The normal road construction process involves subgrade (the existing soil) preparation followed by placement of an aggregate layer, and then the pavement surface layer (asphalt or concrete). In practice, most highways are designed using a set of assumptions that is sufficiently conservative to ensure that the roadway will meet its minimum specifications over its design life. The cost of the pavement layer can be optimized by testing the firmness of the prepared roadbed before paving. The testing process validates the basic design assumptions, and enables the amount of pavement to be tailored or optimized for each segment of the road. Ingios Geotechnics (Northfield, MN) was formed to develop and manufacture equipment to automate the process.

Figure 1. The Ingios Geotechnics roadbed test system is trailer-mounted for portability.

The Opportunity

Historically, the testing of highway roadbeds has been done by applying a static force (i.e., weight) to the compacted foundation for about an hour, and measuring the response (i.e., measuring the compressibility, or modulus, of the substrate). “We’ve discovered that the traditional approach doesn’t adequately simulate the dynamic loading from vehicles,” said David J. White, PhD, P.E., and President and Chief Engineer of Ingios. “A better solution is to impart dynamic loading of the roadbed using controlled load pulses. If we can verify that the target modulus values have been met and the variability is low, we can reduce the pavement cost by millions of dollars for a large project.”

To deliver the desired loading conditions to simulate vehicles, Ingios designed and patented a test trailer that’s pulled behind a tow vehicle (Figure 1). The trailer has a hydraulic power unit that operates two main hydraulic cylinders: a high-force cylinder and a low-force cylinder. Only one cylinder is used at a time to press a steel plate (Figure 2) onto the ground under a controlled force, while a data acquisition system measures the displacement of the ground. Since the impact pulses must be tightly controlled, the Ingios engineers needed an electro-hydraulic motion controller capable of controlling force with a high degree of precision. To help with the design of the hydraulics, Ingios turned to system integrator and distributor JM Grimstad (Muskego, WI).

Selecting the Motion Controller

Figure 2. The business end of the high-force cylinder in the test system. The black cable coming from the load cell is visible, as are the gray cables from the position sensors.

Grimstad combined their hydraulics and control systems experience to select the RMC75E (Figure 3) two-axis electro-hydraulic motion controller from Delta Computer Systems (Battle Ground, WA) to control the two hydraulic cylinders. The controller has the capability to synchronize axes and precisely control pressure/force being applied, as well as meeting any precise axis position/velocity/acceleration requirements.

The Ingios equipment is set up to support two testing procedures: a plate test and a cone penetration test. The plate test is designed to apply a force pulse to the ground that goes from zero to 12,000 pounds, and back to zero in 0.15 second, with a force accuracy of about 1 percent. The cone penetration test involves pushing a rod with a cone tip into the ground at a constant velocity and measuring the displacement. To support these operations, the motion controller is programmed for both position/velocity and force control of the large cylinder, and force control only of the smaller cylinder. Force feedback is provided by load cells mounted on the cylinder rods, and position/velocity feedback comes from an SSI (synchronous serial protocol) probe mounted inside the large cylinder. The system also uses lasers to achieve ultra-precise measurements of the plate positioning.

Figure 3. The motion controller is mounted in the system’s panel enclosure.

“Our main goal was to produce 0.15- second load pulses consistently on the load profiles,” said Ryan Behnke, a Grimstad engineer. To accomplish this, Behnke came up with an advanced control concept that he knew could be implemented within the motion controller. The solution is to employ “gain scheduling” — changing motion control closed-loop gains on-the-fly, based on where the load pulse is in the cycle. He programmed the controller to switch among three different gain settings during each force pulse. “We needed the three gain sets to get better control of the system,” said Behnke. “When you push down, the ground is compressing and requires a different gain set to accurately control the quick decrease in force. A third gain set is needed to control the holding of force during the dwell period between pulses. We first tried to use just one gain, but weren’t getting the control that we wanted.”

How the test proceeds depends on temperature, and gain scheduling also allows the operator to change the gain sets, depending on actual field conditions; for example, spongy ground requires more aggressive gains. Ingios designed the apparatus so the test operator can change the gain sets sent to the motion controller by making changes on the HMI screen.

Tuning the gain sets for optimal performance was a challenge in order to compensate for the mechanical, hydraulic, and load conditions posed by the trailer-based physical system. According to Behnke, the key is to minimize trailer movement and control the temperature of the oil in the system during tests. The more the trailer moves during a test, the harder it is to control force because the trailer also induces a force by its movement. High variations in temperature affect the viscosity of the oil. Thicker oil requires more gain while thinner oil requires less gain. Programming the motion controller was done using Delta Computer Systems’ RMCTools software that includes graphical tools to simplify the tuning of control loop gains.

Results

“The speed and crispness of the pulses that the system produces are amazing,” said White. “And the ease of setup and using the plot tool helps tremendously with the setup and tuning process,” said Behnke. In addition to being used to control the force being applied, feedback from the load cell is supplied to a data acquisition system that runs on a PC. “We developed our own software to do the analysis of the collected data,” said White.

“Our system has been in use for about a year now, and we’re reinvigorating the industry that we work in, with worldwide use implications,” continued White. “We’ve pushed the hardware control envelope, but we want to go even further in applying dynamic loading in the field.”

This article was written by by Bill Savela, P.E., Delta Computer Systems. For more information, Click Here .