Proceedings of the Institution of Mechanical Engineers

A History of the development at Cambridge of a notch tensile test for the investigation of brittleness in structural mild-steel is described. The test procedure and the methods of assessing the temperature limit at which the material becomes brittle are outlined in Part I.

Investigations of material from an oil storage tank and from ships which fractured in service show that this notch tensile test gives good correlation with these casualties whatever the nominal stress in the plates at the time of the fracture. The results of Charpy V-notch and Izod tests are given for comparison, and it is shown that the energy absorption criterion used in these tests may be misleading (Part II).

In Part III experiments are described in which the tensile test has been used to predict performance in test-pieces of more complicated geometric form.

Criticisms of the notch tensile test are dealt with in Part IV.

In the conclusions, attention is drawn to the fact that the designer of orthodox fabricated structures must be provided with material which is ductile at all operating temperatures and at all stress levels.

To investigate the incidence of air-entraining vortices in suction wells, experiments have been made with a suction tube of 2 1/2 inches diameter, fitted with a bellmouth intake, mounted in a series of wells of different shapes.

With the tube set on the axis of a cylindrical sump into which water flowed radially between baffles at the bottom, only slight and transient air entrainment was noted, and this is ascribed to temporary irregularities in the flow near the water surface. Air entrainment was completely prevented by imparting a swirl to the flow at entry to the cylinder, and this swirl was found to pass into the suction tube with only slight reduction except near the axis.

With the tube drawing from a rectangular well, air-entraining vortices formed at the edge of the wake of the external surface of the tube in the general flow in the well. The critical submergence (submergence of the bellmouth required to prevent air entrainment) is greatly affected by non-uniformity of the flow along the well.

Any measure which reduces or diffuses the vorticity shed from the outer surface of the tube into the well has a beneficial effect; reducing the clearance between the rear wall of the sump and the suction tube reduces the streaming flow past it and so reduces the critical submergence, while a streamlined fairing fixed round the tube, or a cylindrical shroud placed within the well and coaxial with the tube, have both been found virtually to eliminate air entrainment.

The factors affecting the formation of air-entraining vortices in sumps are enumerated, and their relative importance assessed from experiments with a number of different sumps. It is established that vortices are caused by rotation of the mass of water within the sump arising primarily from the entry conditions. Their severity, measured in terms of the minimum safe submergence to avoid air entrainment was also found to depend greatly on the velocity in the suction inlet; air entrainment rarely occurred below a velocity of 2 ft. per sec., and above 15 or 20 ft. per sec. the severity of the vortices did not increase substantially.

The shape and disposition of the suction inlet only slightly affected the vortices, but sump boundaries were found to reduce appreciably the tendency for vortex formation when closer than 8 diameters to the suction pipe: with the walls very close to the pipe, vortices did not occur. Small sumps may thus be more satisfactory than spacious ones.

Tests with scale models of existing or proposed pump installations are described, and for one sump, where vortices formed within the working range, it is demonstrated that suitably placed boundary walls would be considerably more effective in preventing vortices than modifications to the suction inlets.

Measurements of the quantity of air drawn into the pump through a vortex showed that air entrainment of this type could seriously impair pump performance. Severe vortices could lead to depriming of pumps operating with suction lift.

The author submits that current technical education falls far short of requirements in training the student engineer in the skills of applying knowledge and using his mind, and he proposes how the means to remedy the situation may be sought.

In this paper, the author emphasizes the rapid development and change which have taken place in the preceding ten years in the sheet metal manufacturing side of the aero-engine industry.

He singles out the development of mechanized welding as one of the more interesting features and gives an illustrated review of a wide range of applications of fusion and resistance welding in current use at the Rodney Works of his company, in the Engine Division.

In the paper are described the principal features of the weight-loaded and gas-loaded types of hydraulic accumulator together with the applications where they are of particular service.

A distinction is drawn between those gas-loaded accumulators where the gas and the pressure liquid are separated, and those where they are directly in contact with each other, and the separated type is further subdivided according to the method of separation.

The appropriate control gears for the various types of accumulator are described and special reference is made to the particular problems associated with the control of the non-separated type of gas-loaded accumulator.

