are quoted showing the avai labi l i ty factors for boilers to be lower than for turbines and from which the conclusion is drawn that for unit type plants avai labi l i ty lies within the probable l imits of 80 to 87 per cent. While granting the premises, I cannot agree with the conclusion, because the percentages of nonavailability include "scheduled cleaning and overhaul" and all outages are assumed as noncoincident.

In the case of schedule maintenance the length of outage is often governed by the need or lack of need of the unit for service. A job requiring 160 crew-hours m a y be done in 4 weeks by a single crew working 40 hours a week, but if the equipment is real ly needed, 3 crews will be put on, working 24 hours a day, and the job will be completed in one week. The reported avai labi l i ty factors, while based on experience, are low because in many cases there was no need to speed up the repair job.

Also, it will generally be the practice to overhaul both the boiler and turbogenerator units and accessories during the same period, so that even with overlap the total t ime un­available will be the duration of the com­bined jobs from start to completion of both, but not the sum of their individual dura­tions.

From table I the total elapsed time for both generators is found to be 68,841 hours, or an average of 34,420 hours per unit. From table II we find that the boilers were on bank for a total of 61,475 hours, and active for 76,996, hence avai lable for service 138,471 hours, or an average 34,618 hours per boiler, which indicates apparent ly a 100 per cent avai labi l i ty factor.

Since 3 boilers will carry the full plant capacity the installation of the fourth does not seem justified in the light of this ex­perience.

Low installation cost per capacity. This has been achieved. The total and unit cost is available to a n y one who will dig it out. Nevertheless, had the fourth boiler been omitted the plant cost would have been lowered by almost 5 per cent. However, the heat cycle chosen and previously re­ferred to is a complex one for a low cost plant. A plant of comparable magnitude, 3 50,000 kw units, designed with but 2 stages of bleeding, no evaporators, no air preheaters, and working a t 375 pounds a t the throttle, 50 pounds less than Huntley No. 2, operated during the 12 months of 1934 under sl ightly less favorable conditions than did Huntley No. 2 during the last seven months of 1933, with better thermal economy. The average load was 58.8 megawatts a s against 64.5, the capaci ty use factor 1.2 per cent lower, the net turbine water rate 145 pounds lower, and the Btu per net kilowatt-hour output 13,576 as against 13,812 for the Huntley No. 2 station.

The 4 multi-drum boilers have a maxi­mum steaming ra te equivalent to S of the Huntley No. 2 boilers. The costs per kilo­watt of these 2 boiler plants are practically identical, the actual investment being somewhat less than tha t of the Buffalo plant. In this plant the saving of the heat loss from the generator air cooler was deemed worth while although the heat loss from the oil cooler is not recovered.

Low operating and maintenance costs. Figure 15 shows total relat ive costs for various monthly outputs and, making

allowances for the fixed charges on the plant investment, th is requirement has been reasonably met.

It must supply an expected 100,000 kw load in Buffalo. The criterion that 100,000 kw of load in Buffalo must be carried on this plant could be met with a 3 boiler plant of the type selected, 2 of which could furnish 1,056,000 pounds of s team, so here again the installation of a fourth boiler seems super­fluous.

Relatively large units should be used. A choice of uni ts of an individual capaci ty of about 15 per cent of the system peak load is in line with conservative engineering.

Initial ties of 210,000 kw for local load and 120,000 kw for interconnection. The initial installation of local subtransmission circuits of over 200 per cent of maximum load when these circuits are not only underground but in many cases feeding low voltage networks seems excessive. However, i t affected the plant design only in the initial space re­quirements necessary for housing the 22 kv breakers and in the cost of both. The in­vestment for the 22 kv and 110 kv feeders was kept low and the switch house layout made simple.

Building design must not restrict choice of future equipment. This requirement seems to have been met by the building design provisions which will permit of 200,000 k v a tandem units being installed if found de­sirable so to do.

