Introduction

In recent years, China's industrial automation construction, large-scale mixer is in urgent demand. Compared with small and medium-sized mixer, large-scale mixer has a large output, high efficiency, stable performance, low energy consumption and other characteristics, representing the development direction of the batching equipment. However, the existing large-scale mixer control system technology is backward, using a single machine, a single screen, a single control, unstable performance, low reliability of the host, line and other key equipment, the control process of the information record is not perfect. For the defects of the current mixer multistate control system, developed a new mixer multistate dosage control system, so that it has a dual-machine full-process redundancy backup, dual-redundant industrial Ethernet bus, the lower computer using a cheap and reliable high-speed 32-bit ARM architecture with a 100Mbps Ethernet interface embedded system, and can be run to record data.

1 mixer multistate control system design

The dosing machine mainly needs to be accurately controlled according to some analog weighing module and digital weighing module return data, its collection includes 6 levels of material delivery, 6 levels of material weighing weight, drying cylinder temperature, amount and temperature of 20 analog quantities. Switching quantity inputs such as feeding, material door interlock, powder door interlock, bucket door switching position, mixer, inclined belt, flat belt operation, system allowable weighing, unloading and so on. Switching outputs include each material gate feeding motor relay, each screw machine switch, pneumatic valve, plus dose pump, each weighing hopper door opening, mixer door switching and so on. The system is configured with 20 analog inputs, 7 analog outputs, 110 switch inputs and 108 switch outputs.

The dual redundant central control computer, the highly reliable embedded lower computer and the industrial Ethernet that connects them together form the framework of the whole control system.

The central control computer is responsible for the management and sending of overall measurement and control information, and under the control of the operating system and application software, it completes the functions of human-machine interface display, data analysis and processing, data recording, monitoring, alarming and printing.

The central controller is composed of two highly reliable embedded 32-bit computers. The board integrates AMDGX2-400MHz CPU, 2 RS-232 and 1 DB-15VGA interface, 2 100Mbps end enclosure network card, 2 USB interfaces, 1 TypeI/IIPC card slot, using embedded Linux operating system and Xwindows graphical user interface.

The feeding system of the central control machine is used A, B machine dual-redundant embedded 32-bit computer, the two machines work at the same time, due to the directionality of the Ethernet frame, the lower computer to send frames to the A machine generally can not reach the B machine, so the system uses programmable Ethernet switches, such switches have a port copy of the function of the packet to be sent to the A machine is a copy of a copy of the packet to the B machine. Nevertheless, as a backup machine, machine B should still be isolated from the network through a dedicated packet filtering routing software, this route by splitting the packet to determine whether the packet can pass, for machine B, normal operation only allows the measurement and control of the Ethernet information on a single transmission to the B, i.e., the route closes the channel of machine B to the measurement and control of the Ethernet channel. A and B monitor each other's work through a dedicated Ethernet channel and heartbeat algorithm. In case of abnormal work of the other machine, it instantly replaces the work of the other machine.

The heartbeat algorithm adopts a two-way pull model, in which the monitor only sends request packets and receives reply packets, while in the two-way pull model, the supervisor (the centralized control machine A in this project) and the supervisee (the centralized control machine B in this project) actually have equal identities as each other's supervisor and supervisee, send request packets and receive reply packets to each other and insert the CPU load and time synchronization information of the local machine in the packets. with time synchronization information. In normal operation, an aliveness packet is sent to the other party at certain intervals δ . When A fails, the reply packet time exceeds the pre-set timeout value T0, which triggers the A-side failure message of machine B, and machine B starts to take over the work of machine A. The packet filtering route (also called packet filtering firewall, packet filtering gateway) of machine B opens, and the packet filtering route of machine A closes, and the open/close time occupies only a few clock steps of the centralized The open/close time only takes up a few clock steps of the central control machine, so it is all in microseconds. While allowing the packets input to the measurement and control Ethernet to pass through to the B machine while blocking the path to the A machine, the B machine records and alarms in a timely manner, and continues to receive test quantities sent by sensors and sends commands to the actuator.

Unlike the failure of machine A, there is no need to open the packet filtering route when machine B fails, as long as machine A alarms to notify the failure information and records it.

The dual redundant Ethernet switches are industrial Ethernet switches, which are in the position of network switching center and are used to transmit network frames in the shortest possible time and ensure as few frame conflicts and packet loss as possible.

The lower node machine with data acquisition, conversion module and Ethernet controller adopts the architecture of ARM7 CPU plus DM9000A, which is responsible for collecting analog and switching quantities as shown in Fig. 2, and adding the collected data with the sensor flag (with IP address as the flag), and then packaged and uploaded. And the control information is received and unpacked, distributed to the control terminal according to the control flag. The Ethernet communication circuit is mainly composed of network control chip, network transformer and E2ROM memory.