The normal method used in calculating the volumetric ratio between the compressed gas and the usable liquid in gas-loaded accumulators is quoted, together with the effect of special conditions upon the amount of pressure variation. Figures are also given to show the effect of temperature changes upon the working pressure in this type of accumulator, and the normal maximum allowable working pressure in the non-separated type.

The experiments described in this paper are part of a series of researches being carried out to obtain data for improving the design, construction, and utilization of production equipment, having regard to the increasing demand for such development which will be occasioned by the growth of automation. Research techniques and equipment were developed for examining the deflection and vibration characteristics of various drilling machines. Some comparisons were made between radial-drilling machines, vertical-drilling machines, and portal-frame drilling machines, the latter type being included particularly because of the increased possibilities for its employment in unit construction automated lines. An account is given of the general experimental procedure and the special apparatus used. An analysis is made of the deflection characteristics of each type of machine, and examples are given of the magnitudes of the deflections. The principal natural frequencies of vibration of the machines were determined, and the vibration characteristics of the machines in drilling and facing operations were examined. Recommendations are made for the reduction of machine deflections and vibrations, and a basis for some machine design calculations is developed.

Machine-tool chatter is essentially a problem of dynamic instability. A drilling machine under working conditions can be regarded as a dynamic system which for certain values of its system parameters becomes unstable and bursts into oscillations. The system parameters are determined by the static and dynamic characteristics of the machine-tool structure, the geometric shape of the drill, certain features of the work-piece material and the working conditions, and it is the interrelation of these which results in chatter or chatter-free machining. The paper commences with a discussion of the dynamic characteristics of radial-drilling machines. This is followed by an experimental investigation of the behaviour of these machines under working conditions, and finally a theory of drilling chatter is developed. The results of this theory are presented in the form of a stability chart which shows those drill speeds at which, for given values of the system parameters, chatter-free drilling is ensured. Although the theory presented is based on an experimental investigation of radial-drilling machines drilling into solid material, it is directly applicable to drilling operations carried out on any other type of machine and yields also results concerning chatter occurring at drill break-through and when drilling up pre-drilled holes.

The paper presents a mathematical theory of stability of small oscillations in surge tanks of the type where two water surfaces are present. The name ‘parallel-branch’ surge tank is coined to describe such a system, and it is shown that results for orifice and differential surge tanks can be obtained from it.

On the assumption of ideal governing and small oscillations of flow and water levels, equations giving the limits of stability are deduced for all cases. It is an easy matter from these to test the stability of any particular design by inserting appropriate numerical values.

In general, the results confirm the well-known Thoma formula for the case of a simple surge tank and give results identical with those obtained by Professor Escande for the case of an orifice surge tank. For differential surge tanks, the limits of stability are given in an equation covering the area factors of the riser and the outer chamber as well as the resistance factors of the main conduit and orifices respectively. Graphs showing this relation between the above quantities indicate that:

(1)

if the combined area of riser and outer chamber is equal to the Thoma area calculated for a simple surge tank, then the system is stable provided the resistance factor of the orifice is below a certain limiting value;

Methods of calculating the frequency and amplitude of permanent oscillations are presented in the paper. The general case of a true parallel-branch system is best solved numerically, but the results indicate that in all practical cases the phenomenon of finite amplitude oscillations will exist. Attention is drawn, however, to the necessity of considering the period at which instability is possible as, on occasion, this is too short for the assumption of ideal governing to be applicable.

The applications of liquid metals in heat-transfer problems are discussed, and comparisons are drawn between those metals whose melting points are sufficiently low to make them of interest in this respect. A number of sodium and sodium-potassium circuits which have been built and operated over the past three years are described; some of these were built in order to gain experience in handling the liquid metals, whilst others were for heat-transfer experiments. A general discussion on the design and construction of liquid metal circuits and the components which are commonly used in them is given. A brief discussion of forced-convection heat-transfer with liquid metals follows, and the results of some work on cavitation in sodium-potassium eutectic flowing through a constriction in a pipe are summarized.

Some of those aspects of the auxiliary services in large power stations on which decisions have to be taken at the planning stage are discussed. The effects on the planning of auxiliary services of recent changes in generation practice are noted. The aspects discussed include alternative means of providing auxiliary power, the economic justification for providing standby for various auxiliary-plant items, the diagrammatic arrangement of auxiliary services, the economic and practical importance of certain basic design features and performance figures and sundry practical details that require thought at the planning stage.

