SIMULATION STUDY OF A MOBILE MACHINE WITH SPECIAL REFERENCE TO ENERGY EFFICIENCY Timo Käppi, Asko Ellman and Matti Vilenius Institute of Hydraulics and Automation, Tampere University of Technology P.O. Box 589, FIN-33101 Tampere, FINLAND email: tkappi@butler.cc.tut.fi ABSTRACT Computer simulation is a powerful and generally accepted practice to carry out research in the area of fluid power. This is especially true with heavy machinery. Production series are relatively small and unit prices are high. Due to this proypes are rarely available for machine designers and researchers. There are also a lot of variables to be measured. This means that field measurements are difficult to arrange and also very costly. It gets even more time consuming if several machine configurations, mechanic or hydraulic, are to be tested. Environmental circumstances typically cause delays, too. Based on these facts simulation study can be very beneficial by technical and economical means. As an example a forest machine is discussed. Software implementation of the simulation model is presented. The simulation model is utilized to study the energy efficiency of the machine while undergoing a work cycle. Different hydraulic configurations are tested by means of the simulation model. The results are shown for qualitative comparison and decision-making. KEYWORDSModelling, simulation, energy efficiency, mobile hydraulics 1. INTRODUCTION Energy efficiency is one of the main topics while considering the product development of mobile hydraulics. This can be motivated by several reasons. Naturally, since the primary power source of the machine is usually a diesel engine, fuel economy is one of the reasons. An even more important thing is the system temperature balance. The power losses in the systems are converted into heat. The oil tanks of the machines are relatively small and the system temperature rises fast while hydraulic power is utilized inefficient way. This has many unwanted effects. The oil cooler must be dimensioned larger which increases price of the machine and may cause problems in the machine layout. In the worst case the primary motor must also be larger than needed in a well designed and energy efficient system. This creates a chain reaction where heavier structures must also be used. The pressure and flow requirements vary remarkable during the machine utilization. This sets demands for the system design in order to minimize the losses and thus maximize the efficiency. The traditional practice in mobile hydraulics is to use fixed displacement pumps with opencenter valves. Constant pressure systems are also used. However, the current trend is towards load sensing (LS) systems with variable displacement pumps. The basic reason for this is the supposition for improved energy efficiency. Nevertheless, this is not unambiguous in case of mobile machines. In load sensing systems the supply pressure is determined based on the highest load pressure in the system. The best system efficiency is achieved when only one function is carried out at a time or when two or more simultaneous functions have the same pressure requirements. However, typically the pressure levels of simultaneous functions vary. In order to prevent load dependence and load interactions between the simultaneous functions, pressure compensation is utilized in modern machines. A pressure compensator maintains nearly constant pressure loss across the main valve spool. So the speed of an actuator remains constant and is a function of main spool opening only. This makes the machine control easier for the operator and productivity is increased. Nevertheless, it usually means remarkable pressure losses i.e. heat generated into the system due to pressure losses occurring in pressure compensators. The amount of possible hydraulic configurations depends on the number of hydraulic actuators and pumps in the system. For a mobile machine typically tens or even hundreds of possible combinations can be found. For the machine designer one of the main questions now is how many pumps there should be the fluid power system and what functions should be grouped together for each pump system. The most advantageous design is achieved when the actuators with the approximately equal average pressure level are placed in the same pump system. Increasing the number of pumps in load sensing system does not make the system worse considering energy efficiency aspects but the advantages may be so marginal that is questionable. Considering the number of alternatives the simulation approach compared to the empirical tests is well motivated while aiming to the optimal design. 2. SOFTWARE IMPLEMENTATION Commercial software are almost without exception implemented in a certain way which limits their suitability. They either lack of readily available component models or the software itself is not flexible enough for required
