New Page 1

This report describes the research carried out at Lancaster University and offers proposals for the next stage of research. (The proposed research will cost about £120 k and take about six months.) Enquiries from potential partners welcome. Please contact us.

Our original aim was to develop a new type of power generator that harnessed the low grade heat currently dumped into the environment at the end of the steam turbine cycle. However, what we ended up with was two “canned wind turbine” designs that used local environment heat or saturated water vapour as fuel.

Dry air turbine This is the simplest design. It exploits local environment heat or the release of latent heat on the outside walls of the turbine unit and has a maximum size for effective performance.

Moist air turbine This exploits the release of latent heat when steam condenses inside the turbine unit and can be built on a larger scale.

Lancaster University research demonstrated that both designs have a surprisingly high thermal efficiency.

It soon became obvious that many evaporation processes could be used as the basis for manufacturing LP Turbine fuel. This opens up the prospect of a new class of heat pumps for extracting energy from low grade heat sources.

What stage are we at?

The technical literature search and proof of concept experiments at Lancaster University verify that the LP Turbine theory is sound but there is still a question mark concerning viscous drag.

The key question to be answered

Is it possible to drive air through the LP Turbine blades at a sufficiently high speed to produce a useful output of electricity, given the low pressure gradients inherent to LP Turbine design?

The small 82 mm diameter turbine used for the proof of concept research offered far too much drag to deliver a net power output. But viscous drag falls as turbine size increases. We propose building a modified version of the Lancaster turbine having a diameter of about 1.0 metres. This will increase the area of the gap between each turbine blade by a factor of about x150.

They borrow characteristics from wind turbine design but operate inside lengths of conduit.

We use this term broadly to embrace pure water vapour and water vapour as a component of moist air.

Fuel LP Turbines can use the latent heat released when steam condenses as fuel.

Steam turbines convert the energy stored in hot, high pressure steam into shaft power. Cool low pressure steam emerges as a waste product.

It is difficult for steam turbines to achieve an efficiency greater than 50% because of the enormous amount of latent heat lost when steam condenses at the end of the turbine cycle. Each kilogram of waste steam squanders about 2.6x 10⁶ Joules of energy.

We produced a design for a new type of turbine that harnessed latent heat instead of wasting it. The UK Technology Strategy Board provided the funding for a small proof of concept turbine to be built and tested at Lancaster University.

The results were encouraging but severely limited by the small size of the turbine.

Figure 1. Two views of the bench top sized “canned wind turbine”. This produced results in line with our expectations, but it was far too small to deliver a net output of power. An open loop turbine was used for the proof of concept research. For our proposed future experiments, a closed loop system is required.

Figure 2. Latent Power Turbines are not pure heat engines. They can have a higher efficiency than predicted by the Carnot equation, without defying the laws of thermodynamics.

To help us break free from “Rankine cycle thinking” we looked for clues about how nature exploited the release of latent heat to produce powerful air movements.

Figure 3. The Foehn wind is a warm wind that occurs on the downwind side of a mountain range.

The warm wind in the rain shadow is mimicked inside a “canned wind turbine”.

Figure 4. Hurricanes can grow when they are fed by moist air having a temperature in excess of 26.5^oC. Nature tells us that high temperatures are not essential for latent heat to be used as fuel.

Figure 6. The essential features of our original “canned wind turbine” design.

[This was our second design but the discussion runs more smoothly if it is explained first.]

Our original intention was to develop a novel type of moist air turbine fuelled by the releases of latent heat. However, in the course of our university research we also carried out tests on a dry air turbine. This turbine displayed an interesting thermodynamic characteristic that has potential industrial applications. An additional patent application to protect the new intellectual property has been filed [8].

Based on our research, we make a number of predictions on how a future dry air closed loop system will behave.

Figure 7. Prediction 1: Overall, the conduit walls will warm up because work is done against friction.

Prediction 2: If the Energy output exceeds Energy input 1, the conduit walls will cool.

The dry air turbine has the potential to convert low grade heat from the environment into electricity.

(i) Increase the stagnation air pressure so that the density of the air increases and, for a given velocity, transfers more momentum to the turbine blades.

(ii) Increase the molecular weight of the fluid circulating round the system. The table below illustrates the benefits of using carbon dioxide for example.

HFO-1234yf is an air conditioning unit refrigerant developed to help automakers meet 2011 European regulations.

The University research suggests that dry air turbines will be more efficient than the Carnot efficiency equation for heat engines allows.