It is shown that decisions at this stage should be based as closely as available data permit on quantitative assessments. Among these, particular attention should be given to assessments of the reliability of individual plant items, of the cost of any measures such as duplication for increasing reliability, and of the total cost of forced outages. Methods of basing decisions on such assessments are given. On the basis of experience with a large number of power stations recommendations are made for obtaining the greatest possible ease of operation and maintenance.

Measurements have been made of adiabatic friction and of heat transfer for air flowing in pipes at speeds up to 1·6 times that of sound for conditions closely approaching established pipe flow in the range of Reynolds numbers Re = 1·24 to 4·35 × 10⁵, free stream air temperatures T_b = 190 to 286 deg. K., and wall to free stream temperature ratios T_w/T_b = 1·08 to 1·73.

The heat transfer results agreed within ±5 per cent with the formula

and frictional data were some 4 per cent above

The above formulae were obtained by a method, first suggested by von Kármán (1935), of allowing for compressibility effects by replacing, in formulae for incompressible fluids, the fluid properties computed at bulk temperature by those at wall temperature, and then transforming to terms of these properties at bulk temperature, using the laws of variation of physical properties with temperature as obtained from N.B.S.-N.A.C.A. Tables (1949-50).

Separate tests were made of the heat transfer in the presence of a plane shock-wave in the heated length of the test pipe, and the values of Stanton number were found to be up to 30 per cent higher than for shockless flow. A subsonic test with a similar adverse pressure gradient over the heated length also showed considerably increased heat transfer.

The recovery factors calculated from the adiabatic runs were close to √Pr, that is, 0·84.

The method of operation of water-ring self-priming pumps is briefly described, with performance data of three designs of such pump. The mechanism by which these pumps produce their effect is discussed, and the conclusions reached indicate that it is impossible to anticipate their performance from their physical dimensions as can be done with a centrifugal pump. If, however, the head-discharge curve of any pump is established, and the horsepower required to overcome mechanical and disk friction is known or can be closely estimated, the horsepower taken by the pump at any point of its range can be calculated with a high degree of accuracy. Amongst the information included are graphs of air-pumping capacity and details of experiments to determine the best designs of rotor vane. Small changes in the shape of the vanes produce remarkable differences in performance.

In 360-deg. journal-bearing theory, the load-carrying film is assumed to be partial as a result of cavitation in the diverging outlet film, but the friction-producing film is taken as complete. The effect of oil inlet conditions on film extent and the state of affairs prevailing in the cavitated film have been matters for speculation. A method has been developed for photographing the oil film in glass bearings of standard clearance ratio (0·001 and 0·002) and the results of a qualitative and quantitative examination of film extent are described. The measured oil outflow rates are analysed by separation into the hydrodynamic and supply-pressure-induced components, and explanations for an apparent deficit in the former are discussed.

An experimental investigation into the operation of stepped thrust-bearings has been carried out using an experimental machine and measuring equipment specially designed for the purpose, but adaptable to other types of thrust bearing. The running oil-thickness and friction characteristics have been determined for a range of speeds, loads, and inlet temperature for several different angular positions of the radial step and for two curved steps. Results are given in the usual manner using ZN/P and also in non-dimensional form and are compared with theoretical predictions based on constant viscosity.

The experimental results are compared with the corresponding results obtained on tests of a Michell bearing of identical size and they establish the stepped bearing as being superior to the Michell bearing so far as load capacity and friction are concerned.

Wilcock (1955)† describes a thrust bearing in which the power loss is very much less than that of the existing conventional thrust bearings. This bearing consists of a stepped oil film with side lands to restrict side leakage, and a pocket of large depth over which the friction is negligible. The analysis described, although allowing for side leakage, is not very exact.

Kettleborough (1955) has described an electrolytic tank investigation into stepped thrust-bearings. The method has been modified for the bearing under discussion and operating characteristics have been calculated.

The results are compared with those for the tilting-pad bearing and stepped bearing and verify that the latter is superior to the others so far as hydrodynamic performance is concerned.

The paper refers to the seven years' period following the nationlaization of electricity supply. The various reasons for plant testing are stated but such work is never an end in itself.

