【正文】 ncy in current systems is due to the link between the flows of the two ports of the cylinder. This is because most valves use a single spool to control the flow in both ports. Because of this link, it is impossible to set the pressure levels in the two sides of the cylinder independently. Therefore, the outlet side will develop a back pressure which acts in opposition to the direction of travel, which increases the pressure required on the inlet side to maintain motion. Since the force generated by the actuator is proportional to the pressure dif ference between the two sides, the actual pressures in the cylinder don39。Virvalo 1997). They are also based on rather plex control algorithms. It is the goal of this thesis to start with a control strategy which is based on simple PI controllers and makes no demands for position and velocity of the cylinder. The performance of this system will be lower than a plex control strategy, but it may be easier to implement mercially because it has no need for special sensors and is easier to understand for the average engineer. Figure 3 Different Situations in Crane Operation Another feature which needs to be acknowledged when designing a control strategy, is thetype of valve used. Mobile hydraulic valves demand low leakage and since most mobilevalves are spool valves, they usually have large overlaps. In addition, to make the cost of thevalve acceptable to industry, the actuation stage on the spool is usually quite slow. This bination of large overlap and slow actuation makes it hard to implement many of the strategies that have been presented. Pressure control especially bees difficult when there is an overlap and a slow actuator. One example of a new strategy which is simple and robust is described as follows. Flow control is implemented on the inlet side and pressure control is implemented on the outlet flow control is based on the Bernoulli equation. Pressure control is done by a PI controller which maintains a low constant pressure to increase the efficiency and prevent work around large overlaps and slow actuation stage, the pressure controller only does meter out control. This means that if the controller wishes to raise the pressure, it can’t add flow to the cylinder, it can only decrease the opening of the meter out port. The benefit of this is that the only time that the spool has to cross the zero position is when the operator wishes to change the direction of motion of the cylinder. For the case where the load force and the velocity are in the same direction, this strategy has to be modified. In this case, the pressure reference of the pressure controller at the outlet is increased to a value which opposes the load force. The pressure reference is increased when it is noticed that the pressure of the inlet side is dropping. The pressure reference is also controlled by a PI controller. A schematic model of the controller system for the load lowering case is shown in Figure 4. At the time of writing this paper the initial experimental tests had been performed on the real crane shown in Figure 1 . Stability was not achieved because the crane is equipped with a load holding valve. However, the load holding valve will be replaced with a pilot operated check valve, which should make it possible to stabilize the system. In current systems, the load holding valve serves two functions, load holding and runaway load protection. Due to the use of a SMISMO valve setup, the runaway load protection is built into the control strategy,therefore the only function which is necessary for the load holding valve to perform is load holding. A pilot operated check valve will be able to do this, without adding plex dynamics which upset the stability of the system. Figure 4 Controller Strategy for Lowering of Load 5 CONCLUSION Even though not much experimental work has been finished, a good start has been made and initial tests have been promising. The outline of the thesis has been developed and organized in a logical manner. The work is split into five parts, requirements analysis, analysis of current systems, analysis of different topologies, development of a near future solution, and development of a mo