Numerical Evaluation of Fourier Integrals

A method is developed for evaluating Fourier integrals of theform A() = ¹_–1f(x) e^fax dx, 0. The method consists of expanding the function f in a seriesof Chebyshev polynomials and expressing the integral A() asa series of the Bessel functionsJ_r+(), r= 0, 1, 2,.... A partialsum A_N() of the series provides an approximant to A(). The principalfeature of the method is that one set of N+1 evaluations off(x) suffices for the calculation of A_N() for all , and alsothe truncation error A()–A_N() is essentially independentof . Numerical tests show that the method is accurate, economicaland reliable. An application to the inversion of Fourier andLaplace transforms is briefly described.     

Blow up of solutions of a generalized Boussinesq equation

Consider the Cauchy problem u_tt = (f(u))_xx + u_xxtt x R, t 0, u(x,0) = u₀(x), u_t(x,0) = u₁(x),7rcub; where f : R R C, f(0) = ). After treatment of the local existenceproblem, we show the blow up of the solution of the equation(1) under the folowing assumptions. Let > 0 be real, such that 2(l + 2)F(u) uf(u), (v₀, Pv₀)_l² + _- F(u₀)dx < 0 where P = 1 - d²/dx², F'(s), and v₀ is given by u₁(x,0) = (v₀(x))_x. Then we focus on various perturbations of the question. We alsostudy the vectorial case in the same way, and finally we giveexamples.     

The Relative Sizes of the Terms in Chebyshev and other Ultraspherical Expansions

Lanczos (1952) claims that expansion of a function in a seriesof Chebyshev polynomials is usually superior to expansion ina series of ultraspherical polynomials for any possible non-zerovalue of the parameter . We look for conditions on the functionunder which this superiority can be proved and measured; weachieve some success on the assumption that is positive, butthe results for negative are less satisfactory.     

Asymptotic Error Formulae for Numerical Quadrature

A method, due to Fox, is used to derive asymptotic error formulaefor numerical procedures having the form (z+h, h).–(z,h)=(f,z,h).These procedures correspond to numerical quadrature for theintegrand (/h)(f,z,0)and compact expressions are given for determiningthe order of convergence as h 0, and the leading term in theerror. It is shown that a natural generalization of the Euler-Maclaurinexpansion is available. These results are applied to the particularcase where the Pt are polynomials in the differentiation operator.A related interpolation problem is also studied, and it is shownthat in certain cases higher order quadrature formulae are possiblewhen this interpolation problem is not poised.     

Computer extended series solution for unsteady flow in a contracting or expanding pipe

Unsteady flow in a semi-infinite contracting or expanding pipeis reinvestigated using long series analysis. The proposed seriesmethod is useful in analysing the problem for a moderately largeconstant ( = aa/, where a = a(t), the radius of the pipe isa function of time, a(t) is the velocity of the wall, and iskinematic viscosity). For positive values of (expansion ofthe pipe) accuracy of the series representing shear stress andpressure gradient is increased from = 2.89 to = 6.0 by extractingthe singularity followed by completion of the series. For negativevalues of (contraction of the pipe), we revert the series whichresults into the increase of the region of validity of the transposedseries from = -25.0 to = -2.89. Later we use Padé approximantsfor summing them. Also, the asymptotic solution for large valuesof is obtained and it agrees closely with pure numerical valuesof shear stress at the wall and pressure gradient.     

Exact Remainders for Asymptotic Expansions of Fractional Integrals

This is a continuation of work begun in an earlier paper inwhich we used the theory of distributions to derive explicitexpressions for the remainder terms associated with the asymptoticexpansions of the Stieltjes transform. In this paper similarresults are obtained for the fractional integral of order definedby 1f(x)=1/f()^x_o(x-)^x-1 f(t)dt, >. Heref(t) is a locally integrable function on [0, ) and satisfies f(t) a_st-5–0(ó >0), s=0 as      

Short-time Asymptotic Solutions of the Heat Conduction Equation with Spatially Varying Coefficients

An asymptotic solution of the heat conduction equation withspatially varying coefficients is developed for small times.The method followed consists of an application of the Laplacetransformation and use of the Liouville-Green approximationin the subdominant solution of the resulting second-order differentialequation. The approximate solution is inverted by contour integration.The resulting asymptotic expression has a time-dependence identicalto that applying to the case of constant properties providedthat an appropriately averaged value of the thermal diffusivity is used, namely, The error term is of the order of t where is a measure of spatialvariability of the coefficients in the heat equation.     