The average heat rates for 33 sets each of 30 MW. and 30 sets each of 60 MW. are stated, together with their standard deviations. Heat rates as low as 7,800 B.Th.U. per kW.-hr. may be achieved in the foreseeable future.

In the seven-year period 77 per cent of new sets were tested, which yields a fair cross-section of the performance of the installed plant. It is the intention to introduce a system of routine testing for the more important machines. In dealing with the subject of boiler tests, available methods are reviewed and consideration is given to recent simplifications of the deferred efficiency test as agreed with the Watertube Boilermakers' Association (W.T.B.A.).

A table of average boiler losses for stoker and for pulverized-fuel firing is presented and from this reliable estimates of boiler performance can be deduced so as to allow for several variables which could affect the answer.

Water flow enters into the testing of both boilers and turbines. The paper refers to standardization of flow metering in the Central Electricity Authority (C.E.A.) and states that five hydraulic calibrating laboratories are being constructed. The Authority is constantly investigating improved means of flow measurement and hopes thereby to contribute to an improvement in technique.

With regard to electrical metering, and in view of the proposal to introduce routine turbine testing, meters should be of the highest quality. N.P.L. certificates have been obtained for a super grade of instrument and results are set out in tabular form. The paper ends by a foreshadow of some possible improvements in electrical integrators.

In this paper two aspects of gas-turbine design and operation are dealt with. First, an account is given of the development of special types of ‘kinematic’ construction suitable for accommodating the very large temperature changes encountered when a gas turbine is started and put on load rapidly. Examples are given of the application of this type of construction to representative components of a 750-1,000 kW. gas turbine.

Secondly, a description is given of the development of combustion chambers to burn fuels ranging from natural gas and gas oil to residual fuels, coal tar fuel and peat. A range of chambers to burn these fuels has been designed to fit interchangeably on the basic components of the standard turbine. A detailed account is given of the development of a combustion chamber of elbow type which takes advantage of the fact that a number of changes of direction through 90 deg. or more are necessary in the ducting of nearly all industrial gas turbines. Development of this chamber is described from early atmospheric test models to the final pressurized combustion chambers which have been fitted to production gas turbines. With only slight adjustments, this type of chamber has shown itself capable of burning hydrocarbon fuels, ranging from natural gas to residual fuel oils, and a modified form of chamber has also successfully burnt low calorific value gases.

In this paper the reasons for attempting to investigate the effect of repeated triaxial stresses are discussed briefly. In particular the case of the thick cylinder subjected to repeated internal pressure, which is of some practical importance, is considered.

The design and development of a machine capable of repeatedly applying to cylinders internal pressures of the order 20 tons per sq. in. (3,000 atm.) are described in some detail. Results are presented for such tests on one particular material and these are regarded as presenting an interim statement, but they are compared with detailed information about the properties of this material under a single and repeated application of uniaxial and biaxial stress. In all cases some attention has been paid to the effect of anisotropy.

The thick cylinder tests give results much lower than might be expected from a knowledge of the properties of the material. A number of tentative explanations are put forward: some of these are at present being investigated.

The initial practical training of the graduate apprentice has been largely a matter of workshop experience and training in craft skills. Such activities do not continue the training of the graduate in the way to which he is accustomed and by which he could best absorb the new knowledge required in industry. The paper describes experiments in methods of instruction for the topics which could be included in the initial three months' graduate training period in any works of which a machine shop is a dominant feature. The methods used followed the typical university teaching pattern of lectures, exercises, and preparation of reports. The topics included the study of the machine shop technology and the study of the machine shop as an operating unit together with the office work, or pre-production activities, concerned with its operation. The experimental methods have been tried out in a factory with a group of twelve undergraduates during a total period of twelve weeks.

It is concluded that training which gives the graduate an opportunity to exercise the special learning abilities he has developed at the university is not only possible but the quickest way to give him the basic knowledge of the technology and organization of production necessary for any mechanical engineering career.

In Part I the relation between the relevant dynamic variables and the power requirements to drive a ball mill, containing the ball charge but without powder charge, has been investigated by means of experiments upon small-scale models.

Dimensional analysis is used to obtain the equations in the form of dimensionless groups, and the functional relations between the groups are given analytically, or by means of graphs.

This analysis indicates that, so far as power input is concerned, the mill with lifters and the mill without lifters are best treated as separate cases, since the motion of the ball charge is different in the two systems, but that the variation in the number and size of the lifters has little effect upon the power required to drive the mill.

