For machine builders, accelerating the process of product development can save time and money by enabling a design to be verified for maximum performance. A sure-fire way to get the job done more smoothly and quickly in these situations is to make use of modeling and simulation.

Figure 1. Hawkeye Pipe concrete pouring machine was built using operation simulation software.

Machine designers with many years of experience can rely on that experience to guide future designs. Engineers without such experience — or possibly those who are developing something new — can gain experience quickly and economically by modeling the systems and then testing the model by running simulations while changing system parameters. Simulating can be done at many levels, but three levels of simulation will be covered in this article, including how each level applies to hydraulic applications.

Design modeling and simulation
The first level is design modeling and simulation. Most designers have a pretty good idea of the components required for the application at the start of a project, but getting to optimal component sizes is often not easy. At this stage, trial and error can be expensive — and even dangerous — if experimentation is done with real hardware or without a clear idea of how the choices might impact the motion of the machine. It’s better and safer to simulate the machine’s operation before making final component selections.

The designer starts by making a model of the system. Trying several combinations of components and component sizes may be necessary before settling on a best practical design. Spreadsheets, such as Excel, often are used. Once the appropriate equations describing how system elements interact are programmed into the spreadsheet, the designer can easily plug in different numbers for pump, accumulator, valve, and cylinder sizes and get an idea of what will happen during steady-state conditions.

Understanding steady-state motion is a good start, but designers of modern applications are looking for more realistic models of machines that also consider the dynamics of their machines’ motion. Doing so can increase production and accuracy while reducing wear and tear on the system. Systems are always accelerating and decelerating while quickly moving from point to point. Very little time is spent at a constant velocity or steadystate condition.

The VCCM equation presented in “Basic Electronics for Hydraulic Motion Control,” by Jack L. Johnson, is a useful equation. However, it only calculates steady-state velocity and does not consider changes in velocity (accelerations) as a function of time. To predict accelerations, you must also predict forces, pressures, and flows as a function of time.

Force divided by the mass equals acceleration, and integrating acceleration over time yields velocity and position. These dynamics are implemented in models using first-order linear and nonlinear differential equations. You can verify that peak accelerations and velocities can be achieved and the system is controllable by plugging in real values.

Some simulation software packages include libraries with basic designs that can help the user make an educated estimate for parameter values. Often, estimations are the best parameters you have to work with because hydraulic component manufacturers typically don’t provide this information. Many iterations or simulations of the model must be run to cover the anticipated ranges of each parameter. The simulations will usually provide information about which component parameters are most critical.

For easier-to-use alternatives to spreadsheet models, look to engineering software that allows you to simulate the performance of a hydraulic system. These include products such as Automation Studio, Scilab, Matlab, and 20Sim. The simulation software works by breaking down time into very small increments and computing the change in flows, forces, pressures, velocity, and positions. These small changes (called numerical integration in mathematical terms) then are added to the previous values to get new current values. Because the rates are always changing, better accuracy is achieved by updating the rates more frequently. There may be as many as 10,000 to 100,000 small time periods per second.

A reasonably accurate model should include:

• a control signal to the valve,
• how the valve responds to the control signal,
• the flow through the valve as a function of the spool position,
• the pressure drop across the valve,
• how the pressure changes on each side of the piston as a function of position, velocity, and flow
• pump flow response to pressure changes, and
• accumulator pressure as a function of changes in oil volume.

The resulting simulations generally produce a reasonably accurate estimate of how the flows, forces, positions, velocities, and accelerations will change as a function of time. The limiting factor is the quality of the model and lack of availability of information from the manufacturers.

Operational simulation
The second level is operational simulation. Operational simulation allows those individuals who will be controlling or interfacing with the hydraulic system to get an idea of how the system will perform. This allows them to write and test their control system or operator interface early in a project cycle and in parallel with the hydraulic and mechanical system design. Using an operational simulator, the control engineers and human-machine interface designers can test the system at their desk, making them better prepared for when the physical system is finally available.

An example of an operational simulator is built into Delta Computer Systems’ RMCTools motion control software package. Information describing how the system will be constructed is entered into the software. The simulated motion of the system then can be plotted by the software in the same manner that real motion of the system may be plotted after it is built.