In conclusion, the most serious defect is the limited capaci ty of the s team drums of the boilers installed, restricting quick pickup of load, which defect can be remedied by the installation of auxi l ia ry drums. If such a program were undertaken, the presence of the fourth boiler would grea t ly facilitate the change since 3 boilers would be avai lable for normal operation, 2 of which could handle the maximum load of the Buffalo area.

Rehabilitation of the Conners Creek Plant

Discussion of a paper by R. E. Greene, pub­lished in the June 1935 issue, pages 610 -17 , and presented for oral discussion at the power generation session of the summer convention, Ithaca, Ν . Y . , June 2 5 , 1 9 3 5 .

C. M . Gilt (Brooklyn Edison Co., Brooklyn, ' N . Y . ) : The problem presented by the

existence of more or less obsolete generating plants and the need for increased capaci ty is one that will become more pressing for many power companies a s load increases during the next few years . The choice must be made between building a new plant and retaining the old for reserve and peak duty, building a new plant and scrapping the old one, or rebuilding the old plant . There is no general solution to the problem but each case must be settled on i ts merits.

Quite frequently an old plant can ad­vantageously be kept for s tand-by and peak load service, thereby postponing the in­vestment necessary for new capaci ty . I t s cost per kilowatt-hour generated m a y be extremely high and y e t i t s annual cost m a y be mater ia l ly less than the carrying charges on the equivalent new capaci ty . When the t ime comes tha t the old plant must be

scrapped, i t would usual ly be cheaper if the old plant were nonexistent and, one could star t fresh with no encumberances. Some­times the old plant can be rehabil i tated for a t ime a t a low cost by changes in par ts of the plant, or by superimposing high pressure units, that improve economies or increase the capaci ty or both.

Rebuilding an old plant usual ly requires retiring part or al l of the old capaci ty with the result that while the net addition to capital for the modernized plant m a y not be great, the new money required for the addi­tional capaci ty m a y be quite high.

The figures given in R. E. Greene's paper for new money plus the value of equipment from other locations amount to $20,826,610 for an.increase in capaci ty of 150,000 kw, or $139 per kw. However, Mr . Greene s ta tes that the old plant would have to be retired shortly under any program, and would be of l i t t le value in the meantime. Recognizing the necessity for this retirement, 150,000 kw increased capaci ty has been obtained for $13,884,755 net addition to capital , or a t $92.50 per kw, and a modernized and more efficient plant of 330,000 kw capaci ty has been obtained with a total book value of $30,367,200 or $92 per kw, a cost per kilo­wat t about the same a s that of the old plant .

Comparing these figures with costs of total ly new plants, i t would appear that the choice of obtaining the added capaci ty by rebuilding the old plant rather than building a new one to supplement it was la rge ly determined by the judgment that the old plant should be retired under a n y condition, plus the fact that by gradual ly rebuilding the old plant, the large initial expenditures associated with start ing a new station would be avoided.

I t would seem evident that a load district method of operation in which each generat­ing station largely supplies i t s own district load, furnishes an incentive toward more rapid rebuilding or scrapping old p lants than one in which new and economical plants feed the system a s a whole and old plants can be relegated to short hour use a s reserve or peak du ty for the system a s a whole.

C. A. Powel (Westinghouse Elec. and Mfg. Co , E. Pittsburgh, P a . ) : Rehabil i tat ion i s a lways interesting because the problems it presents are much more difficult than in building a new plant. R. E. Greene's company selected a method of rehabil i tat ion which worked out part icularly well in then-case because they were able apparent ly to utilize a great deal of the existing par ts in the rebuilt assembly. I wonder if, under the circumstances, the statement is justified that high pressure units superimposed on the old machines would have resulted in higher maintenance and lower re l iabi l i ty , inherent in old equipment. In any scheme of rehabilitation if it is to be economical, considerable old equipment must be re­tained and the weak member in rel iabi l i ty m a y not be the main units . Wi th super­imposed turbines the life of the existing equipment enters the problem to no greater degree than i t would if the station were left a t i ts original pressure. When thé decision has been made to replace the old boilers b y higher pressure units, the question of the old turbinegenerators should be t reated a s a separate problem. The turbines need

SEPTEMBER 1935 1 0 0 1

not be rebladed or the generators rewound a n y sooner because of the superimposition of high pressure equipment, and these opera­tions are about a l l that is necessary to main­tain the life of the units almost indefinitely.

