Proceedings of the Institution of Mechanical Engineers

The importance of radiation damage in materials for use in nuclear power reactors is emphasized and an account is given of various types of radiation damage in such materials. The displacement of atoms by nuclear radiation is briefly discussed and examples are shown of recent electron microscopical photographs of radiation damage in materials. A discussion is given of radiation damage in graphite and the importance of stored energy as a reactor problem. The effect of displaced atoms in producing radiation hardening in metals and radiation embrittlement in structural steels is emphasized and discussed in relation to reactor pressure-vessels. Problems of growth, wrinkling, creep, and inert fission gases in uranium are considered in relation to the life of reactor fuel elements.

The paper is concerned with industrial boiler plant generating steam for power and process at pressures of 1000–2000 lb/in² and having a high make-up of treated feed water. The capital and running costs of the water treatment plant and process are therefore much more important than in power station boilers where make-up is small. A few figures are given to illustrate this point and a plea is made that water treatment should be studied carefully at an early stage in any project for new boilers.

The quality of feed and boiler waters now in use, or recommended, is shown in two tables and it is clear that above about 1500 lb/in² the trend is toward ‘pure’ feed water, though this may not be essential.

The next section discusses the control of steam purity to prevent corrosion and deposits, by elimination of carbon dioxide and oxygen, addition of volatile amines, control of the concentration of salts and silica in the boiler water, and the possibility of steam washing.

A short description is given of the circumstances which, it is believed, have caused numerous cases of corrosion in high-pressure boilers, and countermeasures are suggested. These comprise strict deaeration of the feed water, addition of alkalis to the boiler feed water, to prevent corrosion in the pre-boiler equipment and in the boiler, and the avoidance of hot spots or sluggish flow by boiler design and operation.

The paper concludes with a reference to the prevention of corrosion in boilers which are standing during erection or normal operation, and to acid cleaning.

In April 1956 the authors presented to the Institution of Mechanical Engineers a paper entitled ‘Fatigue under Triaxial Stress: Development of a Testing Machine and Preliminary Results’, and in September 1956 a supplementary paper was presented at the International Conference on Fatigue of Metals. In these papers the authors reported tests carried out on cylinders made from a 2 1/2 per cent nickel-chromium-molybdenum steel, which were subjected to (up to) 10 million repetitions of internal oil pressure of (up to) 20 ton/in². Since these papers were published a considerable amount of testing has been carried out on cylinders made from a mild steel, a 3 per cent chromium steel, an austenitic stainless steel, a light alloy, a nearly pure titanium, the nickel-chromium-molybdenum steel in a harder state, and both the nickel-chromium-molybdenum steel and the chromium steel in the nitrided condition. In addition, tests of more academic significance have been carried out on the nickel-chromium-molybdenum steel in an attempt to achieve a better understanding of the extraordinary results which have been obtained.

The present paper is concerned mainly with the presentation of results (supported, of course, by ancillary tests on each material) which are of importance in design. Points of academic interest are discussed only when they are relevant to the practical problem. To make the paper reasonably self-contained, a brief summary of the previous work has been given.

The paper outlines some of the problems encountered in the design and testing of the large ducts associated with present-day gas-cooled nuclear reactors. Various ways of providing ductwork with flexibility are analysed and it is shown that a design involving the use of a combination of tied expansion bellows and cascaded corners is the most economic way of meeting the need for high flexibility and low flow pressure loss. The design of large tied expansion bellows is discussed and an account is given of the tests carried out on a prototype of the bellows now being installed in the gas circuits of the Bradwell nuclear power station in Essex. A brief reference is made to the difficulties encountered in the development of large butterfly valves for reactor ductwork and performance figures are quoted for one of the valve designs which eventually proved successful under test.

There are two distinct modes of operation of a stuffing-box seal. One mode occurs when the axial pressures produced in the packing by tightening the gland-bush are greater than the sealed fluid pressure and the other when they are less. Under the former conditions the fluid pressure distribution is an exponential function of the packed length and the frictional torque increases rapidly with this quantity. With high fluid pressures, however, the relation between fluid pressure and packed length is no longer exponential and most of the fluid pressure drop occurs over the last 10 per cent of the packed length. In addition very little increase in frictional torque occurs when the packed length is increased.

