HHO and Diesel Technology

by Charles Ware

December 8, 2014

�lename: wp12.pdf. Previous version of this paperis now wp12a.pdf

Abstract

Background: HHO is a mixture of hydrogen andoxygen gas produced by water electrolysis that is pur-ported to increase e�ciency of internal combustionengines when fed into the air intake. Available dataindicates that this is true. This report attempts todetermine whether HHO gas injection can serve as acost-e�ective method for reducing vehicle fuel costs.Methods: The results of eight investigations are

analyzed. In all studies, the method of determiningthe e�ect of HHO on fuel consumption is roughlythe same. Fuel consumption of a Diesel engine ona dynamometer test stand is measured at di�erentHHO gas feed rates. In this way, tests are performedunder a relatively controlled set of conditions.Results: The additional amount of energy pro-

duced per mass units of hydrogen gas was calculatedat various speeds and loads. This determination,called a yield value, was useful for comparing resultsagainst various thresholds. A yield of approximately9 megajoules per gram of H2 is needed for cost e�ec-tive reduction of vehicle fuel costs. Only three of theeight investigations met this requirement.Conclusions: A restricted HHO feed rate was one

factor common to the three investigations with higheryield values. A wet cell reactor of similar design wasalso used. The issue of whether there might be somevariability in the composition of the HHO gas itselfis also addressed.Practical Implications: Yield values for some

of these studies probably do not exceed an economic

break even point. Thus, they should not be used toindicate that HHO injection can serve as a cost ef-fective technology for reduction of vehicle fuel costs.Also, the studies that do indicate a cost e�ective fuelcost reduction are not found in peer reviewed litera-ture. The validation and development of HHO tech-nology requires much more data of better quality. Italso requires better evaluation of this data.

Key Words: Brown's gas, oxyhydrogen, HHO,diesel, fuel e�ciency, fuel consumption

Background.

Because of its energy e�ciency and widespread fuelavailability, Diesel technology serves as the primarypower source used for surface transportation of bulk,freight and containerized cargo. Therefore, the sur-face network of the global logistics system is pow-ered largely by Diesel technology. It also powers themajority of industrial construction, agricultural andmining equipment.

HHO gas injection is a a largely undeveloped tech-nology that could very possibly be used to increasee�ciency of Diesel technology resulting in billionsof dollars in reduced fuel costs. In this study, weshall evaluate available data obtained from labora-tory tests of the e�ect of HHO on Diesel engine ef-�ciency to determine how e�ective HHO injectionmight be for reducing vehicle fuel costs. The appli-cation used to calculate a performance threshold forcost e�ectiveness shall be a Class 8 truck or �semi�traveling over 100,000 miles per year.

1

2

Methods.

Brake-Speci�c Fuel Consumption.

A laboratory evaluation of the e�ect of HHO gas in-jection on fuel consumption requires a very speci�cde�nition of fuel consumption. Brake speci�c fuelconsumption (BSFC) is expressed in units of gramsper kilowatt hour (g/kW hr) or in the US, poundsper horsepower hour (lbs / hp hr). A typical BSFCvalue for a Diesel engine is 0.35 lbs / hp hr. [1] Asthe name implies, BSFC is speci�c to engine load.The engine is set on a dynamometer test stand thatapplies a load to the engine and regulates the speed.Engine load in horsepower is given by the equation:

RPM × torque

5252.113= brake horsepower (1)

where RPM is engine speed in revolutions perminute and torque is output torque of the engine mea-sured in foot pounds. The exact value for the denom-inator of the horsepower equation is 220 × 150 ÷ 2πwhich comes from the observation of James Watt thata mine pony harnessed to a winch could haul 150 lbs.up a mine shaft at 220 feet per minute.The quantity of fuel is given as weight as opposed

to volume. The amount of fuel in a unit of volumevaries with density which is a�ected by temperature.Also, weight can generally be measured more pre-cisely. A common procedure is to measure the weightof the fuel container before and after a timed test runand calculate the di�erence.

Evaluations.

The evaluations analyzed were published either inpeer-review literature or on-line.1 All analysis wasperformed using Open O�ce 3.3 spread sheet func-tions. These spread sheets can be found on-line athho-research.org . In some cases, data was givenin tabular form and could be copy/pasted into the

1In the case of the on-line studies, both Univ. of North-

west Ohio and Fox Valley Technical College were contacted by

email and they did con�rm that such evaluations were actually

performed. The Purdue study was a student project and could

not be con�rmed because of privacy policies.

spread sheet. In other cases, data was presented asa graph, in which case, pixel positions of the datapoints were used to calculate the value of the datapoint. Modi�ed images can be found on-line at hho-research.org .