D. F. Pennell (nonmember; Brooklyn Edi­son /Co . , Brooklyn, Ν. Y . ) : The most interesting point for consideration in the paper on Connors Creek plant is the amount of old equipment it was possible to use in the rehabilitation. The following pieces of equipment were used of the original plant :

a. Building and foundations (with addition to turbine room) .

b. S tacks .

c. Coal handling equipment. d. Condensing equipment including condensers originally installed with 2 0 , 0 0 0 k w units but adapted t o 3 0 , 0 0 0 kw units b y use of improved tube layouts and extract ion heaters , also condenser auxiliaries

e. Circulat ing water tunnels .

In addition to this, 2 3 0 , 0 0 0 kw gener­ators were used that had been salvaged from Connors Creek and Delray No. 2 .

Included in the construction of the new machines are the exhaust casings of 3 of the 2 0 , 0 0 0 kw units and the pillow blocks of the old 3 0 , 0 0 0 kw units. Nearly all the piping, including service piping, had to be replaced to meet the new requirements.

In spite of the amount of this equipment it was possible to salvage, the installed cost per kilowatt-hour does not reflect any great saving.

Original capaci ty 1 8 0 , 0 0 0 kw Expected capaci ty 3 3 0 , 0 0 0 kw

Increase 1 5 0 , 0 0 0 k w Net increase in book va lue = $ 1 3 , 8 8 0 , 0 0 0 Cost per kw increased capacity = $92 .50

It appears that with the same investment 2 large units might well have been installed giving the same increase in capaci ty and thermal economy and that these units could have been added when, as , and if needed without interrupting normal operation of the old plant.

One statement made by R. E. Greene is interesting to note. He says , "At the s tar t it was recognized that certain predi­lections existed in the minds of the oper­at ing staff, and it was agreed that these should be tested and followed out only when found to be sound." As a matter of practical design, it might be better to re­verse this and follow these so-called pre­dilections unless they prove on analysis to be unsound.

The author also states "the electrical system of the Detroit Edison Company is divided into load areas interconnected by ties for mutual help in emergencies." This electrical layout of the Detroit system appears to give the plant designer a some­what different angle of a t tack than we would have in Brooklyn. It complicates the problem by making it necessary to pre­dict not only changes in system load but to predict where these changes will occur. In New York, however, we build for the system load and due to adequate interconnections have the economy resulting from inter­change of power. B y this method we are able to supply the system base load with the most modern generating equipment, taking advantage of the economy of first cost and operation that goes with large generating

units and using older stations for s tand-by and peak load capaci ty .

One reason given for revamping the old plant is that in addition to high maintenance costs the "re l iabi l i ty would be not of high degree." This is a minor point but I feel that our experience is that old equipment properly maintained gives a satisfactory degree of rel iabil i ty. Oftentimes it is more reliable than new equipment that tqnds toward the experimental.

Complex Hyperbolic Function Charts

Discussion of a paper by L . F. Woodruff pub­lished in the May 1935 issue, pases 550-4.

A. E. Kennelly (Harvard University, Cam­bridge, M a s s . ) : The charts offered in the paper will be serviceable to transmission engineers and to all those who are interested in alternating current lines having a t oper­ating frequency an angle not exceeding 0 . 4 in size.

The charts possess an unusual feature from a mathematical standpoint. In order to find certain desired trigonometrical

functions of the line angle 0 ^ n a m e l y cosh 0,

sinh 0 Λ Ι tanh 0/2 \ - , and/or Λ / 0 )

θ/2 one enters the

curvilinear system of each chart with the size and slope of 0 2 ; i. e.,| ZY | and ZY, then reading off the size and slope of the function on the semirectilinear ruled back­ground. This unusual procedure reason­ably claims a lessening of t ime and effort in the process.