Theoretical expressions have been derived for pressure distribution and friction torque and good agreement has been obtained between these and experimental results.

The paper presents a survey of a number of aspects of the design and construction of feed heaters for present-day steam-turbine cycles.

Many papers have been published in recent years which cover a variety of the detail factors which can enter into the design of a feed heater (for example, the design of tube plates). No attempt has been made to add to such information. The aim of the paper is to present the overall picture and to confine detail description to those items where existing literature is scanty or where the author feels that information of interest can be presented.

Firstly, a methed is discussed whereby the optimum economic size of the individual feed heater can be investigated. This is followed by consideration of the thermal design of the heater nest, and of the pattern of temperature difference introduced by the advent of heaters containing desuperheating and drain-cooling sections. Several methods are shown by which such sections may be arranged in a feed-heater nest.

The remainder of the paper deals with particular aspects of detail construction. The choice of tube materials and their safe working stresses is discussed. The problem of the production of a satisfactory roller expanded tube to tube-plate joint is examined in some detail, together with methods for the consistent production of this joint. An account is given of development work on means for evaluating the soundness of expanded joints by both non-destructive methods and by tube pull-out tests.

A description is given of the investigational work which has taken place and has led to the introduction of welded tube to tube-plate joints with both ferrous and non-ferrous tubes.

Information is presented regarding the amount of torque necessary to achieve desired loadings in heavy bolted constructions and the problem of achieving uniform bolt loading is discussed.

A method of investigation is described which enables the resilience of compressed asbestos jointing to be examined when subjected to a cycle of loading, including heating which simulates working conditions as closely as possible.

The main problems occurring in the design of tube plates and heater headers are examined in a brief discussion.

The thermodynamic principles on which the flash evaporator operates are described and some of the reasons for the recent emergence of this type of distillation plant are explained. It is shown that the performance of a flash evaporator is much less dependent on the number of stages used than is the case for a submerged coil unit. The thermodynamic losses involved in using a finite number of stages, and the effects of other losses which are unavoidable in practice, are described and assessed.

The actual designs of evaporator vessels are described showing means of avoiding interstage piping or ducting despite the use of a large number of stages.

Reference is made to two large-size test rigs built to confirm the basic ideas underlying the development. One was devoted to the exploration of deposition problems and the measurement of heat-transfer data, while the other test rig was used to develop various methods of separating moisture from the vapour produced in the evaporator, and to investigate the flashing-off process. Some of the test results achieved on the rigs are described.

The paper illustrates work carried out on motored and non-motored underground train bogies and the experiments carried out in the use of light-alloy bogies and vehicles.

This three-part paper describes recent work on rails carried out by the Research Department of British Railways.

Part I details the rail sections considered, theoretical considerations of longitudinal bending strength and an analysis of failures in service.

Part II describes a site investigation which was made into the dynamic stresses experienced by rail joints under traffic. Stresses in both flat bottom (F.B.) and bull head (B.H.) rail joints have been measured with both normal and abnormal conditions of sleeper packing. Electrical resistance strain gauges were used at the calculated positions of maximum stress in each fishbolt hole, and at other positions in the joint.

The work initially covered the effects of steam locomotives and steam hauled stock, but was later extended to a second site so that electric locomotives and multiple unit electric stock could be similarly covered.

The investigation brought out the marked influence of speed, wheel diameter and axle load on the magnitude of the stresses induced; and in view of the prospective increase in speeds, and the employment of diesel electric and electric locomotives with heavy axle loads on smaller wheels than as is usual currently, importance attaches to these results and their implications on the life expectancy of rail steel under fatigue loading.