The types of engine and test conditions for eachstudy are listed as follows:

• Yilmaz [2, 10] Engine: Four-stroke, four cylin-der, direct injection with glow plug, 3.57 L, HHOreactor: specially made parallel plate type. Flowrate: 4 lpm.

• Milen and Kiril Engine: [3] Single cylinder, 98mm bore, 130 mm stroke, direct injection. HHOreactor: not speci�ed. Flow rate: 4 lpm.

• Purdue University [4] Engine: 4.5 L John Deeretractor engine. HHO reactor: Commerciallyavailable SS-20. Flow rate: Approx. 2 lpm.

• Fox Valley Technical College [5] Engine: Cater-pillar C15 turbocharged truck engine. HHO re-actor: Commercially available SS-40. Flow rate:approx. 4 lpm. Estimate based on measuredreactor current and reactor speci�cations.

• Univ. of Northwest Ohio [6] Engine: 2004 De-troit Diesel 14 L Series 60 truck engine, 515 hpcapacity. HHO reactor; Commercially availableSS-40. Flow rate: Approx. 4 lpm. Estimatebased on 32 amp and 46 amp supplied by a cur-rent regulator and reactor speci�cations.

• Bari and Esmaeil [7] Engine: Direct injection, 4cylinder, 4.009 L. HHO reactor: Epoch EP-500.Flow rate: 10-60 lpm.

• Wang, et. al. [8] Engine: Cummins B5.9-160,six-cylinder four strokes, direct injection, 5.88L HHO reactor: Epoch 560A. Flow rate: 10-60lpm.

• Birtas, et.al., [9] Direct injection, 4 cylinder inline. 3.76 L tractor engine. HHO reactor: notspeci�ed. Flow rate: 0.111 to 0.153 kg/hr.

3

Flow Rate Measurement.

Measurement of HHO gas �ow rate is relevant to thisanalysis. A mass �ow measurement is preferred sincea volume measurement is a�ected by pressure andtemperature. Mass �ow measurement was used inonly one of the investigations. Grams of hydrogenwere estimated for the other studies assuming 760mm Hg pressure and 298 degrees Kelvin temperatureusing the ideal gas equation and assuming an averagemolecular weight of 12 amu. Although this is an ap-proximation, any error introduced would be compar-atively small and could not account for the disparityof results obtained by this analysis. In one investiga-tion [11] average molecular weight of HHO was foundto be 12.3 amu. The study states that the same valuewas obtained for 2 di�erent samples at the Adsorp-tion Research Laboratory, Dublin Ohio on June 30,2003 although it does not state the method used tomake the measurements. In future investigations, aprecise measurement of the rate at which water isused by the HHO reactor as well as measurements ofrelatively humidity, pressure, temperature and volu-metric �ow rate might be used to estimate mass �owand the average molecular weight of the gas.

Calculations.

The equation used for the energy yield value whenthe dynamometer test stand had a regulated load ca-pability is given by:

E

mg[1− r] = energy yield (2)

where E is total energy produced in a unit timeperiod in mega joules (MJ), mg is grams of hydrogen(H2) injected during the same time period and r is adimensionless ratio equivalent to m1

m2where m1 is the

mass �ow rate of fuel with gas injection and m2is themass �ow rate of fuel without HHO gas injection.In some cases, the dynamometer load was ad-

justable and measured but it did not have a feedbackloop to regulate it. In that case, yield value is givenas:

Eavg

mg[1− k r] = energy yield (3)

Study Average Max.Bari [7] 0.081 0.257Birtas [9] 0.034 0.042Milen [3] 0.149 0.233FVTC [4] 6.545 13.263Purdue [5] 2.639 12.415UNOH [6] 6.529 19.797Wang [8] 0.034 0.044Yilmaz [10, 2] 0.541 1.054

Table 1: Energy yields (MJ/grams H2)

where k is a dimensionless correction ratio equiva-lent to E2

E1where E1 is energy output with gas injec-

tion, E2

is energy output without gas injection and Eavg isthe average of E1 and E2 at a given engine speed.This is generally acceptable if E1 and E2 are closeenough together. In all cases, engine speed is regu-lated by an automatic feedback loop. These Eqns. 2and 3 are derived in Appendix A.

Results.