The Sparkless Sphere Gap Voltmeter

Discussion and authors' closure of a paper by R. W. Sorensen, J . E. Hobson, and Simon Ramo, published in the June 1935 issue, pases 651-6 , and presented for oral dis­cussion at the selected subjects session of the summer convention, Ithaca, Ν . Y., June 27, 1935.

A. O. Austin (consulting engr., Barberton, Omo): The authors have carried out a very effective method of checking root mean square values with means a t the dis­posal of practically a n y high voltage labora­tory. Anyone making high voltage meas­urements is continually running up against conditions upon which it is desired to make a check, and since the method involves no elaborate equipment i t can be readi ly car­ried out. Where the wave form is good the method m a y be used for checking crest values. I t would seem tha t the method is primarily applicable for calibration pur­poses. The capacitance calculations of the sphere gap can also be used very effectively for determining crest voltages which I believe have a wider application than the determination of root mean square values.

Wi th the close control of frequency the determination of voltage with an ammeter in series with a capacitor of known capaci­tance is very simple.

While the capacitance tap on bushings or condensers is one of the most useful methods in determining high voltages in the labora­tory, i t is necessary that the capacitance of the condenser supplying current to the insulating instrument does not change with voltage due to corona from the leads or field et up by external objects. It is hoped that

the calculations in connection with the capacitance of spheres will be carried further and l imits established so that the sphere gap may be used a s a capacitor for direct checks upon the crest voltage. This is based upon the fact that / = ECœ, hence

Ε — —— This method has proved to be

very useful and makes it unnecessary to subject equipment under test to an oscilla­tion following a discharge of a parallel gap.

The accuracy of this method of determin­ing crest values depends pr imari ly upon the accuracy of the capacitor. R. W. Soren-sen's calculations would indicate that the sphere gap m a y be used to supply the needed capacitance. Irregularit ies which affect the discharge of a sphere gap also have an effect upon i ts capacitance, and will have to be determined within fairly close l imits . It is to be hoped that R. W. Sorensen will continue his work and determine the neces­sary l imitations. Wi th a capacitor of known capacitance it is possible to use a ballistic method to determine transient voltages to advantage in many cases. Wi th R. W. Sorensen's method it would be possi­ble to show some of the irregularit ies of the sphere gap and to obtain a further check upon calibration curves.

For normal frequency tests it has been evident that the needle gap gave far more consistent results based upon transformer ratio than sphere gaps for normal frequency determinations up to 3 5 0 or 4 0 0 kv. The capacitance measurements for the sphere gap calls attention to the errors which m a y occur where an a t tempt is mane to measure crest transient voltages with a resistance in series with a gap.

It is to be hoped tha t the work outlined in the paper will be continued, and tha t the advantages and limitations of the various methods will be given due consideration a s a single method will not give the best results for al l conditions.

J . R /1 l leador (General Elec. Co., Pittsfield, Mass . ) : The authors have presented a very interesting paper on voltage measurement. It is especially t imely since the A.I.E.E. s tandard sphere gap calibration curves a re now being revised. I t is important to have da ta taken by a s many different methods a s possible. Using the method of voltage measurement described by the authors, a check can be made on the 6 0 cycle calibra­tion curves for sphere gaps, which have been recently recommended ("Calibrat ion of the Sphere Gap," J . R. Meador, ELEC. ENGG. (A.I.E.E. TRANS.) , V. 5 3 , June 1 9 3 4 , p. 9 4 2 -8 ; "Impulse Calibration of Sphere Gaps," P. L. Bellaschi and P. H. McAuley , Elec. Jl., v. 3 1 , June 1 9 3 4 , p. 2 2 8 - 3 2 ) .

Although the experimental curve in figure 4 does not agree closely a t large spacings

1002 ELECTRICAL ENGINEERING