Design of Networked Control Model

0 Introduction
With the development of network technology, the Internet is gradually integrating computer systems and communication systems around the world to form an information highway and a public data network. On this basis, the traditional industrial control field is also undergoing an unprecedented transformation, from traditional control systems to control characterized by networking, forming a new control network.

The structure of control system has evolved from the initial computer centralized control system (CCS), to the second generation distributed control system (DCS), to the now popular fieldbus control system (FCS). Ethernet has gradually combined with fieldbus and entered the field of industrial control.

The development of the network has brought about major changes in the architecture, control methods, and human-machine collaboration methods of automation systems and industrial control systems. On the one hand, automation and industrial control require deeper penetration of communication and network technologies. On the other hand, the management and control of communication networks also require more control theories and strategies. At the same time, it also brings new topics, one of which is the need for innovation in control methods and algorithms in network environments.

Therefore, this paper studies the design of a model for networked control technology. This model has good openness and reliability. It can not only simulate some common controlled objects in industrial sites and design specific control loops, but also conduct specific research on the application and improvement of control algorithms in a networked environment.

1 Hardware system design
The hardware system of this networked control technology model is mainly designed in two parts: the design of the field control model device and the construction of the network control system.

1.1 Design of the field control model device
The field model device is a multifunctional experimental device based on the physical simulation object of the industrial process, which is used to simulate the actual controlled object of the industrial production site. The system parameters are comprehensive, covering typical thermal parameters such as liquid level, flow, pressure, temperature, etc. in the industrial production site, and can realize various control forms such as system parameter identification, single-loop control, cascade control, feedforward control, ratio control, etc.

The field control model device is mainly composed of a heating furnace, an upper water tank, a middle water tank, a lower water tank and a water storage tank, as well as solenoid valves, water pumps, temperature, pressure, flow and other field instruments. The structure of the field control model device is shown in Figure 1.

1.2 Construction of network control system
Ethernet has the advantages of high transmission speed, low consumption, easy installation and good compatibility. Since it supports almost all popular network protocols, it is widely used in commercial systems. In recent years, with the rapid development of network technology, due to the great increase in switching technology and network bandwidth, Ethernet has begun to combine with fieldbus technology to enter the control field, forming a new type of Ethernet control network technology. This platform network control system adopts Ethernet control system.

Using the FieldPoint network module with Ethernet interface produced by NI, with appropriate I/O modules and analog instruments, the temperature, liquid level and other parameters of the field experimental model device can be controlled by Ethernet. The Ethernet control system structure is shown in Figure 2.

2 Software Platform Construction
2.1 Monitoring Software Design
The upper-level monitoring software of this system is designed using the LabVIEW graphical development environment, which not only realizes the computer monitoring of the field parameters of the model experimental device, but also specifically realizes the computer simulation of the process control experiment. Figure 3 is one of the monitoring interfaces of the system.

2.2 Experimental Research
The liquid level of the upper tank is taken as a simple regulation object to carry out the liquid level control experiment.

The upper tank level control is a single-volume self-balancing liquid level controlled process. As shown in Figure 4, the inflow of the upper tank is g1, and changing the opening of valve 1 can change the size of q1. The outflow is q2, which depends on the user's requirements and the height of the liquid level h. Changing the opening of valve 2 can change q2; the higher the liquid level h, the greater the static pressure of the water in the water tank, and q2 also increases. The dynamic equation is:

where T=FRs, K=KμRs. Rs is the liquid resistance; Kμ is the proportional coefficient; F is the liquid capacity. Rewrite the formula into a Laplace transform formula:

This is the transfer function of the upper tank liquid level object, where T is called the time constant of the object, and K is called the object amplification factor.

Based on this, the working principle diagram and liquid level control program of the upper tank liquid level control experiment are designed, as shown in Figures 5 and 6.

2.3 Application of neural network in the parameter tuning and optimization process of this control system
Figure 7 is a block diagram of the neural network fuzzy adaptive PID controller written in LabVIEW, which mainly consists of three parts: adaptive wizard, adaptive module, and PID control module. The adaptive wizard provides users with a user-friendly interface, and users can intervene in the adaptive adjustment process and set some parameters. The adaptive module automatically tunes the PID parameters online according to the neural network adaptive algorithm based on the set process parameters. The PID parameters after tuning are used to correct the parameters in the field equipment.

However, the actual object step response experimental results show that the actual characteristics of the upper tank liquid level object are quite different from the ideal first-order inertia link. This is because the linearity and sensitivity of the water inlet valve are not high enough due to hardware limitations, and on the other hand, it is also due to interference, such as the impact caused by the start and stop of the water pump. In addition, due to the transmission time delay caused by the communication network, the object also has to add a pure delay link.

When the Ziegler-Nichol tuning method is used directly, the tuning effect is not ideal because the object step response curve is not accurate enough and the actual controlled object is not a first-order inertia link. The closed-loop step response curve measured in the experiment is shown in Figure 8.

As can be seen from Figure 8, when the conventional tuning method is used, the tuning effect is not satisfactory, the system overshoot is too large, and the adjustment time is too long.

Moreover, when the characteristic parameters of the system are affected by other interference factors, such as rising temperature, it is necessary to re-measure the open-loop step response curve of the object, which is time-consuming, labor-intensive and lacks automation.

When the online tuning scheme based on neural network fuzzy PID control algorithm is adopted, the closed-loop step response curve of the system is shown in Figure 9.

From the response curve measured in the above experiment, it can be seen that after adopting the neural network fuzzy adaptive control algorithm, the control index has been significantly improved, the overshoot has been reduced, and the steady-state accuracy has also been improved. More importantly, the neural network fuzzy adaptive control algorithm can adapt to the changes in the dynamic characteristic parameters of the object, and can be automatically adjusted online, so it has good application value.

3 Conclusion
Network technology, as a representative of information technology, its combination with control systems will greatly improve the level of control systems. As a product of the cross-integration of control, network and computer technologies, the development of networked control technology is a reflection of the increasing complexity of control systems. Its theoretical basis spans multiple disciplines and its application range covers multiple fields. As an emerging research field, the research on many issues of networked control systems is only a beginning, and there are still a large number of topics that need further in-depth research.

This networked model device integrates Ethernet with traditional industrial control to form a new control network. While ensuring the original stability and real-time requirements of the control system, it also enhances the openness and interoperability of the system and improves the system's adaptability to different environments.