The asymptotic solution of a family of boundary value problems involving exponentially small terms

In this paper, the authors consider the family of boundary valueproblems in the limit || 0. This problem has recently appeared as a modelfor magnetic field annihilation but the equation itself, withvariously different boundary conditions, has an extensive literature.Using a combination of asymptotic and numerical analyses, thepaper gives a comprehensive treatment of the small || problem,paying particular attention to the question of duality of solutions.For |0, this is intimately connected with the occurrence ofexponentially small terms in the asymptotic solution. When =0(1) these termsz are forced by the boundary layer at y = 1,and the techniques used to deal with this case are well knownfrom previous work on the equation. However, for small ||, acase which reveals the true nature of the duality propertiesof the asymptotic solution, these well-known methods are notapplicable, and a new approach via the initial value formulationof (*) is used. The approach is based on a scaling method whichenables the problem to be reduced to a one-parameter familyof problems of initial value type. This considerably simplifiesthe search for and construction of numerical solutions thatare used to support the asymptotic analysis. For 0, it is shownthat convergence to the =0 solution only takes place for a restrictedrange of values of a and that, for sufficiently small || thereis only one solution to the given boundary value problem.     

Elliptic Problems Involving Phase Boundaries Satisfying a Curvature Condition

Two theorems related to equilibrium free-boundary problems arepresented. One arises as a time-independent solution to thephase-field equations. The other is the relevant time-independentproblem for the Stefan model, modified for the surface tensioneffect. It also serves as a preliminary result for the phase-fieldformulation. Under appropriate conditions, we prove that, givenan appropriate positive constant and a smooth function u: R;,where is an annular domain in R², there exists a curve suchthat u(x)=—K(x) for all x , where K is the curvature.Using this result, we prove the existence of solutions to O=²+ ?(—³) + 2u that have a transition layer behaviour (from=—1 to =+1) for small and make the transition on thecurve . This proves there exist solutions to the phase fieldmodel that satisfy a Gibbs-Thompson relation.     

Application of JMN Approximants to Numerical Initial-value Problems in Linear Differential-algebraic Systems

For f: [0,) R, the J_MN approximant of f(t) is where M and N are non-negative integers and ß_k, _i,K_iare defined constants. Under appropriate conditions on f andprovided Re(_i) > 0 The approximants are the bases of recursions for numerical initial-valueproblems in linear differential-algebraic systems with constantcoefficients. The recursions are stable when N M N–2.Each step of a recursion involves mainly the solution of N/2uncoupled algebraic systems.     

An Asymptotic Analysis of the Delayed Logistic Equation when the Delay is Large

We show how to construct an asymptotic solution to the delayedlogistic equation = y(1–y₁), corresponding to the asymptoticlimit . The results of the analysis are compared with a numericalcomputation, and found to be comparatively accurate for >2. Since the approach adopted is novel, we comment on some featureswhich may be relevant in other problems.     

An analogue of Berry's conjecture for the phase in fractal obstacle scattering

In this paper, the authors consider the high-frequency asymptoticsof the phase s() of acoustic waves scattered by an obstacleRⁿ with fractal boundary. Under certain conditions, it is provedthat if is –Minkowski measurable with –Minkowskimeasure µ then there exists a positive constant C_n, dependingonlyon n and such that where     

The Model Plasma Problem: Existence of a Solution Branch

We are interested in the model plasma problem –u = u⁺in ,u = –d on , _au⁺ dx=j where is a bounded domain in with boundary ; here, j isa given positive number, the function u and the positive number are the unknowns of the problem, and d is a real parameter.Using a variant of the implicit function theorem, we can provethe existence of a global solution branch parametrized by d.The method has the advantage that it can be used for analysingthe approximation of the above problem by a finite-element method.     

Turning Point Problems and Resonance

Consider the boundary value problem: ²yⁿ + (xp(x) + ²f(x, ))y'+ g(x, )y = 0, y(a) = A, y(b) = B, where a < 0 < b, p(x)< p(x) < 0, and p, f, and g are analytic. We investigatethe solution of this problem for small positive values of theparameter . If-g(0, 0)/p(0) c where c N = {0, 1, 2, 3,...},then so-called resonance does not occur, and y = o(ⁿ) on closedsubintervals of (a, b), for any n N, with expected boundarylayer behaviour at the end-points. If -g(0, 0)/p(0) = c, c N, then further transformations of dependent and independentvariables may still expose resonance or non-resonance. The setof necessary conditions that is developed is compared to otherauthors' criteria, most notably, Olver's sufficiency condition,and the necessary conditions of Cook & Eckhaus, Lakin, andMatkowsky. Finally, it is proved that these conditions are necessaryfor resonance.     