Table 1 gives average and maximum energy yieldvalue estimates for the various evaluations. Eqn. 3was used to estimate yields for the Milen, FVTC, andYilmaz studies. Eqn. 2 was used for the rest. )

There is an implied protocol here that the testwith and without HHO injection should be run un-der the same conditions, namely engine speed andload. If the test conditions are substantially di�er-ent, then fuel consumption measurements are mean-ingless since fuel consumption is a�ected mostly byengine speed and load. In fact, it is common to createmaps of BSFC as a function of engine speed and load.These diagrams always show a great deal of variationin the BSFC over the mapped range of speed andload.

Particularly in the case of the studies done byMilen and Yilmaz, the authors apparently did notunderstand the need for such a protocol. With thesestudies, it was necessary to correlate two di�erent ta-bles, one for speed and one for load. It is not entirely

4

clear that the tables were necessarily correspondent.However, it had to be assumed that they were in or-der to do the calculation. Regardless, the yields wererather meager and there is no reason to suppose thatbetter testing would have had much of an e�ect onthe results.

Discussion of results.

Break even values.

Yield values can be compared break even values. Thepractical break even value is the amount of yieldneeded to exceed the total amount of energy neededto generate a gram of hydrogen. The heat of com-bustion of hydrogen is 0.118 MJ/gram. The SS-40HHO reactor used for the FVTC and UNOH stud-ies is about 70% e�cient. An automotive alternatoris about 55% e�cient. So there is a practical breakeven value of 0.31 MJ/gram = 0.118 / (0.7 x 0.55).An economic break even value is the amount of

yield needed to cost justify use of the HHO injectiontechnology. This is a more complex value. Cost ofinstallation of the technology, type of vehicle, fuelcost, vehicle use conditions and required pay backperiod should all be identi�ed. A typical set of inputsmight be:

• Installation cost: High end HHO injection sys-tem labor and parts = 5,000 USD.

• Vehicle: Class 8 truck. Gross Vehicle WeightRating = 33,000 lbs. Typical = 65,000 lbs.

• Use conditions: long haul. 100,000 miles peryear. Typical mileage: 6 mpg or 0.167 gallonsper mile. Average engine load: 200 Hp. or 520.5MJ/hr. (200 x 723 x 3600)

• Fuel type and cost: No. 2 Diesel. 4.00 USDper gallon.

• Pay back period: 6 months

Total annual fuel cost would be 100,000 x 0.167 x4 = 66,800 USD. This requires a 15% reduction infuel costs since 0.15 x 66,800 = 10,000. Yield value isbased on a fuel weight ratio so yield needed to achieve

a 15% reduction of 520.5 MJ/hr would be 520.5 x 0.15= 78 MJ. A feed rate of 8 grams of hydrogen perhour would require a yield of 9.76 MJ/gram. This iswithin the range of the SS-40 used in the FVTC andUNOH evaluations. It is also helpful that both theseevaluations were performed on Diesel truck enginestypically used in the example given here. This is arather approximate estimate. More detailed model-ing has been done indicating that a 30% reduction infuel cost is perhaps possible. [12]

Yield values.

The range in the values listed in Table 1 is notable. Acommon characteristic of the high yield evaluationsis a lower feed rate in comparison to the size of theengine, about 0.5 to 1 lpm per 100 hp of engine ca-pacity. Such a large e�ect on engine performance incomparison to the small amount of gas injected sug-gests a free radical chain reaction mechanism that isperhaps similar to the e�ect of tetra ethyl lead onengine knocking in spark ignition engines. See Ap-pendix B for further discussion of reasons why HHOincreases the e�ciency of Diesel engines.A commonly held idea is that HHO is in some way

di�erent from a 2 to 1 molar mixture of molecularhydrogen and oxygen gases. This is suggest by theappearance of the �ame of HHO (Fig. 1) [13] versusthat of hydrogen (Fig. 2) [14] .While a hydrogen �ame is almost invisible, an

HHO �ame is brightly luminous. The di�erence is noteasily explained by conventional factors such as con-tamination due to sodium or potassium electrolyte.The presentation of the �ame is subjective. How-

ever, it can be converted to data by means of a spec-troscopic measurement which is a plot of light inten-sity as a function of its wavelength. It is generallyobtained using an electronic spectrophotometer in-strument. At this time, an HHO spectrograph is notavailable. Fig. 3 [15] shows a spectrograhic signa-ture of a butane �ame. This illustrates how variousfree radicals that form momentarily during combus-tion produces emission peaks at characteristic wave-lengths. In the case of hydrocarbon combustion, theC2 radical is particularly abundant.The e�ect of HHO could be caused by radicals that

5

Figure 1: HHO �ame

Figure 2: Conventional hydrogen �ame

Figure 3: Flame spectrograph of a butane �ame

are produced momentarily when HHO burns. Theconcentration of these radicals could be measuredby analyzing �ame spectroscopic data. This mightenable calculation of a baseline independent of en-gine performance that could be used to predict thestrength of the e�ect of HHO on engine e�ciency.