Network control chip DM9000A is the core of Ethernet communication of the lower node machine. It has 10M/100M communication rate. Network transformer PH163530 is to convert the data transmitted from the Ethernet into level signals matching with the Ethernet control chip.E2ROM memory mainly stores some parameters that need to be saved in the power-down of Ethernet communication, such as the local P address. The output control part consists of D/A analog output and digital switching output. the D/A output signals mainly control some analog input devices, and these signals are generated by the 10-bit D/A converter inside the CPU under the control of software. The digital switching output circuit can control external devices with higher power, such as conveyor belt motors and fans.

The LPC2132 microcontroller is based on a 32-bit ARM7TDMI-SCPU that supports real-time emulation and embedded tracing, running at a clock frequency of 50MHz up to 70MHz with 64kB of embedded high-speed ash memory. 128-bit wide memory interfaces and a unique acceleration architecture enable 32-bit code to run at maximum clock rates. Applications with tight control over code size can use the 16-bit Thumb mode to reduce code size by more than 30% with minimal loss of performance. The small package and very low power consumption make the IPC2132 particularly suitable for small applications such as access control and POS; with a built-in serial communication interface and 16kB of on-chip SRAM, it is also well suited for communication gateways, protocol converters, software modems, voice recognition, and low-end imaging, providing massive buffers and powerful processing for these applications. Multiple 32-bit timers, one or two 10-bit 8-way ADCs, 10-bit DACs, PWM channels, 47 GPI0s, and up to nine edge- or level-triggered external interrupts make them particularly suitable for industrial control applications and other systems requiring high reliability.

The data port SDx of the system's Ethernet chip controller is connected opposite the POx of the CPU. When the lower computer starts to work, the software control IPC2132 sends out the initialization command, and the DM9000A initializes, loads the IP address, MAC address, assembles the standard ARP packet and UDP packet (the UDP packet is used here in consideration of the system transmission efficiency, and TCP packets can be used when the system nodes are less burdensome) and sends the initialization packet to the upper computer continuously, until the upper computer receives and send back MsgOK packet, and the lower computer receives and replies MsgOK packet, these two packets are distinguished by the different positions of P address in the Ethernet frame. At this point, the initialization and self-test are complete. The lower computer starts transmitting/receiving data and commands to/from the upper computer at a frequency that realizes the predetermined frequency.

The lower unit requires 2 operating voltages, +5V and 3.3 V. The 5V power supply is designed to provide power to the transmitter of the sensor: the 3.3V power supply is isolated into digital and analog parts by inductive components to improve system reliability.

The power supply of this system adopts the IEEE802.3af Power over Ethernet (POE) method, and the 48V power obtained from the idle pairs of the Ethernet data lines will be converted by DC/DC and regulated by the voltage regulator to output the working voltage required by this system without the need for a special AC/DC power socket. The power supply part uses the DF1501S bridge rectifier to rectify the 48V voltage, and then converts the power supply to +5V after passing through the DC/DC module, and the LM1117-3.3 module converts the +5V power supply to the 3.3V voltage required by the CPU and the A/D conversion part.

2 System Software Configuration

The polymorphic control software of this dosing equipment includes the modular configuration of control interface and control program, the configuration of operation, process monitoring, report and event record, and the configuration of remote transmission, dual-machine redundancy and software encryption.

The whole feeding system software is divided into two major parts: the central control machine software system and the lower node machine software system. Among them, the central control machine is supported by the backend service provided by MySOL, which carries out a variety of tasks such as system-wide control interface, heartbeat maintenance, network communication maintenance, operation data recording, remote control and so on. The lower node machine is responsible for receiving/following commands and data, network transmission services, data solving (e.g., FFT of vibration signals, etc.) services, etc.

3 Conclusion

1)The large-scale blender control system with industrial Ethernet as the architecture, dual redundant controller as the core and 32-bit ARM system embedded system as the node is realized, and the IP-based Ethernet one-network structure is realized, and the industrial Ethernet bus can fully meet the needs of large-scale blender;

2) High-speed 32-bit ARM embedded systems and sensors composed of 100Mbps bandwidth intelligent sensing/execution nodes have been applied in the large-scale control network of this project and its transmission efficiency and reliability have been fully verified;

3) Dual redundant on-line main controllers are used to work simultaneously, using packet filtering gateway to control failover;

4) The use of POE power supply removes redundant power supply units, making the system more concise and reliable.

4 Summary

TTCAN bus network formed using redundant dual-channel transmission and bus protector. Not only improves some defects in the reliability of CAN itself, and combines the characteristics of TTCAN itself, improves the reliability of transmission and real-time, so that it can be adapted to the control system of the modern complex network, especially in the need for high security system, has certain practical significance.