Figure 10. The dry air turbine acts as a hybrid engine. It loses thermal and bulk kinetic energy as the turbine shaft does external work.

The Carnot efficiency equation only takes into account the loss of thermal energy, so a hybrid engine can be more efficient than a Carnot engine without violating the laws of thermodynamics.

The experimental work described below in section 7.2 confirms our claim. To improve the accuracy of the temperature readings an open looped system was used so that the air only made a single transit of the turbine.

The Lancaster research was carried out using a very small (82 mm) diameter turbine, so the drag losses were high and the efficiency low.

A large array of dry air turbines could be used instead of a condenser and cooling towers for extracting the latent heat from steam turbine exhausts. Such arrays could be retrofitted to existing fossil fuel power stations, allowing their working lives to be extended in the face of increasingly stringent EU emission control legislation. Retrofitting could also avoid an energy famine following the anti-nuclear reaction to the Fukushima disaster.

The heat released by condensation acts as turbine fuel instead of being dumped into the environment. This type of power production will earn carbon credits instead of costing them.

Nuclear reactors could run cool for the limited purpose of producing low pressure steam. Compared with conventional reactors, cool nuclear would be very safe and produce less nuclear waste per unit of electricity.

This design would be useful in environments where there is also a demand for moist air cooling.

Combined micro-power and solar desalination plants are a feasible option. Solar radiation would be used to produce water vapour from brine and a dry air turbine would act as the heat sink for condensation.

Disadvantage compared with dry air: They are more complex because condensation occurs inside the conduit.

Advantages: (i) They increase in efficiency as turbine size increases; (ii) They have many more applications.

Figure 12. The steam condenses inside the turbine unit, so unlike the dry air turbine, efficiency should increase with size.

Here is a summary of our research findings for lagged open loop experiments. Full details are provided in Section 7.

The moist air turbine produced a larger power output for a smaller temperature drop across the turbine.

This can be explained if the moist air turbine acts as a pure heat engine in parallel with a latent heat release process.

Figure 13. This flow chart is a modified version of a standard Carnot engine diagram. It takes into account the heat liberating effects of condensation.

Q. Wet steam is known to cause pitting problems in steam turbines. Is the wet air impacting on moist air turbine blades likely to cause similar problems?

A. This is unlikely because moist air turbines will operate under different temperature and pressure conditions to steam turbines. The air will also hit the blades at lower, sub-sonic speeds. A better analogy is to compare moist air turbines with wind turbines.

Figure 14. The moist air turbine will be made from similar materials and operate at similar pressures to marine wind turbines. Marine turbines are expected to operate in corrosive salt water conditions and suffer impact from heavy rain and hailstones for many years. By analogy, we do not expect wet air pitting of the LPT blades to be a problem.

Moist air turbines use saturated steam as fuel. In theory, the fuel can be manufactured by evaporating water at any temperature above 0^oC. In reality, the lowest temperature for cost effective fuel production is probably about 20^oC.

Figure 15. The fuel will need to be injected into the turbine chamber at a pressure slightly higher than chamber pressure, but it can be manufactured at lower pressures and at a temperature of about 20^oC or higher.
Work has to be done on the fuel compressing it to injection pressure, but this is partly recovered when the steam mixes with the air and the steam temperature drops.

In order to convert low temperature heat into steam fuel at injection pressure three types of heat pump are required.

These convert heat of compression into steam fuel at the required injection pressure.

Figure 16. If the stagnation pressure inside the turbine unit is one atmosphere, the steam fuel will have to be injected at a slightly higher pressure. So the water inside the insulated jacket must boil just above 100^oC.

5.2 Type 2 heat pumps
These convert heat from fluids warmer than about 20^oC into steam fuel at the required injection pressure.

Figure 17. If the warm gas or vapour enters the first manufacturing chamber at a temperature lower than the fuel injection temperature the fuel produced will need to be compressed.

In principle, a single manufacturing chamber could be used to produce all of the fuel at 20^oC, but this would be inefficient because the compression pump would need to compress very large volumes of steam from very low pressures.

5.3 Type 3 heat pumps
These convert heat extracted during refrigeration process into steam fuel at the required injection pressure.

Figure 18. The Type 3 heat pump can manufacture fuel from any waste heat produced by a refrigeration process. The illustration shows a fluid being refrigerated.

The CO₂ leaves the top of the stripper at about 80^oC. It is mixed with saturated water vapour. We propose passing this mixture through a chain of LPT heat pumps.