Part III describes fatigue tests which have been made on B.H. and F.B. rails, both plain and when drilled with fishbolt holes. The susceptibility to cracks starting at bolt holes, due to shear stresses in the web, is compared for the two rail sections, and some consideration is given to various methods of increasing the resistance to cracking of this type. It is shown that ordinary atmospheric corrosion greatly reduces the fatigue strength of rails, and any proposed method of reducing bolt hole failures must therefore take this into account.

On the basis that the future development of the diesel engine is inevitably dependent upon its association with a gas turbine, this paper explores three known methods by which the association between the engine and the gas turbine may be achieved, namely, turbocharging, compounding, and the gas generator.

Potential performances attainable from these three arrangements for both 2-cycle and 4-cycle engines with and without charge cooling are derived by methods which are explained, and it is demonstrated that high specific power outputs become possible at high air-manifold pressures without exceeding established limitations for maximum cylinder pressure, piston temperature, or turbine inlet temperature.

The characteristics and limitations of the three alternative arrangements are compared and the conclusion is drawn that a highly supercharged 2-cycle engine used as a gas generator feeding a power turbine would provide an efficient and versatile power unit.

A centrifugal pump of a simplified and unorthodox design has been developed for the immediate satisfaction of special urgent requirements. The experiences gained were encouraging, and this type of pump has been successfully applied to a number of different purposes. Results of tests and recent investigations are given which suggest that some aspects of common centrifugal pump design practice are of lesser importance than are generally considered.

A few more items developed in connection with this pump are described, such as shaft seals especially suitable for high rotational speeds, means to reduce hydraulic friction losses of pump impellers, and a novel pump impeller designed to overcome cavitation difficulties.

The authors of this paper have analysed the relevant heat-transfer processes occurring in gas-turbine combustion chambers and have derived a general equation for flame-tube temperatures. This equation is used to determine the effect on flame-tube temperature of changes in the operating conditions of pressure, inlet temperature, and air mass flow. Changes in aircraft flight conditions, involving simultaneous changes in all these operating variables are also considered. Experimental evidence confirming the predictions of the general equation is provided. The paper concludes with a discussion on the influence on flame-tube temperatures of combustion-chamber design.

To operate at high temperatures, up to 1100°F, bearings must be protected against two sources of early failure: formation of red rust by oxygen attack on the surfaces in either rolling or sliding contact, and rapid wear by local seizure (scuffing) of the cage surfaces in sliding contact with other elements. The first function can be filled for all-ferrous bearings by an atmosphere consisting of a mixture of air and an organic vapour which acts as a reducing agent at elevated temperatures. The performance of the second function requires constituents or additives that form anti-scuffing films.

An olefine oxide polymer was found to be a satisfactory reducing agent over the whole temperature range. Conventional extreme pressure (e.p.) additives are effective in preventing scuffing up to about 900°F. Above 900°F the polymer forms an organic film which itself gives some protection against scuffing. Size 206 tool steel bearings have been operated for 100 hours at 1000°F, 10 000 rev/min under considerable radial and axial load; and for 22 hours at 1100° F. With the bearings available thus far, maximum loads are dictated by softening of the metals above 900°F.

These high temperatures create special problems in the design of the test equipment with respect to fits, warping and dimensional changes. The rig and procedures used in this work are described.

The measuring methods available for obtaining pump performance are examined in the light of experience obtained at the National Engineering Laboratory (N.E.L.). Consideration is given both to experimental procedure and to the interpretation of results.

Since it has been found that pump performance can be markedly influenced by installation conditions, some typical effects are examined.

The number of explosions and failures in steam tubes of ovens remains extremely high, and measures for a reduction of this number are overdue. The lack of safety controls is the main reason for the majority of these accidents, and it is usual, following an explosion, to state that the tube was ‘overheated’. Experimental results or data on metal temperatures on enclosed steam tubes do not appear available, and this paper embodies experimental results in this direction. A great deal of further research is necessary and should be carried out.

