Meco 2017 Zaev Final
Meco 2017 Zaev Final
Meco 2017 Zaev Final
Abstract—Hydraulic equipment suffered a great change during systems allowed development of hydraulic power schemes
the last years advancing from hydro-mechanical devices to which offer more than one I/O variables. Separate meter-in
digitally controlled electro-hydraulic systems. There are also separate meter-out (SMISMO) hydraulic systems are one
changes in hydraulic schemes permitting increased number of example of them (Fig. 2). Here, meter in and meter out valves
control variables (e.g. separate metering systems - SMISMO). are controlled independently allowing simultaneous control of
Available control strategies which require high computing the speed and backpressure of the actuator, which makes them
performances, like optimal control, must be examined for multi-input multi-output systems (MIMO). Introducing second
implementation on real hardware. For this purpose digital control variable, back actuator pressure, allow improvements
controllers must be tested in real-time before they are
in: 1) Cylinder movement, making the movement stiffer, more
implemented in the real machine. Hardware-in-the-loop (HIL) is
technology which is available for this. In this paper a procedure
stable and lowering the vibration [7], 2) Energy efficiency of
for development of HIL simulator for servovalve driven system the hydraulic system, making possible to use power of over-
and SMISMO hydraulic systems is shown. First, mathematical running loads to move the load itself [7][8].
model of hydraulic systems is developed using graphical
environment. Then, real-time simulations are performed with a
platform consisted from two industrial PLC and one PC. One
PLC is used to simulate the real system (servo valve-cylinder)
and the other one is real controller under test. PC is used for
changing controller parameters and visualization.
I. INTRODUCTION
Hardware-in the-loop (HIL) is a widely used new Figure 1. Typical hardware-in-the-loop configuration.
technology for testing controllers before their implementation
in the real system. They drastically decrease the costs of Advanced control strategies like optimal control require
required real-time simulations, changing physical models of the high computing performances and must be examined for their
controlled system (plant, machine. process) with virtually real-time capabilities once implemented on real hardware. This
simulated ones (Fig. 1). In those simulations, part of the system is necessary because all computing must finish before next
is real, in his final stage of production (controller) and part of cycle starts. If computing is not finish, the controller (hard real-
the system is simulated in real-time (controlled system) [1][2]. time hardware- PLC), gets stuck.
Virtual models are running in real-time (same as physical ones)
and can have same I/O interface for communication with In order to be able to test advanced control strategies for
control system as real plants. This technology is especially hydraulic servo systems and SMISMO hydraulic system, HIL
used in automotive industry for development and testing of simulation platform is developed. This platform is available for
ECUs [3][4]. commercial use as well as for educational purposes in the
Laboratory for automation at the Faculty of Mechanical
Hydraulic power systems have been, for a long time, used Engineering in Skopje. The development process is described
in industrial manufacturing and in heavy machinery. Classical in this paper. The goal of the HIL simulation is to analyse the
hydraulic power systems (e.g. cylinder driven with servo valve) influence of the real control system against the controlled
represent single-input single-output (SISO) systems [8]. There, object, in our case a hydraulic drive system, represented with a
operator’s joystick signal is input and speed of the actuator is simulation model that includes dynamics of the system.
output variable. Introducing digital control in hydraulic
6th Mediterranean Conference on Embedded Computing MECO’2017, Bar, Montenegro
II. MODEL DEVELOPMENT flow equation where flow coefficients are calculated using the
catalogued data of the valve (QN, pN, and y [V]):
A. Model of Servo Valve controlled Cylinder
For the schematic on Fig. 2, the Newton’s Second Law of Qv Kv ·y· pv ; Kv
Qn
motion equations is: ymax · pn
dxa
va
dt (1) Nominal flow and nominal pressure drop are very usual
parameters for the servovalves and can be find in eny valve
dva pA A1 pB A2 FB Fext catalogue. y[V ] represents the input voltage of the valve and is
(2)
dt ma proportional to valve opening and pv is the pressure drop
Load forces are represented as external force and in here over the valve.
presented simulations it is assumed that they are constant. A fluid power system using a liquid such as a hydrocarbon
Friction forces are directly related to velocity. If cylinder seals liquid operates at pressures where the compressibility of a
exhibit viscous friction characteristics, then friction forces can liquid has a noticeable effect on the operation of a system. Bulk
be represented with the simple linear equation modulus is an elastic constant giving the amount that the oil
FB B va volume is reduced for a given application of pressure.
2( ps pcyl )
Qv Cd wxv (t )
This equation can now be rearranged as the first building if a linear model of the hydraulic system is used (for example
block in the development of dynamic equations. cyclic computing in 10 ms .
dp E dV
dt V dt
At any instant of time, the rate at which the fluid in the
volume V is being compressed (i.e. dV/dt < 0, hence the sign
adjustment discussed previously) is given by:
dV i m i m
dt i 1
Qin Q
j 1
out
Using the above equations, this volume change can be
expressed as a rate of increase in pressure:
dV E i m i m
dt V (t ) i 1
Qin Q
j 1
out
Figure 4. Simulink model of the system servovalve-cylinder
dp A Ee
QA A1va (3)
dt V1
dpB Ee
A2 va QB (4)
dt V2
Thus this system requires 4 ordinary differential equations
to describe its operation (1,2,3 and 4). A state variable is
simply a quantity that appears as a first derivative on the left-
hand side of an ordinary differential equation, in this case, xa,
va, pA, and pB.
The dynamics of the valve can be represented with the
following equation [6][7] for a standard first order system:
Xv 1
Gv ( s) ( s)
Uv Tv ·s 1
dxv
vv
dt (6) Figure 5. Servovalve controlled actuator. Displacement and cylinder
The system has 6 state variables: xa, va, p1, p2, xv, vv. From pressures vs. time.
equations 1-6 we have 6 differential equations describing the
state-space form of the system servovalve-hydraulic cylinder. B. Real-Time Model Of The System Servovalve-Cylinder
These equations are implemented in the Simulink model of the A real-time model of the system servovalve-cylinder is
system servovalve-cylinder (Fig.4). For simulating SMISMO developed in such a way that B&R blocks are introduced in the
hydraulic systems, different model of the valve has been Simulik models on Fig. 4. Those blocks surve for comunication
created (Fig.3). Results from a simulation of a cylinder motion with the simulated system. This is shown on Fig. 6. Thanks to
can be seen at Fig. 5. Simulations have shown that for real-time the B&R Automation Studio target to Simulink, this model can
simulations it is the best to use the Matlab solver Ode14x with be directly downloaded from Matlab/Simulink (Real-time
1 ms time step. Simulations with bigger time steps are possible Workshop), as C program, to the Automation Studio as a cyclic
program.
6th Mediterranean Conference on Embedded Computing MECO’2017, Bar, Montenegro