Figure 19. LP Turbines could be used to convert the bulk of the heat released by condensing CO₂ and water vapour into electricity.

(ii) Improving the energy efficiency of the cryogenic carbon capture process

The cryogenic process can be used to capture CO₂, SO₂, and NO₂ plus any Hg vapour present. This process requires very higher compression ratios because (for high extraction efficiency) the partial pressure of the CO₂ should preferably reach its critical pressure.

Figure 20.The cryogenic method of flue gas capture requires higher compression ratios but offers greater scope for improving overall efficiency.

The compressed nitrogen can be liquefied by passing it through a further Type 3 heat pump and then stored.

During periods of peak power demand unwanted heat can be added to convert the nitrogen back into a highly pressurised gas for driving a secondary turbine. The required heat could come from a wide range of industrial processes including server centre cooling, food chilling, freeze desalination of sea water and the production of synthetic snow and ice for year round winter sports centres.

One option would be to use the liquefied nitrogen to cool the power transmission cables and then use the nitrogen to drive turbines at remote sub stations.

Figure 21. Ideally, the power cable would be superconducting, but this is not essential because heat losses from conventional cables would be employed usefully, evaporating the nitrogen.

The Lancaster University research database search encompassed academic journal and conference papers as far back as the early years of the last century. No work into the deliberate provocation of condensation inside turbines, with the intention of increasing efficiency, appears to have been done. These negative findings are confirmed by three patent literature searches carried out by the UK Intellectual Property Office, in response to patent applications filed by the present inventor, Courtney. [1, 2, 3.]

Prior to the Second World War, research was done into condensation inside turbines, but with the view to eliminating it as a nuisance. In 1937 Binnie et. al. [4, 5] investigated the steam pressure drop in convergent-divergent nozzles. They noted that under certain circumstances condensation occurred close to the throat of the constriction, with the release of latent heat and a sharp increase in pressure.

Figure 22. This graph from Binnie [4] shows an increase in pressure, as a result of condensation. A second paper by Binnie [5] describes an elegant electrode method for determining the onset of saturation.

The pressure spike is in line with our own theory [2]. It also supports our prediction that a correction term will be required for Bernoulli’s equation when a phase change takes place inside a Venturi throat.

Here is our argument: The LPT working fluid is compressible; nevertheless, changes in pressure can be estimated with a fair degree of accuracy using Bernoulli’s equation. This states that for an incompressible, non-viscous fluid undergoing steady flow, the pressure (p) plus the kinetic energy per unit volume (1/2x density, r x velocity, v squared) plus the potential energy per unit volume (density, r x acceleration due to gravity, g x height h) is constant at all points on a streamline.

If the working fluid includes a saturated vapour then we predict that Bernoulli’s equation breaks down when flowing through a nozzle, Venturi throat or constricting gap between turbine blades. Any tendency to cool on passing through the Venturi throat will result in the production of small condensation droplets and the release of latent heat. Consequently, the temperature and pressure drops will be reduced compared with the flow of dry fluid. On passing through the flared section, the latent heat processes are reversed, with heat being absorbed as the water droplets evaporate. The rate of mass flow remains constant through all sections of the conduit perpendicular to the streamlines. In the case of a saturated vapour, the rate of volume flow drops as a consequence of condensation in the constriction, then increases as a consequence of evaporation as the gap widens. For the process to be reversible, the condensation droplets must continue to move forward as an aerosol and not be centrifuged out by the turbine blades.

In order to produce an equation that allows for the release of latent heat an additional term
dQ_l/dV needs to be added. The term dQ_L/dV represents the latent heat lost/gained per unit volume of static fluid.

Volume is used as part of the correction term, to ensure dimensional consistency.

When condensation occurs and latent heat is liberated, the minus sign is retained in front of the latent heat term. A positive sign is used if evaporation occurs and latent heat is absorbed.

The literature search did not reveal any references to a Bernoulli’s equation correction term for latent heat, but discovered several references to a correction term for sensible heat changes, for example, Segletes and Walters [6]. We corresponded with the authors of this paper, who considered our arguments to be valid. They went further, by offering suggestions on how our correction was consistent with Van der Waals’ equation relating to inter-molecular forces.

Latent Power Turbines are designed to maximise the release of latent heat inside the turbine but his is not always desirable. For example, Roumeliotis and Mathioudakis [7] have reported on the problems caused by the release of latent heat inside air cooling turbines when the air enters the turbine under saturation conditions. The turbines strip out part of the water vapour, but the thermal rejuvenation caused by the release of latent heat means that the exit temperature of the air is too warm for personal comfort. What Roumeliotis and Mathioudakis saw as a nuisance, we now recognise as a merit.