Torsional vibration of long rollers stimulated by stick-slip frictional behaviour constitutes a problem in the drafting of fibrous textile materials. A study of the behaviour of long rollers, including the measurement of the periods of stick and slip during vibration is described. With the object of finding a combination incapable of maintaining stick-slip vibrations the detailed frictional behaviour of a number of bearing materials and lubricants was investigated and a laboratory instrument was made to simulate service conditions and indicate the liability to stick-slip vibration of single bearings. Good correlation was observed between the liability to vibration of single bearings in the tester and the performance of multiple bearing assemblies in service. No completely satisfactory solution has been found, although many combinations have been tried.

This paper gives an approximate method of analysis for corrugated pipes and ducts, of the type in which the corrugations have a constant radius of curvature. The method is applicable to corrugations of any included angle, provided that the radius of curvature of the corrugation is small compared with the radius of the pipe. The analysis is carried out using the principle of minimum strain energy, in which the smallest possible number of terms are employed.

The advantage of the method is that results giving extension, bending stress and circumferential stress can be evaluated for corrugations with different included angles, and simply presented in graphical form. These results are then readily available for design considerations covering corrugated pipes with internal pressure, axial load, and bending moment.

The approximate results are compared, where possible, with some results from more precise methods of analysis, and also with some experimental results, and the comparison shows the present method to be sufficiently accurate for design considerations.

With the pressures now being used for central station boiler plant, the feed water pipelines may be excessively stiff and distortion of feed pump bodies is possible owing to the thermal expansion of the pipes when carrying feed water at temperatures up to 500°F. Everything must, therefore, be done to increase the flexibility of the feed pipelines by providing an arrangement which will reduce the pipe thrust and bending moment on the pump terminals, by reducing the pipe bore and increasing the velocity consistent with economic pressure drop, and by reducing the pipe wall thickness to the limit imposed by the safe working stress for the available materials.

The movements of a pair of wheels and axle with and without an electric traction motor passing over a rail joint are investigated. The track is assumed to be resiliently supported and two basic idealized types of joint are considered. In the case of there being a traction motor on the axle, the rotating mass of the armature is regarded as a separate mass resiliently connected to the unsprung masses of the wheels and motor body and the dynamic gear loads are calculated. It is shown that further approximation of actual conditions was possible by a combination of the basic types of rail joint. A concrete example is evaluated and it is seen that the tooth loads are greater with a limited amount of resilience than with an ideally rigid drive, the maximum occurring in the region of the inherent structural resilience of a normal ‘non-resilient’ gear drive. Gear-tooth separation due to dynamic actions is investigated and it is found that the gear loads normally attending such separation are not excessive.

The paper describes an investigation of the influence of belt design on the discharge process of a two-stroke cycle engine. Theoretical methods are developed for calculating the pressure changes in the cylinder, exhaust belt and pipe. A comprehensive experimental investigation of ten exhaust-belt configurations is then described. Pressure measurements were taken in the cylinder, belt and pipe. The experiments showed that the pressure was not uniform within the belt and decreased from the belt to the pipe. It was shown that an increase in belt size reduced the restrictive effects of a belt, shortened the blowdown time and increased the pulse amplitude in the exhaust pipe. There was an optimum belt size above which no further improvement was obtained unless the belt profile was altered, at this size the belt could be made within the normal engine dimensional limits. Data are given for designing exhaust ports and belts based on these experiments. The detailed development of the boundary equations for the gas exchange process is given in an appendix.

This work is concerned with an experimental analysis of tandem mill control equations as developed by Hessenberg and Jenkins. A 3-stand model scale mill was used to study the effect on steady state conditions of three different forms of disturbance, namely front tension changes, screw adjustments, and variations in ingoing gauge. The object of the experiments was to investigate the conditions for maximum inherent stability of the mill as a whole unit. The effect of changes in stand rigidity and motor speed regulation are discussed in detail and conclusions drawn regarding the settings for optimum performance when the gauge is controlled by screw movements in the early stands combined with tension variations in the latter.

In earlier N.P.L. work on the strength of carbon steels at high temperatures, steels showing ‘normal’ and ‘abnormal’ creep properties were assessed from tests extending to 1000 hours. Further results on the same steels indicate that the order of merit of these particular steels is not altered in tests extending to 100 000 hours.