By means of the results of this work, the power input required to drive a mill grinding powder may be calculated with reasonable accuracy, and examples of the application of these calculations to large mills are given.

In Part II the relation between the relevant dynamic variables and the power required to drive a ball mill containing both balls and powder charge is established; this being an extension, to the practical case of a mill grinding a powder, of the results, given in Part I of the paper, of an investigation into the dynamics of a mill containing a ball charge only.

By means of the results of Part II the power required to drive a mill, in which a granular material is being dry ground, may be calculated with an accuracy sufficient for design and operational purposes.

The problem of surging in ball mills has been investigated by the use of small-scale models and, although a complete analysis of all the variables has not been made, a criterion has been obtained by use of which it is possible to determine, whilst the design is still on the drawing board, whether the charge of a proposed mill is likely to surge.

The validity of the proposed criterion has been demonstrated by comparison with the published results relating to industrial mills; this comparison showing that all the mills considered are operated under conditions which are in accordance with those laid down in the present work.

So far as is known, such a criterion has not previously been propounded, and it is believed that it is of considerable value to the designer and to the mill operator.

For the purpose of predicting the existence or non-existence of surging in a mill, it is only necessary to establish within which of two possible zones on a certain graph the point corresponding to the operating conditions of the mill falls; the relevant graph having for the abscissae the value of the ratio of the mill diameter to the ball diameter and for the ordinate the product of the mill filling and the mean coefficient of friction of the ball and powder charge of the mill. Thus, in order that a mill shall not surge, it is necessary that the mill filling and the ratio of mill diameter to ball diameter shall be so chosen in relation to the coefficient of friction, which is fixed by the nature of the material being ground, that the operating point for the mill shall fall within the non-surging zone on the graph mentioned.

It is suggested that if a mill can be operated under ‘non-surging’ conditions, then, for lower first costs and maintenance costs, plain liners should be adopted, but if the mill cannot be operated under these conditions, then heavy lifters should be provided to suppress surging—even though the cost is thereby increased.

Remedies which have been suggested for crankcase explosions include the use of warning devices to detect over-heating in the crankcase, and the provision of relief vents to limit the explosion pressure to a safe value.

Data were obtained experimentally at the Thornton Research Centre for the conditions necessary for the ignition of mists of lubricating oil in air in order that the margin of safety offered by a detector of overheating might be more fully assessed.

The venting of explosions was also studied both on a rig scale and in the crankcase of a Diesel engine of 1,100 h.p. The results indicated that it would not be practical to provide sufficient venting area on an engine of this size, to limit the maximum explosion pressures to a safe value for the more violent explosions which might occur. For venting to provide complete protection, it would seem necessary to take measures to prevent the flame from spreading throughout the whole crankcase.

Tests of a crankcase explosion relief-valve 6 inches in diameter, provided with various wire-gauze flame traps showed such flame traps to be ineffective when ***severe explosions were made with mixtures of town gas and air. It was discovered that coating the wire gauze with lubricating oil greatly increased the effectiveness. of the flame trap. An internal, oil-wetted flame trap of less ***than half the original size was fully effective with the most severe explosion that could be produced.

An oil-wetted flame trap, covering an aperture in a dividing partition of a cylindrical explosion vessel fully charged with gas-air mixture, was found to limit combustion to the chamber in which the mixture was ignited, and to limit the maximum pressure to one quarter of the value reached without the partition flame trap.

Internal oil-wetted relief-valve flame traps and partition flame traps were then successfully tested on a ***three-cylinder engine of 9 inches bore. From the results obtained, the requirements of a large engine are estimated and it is concluded that the use of the two new devices would provide adequate protection at acceptable cost.

Detailed surveys have been made of the flow through a low-pressure-rise axial-flow impeller (hub/tip diameter ratio 0·33), which did not produce ‘free vortex’ whirl distribution.

It was found that (1) at all flows, centrifugal forces caused large variations in axial velocity component through the rotor; (2) the angular direction of the flow varied considerably within a short distance of the rotor, and the flow did not stabilize until about two blade chords downstream; (3) as the tip clearance was increased, secondary motion was induced opposite to the motion caused by centrifugal effects and, for clearances greater than 1 per cent of the blade height, seriously restricted the use of radial equilibrium theory.

