read and write data with the server in addition to operating the operating system.

The server's CPU and hard drive are subject to service requests from all clients on the network. The server requires a lot of memory, which is much higher than the client memory capacity.

It is best to know the number of memory slots on the motherboard and the maximum amount of memory supported. Check how the memory that comes with the DCS server is configured. A single 64MB DIMM machine is easier to upgrade than a machine with a 16MB module full of 4 slots.

The server preferably uses a memory that supports ECC memory. When the ECC memory is used with an ECC-enabled motherboard chipset, it can correct single bit memory errors and detect multi-bit memory errors.

Hard drives are also important for servers. Computers connected to the Internet share server files. Hard drives should be durable, reliable, and suitable for serving multiple users at the same time. Therefore, it is more appropriate to use a SCSI hard disk drive. SCSI drives have intelligent and fast speeds, and UtraWideSCSI has a very high data transfer rate from drive to system. Whether it's an IDE hard drive or a SCSI hard drive, RAID technology can be used to save data more securely and improve server quality. The SCSI bus can connect multiple hard drives, using a 9GB hard drive without the efficiency of nine 1GB SCSI hard drives.

There is a disk array on the market that is designed for servers and is a mass storage product. The hard drive array can be loaded with a large number of drives, providing fault tolerance with disk mirroring or RAID, automatically storing multiple copies of server data to different hard drives. When the drive fails, all data is still useful to the user. Some array drives can even be hot swapped to replace faults while the machine is running.

The best version of RAID used by the server is RAID5, which is supported by all current versions of the C/S network operating system. RAID5 data "strips" onto multiple SCSI drives so that the contents of a single corrupted drive can be rebuilt from information from other drives in the array. When a server with a hot-swappable drive and a powered server is combined, RAID 5 allows the server to operate at approximately 24×7.

If the DCS is not an operation station with a C/S structure, since each operation station is independent of each other, only a part of the data is saved. If the reliability is already high, the RAID technology is not necessarily used.

The DCS operator station is a file backup and a SCSI tape drive can be installed on the operator station. In the C/S structure, it is better to install a tape drive on the server, and the user-configured application software can be copied on the tape drive.

The application can be re-entered when the operator station or server fails. Tape drive product updates are also faster. Tape drives equipped with imported DCS around 1990 are no longer available for spare or tape, and old and new tape drives are not compatible with tape.

The S9000 system controller consists of two parts, one 3C905 card completes the analog control, the other plc host LPM620-0072, and the host has an Ethernet port. When it is composed of the system, it adopts the Ethernet connection to form the C/S structure. There are two ways to connect an Ethernet cable. In many cases, a hub connection is used.

If only one or two operating stations and one or two controllers are used, all the operating stations and controllers are connected to the public cable, and there are few online connecting devices. T-connectors can also be used.

The server contains more drives than the client, and its power and reliability are also considered. Power supplies are easily overlooked, such as ignoring the stability of the power supply's DC output, as well as noise, spurious signals, spikes, and surges. The power supply is one of the most prone to failures. The electronic circuit uses +3.3V or 5V, and the hard drive and fan use +12V.

There are Ethernet network adapters and other network adapters, but they are not as much used as Ethernet, so the price is high. Using Ethernet can reduce DCS costs.

Third, the system instance is prone to crash

Design a system, the operation station adopts C/S structure, and the connection is as shown. For example, if 8 operating stations are configured, 2 servers, and the lower level is 4 S9000 controllers, the number of I/O points to be displayed in the operating station is 2000, and the controller and the server communicate via Ethernet. The physical connection of controllers, servers, and controllers is through a 16-port hub (SwitchHub). The default rate for a 16-port hub is 100BASE. If the adapter of the server and the operation station is 10BASE, the memory of the server and the client are both 64kB, and the system crashes after the system runs.

Replace the server's adapter with 100BASE. The client still uses the 10BASE network adapter, increasing the server memory to 256kB and the client to 128kB. The situation has changed a lot.

From the actual situation, the cause of the server and operation station crashes has a great relationship with the server memory capacity. 64kB of memory is the most basic, its memory should be related to the number of operating stations. The operation shows that adding at least one operating station server requires at least 10 to 30 kB of memory. There are two network adapters in the server, 10BASE for communication with the controller, 100BASE for communication with the operator station, and 10BASE adapter for the operator station. Use a normal cable for the connection.

If the server software and the operating station software are fix, the original S9000 operating station uses the software, the operating system is WINOOWs3.2, and the crash is not serious. At present, the S9000 system monitoring software is dedicated. The software may have some bugs. In addition, the network and memory configuration are not reasonable, and the crash phenomenon is very serious.

The server that the client should configure is the primary server and the other is the secondary server. Otherwise, when one server fails, the other server does not work well. Configuration is even more important if there are 3 servers.