The rupture properties of carbon steels are reviewed in the light of the N.P.L. results and other published data. A wide range of strength properties is shown to exist. The pattern of results, however, appears reasonably consistent. At the lower end of the strength scale are almost pure irons and at the upper end are steels of moderate carbon content and not low in manganese and silicon. Heavy deoxidation with aluminium of conventional carbon steels results in lower creep and rupture properties than are observed in similar steels deoxidized with silicon; the part played by nitrogen in this effect is considered.

The rupture ductility of carbon steels is discussed and in the types covered by the N.P.L. researches, low ductility in long-time service appears unlikely.

Further work required in this field is considered and the need for long-time, low-strain creep properties of mild steels for use in nuclear power plants is emphasized.

The author discusses the exact relationships obtained for the optimum distribution of heater enthalpy rises in a contact heater train for the simple regenerative subcritical cycle, the single reheat cycle with one and several heaters at and above the reheat point, and also for the double reheat cycle. The problem of supercritical cycles and the method of taking exact account of the influence of the feed pumps are also discussed.

The object of this work has been to obtain agreement between the existing theory of tube drawing with a plug, and experimental results, and thus to be able to predict the behaviour of the metal when conditions of drawing are altered. As the basis of theoretical study Sachs's approximate theory has been chosen, and mild steel containing 012 per cent carbon has been used as the test material.

The use of the theory requires the knowledge of the true mean yield stress of the metal undergoing plastic deformation, as well as that of the value of the coefficient of friction. The method of establishing those two quantities is discussed in detail. Drawing stresses are calculated and it is shown that close overall agreement exists between theory and practice and that, therefore, the mathematical formulae can be used with confidence. The components of the total work done in drawing a tube are determined and discussed, and the adverse effect of sink on the magnitude of redundant work is indicated.

The operation of the first reactors at Calder Hall revealed a significant leakage of carbon dioxide coolant. Tests have been carried out on subsequent reactors to establish the leakage rate and to trace and minimize this loss of coolant gas. This paper deals with the tests carried out on Calder reactors 3 and 4 and Chapelcross reactors 1 and 2. The authors outline the various methods used to determine the rate and to locate the origin of gas leaks.

A number of conclusions have been reached on design aspects and several suggestions are made for improving gas tightness of future plants of a similar construction.

A review is given of existing theories of drawing under conditions of plane strain†, and these theories are extended in certain respects.

Comparison between accurate and approximate theories in which work hardening is neglected, shows that one of the approximate theories is reasonably accurate in predicting the distribution of pressure on the die, and the effect of friction on the drawing load.

A new approximate method is proposed to allow simultaneously for work hardening, friction, and redundant work, in calculating die pressures and drawing loads. Redundant work is allowed for by suitably modifying the stress-strain curve of the metal. Formulae are derived for drawing through a wedge-shape die, using analytical approximations for the stress-strain curve. Similar formulae have not yet been found for dies with a cylindrical profile.

The possible errors involved in calculating a mean coefficient of friction from the measurement of forces acting on each half of a wedge-shape die are analysed.

An analytical equation is derived for the prediction of maximum reductions possible as dependent on the process variables. The results of this equation are presented in graphical form for a wide range of the process variables. A simple experimental procedure is suggested for the determination of the coefficient of friction.

The process variables are: The roll radius R₀, initial t _i and final t _f strip thickness, material yield limit σ₀ at uniaxial tensile test with yield limit σ_oi at initial thickness and σ_of at final exit thickness, front pull stress σxf and back pull stress σxb, the coefficient of friction μ, and the rolling speed U.

The power balance is set for the rolling operation. The powers computed are:

(1)

the useful power consumed as internal work of deformation of the strip, called sometimes ‘The homogeneous power’;

From the equation: sum of the powers = 0, a dependence of the process variables one on each other is established. From this dependence the coefficient of friction can be computed at maximum reduction, and this gives a simple procedure for the experimental determination of the average coefficient of friction. With the coefficient of friction, yield limit, dimensions of the strip and roll and the values of the front and back pull known, from a given graph the maximum reduction possible can then be computed or picked.