modifications. In house software equipped with comprehensive component library forms an effective engineering tool. WinSIMU is a fluid power simulation package developed at Tampere University of Technology during the last ten years [1],[2],[3]. It has a component library including all the basic fluid power components. The user is allowed to make his own extensions based on already existing model classes such as orifices and fluid fields or more complex models. The source code of WinSIMU is implemented with 90 standard FORTRAN. Then fluid power system model itself consists of simple FORTRAN subroutine calls to the library or userwritten subroutine models. In that case even a very complex system entity becomes easy to control and modify. Due to the nature of mobile hydraulics, a special model library for mobile hydraulics components is also created including variable displacement pump with LS-regulator and power limiter mobile valve including integrated functions models for passive and active oil cooling non-linear mechanism model models for load conditions and working process The software implementation is based on object-oriented programming while each component forms an object of its own. By employing indexed structures supplied by FORTRAN programming language the number of fluid power components in a simulation model can vary. The major benefit of this implementation is that single simulation model can cover a large number of possible hydraulic configurations. Commercial software packages require that the fluid power circuit is constant i.e. the number of hydraulic components and connections (pipes/hoses) between components are fixed. This means that separate simulation models for each configuration of the fluid power circuit have to be created. In the presented software concept the number of pumps is given as a parameter and each hydraulic actuator is parameterized by giving a number corresponding to the pump system it belongs. In the initialization phase of the numerical simulation the hydraulic circuit equivalent to the current parameterization is formed. number of pumps 1 cylinder 2 1 cylinder 1 1 cylinder 3 1 number of pumps 3 cylinder 2 2 cylinder 1 1 cylinder 3 3 Fig. 1 Example of two systems generated by different model parameterization In the simple example illustrated in Fig. 1 two simple hydraulic circuits are generated based on the four parameters given. Considering the number of components and possible different connections in the real machine systems the presented implementation saves a great deal of programming work while the performance of different configurations is studied. The software implementation concept makes the model especially advantageous for the energy efficiency study which is presented as an example in chapter 4. 3. SIMULATION MODEL The simulation model covers a forest machine having 7 degrees of freedom, 9 hydraulic actuators, variable number of LS-pumps and oil cooling system. The modelling of oil cooling system and energy efficiency calculations are discussed next. 3.1 Oil cooling system The oil cooling in the entire machine system consists of active and passive cooling, oil cooler and passive heat transmission, respectively. The cooling power of the oil cooler is a function of oil flow across the cooler, air flow from the fan and the temperature difference between oil and environment. The flow across the cooler depends on the utilization rate of the hydraulic system. Typically the return oil of a mobile valve is driven through the oil cooler. Since the analytical model of cooler performance is difficult to parameterisize and cooler measurements can be received from the manufacturers, empirical model for oil cooler is suggested. The oil cooler performance is assumed to be linear between the measured points. Due to the empirical nature of the oil cooler model the accuracy of model is very good. Therefore, when it comes heat generation study the dominant
sources of error are the cooler measurement accuracy and the description of the work cycle. In fig. 2 the cooling power of a oil cooler is presented as a function of temperature difference and oil flow. Air flow is assumed to be constant which is typically the case with mobile machinery. i k Peff = max i j = 1 j ( F x &, 0) + max ( M θ&, 0), k 1...n 1 The eq. (2) takes into account the negative work i.e. negative work is not considered to reduce the amount of the effective power used. The difference between input power and effective power is therefore the loss power (2) P k k k loss s eff = = P - P, k 1...n (3) Fig. 2 Oil cooler power Passive heat transmission results from the heat energy that is emitted from the system via surface areas of oil tank, pipes and actuators. It is a function of the surface area and temperature difference between the system and ambient temperature (fig. 3). The energy loss made in the complete system can be calculated by integrating the power loss over the cycle time t c in each pump system and summing up individual losses. While the effect of active and passive cooling is also taken into account the heat energy cumulated into the system is n t c t c k = E loss Ploss dt - P cool dt (4) 1 0 0 The temperature rise in system during one work cycle can be calculated as follows when the energy losses are assumed to be completely converted into heat. E loss E loss = c o m o T T = (5) c o m o The temperature progress during consecutive work cycles can be determined by repeating calculation phases (4) and (5) and considering the change in temperature difference between fluid power system and ambient temperature which affects to the cooling power. 