Conclusion.

This sort of analysis highlights the state of uncer-tainty regarding laboratory evaluation of the idea ofHHO. To resolve these issues, engine lab evaluationsof HHO might follow these guidelines:

• An investigation should be a diligent attemptto replicate the best results that have been ob-tained.

• There is a wide variety of con�gurations of reac-tors that are purported to produce HHO. If thee�ect of HHO on engine e�ciency is the resultof some product of the combustion of HHO, itis possible that this factor can be identi�ed andquanti�ed by means of �ame spectroscopy. Thenall of the variability associated with reactor op-eration can be reduced down to a few variablesor perhaps one variable. It would also be impor-tant that this factor remains constant for eachseries of tests.

• Mass �ow rate of HHO should be measured. Theissue of the average molecular weight of HHOshould also be investigated further. For mosttypes of mass �ow measurement of gas, averagemolecular weight is a relevant variable.

• Test conditions include load and speed. Load onthe engine should be regulated by means of a feedback control loop. E�ective compression ratioshould also be included among test conditions.Most Diesels for vehicles are turbocharged. Bycarefully o�setting the turbocharger waste gateactuator, e�ective compression ratio can be ad-justed.

• By means of combustion pressure analysis, engi-neers can essentially see what is going on inside

6 REFERENCES

the cylinder through the course of the compres-sion/power stroke. Almost none of this sort ofwork has been done with HHO but it could bevery helpful in developing a mathematical modelof the e�ect of HHO on the Diesel engine cycle.This could be helpful in improving the perfor-mance and �exibility of HHO technology.

References

[1] Dempsey, P.; Troubleshooting and RepairingDiesel Engines. Pg. 18, ISBN: 978-0-07-149371-0.

[2] Yilmaz, A.C.; Uludamar, E.; Aydin, K.; Ef-fect of hydroxy (HHO) gas addition on perfor-mance and exhaust emissions in compression ig-nition engines. International Journal of Hydro-gen Energy, Vol. 35, Issue 20, Oct. 2010, Pgs.11366-11372. See �les: yilmaz_�g3_4.xls, yil-maz_�g3.png, yilmaz_�g4.png.

[3] Milen, M; Kiril, B; Investigation of the ef-fects of hydrogen addition on performanceand exhaust emissions of diesel engine,FISITA 2004, World Automotive Congress.See �les: milen_�g7_8.xls, milen_�g7.png,milen_�g8.png.

[4] Purdue University evaluation. See �le: pur-due.xls

[5] Fox Valley Technical College evaluation. See �le:fvtc1.xls

[6] Univ. of Northwest Ohio evaluation. See �les:unoh.xls and unoh.png

[7] Bari, S.; Esmaeil, M.; E�ect of H2/O2 addi-tion in increasing thermal e�ciency of Dieselengine. Fuel, vol. 89, pgs. 378-383. See �les:bari_�g4.xls, bari_�g4.png.

[8] Wang H-K.; et. al., E�ect of regulated harm-ful matters from a heavy-duty Diesel engineby H2/O2 addition to the combustion cham-ber. Fuel (2012), Vol. 93 pgs. 524-527. See �les:wang_�g3.xls, wang_�g3.png.

[9] Birtas, A.; et. al., E�ects of Air-Hydrogen In-duction on Performance and Combustion of aDiesel Engine. SAE Paper 2011-24-0094.

[10] Yilmaz, A. C.; Design and Applications of Hy-droxy (HHO) Systems

[11] Santilli, R. M.; A new gaseous and combustibleform of water. Hydrogen Energy, Vol. 31, (2006)Pgs. 1113-1128.

[12] Ware, C.; Optimized HHO injection for Class 8Diesel Vehicles. wp14.pdf

[13] Prisecaru, T., et. al.; Experimental validationof an HHO gas cutting �ame CFD model. 9thInternational Conference on Quantitative In-fraRed Thermography. Politehnica University ofBucharest, Romania

[14] Simon, J. M., et. al.; Guidelines for use of hydro-gen fuel in commercial vehicles. Dept. of Trans-portation document

[15] wikipedia image. Spectrum_of_blue_�ame.png

[16] Petchers, N., Combined Heating, Cooling &Power Handbook: Technologies & Applications,Fairmont Press, ISBN 0-88173-433-0, pg. 350,Table 17-2.