We corresponded with Roumeliotis and Mathioudakis, who expressed interest in participating in any future European Latent Power Turbine project.

Figure 23. The open loop design made it easier to take temperature measurements when testing our dry turbine theory.

Note: Atmospheric air (which naturally includes some water vapour, but is unsaturated) is referred to as “Dry air”. Air which includes just sufficient added steam to produce a trace of condensation inside the glass wall of the flow meter is referred to as “Dew point air”.

To ensure that the dry air was well clear of its dew point a small amount of air from a hot air gun was added. The consequent warming of the rig imposed a small temperature drift during the course of the experiments.

The turbine was first spun with the generator on open circuit and then a resistive load (car sidelight bulb) added.

Figure 25, Dry air experiment. The temperature drop across the turbine increases by about 0.3 K when the resistive load is added.

Figure 26, Dew point air experiment. The temperature drop across the turbine increases by about 0.06 K when the resistive load is added.

Figure 27, Combined results from dry and dew point air experiments. The temperature drops across the turbine are far lower for dew point air because latent heat has been released.

The turbine superficially appears to be more efficient than a Carnot engine. But this is only an illusion, because the parallel sided turbine system is not a pure heat engine.

Figure 28. The turbine is not a pure heat engine: it acts as a heat engine in parallel with a device that converts translational bulk kinetic energy into external work.

At a distance from the turbine blades the air temperature varies in a complex manner. This is discussed in section 8.2.

Figure 29. When dew point air passes through the turbine, the bulk of the additional efficiency comes from the liberation of latent heat.

The turbine acts as a sensible heat engine in parallel with a device that converts latent heat into external work.

Figure 30. This chart summarises the illusion that a latent power turbine can be more efficient than a Carnot engine.

The purpose of these experiments was to check for differences in the behaviour of dry and dew point air as the air moved through the constricting gap between the turbine blades, but without the complications of energy conversion into electricity.

Figure 31. The turbine includes two levels of constriction. The Venturi analogy simplified this to a single constriction.

For this experiment a Venturi constriction was made from polyurethane foam. The pressure drop between the mouth and throat was measured in a series of experiments, in which the throat was gradually widened from 12 mm to 22.5 mm. The mouth of the Venturi tube had the same diameter as the LP Turbine conduit.

Figure 32. The diameter of the throttle was increased for successive experiments.

Figure 33. The assumed benefit of constricting air flow was that it would speed up the air as it impacted on the turbine blades. This experiment suggests that the gaps between the turbine blades was so small that the point of maximum returns had been exceeded and that viscous drag was slowing the air down more effectively than the narrow gap was increasing it.

We concluded that for future experiments the diameter of the turbine would need to be increased by about an order of magnitude, to ensure that viscous drag did not dominate the outcomes.

8.2 Differences in the behaviour of dry and dew point air passing through a small gap at speed.

Our original plan had been to look for deviations from the Bernoulli equation when moist air passed through a constriction. The first Venturi results indicated that viscous drag would swamp the effect we were looking for, so we revised our predictions to those summarised in the following diagram.

Figure 34. Our revised predictions on how the constriction system would behave under significant drag conditions.

The temperature at three points along the length of the Venturi tube was measured when air from the turbine blower was passed through it.

Figure 35. The results demonstrate that the presence of saturated water vapour significantly modifies the behaviour of air as it passes through a constriction or nozzle.

To simplify the Mk 1 design, operation at a stagnation pressure of one atmosphere is suggested. The prototype will be suitable for dry and moist air experiments.

9.1 Estimated power output from a moist air turbine inside a 1 metre diameter conduit

(i) Dry air with a partial pressure just less than one atmosphere is used as a carrier fluid.

(iii) At the turbine mouth, the mixture has a total stagnation pressure of one atmosphere and the air is saturated at 40^oC.

(iv) The values for pressures and densities are extrapolated from the tables in the Appendix below. The table data is limited, so the calculations are only first approximations.

Saturated air at 40^oC has (approximately) 1.83 x density of saturated steam at 100^oC.

