Research on Networked Hydraulic Synchronous Control System

Networked synchronous control systems (NSCS) are developed rapidly. Networked synchronous control system refers to the distributed subsystems whose outputs need to be synchronized are connected through the network to form a distributed control type, and multi-node coordinated synchronization. Subsystems of hydraulic synchronous systems are normally distributed in a wide range of industrial area. The requirement of low-cost synchronous inter-subsystem data collection, communication and system distribution control are achieved by applying internet connection. In order to transmit sensing and control information, each subsystem inevitably generate network delay issues for networking synchronous control system. Therefore, setting up a hydraulic synchronization control system error model with network delay function to solve the network delay problem. Meanwhile, the robust asymptotic synchronization controller is designed by robust control theory, which could adopt a series of relevant simulations and experiments to verify the effectiveness. It has great engineering application value.


Introduction
With the improvement of mechanical system cooperation capability, the increase of load and the rapid development of computer network technology, various types of networks are widely used as communication media in engineering to connect control units with specific functions in the control system into networked control systems (Networked control). Systems: NCS) [1]. Networked synchronous control systems (NSCS) were developed in this context. The so-called networked synchronous control system refers to the distributed subsystems whose outputs need to be synchronized are connected through the network to form a distributed control type, and multi-node coordinated synchronization.
The offshore wind turbine lifting installation process includes hoisting, rough guiding, buffering, synchronous lifting, accurate-positioning automatic centering, flange connection and removal of the lifting system and so on. The synchronous lifting process is used to adjust the posture of the upper hanger and the fan after the buffering process, so as to facilitate the operation of accurate-positioning automatic centering process. Since the upper hanger and the fan have a total weight of several hundred tons, the synchronous lifting process is completed by using a hydraulic synchronous control system [2]. The entire hydraulic synchronous control system consists of four subsystems which are connected via a network. From the view of close-loop control, the sensors, controllers, and actuators in the hydraulic synchronous control system form a close-loop through the control network. These nodes share the network and send them in a time-sharing manner, which unavoidably generates network delays. Since the forward channel and the feedback channel have pure hysteresis links related to the network delay, it is difficult to analyze and design the synchronous control system [3,4].
In this paper, the network induced delay of the hydraulic synchronous control system used in offshore wind turbine installation is established. The system error model with time delay is established. The robust asymptotic synchronous controller is designed and verified by simulation and experiment. It has great engineering application value.

Offshore Wind Turbine Installation Hydraulic Synchronous Control System
The four subsystems of the Hydraulic Synchronous Control System are mounted on the four seating arms of the lower seating structure (located on the base platform), each subsystem controlling two hydraulic cylinders by one pumping station. During synchronous lifting, the eight hydraulic cylinders simultaneously support the synchronous platform on the upper hanger (the upper hanger is consolidated by the inner flange on the lower ring beam and the outer flange of the fan), and the synchronous platform is raised or lowered simultaneously to adjust the attitude of the upper hanger and the fan. Each pump station of the Hydraulic Synchronous Control System has a local controller (with CAN network interface). In order to realize the data acquisition, communication and system distributed control between subsystems, this project uses CAN bus to connect four subsystems. The network connection is shown in figure 1. Since the inconvenience of using wired connection, ZigBee wireless network communication is used to connect the upper computer (on the operating ship) and the main controller (on the base platform). The upper computer is only used to send the operation command and the operation of the monitoring system. The structure of the hydraulic synchronization control system is shown in figure 2. The main controller is used to receive the sensing information of each subsystem and calculate the control amount according to the set algorithm. The local controller is responsible for sampling the sensor signals of the subsystem, receiving the control volume sent by the main controller and controlling the pump station.

Hydraulic Synchronous Control System Modeling
A hydraulic cylinder is used as the master cylinder, and the other hydraulic cylinders (slave cylinders) track the output of the master cylinder for synchronization purposes. As long as each slave cylinder can accurately track the master cylinder, the synchronization performance of the whole system can be guaranteed. Therefore, the synchronous control problem of multiple subsystems can come down to a single synchronization between the slave cylinder and the master cylinder. Its equivalent control structure is shown in figure 3. Figure 3. Structure of the master-slave hydraulic synchronization system.
The hydraulic system principle of the subsystem is shown in figure 4. The two parallel hydraulic cylinders use variable frequency speed regulation. When the displacement between the two parallel hydraulic cylinders is not synchronized, the stop valve 21 is used for adjustment. A non-contact displacement sensor 19 is mounted in the hydraulic cylinder, and a hydraulic pressure sensor 18 is mounted on the large cavity of the cylinder. To visually verify the effectiveness of the prediction algorithm, each synchronization subsystem is modeled as a pump station and a hydraulic cylinder.  The following minor problems are neglected when establishing the mathematical model: pressure loss of check valves, oil filters, pipes, etc.; leakage of various hydraulic components except hydraulic cylinders; nonlinear friction and temperature characteristics of hydraulic systems; flow pulsation of pumps. Modeling the ascending process of the hydraulic cylinder, the established motor torque balance equation, flow equation, hydraulic cylinder force balance equation and displacement equation are as follows: In the formula, , uf = , respectively represent the state and input of the system, take s as the output of the system, and obtain the model described by the following equation of state from equation (1).
In order to facilitate digital simulation and single-chip computing, the units of The following can be obtained as follows: the master-slave system error model without considering the influence of network communication.