REFERENCES 7

Figure A.1: Yield energy

Appendix A

Total energy is given as power multiplied by time,in this case, the time interval of the test run. Withor without gas injection, the total amount of energyproduced is the same since load and speed should bethe same for both tests. The additional energy pro-duced is actually the amount produced by a quantityof fuel. That is de�ned by the equation:

E

m1− E

m2(A.1)

where E is the amount of energy produced withinthe time interval, m1 is the mass of fuel used withHHO injection and m2 is the mass of fuel used with-out HHO injection. The relationship can be visu-alized in Figure A1. The total area of the circle isproportional to E

m1. The area of the lighter gray

is proportional to Em2

. Therefore, the area of thedarker wedge would be proportional to the di�erenceEm1

− Em2

which is the additional amount of energyproduced.

The fraction of the total area that the darker wedgeoccupies is necessarily the area of the di�erence di-vided by the total area of the circle. This fractionwould be:

Em1

− Em2

Em1

(A.2)

The E term cancels out of Eqn. A2. Also, the twoterms in the numerator when multiplied by m1 give:

1− m1

m2(A.3)

The total amount of energy produced per gram ofhydrogen would be E

mgwhere mg is grams of hydro-

gen gas injected during the time interval. The ad-ditional amount of energy produced would be thisquantity times the proportion given by Eqn. A3 thusgiving Eqn. 2:

E

mg

[1− m1

m2

]= energy yield (A.4)

In the case where the amount of energy producedwith and without HHO is slightly di�erent, such thatE1is the amount of energy produced with HHO andE2is the amount of energy produced without HHO,the values m1and m2should be adjusted to what theywould be if total energy produced is Eavg whichequals (E1+E2) / 2. Assume that the mass of fuelused is proportional to the energy produced withinthe interval between E1and E2. To set the m valueto what it would be if the E value was equal to Eavg,we would have to increase m if E < Eavg and de-crease m if E > Eavg . The adjustment would have

to be inversely proportional to E so it would beEavg

E .

The adjustment for m1

m2would be

Eavg

E1/

Eavg

E2which

factors out as E2

E1or the correction factor k given in

Eqn. 3.

8 REFERENCES

Appendix B.

Various reasons are commonly given as to why HHOincreases Diesel engine e�ciency. In this discus-sion, an increase in combustion e�ciency shall bediscounted because this explanation is not consistentwith laboratory data.Combustion e�ciency is sometimes confused with

fuel conversion e�ciency so the �rst task will be to de-�ne these terms. Combustion e�ciency is the fractionof energy latent in the fuel that is released by com-bustion. For Diesel engines, combustion e�ciency isgenerally better than 98% as will be calculated belowbased on typical emissions data.Fuel conversion e�ciency is the fraction of energy

latent in the fuel that is converted to useful mechan-ical output of the engine. It is typically about 35%.Much of the heat released by combustion is lost byconduction to the engine block or through the ex-haust gas. Fuel conversion e�ciency is a subset ofcombustion e�ciency.Table B1 [16] gives values needed to determine

combustion e�ciency. The three emissions compo-nents found in Diesel exhaust that contain carbonare particulate carbon (soot), carbon monoxide, andhydrocarbon (principally methane).On this table, MJ per MJ of input from fuel is

obtained by multiplying the heat of combustion ofthe emissions component (MJ/gram) by grams perMJ input. These quantities for each component aresummed to obtain the Ec , the amount of energycontained in emissions components per MJ of fuelenergy. Combustion e�ciency is given by:

ηc = 1− Ec (B.1)

where ηc is combustion e�ciency. This value is

Emissions component MJ/gper MJ of inputgrams MJ

Carbon (soot) 0.032808 0.09 0.00295272carbon monoxide 0.010112 0.05 0.0005056

Hydrocarbon (methane) 0.050009 0.16 0.00800144

total MJ per MJ input (Ec) 0.01145976

Table B.1: Diesel emissions components

approximately 0.98 or 98%. HHO injection reducesemissions by about 10% [6]. Even if it reduced themto nothing, the fuel conversion e�ciency could be in-creased by no more than 0.6% (2% increase in com-bustion e�ciency x 0.3 fuel conversion e�ciency).The �le unoh.xls indicates an average increase in fuelconversion e�ciency of 11% which could not be ac-counted for by an increase in combustion e�ciency.It may be possible to arrive at an explanation for

the HHO e�ect by means of combustion pressureanalysis. Almost no combustion pressure analysis hasbeen done with HHO. Milen and Kiril [3] obtained asingle plot of pressure versus volume with and with-out HHO at 1500 RPM.