Figure 36. The mass rate of flow through a 1 metre diameter conduit = 1.1 x p 1²/4 x v

Constriction ratio =    Cross section area of conduit
                                Area “seen” by fluid as it passes
                                  through the turbine blades

(i) The assumption that the Power output = 50% of KE at blades is an educated guess based on Betz's law. This states that wind turbines have a maximum kinetic energy conversion efficiency of 59.3%. Betz theory does not take into account the heat liberating benefits of condensation. So, moist air turbines may be capable of operating at a higher efficiency.

(ii) Until we have contradicting evidence, the maximum velocity should be kept below the speed of sound (330 m/s for dry air.) If this limit is exceeded the gaps between the turbine blades may act as de Laval nozzles with the air expanding away from saturation conditions.

(iii) In reality, air friction (drag) may limit the maximum working velocity well below this.

One research aim should be to maximise the rate of air flow through the widest pat of the conduit. The following extract from the above table illustrates the benefits of doubling the mean air speed.

Figure 37. “Blowback” caused by drag may limit the maximum working velocity.

The maximum velocity can be increased by
(i) scaling up the turbine unit, to reduce the proportionate effect of drag,
(ii) using a more efficient circulation pump to overcome friction. This option is elaborated below.

It would be helpful to know the power rating where blowback influences performance before the generator and steam supply are chosen.

This can be done by looking for deviations from Bernoulli’s equation, when running the unloaded system using dry air.

Bernoulli’s equation states that for an incompressible, non-viscous fluid undergoing steady flow, the pressure (p) plus the kinetic energy per unit volume (1/2x density, r x velocity vsquared) plus the potential energy per unit volume (density ,r x acceleration due to gravity, g x height h) is constant at all points on a streamline.

If there is no blowback the velocity, v will increase linearly with the rate of rotation of the circulation pump.

At 40^oC, steam density = 7% of atmospheric pressure and the saturated steam density is 0.0535 kg/m³.

For this calculation, we consider one cubic metre of saturated air because this will allow us to work out the drop in vapour density.

15.8 kW. = mass of vapour condensed/ m³/ second x latent heat of vaporisation (2 406 kJ/kg)

Power output = mass of air/second x specific heat capacity of air (1.005 kJ/kg.K) x temperature drop

This cooling will increase the density of the air but it will also reduce its velocity.
The kinetic energy = ½ x density x (velocity)²

Consequently, for a lagged prototype, the air will have a lower kinetic energy when it re-circulates and enters the turbine for the second time.

Figure 39. Our patents are not restricted to specific types of turbines or circulation pumps. We welcome discussion and recommendations on the details of LP Turbine design.

For a given total power output, the constriction ratio can be reduced by using two sets of counter-rotating turbine blades. This offsets the increase in viscous drag caused by using two sets of turbine blades.

Figure 40. Jet circulation pumps simplify the design for multiple turbine units. Several compression pumps working out of phase can be used to smooth the jet pump action.

1 Courtney, W. A., Patent GB 2427249 “Combined power generator and water desalination plant”.

2 Courtney, W. A., Patent Application GB 2459326 A “Saturated vapour turbine system”

3 Courtney, W. A., Patent Application GB 1011794.3 “Phase change turbine incorporating carrier fluid”
This application was filed 14 July 2010. It is being taken through the PCT system.

4 A. M. Binnie and M. J. Woods. The pressure distribution in a convergent-divergent steam nozzle. Proc. Inst. Mech. Engrs. 138 (1937(, pp9-266.

5 A. M. Binnie and J. R. Green, An electrical detector of condensation in high speed steam. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, Vol. 181, No. 985 (December 31, 1942), pp. 134-154.

7 Roumeliotis and Mathioudakis, Analysis of moisture condensation during air expansion in turbines, International Journal of Refrigeration, 29 (2006) pp. 1092- 1099.

8 Courtney, W. A., Patent Application GB 1116309.4 “Hybrid heat engine” Filing date 22 September 2011.

	Molecular weight	Relative density at same temperature and pressure.
steam	18	1.0
Air	28.84	1.6
Carbon dioxide	44	2.4
HFO-1234yf	114	6.3

Conduit velocity, v (m/s)	Rate of mass flow (kg/s)	Constriction ratio	Velocity on entering blade gap (m/s)	Kinetic energy/s on impact with blades (=½ x rate of mass flow x velocity²)(kW))⁽	Power output *Assuming this = 50% of KE at blades* (kW)
6	5.16	5	30	2.32	1.16
6	5.16	10	60	9.28	4.64
12	10.32	5	60	18.58	9.23
12	10.32	10	120	74.32	37.16
18	15.48	10	180	250.7	125.4
24	20.64	10	240	594.4	297.2