The measured head-flow characteristics for various sections along the impeller blades did not agree well with those calculated by the following two-dimensional design methods: (1) ‘slip’ theory, (2) ‘cascade’ theory, (3) ‘aerofoil’ theory. Better results were given by the use of a new method which consisted of two steps: (1) calculating the change in axial velocity by the radial equilibrium theory, involving the use of cascade data; and (2) calculating the head-flow characteristics, using the aerofoil theory based on the mean axial velocity obtained from (1).

By this compromise the calculated local performance over most of the blade agreed within 4 per cent of experiment for a wide range of flows between design point and the stall. The new method may be generally applicable to low-pressure-rise fans and pumps, but further experimental evidence will be required to confirm this.

Tests were made on an 11-inch diameter axial-flow propeller pump with impeller and guide blades designed for free vortex conditions, using as a basis the modified aerofoil theory. The best overall efficiency obtained was 82 per cent. Apart from head-flow and efficiency characteristics, measurements were made of velocities and yaw angles within the pump at the design flow of 6 cusecs. and these showed where departures from the theoretical assumptions occurred.

Head-flow characteristics were obtained for various impeller blade-tip clearances from 0·015 to 0·060 inch (0·6 to 2·4 per cent blade height) and it was deduced that secondary flows were not confined to the tip region alone, but extended across the whole annulus.

The pump was on an open circuit, so that cavitation tests were limited. Nevertheless, methods of increasing the resistance to cavitation susceptibility are considered.

It was concluded that despite the fact that some of the assumptions made in the theory are invalid, this method of design may be used with confidence for pumps in the specific speed range of approximately 8,000.

In this paper, it is proposed, very briefly, to review the traditional methods of manufacture of steam-turbine blades, to comment on the different technical requirements of blades for gas turbines and then survey the history of some of the developments in manufacturing methods of blades for gas turbines with particular reference to the methods developed for the production of blades for an early series of axial-flow jet engines.

As by far the larger proportion of blades in an axial-flow aero-engine are compressor blades, this paper is largely concerned with the development of compressor blade manufacturing methods. The paper concludes with a short appreciation of the past and future competitive positions of the precision forged and machined stainless-steel compressor blade.

An electronically controlled automatic feeder is described. Servo techniques are used to keep a weighing machine in continuous balance. The operation is discussed and circuit details are described.

The object of the experiments was to investigate the influence of free-stream turbulence, which it was considered had been overlooked as one of the variables in the field of heat transfer by forced convection from a plane surface.

In order to cover a suitable range of Reynolds number, and to avoid the complication of laminar, transition, and turbulent boundary layers existing simultaneously over the transferring region, an isolated arrangement was used. The influence of a square nose on the plate, with no turbulence grid fitted in the tunnel, was also investigated.

The results are given in Fig. 5 in the form of Nusselt number against Reynolds number.

The mean heat-transfer coefficient has been corrected for conduction (edge) and radiation losses.

A method is presented of calculating the natural frequencies of plane frames. The technique is a numerical one and is based upon tables which were presented by the author (Bishop 1955)^† in a recent article. The present paper is a sequel to the previous one and illustrates a way of using the data given in it. The frequency calculations are not restricted to the fundamental and they are such that a given frame may be checked for possible resonances over a prescribed frequency range without calculation of all lower natural frequencies. When a natural frequency has been found the appropriate principal mode may be calculated. The natural frequencies of the systems which are used for the purposes of explanation were checked experimentally.

The paper commences with an examination of the forces on a brake block applied to the wheel of a railway vehicle. Servo action and certain implications concerning test figures of block friction and adhesion are discussed. The dynamic conditions of equilibrium are shown, if upset, to cause oscillation or ‘chatter’ of the brake gear; an explanation is given of this phenomenon and, with it, a cure.

The economic basis of the design of steam power plant as compared with the situation recorded in Baumann (1949)† is reviewed, and the character of turbine design is illustrated by reference to two selected groups of problems, one related to the behaviour of rotors in service, the other to the design of diaphragms and some associated questions of efficiency.

Two new machines for the testing of metals to fracture under repeated dynamic loading in both tension and torsion are described. These machines are of the ‘falling weight’ type, the yield stress (or yield torque) being measured electronically and the total extension (or total twist) during a particular loading cycle being limited to any desired value.