The pulsation problems inherent in installations comprising large and relatively slow speed-reciprocating compressors or engines, both on the induction and delivery side, have been recognized for many years. The problem of pulsating flow has recently become particularly acute with the large-scale introduction of free-piston gas generators which, because of their unique mode of operation and the fact that they are usually installed in groups drawing from a common air manifold and delivering to a common exhaust duct, are particularly sensitive to such pulsations and where a rigorous approach to the problem is imperative. The present paper is concerned exclusively with intake systems which usually consist of two distinct elements, namely the individual damping system for each gas generator and a common ducting system for the entire group of gas generators, from which the individual damping systems draw their air supply. The individual damping systems take the form of one or two large smoothing chambers or damping capacities connected by Venturi-shaped ducts to each other (when two chambers are used) and to the common ducting system, the gas generators drawing air from these smoothing chambers through pipes which may be arranged to give some ram effect with resultant improved volumetric efficiency. The object of the individual damping systems is to ensure that the magnitude of the pulsations reaching the common ducting system is sufficiently low to prevent harmful pulsations which could cause damage to air filters, or even to property in the vicinity, particularly if resonances are excited in the common ducting system.

A typical intake system for an eight gas generator installation is shown in Fig. 1. The paper is concerned solely with the method of analysis applicable to the individual damping systems, not with the external ducting system. The latter should present no difficulties if the individual systems are properly designed. The method of analysis is based on mechanical vibration theory, the governing equations for the damping systems—conveniently referred to as acoustic systems—being identical with those of better known mechanical systems. The equivalent electrical systems are given in an appendix, with a view to facilitating the analysis of damping systems with the aid of analogue computers. The object of the analysis is first to put the design of such damping systems on a sound theoretical basis and secondly to facilitate the calculation of optimum dimensions, given permissible pressure drop and pulsation amplitude outside the system, as well as the form of the suction pulse of the gas generator. The latter is most conveniently represented as a harmonic series using Fourier analysis.

The present paper is based on the author's own experience in this field and, though restricted to purely theoretical considerations to avoid excessive length, the material contained in the paper is confirmed by a substantial body of experimental work.

The Summer Meeting of the Institution was held in Northern Ireland from 20th to 23rd June 1960.

The Proceedings commenced at 2 p.m. on 20th June in the City Hall, Belfast, with a welcome by the Deputy Lord Mayor of Belfast, Alderman M. K. Wallace, The President, Professor Owen A. Saunders, D.Sc. (Eng.), M.A. (Member), F.R.S., responded, and he was supported by the Chairman of the Branch, Mr. D. H. Alexander, O.B.E., M.Sc., Wh.Sch.

The shear stress-strain relationship for an initially stress-free 2 1/2 per cent nickel-chromium-molybdenum steel under reversed stress has been determined by twisting thin tubes. The effect of a low-temperature heat treatment given to this material after various amounts of overstrain has been similarly determined.

The information so obtained has been used to calculate by methods outlined in Appendixes I and II the behaviour of thick cylinders of the same material when subjected to one or two cycles of internal pressure, with or without an intervening low-temperature heat treatment, and to calculate the stresses throughout the material of these cylinders at any stage in the processes.

Such cylinders have been made and subjected to the treatment described, the pressure and dimensional changes being measured at all stages. They have then been bored out in steps until all the overstrained material has been removed, the corresponding dimensional changes being again measured.

One of the assumptions used in the calculations has an important effect on the values of the residual axial stresses; moreover the axial dimensional changes during the boring-out process are so small that great accuracy cannot be claimed for the values measured. To this extent the results are inconclusive. It is, however, shown that both the dimensional changes and the shear stress distribution (from which the radial and hoop stress distribution can, of course, be deduced) at any stage in the process can be calculated with satisfactory accuracy by the relatively simple approximate methods described.