**Internal Model of the EFCV**

_{1,d}to the desired spool displacement xv

_{1,d}. The error between the desired and the actual displacement is imposed on the PI controller, whose output will drive the current into the pilot valve voice coil. The controller drives the displacement error to zero. If the actual flow follows the equation above perfectly, then the flow error goes to zero as well, or Q

_{1,c}= Q

_{1,d}.

In the flow controller design, the above equation was used to approximate the actual flow rate. The equation is a widely acceptable quasi-steady-state model for the flow rate across an orifice. In the following section, an experimental study will be discussed that investigated the accuracy of the calculation.

A prototype of the EFCV was built and tested. For the hydraulic test setup, the valve was set on a fixed displacement test stand. An adjustable relief valve simulates a load on the valve. A flow meter in series with the “load” relief valve is used to measure the actual flow from that service. The type of flow meter used depends on the demanded flow in order to improve the measurement accuracy. For flows below 7 GPM, a 0.95 in^{3}/rev meter motor was used; for flows above 7 GPM, a flow turbine was used.

Given a flow command from the supervisory controller, the “load” relief valve is adjusted so that the pressure drop between P_{1} and P_{s} is equal to some predefined value. Then the actual flow rate was measured.

**Application of the EFCV: Self-Sensing Cylinder**

An application for the EFCV is a selfsensing cylinder, which will estimate the piston displacement of a regular hydraulic cylinder by taking advantage of integrated sensors and flow rate calculation of the EFCV.

In the cylinder, two EFCVs connect each chamber, respectively. The sensors in the EFCVs measure the pressures, P_{1} and P_{2}. The flow rates, Q_{1} and Q_{2}, are calculated by using the analytical model with the experimental calibration. In addition, in order to eliminate the integration error, some deterministic displacement information is required (an absolute start position). For instance, a latch sensor could be installed so that the output voltage is high when the piston displacement x_{p} = 0 and is low as x_{p} ≠ 0.

Due to the combined effect of the flow rate and the load applied to the cylinder, the actual displacement is the sinusoidal signal with the higher frequency oscillation. At t=0.1 [sec], the latch sensor is enabled and then enforcing the error to be zero. Without the latch sensor, the observer will still give the similar displacement profile but with an offset error. Integrated pressure sensors and the experimentally calibrated flow rates in the EFCV, together with the proposed observer, make it possible to implement the self-sensing cylinder.

**Conclusion**

Due to the embedded sensors and microcontroller inside the valve, the flow rate can be controlled through the power elements without the need to know the load or the displacement condition from the power elements. The flow controller utilizes the well-known quasi-steady flow rate equation to approximate the actual flow rate in the internal closed-loop system.

Experimental studies show that the one equation model with constant parameters is not accurate enough to cover all conditions. In particular, for the low-flow-rate, high-pressure-drop case, the flow error calculation is significant. An experiment-based calibration method is then presented.

*This article was written by QingHui Yuan, Chris Schottler, and Jae Y. Lew of Eaton Corporation, Eden Prairie, MN. For more information, visit http://info.hotims.com/ 34459-320.*