Robust Asymptotic Synchronous Controller Design
The master-slave NSCS controller is designed to track the output of the primary system asymptotically from the output of the system in the presence of network latency. Regardless of the packet loss that occurs during network transmission, the master-slave NSCS can be simplified to the structure diagram shown in figure 5. The control object in the figure is the synchronous control system error model, / ex is the sensor output, ca  is the controller-to-actuator delay, sc  is the sensor-to-controller delay. is the controller-to-actuator delay, which is a continuous function of time. Assuming that the master-slave system state can be measured, its error state vectors can be obtained. Due to the sensor-to-controller delay, the state feedback synchronous controller can be expressed as Substituting equation (7) into equation (6) can obtain the NSCS error model using state feedback.
In the formula, Considering that under the action of the controller, if the system error state () xt shown in equation (8) is asymptotically zero, the output error of the primary system and the secondary system can also be , if there is a gain matrix K such that the state feedback system (8) is asymptotically stable, then NSCS (6) is asymptotically synchronized. In order to obtain a robust asymptotic synchronization controller, the following time-vary and time-delay liner systems are considered.
where, ( In the formula, (1 ) Multiplying the formula (13) both sides to obtain: *0 ** P X P P X P P N P P X P P N P P ZP is asymptotically synchronized. It should be pointed out that equation (17) in Theorem 1 is not LMI. For this non-convex feasibility problem, it is difficult to find the global maximum delay h that can make the NSCS (6) asymptotically synchronized. Obviously 1 ZP − = can be chosen, equation (17) will become LMI and it is convenient to solve a suboptimal maximum delay h with the minimization problem solver MINCX in the MATLAB LMI toolbox. However, this is not the best way. We can use the method of paper [6] about introducing new variables,  If the solution to minimize the problem is 3n, then NSCS (6) is robustly asymptotically synchronized by the controller Although this LMI-based nonlinear minimization problem is difficult to solve the global maximum delay h , it is easier to solve than the non-convex feasibility problem of Theorem 1. Moreover, the following iterative method can be used to easily solve a suboptimal maximum delay h .
Iterative algorithm: (1) Set h to a sufficiently small initial value 0 h so that equations (16) and (21) (17) is not true and the number of iterations k exceeds the specified number of times, the program is terminated, otherwise, 1 kk =+,back to the third step. Since it is difficult to obtain an optimal solution that 1 1 1

() Tr JJ PP ZZ ++
is exactly equal to 3n , the iterative algorithm uses equation (17) as the condition for iterative suspension, and only obtains a suboptimal maximum delay h . When the maximum delay h of the NSCS is known, the above iterative method is slightly modified, and can also be used to obtain a state feedback controller that can make the NSCS asymptotically synchronized.

Hydraulic Synchronous Control System Modeling
The Hydraulic Pressure Synchronization Control System uses CAN bus to connect each function node in the system, and the baud rate of CAN bus data transmission is set to 200 kbps. The sensor information required for the synchronous control between the single slave cylinder and the master cylinder is two data packets. The control information has only one data packet, and each data packet is 8 bytes. These data are transmitted through the CAN bus. Tested by the CAN bus experimental platform [7], the average total network delay of the forward channel and the feedback channel of the NSCS is 1.8 ms, and packet loss does not occur during the process of the entire data transmission. When the state feedback controller is used, the control variable calculation time can be ignored, so the total network delay of the entire control closed loop is less than 0.05 s.
For equation (4), regardless of the influence of parameter uncertainty, external disturbance and delay derivative, using theorem 1 and iterative algorithm, a robust controller capable of asymptotic synchronization which can enable the hydraulic synchronous control system to asymptotically synchronize with a random network delay of less than 0.05 s can be obtained, and with a gain of = [ 0.0058 0.047 -0.000015 2]. The sensor node and the actuator node in the networked hydraulic synchronous control system are set to time drive [8], the controller node is set to event-driven, the sampling period is set to 0.05s, and the controller adopts the above robust control law, the initial error state is set to [ 0 0 0 20] , the displacement synchronization error simulation curve is shown in figure  6. The result shows that the displacement of the master cylinder and the slave cylinder can be asymptotically synchronized. Experiment with the master cylinder and a slave cylinder, the reference frequency of the main system is set to 35 Hz, the initial displacement of the master cylinder is set to 8mm, the initial displacement of the cylinder is set to 28 mm, and the measured displacement synchronization error curve and the displacement curve are shown in figures 7 and 8, respectively. The results show that in the case where the above-described robust control law is employed, the displacement synchronization error between the master cylinder and the slave cylinder is in the interval [ 1 ,1 ]

Conclusion
Offshore wind turbine installation The various subsystems of the soft landing hydraulic synchronous control system are connected via a CAN network and are a typical NSCS. Firstly, the mathematical model of the system is constructed. The modeling method used has certain universal significance for the hydraulic synchronous control system. Then based on the established data model and the time delay measured by the CAN experimental platform, the robust control law is solved. The effectiveness of the control law is proved by simulation and experiment.