The ZPE is the source for the Van der Waals force. This force is seen as a force that pushes two very flat and smooth plates together when they approach each other at very close distances. It is what makes substances be a liquid or solid instead of a gas and supplies the bonding for colevalently bonded molecules To extract energy from the ZPE via the Van der Waals force is easy, simply allow particles to approach each other, and the ZPE will supply a small amount of energy just before contact. This can be as simple as letting steam condense. However to extract the energy again will require separating the particles, which transfers the energy previously obtained right back to the ZPE so the cycle becomes a zero sum game.
If a substance can be divided up without paying this Van der Waals penalty, yet allowed to reform in such a way that the ZPE will give up energy then this should result in an over unity system. One possibility would be a system in which two plates move along an orbit such that they slide apart, but swing around and approach each other with the faces flat on. However the maximum force occurs just before the plates touch, so even if they could be held apart with a few atoms of space between them, any air would cause viscous drag which would far exceed the energy obtained from the ZPE.
Instead lets look at a fluid. If we can create and maintain a bubble in the liquid that is a vacuum, then one could theorectically spin small bubbles off of this mother bubble, without paying the Van Der Waals price of formation. Once this is done, they could be allowed to collapse, generating heat from the collapse, which will exceed the amount of energy necessary to form them. So the problem may be solved if a method of creating a somewhat stable large mother bubble in a liquid can be found such that small bubbles can be bred off of it with less energy required then creating the small bubbles from homogeneous liquid. To do this would requires a gravitational gradient. Lacking that, then a gradient could be formed via acceleration. A vessel filled with a liquid can form such a stable vacuum bubble if spun at the proper speed. The center of the vessel will form a void, with the liquid being forced to the sides via centriphical force. However, this simple arrangement will not cause huge numbers of smaller bubbles to separate from the void.
Application
If we take a solid cylinder and drill shallow holes around it's outside, and fit this closely inside a hollow cylinder, and spin the inner cylinder, and have a fluid in the area between the cylinders, we can form stable voids in each of the shallow holes once the centrifugal force equals the pressure. Due to shear in the liquid, turbulence will be generated between the cylinders, which will also penetrate the shallow holes and be present along the boundary layer of the void and the liquid This turbulence is exactly what is needed to capture portions of the void and spin them off into the liquid. As the turbulence then carries the small bubbles out toward the outer cylinder, the pressure increases, and the bubbles collapse. When they collapse they gain added energy from the ZPE. This additional energy was not supplied by the ZPE during the formation of the daughter bubble, thus the ZPE ends up supplying energy that was not given up to it previously. The result should be an over unity device.
Can such a device be built, and will it actually give over unity performance? Actually it has. James Griggs' Hydrosonic Pump is already being sold to customers, sometimes providing them with over-unity energy. An energy efficiency consultant from Georgia, Griggs invented the pump as a result of his curiosity about a common phenomenon called water hammer or cavitation. Others have successfully built other over unity devices with similar designs. For more information on this phenomena see:
FUELLESS HEATER NO FUEL NO GAS NO WOOD NO GREEN HOUSE GASES
Hydrosonic Pump
HydroDynamics, Inc
FUELLESS CAVITATION HEATER HOMEMADE WATERPUMP MODIFICATION
With the theory of how such a device works, we can now move from the research to the engineering stage. To maximize COP requires maximizing turbulance, while minimizing viscosity and drag. Thinner fluids such as light hydrocarbons, and chlorinated hydrocarbons should give a higher COP than water. Also designs that minimize drag yet maximize turbulance should also increase the COP. Such design deviations as replace the holes with slots, either across the face of the inner cylinder, or in a "V" shape might prove more effective. Putting groves in the outter cylinder, or increasing the spacing between the inner and outer cylinder and placing ridges on it may also prove to be more effective. If a design could increase whirl pools such that they strip bubbles from the void, instead of chaotic turbulance, the amount of power needed to produce the energy giving bubbles may decrease increasing the COP as well.