A preliminary series of tests in both dynamic tension and dynamic torsion, using solid cylindrical bar specimens of 3/8 inch diameter, was carried out on a medium-carbon steel, a mild steel, a silico-manganese spring steel, and two aluminium alloys. For tension, the average time from zero stress to the dynamic yield stress was in the range 0·18-1·00 milliseconds and, for torsion, 0·65-2·67 milliseconds from zero torque to the dynamic yield torque.

In straining specimens to fracture under such repeated dynamic loading, the yield point was found to vary in a manner similar to that for static loading, though all dynamic yield points showed a progressive increase in value with reduction in loading time.

For equal loading times, a greater percentage increase in yield point value was noted for the medium-carbon steel and the mild steel under dynamic torsion as compared with dynamic tension, thus indicating a migration of the criterion of failure away from the Mises-Hencky theory towards the theory of maximum principal stress.

A new technique for visual examination of the progressive development of fretting is evolved and found to be especially useful in following the progress of fretting from the initial stage to the later stages. The qualitative investigation, co-ordinated with a quantitative study in three different atmospheres, dry air, carbon dioxide, and helium, yielded results supporting the mechanism of fretting suggested by the authors in a previous publication. Visual observation also leads to the conclusion that chemical reaction, for example, oxidation, of metal with the environment takes place when the metal is being removed, or after it has been removed, from the surface.

This paper is concerned with stiffened plating to which a bending moment is applied in the plane of the stiffener webs. Owing to the initial or imposed curvature there is a radial component of force which tends to deflect the plating from its original position, and the effects of this deflexion may be considerable. There is a change in the distribution of longitudinal stress across the plate, and a transverse bending stress is induced which may exceed the longitudinal stress. The effects of initial deflexion of the plating, and of lateral load, are also considered.

A numerical example illustrates the application of the analysis to the transverse bending of a ship's side. An experiment is described which confirms the theoretical conclusions for an initially flat plate.

The effect of an abrupt convergence, diameter ratio 2/1, on the local heat-transfer coefficient for water flowing in a straight pipe has been determined for Reynolds numbers from 800 to 100,000 in the smaller pipe, which had an internal diameter of 1 inch. The pipe was heated by the longitudinal passage of a heavy direct current through the wall. The variation in local coefficient, which was comparatively small, is presented and discussed; it does not differ greatly from the variation found in the entry regions of pipes under more normal entry conditions.

The effect of an abrupt divergence, diameter ratio 1/2, has also been determined for Reynolds numbers from 3,700 to 45,000 in the larger (2-inch) pipe. The variation in local coefficient was considerable.

The work described was carried out in the Mechanical Engineering Department of Birmingham University and is an extension of the investigation into air flow in a naturally aspirated two-stroke engine described in an earlier paper (Wallace and Nassif 1954)‡. In that paper the scavenging process of an unblown, opposed-piston engine with simple parallel-pipe exhaust and inlet systems was analysed, and the results were compared with experimental observations. Only a brief reference was made to the very marked increase in air flow resulting from incorporation of a diffuser in the exhaust system.

The present paper describes first a theoretical investigation of wave effects in the exhaust system using: (1) The small-wave theory; (2) The method of characteristics; (3) An approximate analytical treatment designed to give better accuracy than (1), but to avoid the extremely laborious computations associated with (2); and, secondly, an experimental programme comprising: (1) Air-flow tests with various diffuser arrangements and over a wide range of speeds, and (2) Tests in which indicator diagrams were taken at salient points of the exhaust system.

The general conclusions are: (1) The incorporation of a diffuser in the exhaust system leads to an increase in air flow of up to 85 per cent under favourable conditions.

(2) The very complex phenomena occurring in the exhaust system can be dealt with satisfactorily by the approximate analytical treatment with great saving in time over the method of characteristics and very much better accuracy than can be achieved with the small-wave theory.