Example: Results obtained before hardware is built
Hawkeye Pipe, Mediapolis, Iowa, manufactures machines that pour segments of concrete pipe, Figure 1, and has used the simulation feature of Delta Computer Systems’ motion controllers in the development of the machines. Because complex mathematics are needed to guide a movable chute that pours concrete, it was important to prove out the numerical methods to be used prior to the machine’s construction.

Hawkeye engineer Ben Schmidgall, who did the programming, used Delta’s built-in simulator to generate sample points for the moves and graph out the target and actual positions of the motion arm. He began by tracking simple shapes and added features. To track the motion around a square, Schmidgall started with a consistent linear speed. Then he slowed down the speed of motion near the corners of the box, still just using the simulator.

Figure 2. Plot of the motion of the Hawkeye concrete pipe feeder.

Although he had never done any complex motion control programming before, Schmidgall was able to simulate all the motion of the system while sitting at his desk. Then, once he got the machine and motion controller hooked up on the manufacturing floor, it only needed to be tuned for optimal performance.

Using a prior mot ion cont rol system (before upgrading to use the Delta controller), Schmidgall had no way of monitoring the motion, so there was no way to tell if the system was tuned precisely. Now, not only is he able to know when the system is well tuned, he can show the system graphs to other people to show his progress with the project. A sample screen capture motion plot is shown in Figure 2.

Physical simulation
The third level of simulation is physical simulation. This is where a test system is constructed to simulate real-world conditions and apply them to a machine or production device to test it. “Closed-loop control simplifies cyclical testing,” in the April 2008 issue of Hydraulics & Pneumatics, discusses such applications. A physical simulator can subject an item to the same stresses in hours or days that the product would normally encounter in a lifetime of operation. Testing using a simulation system can be destructive or non-destructive. Physical simulators used in training are less common, but obvious examples are flight or submarine simulators.

Example: Simulating forces on automotive suspensions

Figure 3. ADI’s KD-Rig automotive chassis test system.

Accelerating Developments International Inc. (ADI), Concord, N.C., produces a system called the KDRig, Figure 3. The KD-Rig applies force along different axes to test the performance of race car chassis under conditions that simulate the stresses that actually occur on the race track. Racing teams need the stationary KD-Rig because without it, they would be faced with the extremely difficult task of having to measure the forces being applied to the chassis while a vehicle is moving. Plus, available track time is limited. With the KD-Rig, racing teams can test and tune their cars’ chassis and suspension responsiveness off the track, enabling them to optimize the cars’ performance more precisely and gain a competitive advantage.

It so happens that this case is an example of operational simulation and physical simulation. Operational simulation was used to develop the system, and physical simulation is the role that the finished system plays — simulating race conditions. During the development of the chassis test rig, it was important to understand how the system would work before building the hardware.

The company conducted 3-D modeling of the KD-Rig structure to full scale with a car on it and ran a simulation before ever having one part machined, or fabricated. ADI’s programmer, Mike Messick, used Delta’s motion controller simulation tool many times during the development of the motion programs. The ability to simulate the motion remotely came in handy because Messick typically works at home in his lab doing difficult programming and operational simulations because it is much quieter there than at ADI’s Research and Development Center. After the operational simulations were done to develop the chassis testing machine, the physical simulations using the machine began.

In recognition of the significance of ADI’s technology to the racing industry, the KD-Rig received the Testing Technology of the Year Award for 2007 at the Professional Motorsport Expo in Cologne, Germany, in December, 2007.

Don’t depend too much on models
Although simulation is an excellent tool, don’t depend too much on a model to indicate if a design will work. There are many things that are usually left un-modeled that can make a big difference in the actual system. Even small items like bends in the pipes can have a big effect on the model and therefore the simulation. Use of a model is probably a better way to prove that a particular design will not work rather than the other way around. At the very least, the model will show the design’s weak points and many simulations can be done using the extremes of component parameters to make sure the design is robust.

Don’t get discouraged or abandon a model if it isn’t perfect. Usually any model is better than none. This is especially true for the control engineers that can use the model in their simulations. Models are very likely to improve over time and as the understanding of applications become more defined.

For more information, contact the author at peter@deltamotion.com or visit www.deltamotion.com.