A rivulet of perfectly wetting fluid draining steadily down a slowly varying substrate

We use the lubrication approximation to investigate the steadylocally unidirectional gravity-driven draining of a thin rivuletof a perfectly wetting Newtonian fluid with prescribed volumeflux down both a locally planar and a locally non-planar slowlyvarying substrate inclined at an angle to the horizontal. Weinterpret our results as describing a slowly varying rivuletdraining in the azimuthal direction some or all of the way fromthe top ( = 0) to the bottom ( = ) of a large horizontal circularcylinder with a non-uniform transverse profile. In particular,we show that the behaviour of a rivulet of perfectly wettingfluid is qualitatively different from that of a rivulet of anon-perfectly wetting fluid. In the case of a locally planar substrate we find that thereare no rivulets possible in 0 /2 (i.e. there are no sessilerivulets or rivulets on a vertical substrate), but that thereare infinitely many pendent rivulets running continuously from = /2 (where they become infinitely wide and vanishingly thin)to = (where they become infinitely deep with finite semi-width). In the case of a locally non-planar substrate with a power-lawtransverse profile with exponent p > 0 we find, rather unexpectedly,that the behaviour of the possible rivulets is qualitativelydifferent in the cases p < 2, p = 2 and p > 2 as wellas in the cases of locally concave and locally convex substrates.In the case of a locally concave substrate there is always asolution near the top of the cylinder representing a rivuletthat becomes infinitely wide and deep, whereas in the case ofa locally convex substrate there is always a solution near thebottom of the cylinder representing a rivulet that becomes infinitelydeep with finite semi-width. In both cases the extent of therivulet around the cylinder and its qualitative behaviour dependon the value of p. In the special case p = 2 the solution representsa rivulet on a locally parabolic substrate that becomes infinitelywide and vanishingly thin in the limit /2. We also determinethe behaviour of the solutions in the physically important limitsof a weakly non-planar substrate, a strongly concave substrate,a strongly convex substrate, a small volume flux, and a largevolume flux.     

Some Integrals of Gamma, Polygamma and Volterra Functions

Infinite integrals of the gamma and polygamma functions areexpressed in terms of the Volterra and related functions usingthe Laplace transformation method. A number of infinite, convolution,indefinite and definite integrals of v(x), v(x, ), µ(x,ß) and µ(x, ß, ) has been evaluated.A short table of the Volterra function v(x) in the range 0 x 10–0 is presented.     

The Critical Conditions in a Reactive Slab with Slowly Varying Surface Temperature

Liouville's non-linear partial differential equation is consideredfor an infinite rectangular strip domain with a slowly varyingboundary condition. The equation describes a layer of chemicallyreactive material under conditions where the resistance to surfaceheat transfer is negligible and the ambient temperature variesslowly along the surface. Symmetrical heating by a zero orderexothermic reaction is assumed. If is a small dimensionlesstemperature difference between regions where the surface temperatureis effectively constant, a perturbation series solution in may be determined provided the Frank-Kamenetskii parameter satisfies _c(). It is shown that a plausible value for thecritical parameter is _c() = _c(0) e^–e,where _c(0) = 0.878.The corresponding critical temperature distribution is shownto have a dependence on different from that for subcriticalcases.     

Exact Solutions of the Space Charge Equation Using the Hodograph Method

In this paper we show that a hodograph method first introducedby Budd & Wheeler (1987) to develop a new numerical algorithmto solve the space charge equation •()=0 can be employedto derive all the known exact solutions, which have recentlybeen found by Smith (1988). All these solutions are shown tobe separable solutions of the equations resulting from the applicationof the hodograph transformation. We also briefly describe thisnew numerical method and use one of the exact solutions to demonstrateits accuracy.     

Optimal Parameter Choice in the {theta}-Method for Parabolic Equations

The general first-order method, known as the -method, is appliedto the semi-discrete form of a parabolic equation. It is shownthat to every required local accuracy there corresponds a valueof the parameter that is optimal in the sense of allowing thelargest step for which the error remains bounded below . Anasymptotic formula for in terms of is obtained, showing thatthe maximum step-size for the optimal -method is more than twiceas large as that for the Crank-Nicolson method. A numericalexample is given, showing good agreement between theory andpractice.     

The onset of thermal runaway in partially insulated or cooled reactors

The conditions for the onset of thermal runaway in partiallyinsulated or cooled reactors are investigated. The temperaturein the reactor is taken to satisfy a nonlinear elliptic equationand the reaction is modelled by an Arrhenius heat generationterm with finite activation energy. To determine the onset ofthermal runaway, the method of matched asymptotic expansionsis used to derive expressions for the critical Frank-Kamenetskiiparameter _c() for reactors containing either a small coolingrod or having a small cooling patch on their boundary. The theoryused to determine _c() is an extension of the results of Wardand Keller (1991). These previous results of Ward and Kellerare also extended to the case of finite activation energiesby using a numerical scheme to evaluate the coefficients inthe asymptotic results for _c(). In some special cases, the asymptoticexpansions for _c() are compared with numerical results for _c(),and clear agreement is found.