The performance of a single-cylinder, high-speed, spark-ignition engine has been studied with gaseous methane and propane as fuels. The power output and thermal efficiencies have been determined at two compression ratios, 5·5/1 and 10/1. A comparison for the same engine using liquid fuels is given in Appendix I. Both fuels will knock at high compression ratios; the limits have been determined for all ignitable mixture ratios, and also the effect of inlet temperatures has been studied. Both fuels are ‘temperature sensitive’, the effect being most marked with methane. An experimental comparison has been made of the auto-ignition limits of both fuels by the motored-engine method. Both fuels have been shown to give a stable region of visible cool-flame-type pre-reaction below the temperature for complete ignition. Despite the differing ignition temperature and pressure relationship determined by static experiments, a close similarity is revealed by auto-ignition experiments in the engine. Possible explanations of the differing effects are considered. Infra-red absorption gas analysis has been used to detect the small quantities of methane in the exhaust gas surviving combustion. A very low and constant concentration of methane is present with fast-burning mixtures, owing, it is believed, to valve overlap and scavenge effects. As the rich and weak mixture limits for steady running are approached unburnt methane increases in the exhaust, the effect being most marked near the weak limit. These effects are relevant to problems of atmospheric pollution from engine exhausts, and also assist in assigning the fall in engine efficiency at weak mixtures primarily to combustion inefficiency.

Both methane and propane are shown to be suitable fuels for high-compression-ratio gas engines. Consideration of the combustion properties of these two pure hydrocarbons of simple structure is also of interest in extending the understanding of the general problem of engine knock.

Radioisotopes are radioactive versions of the natural chemical elements, differing from these only in their nuclear mass and in their ability to emit radiations. They are produced mainly by neutron bombardment of ordinary materials in a nuclear reactor.

The most important radiations are beta particles (fast electrons of nuclear origin) and gamma rays (penetrating electromagnetic radiations). These are detected with ionization chambers, Geiger counters or scintillation counters. In production control, the absorption or scattering of radiations is used as a measure of material thickness, and in some cases, of its atomic number.

Radioactive tracers aid in production research in rapid determination of mixing efficiencies and hold-up times. Tracers are also used in the non-destructive testing of materials and components and in more basic research into alloy structure. In this latter research, photographic films are used to produce autoradiographs which show the microscopic distribution of radioactively labelled constituents. Radioactivation by neutron irradiation enables certain impurities to be assayed at sub-microgram levels without chemical separation.

The wear of engine parts can be measured, in a fraction of the time taken for orthodox methods, by making the parts in question radioactive and measuring the worn radioactive debris in circulating lubricant. Cutting tools can be measured similarly by assaying radioactive debris attached to the swarf.

For many purposes gamma radiography supplements X-radiography or replaces it as a cheaper alternative. The greater manoeuvrability of gamma-ray sources is of especial advantage in the radiography of assembled engines.

The paper describes the development of a mechanical-draught water-cooling tower containing a number of novel features. The packing is composed of serrated timber laths down which the water flow is film-wise to avoid splash formation. An earlier paper (Carey and Williamson 1950)§ gave the results of small-scale tests which indicated the theoretical advantage of grid film-flow packings in that the ratio of heat transfer to pressure drop is greater in this type of packing than in any other. A non-splash multi-trough water distributor is used to avoid the carry-over of splash droplets and hence to prevent a local deposition nuisance. Special attention has been given to the even distribution of water and air. Induced draught effects this better than forced draught and, since the fans exhaust heated air, there is no danger of ice formation upon them in frosty weather. The fans are large slow-running and aerofoil bladed, designed to give the highest possible static efficiency and to reduce the loss of velocity head at exit. The main design variables of air velocity and air/water ratio are determined from a balance of capital charges, packing, and power cost.

Several examples, showing the methods of construction, are given of these towers. The main structure and shell is usually in reinforced concrete because of its rigidity in supporting the fans and drives and its permanence and fire resistance, but timber is quite satisfactory and cheaper. The pond is in concrete in all large towers. Timber is the usual material for the packing and distribution troughs, and the water flow is controlled by orifice tubes now made in glass. Methods have been developed for supporting the laths and troughs so that they remain level and in proper relation to each other after wetting. The packing is usually built on reinforced concrete beams across the main structure.

Running experience obtained during six years is given together with the results of a comprehensive programme of tests. The good small-scale packing performance has been confirmed on the large scale. The tests have shown deficiencies in gear-driven fan installations, established the harmful effect on cooling efficiency of recirculation of the warm exit air and, in general, put on record performance details for critical assessment of this new development.