3.3 Work cycle Fig. 3 Passive heat transmission power 3.2 Calculation of loss energy and temperature rise in machine system The calculation of the temperature rise in machine system is based on the supposition that the energy brought into system which is not used for the active change of state of motion of the machine mechanism is considered to be loss energy and thus heat. The input power of separate pump systems can be calculated based on the rotational speed and input torque of the pump An investigated work cycle was developed and programmed in order to take advantage of the simulation model. By controlling the valve openings in the model, the determined machine performance is achieved. The examined work cycle was created based on the impression of typical machine operation, video recordings and opinions of the machine operators. It should represent the average work cycle which is more or less repeated during normal machine performance. The scheme presented in fig. 4 illustrates the machine operations which are partly simultaneous as a function of time. During 20 second work cycle three trees are cut and collected to the attachment accumulating and felled down in the end of the cycle. k k k s p s = P = M ω, k 1...n (1) The effective power used in the systems can solved from the state of motion of the mechanism. The effective power used in each pump system is defined as
Track 1 (tdm1) Track 2 (tdm2) Hoist Stick Tilt Clamps 1 (up) Clamps 2 Tree 1 (1000) Tree 2 (500) Tree 3 (500) Swing Wrist 3.4 Research procedure Time (s) Fig 4. Simulated work cycle When it comes to the energy efficiency study of a mobile machine, the research procedure can be carried out in the following way. 1. The average work cycle is determined to compare the alternative system designs. 2. The work cycle is programmed to the simulation model of the machine. 3. The loss work done per work cycle for each alternative system is determined by means of the simulation model. 4. The temperature gradient of the system as a function of time can be solved when the loss work and active (oil cooling) and passive (heat transmission) cooling power is taken into account. The temperature rise in the system during consecutive work cycles is determined using some numerical mathematics software available such as Matlab. 5. The physical proype tests are carried out with the most promising alternatives. 4. SIMULATION RESULTS AND ANALYSIS Three machine systems having different number of LSpumps were studied. The energy losses of the systems were first investigated during a single work cycle. Fig. 6 Simulated flow across the oil cooler as a function of time The question of main interest is what is the final temperature reached when the work cycle is repeated. In the following the oil temperature progress is presented for different systems. The non-linearity of the temperature development is caused by changing temperature difference between the system and environment. In the beginning the temperature of the oil and the ambient temperature are assumed to be the same (20 C). Fig. 7 Temperature rise in 1-pump system The plots in fig. 7, 9 and 11 (from up to down) are system temperature without cooling system temperature with passive cooling only system temperature with active cooling only system temperature with passive and active cooling The maximum cooling power can be achieved in ideal conditions only. In figures 8, 10 and 12 the temperature developments of different systems with cooling power of 100, 75 and 50 % are presented. Fig. 5 Simulated cumulated energy losses as function of time The loss energies cumulated into the systems (cooling ignored) during one work cycle are 825, 514, 477 kj for 1, 2 and 3 pump systems, respectively. Fig. 8 Temperature rise in 1-pump system
system is only marginal and considering economical aspects hardly worthwhile. 60 50 Temperature 40 30 20 Fig. 9 Temperature rise in 2-pump system 10 0 1 pump system 2 pump system 3 pump system Fig. 13 System temperature balances 5. CONCLUSIONS The modelling and simulation of mobile machines were discussed. The advantages of simulation approach were described. The benefits of software implementation were clarified. As an example the energy efficiency study of a forest machine was presented. Fig. 10 Temperature rise in 2-pump system Fig. 11 Temperature rise in 3-pump system LIST OF SYMBOLS c o specific heat capacity of oil [J/m C] E loss loss energy [J] F force [N] M torque [Nm] m o mass of the oil in the system [kg] P cool cooling power [W] P eff effective power [W] P loss loss power [W] p s supply pressure [Pa] T temperature difference between oil [ C] and ambient temperature t c work cycle duration [s] x linear actuator displacement [m] θ rotational actuator angle [rad] REFERENCES Fig. 12 Temperature rise in 3-pump As seen in fig. 13 the change from 1 pump system to 2 pump system is clearly advantageous and considerable difference between temperature balances can be seen. However, the improvement from 2 pump system to 3 pump [1] Ellman,A.U., Käppi,T.J., Vilenius,M.J. : Simulation and Analysis of Hydraulically Driven Boom Mechanism. 9 th Bath International Fluid Power Workshop, September 9-11, 1996, Bath, England. Published in book Fluid Power Systems, edited by C. Burrows and K. Edge, Research Studies Press, Somerset, En gland. pp. 413-429. [2] Ellman,A.U., Lindberg,I.I., Vilenius,M.J. : Simulation in the Design of Hydraulic-Driven Machines: New Approach and Aspects of Application. Proceedings of the 3 rd Scandinavian International Conference on Fluid Power, May 25-26, 1993, Linköping, Sweden. pp. 29-41. [3] Käppi,T.J., Ellman,A.U. : Modelling and Simulation of Proportional Mobile Valves. Proceedings of the 4 th JHPS International Symposium on Fluid Power, November 15-17, 1999, Tokyo, Japan. pp. 531-536.