TUBES FOR HIGH TEMPERATURE INDUSTRIAL APPLICATION AND METHODS FOR

PRODUCING SAME"

FIELD OF THE INVENTION

The invention relates generally to a tube construction for high temperature industrial application, such as for example in reformer tubes.

BACKGROUND OF THE INVENTION

Pipework in industrial plant which operates at high temperature and is also subjected to stress will experience a progressive damage mechanism known as creep. For example in vertical runs of pipework or tubes (hereinafter referred to as tube(s)) creep can occur downwardly i.e., in-axis due to gravity. This may occur in plant carrying out direct reduction of iron ore (DRI plant) for example. Creep may also occur across the tube axis where the tubes are subjected to internal pressure as well as high temperature, such as in reformer plant such as in catalytic or steam reforming.

By creep is meant slow migration of material of the tube wall so that after a period of operation of the plant, a tube (or tubes) of, for example, constant wall thickness over its length at the beginning of its life, at the end of its life will have exceeded allowable dimensions or even rupture, thus requiring replacement.

The primary mode of failure of tubular creep samples under internal pressure is diametral creep, where cracks are oriented along the longitudinal axis of the pipe. Longitudinal expansion is another mode of failure of tubular creep samples. When diametral creep is restricted by helical wire windings such as that shown in Figure 1, longitudinal creep is greatly accelerated and the failure mode becomes cracking in the hoop direction caused by longitudinal stress.

The invention aims to ameliorate at least some of the problems mentioned above or at least provide an alternative choice for the public.

International patent application publication WO WO2010/101482 discloses a metal alloy tube construction for use in high temperature environments comprising a refractory material which may comprise wire or mesh, around the tube as a reinforcement material. SUMMARY OF THE INVENTION

In a first aspect, the invention broadly comprises a metal alloy tube construction for use in high temperature environments, comprising around the tube a layer of reinforcement material wound around at least a portion of the metal alloy tube,

wherein the reinforcement layer comprises a refractory material provided as a braided sheath, the braided sheath being formed by interweaving wire bundles in an overlapping pattern, wherein each wire bundle comprises a plurality of wires.

Preferably the wires in each wire bundle are arranged side by side so that there is no or minimal gap between adjacent wires. Preferably each wire bundle has a substantially uniform thickness across its width. In another form, the plurality of wires in each wire bundle may be twisted or braided or bonded before they are braided into a sheet form.

In one embodiment each wire may comprise a plurality of filaments. The plurality of filaments may be bonded together or twisted along their length or are arranged side by side to reduce the gap between adjacent filaments.

In some embodiments each wire bundle comprises about 2 to 50 wires, or about 5 to 40 wires, or about 5 to 30 wires, or about 8 to 25, or about 8 to 20 wires arranged side by side.

Preferably the braided sheath comprises about 2 to 80 such wire bundles, or about 2 to 70 wire bundles, or about 2 to 60 wire bundles, or about 10 to 50 wire bundles, or about 20 to 48 wire bundles. In one form the wires each comprise a similar cross section dimension. Preferably the diameter of the wires is in the range of about 0.1mm to 1mm. More preferably the diameter of the wires is in the range of about 0.12mm to 0.5mm, or 0.12mm to 0.28mm.

Preferably at least some wires or wire bundles are provided at an angle relative to the longitudinal axis of the tube.

Preferably the braided sheath comprises at least some wires which extend at an angle of approximately +2° to 88° to the longitudinal axis of the tube, wherein the angle is measured relative to the longitudinal axis (0°) of the tube and defined from 0° to 90°. More preferably at least some wire bundles extend at an angle of about + 10° to 80°, or about +20° to 75°, or about +30° to 70°, or about +35° to 70°, or about +40° to 70°, or about +45° to 70°, or about +50° to 65°, or about +50° to 60°, or most preferably at an angle of about 50° to 55° relative to the longitudinal axis of the tube.

In one embodiment the reinforcement layer has a cover factor of less than 100%.

Preferably the reinforcement layer has a cover factor of more than 70%, or preferably more than 80%, 85%, 88%, 90%, 93%, 95%, or 98%.

The reinforcement layer comprises a refractory material such as stainless steel, tungsten, molybdenum, niobium, tantalum, columbium, hafnium, or rhenium, or metal oxides such as alumina (Al ₂ 0 ₃ ), or carbides such as tungsten carbide (WC). Preferably the refractory material comprises a substantially greater creep rupture life than the tube material.

In one embodiment the reinforcement layer may comprise an oxidation sleeve to substantially isolate the refractory material from a surrounding atmosphere.

By "cover factor" is meant the percentage of the area that is covered by the braided sheath of reinforcement material per unit area, judged from perpendicular to the plane of the braided sheath when laid out flat. Thus a cover factor of 90% means in each unit area, 90% of the tube exterior surface is covered by the wire bundles of the braided sheath whereas 10% of the tube exterior surface is not covered by the wire bundles. By 'high temperature' in this specification is meant typically temperatures above 500°C and typically in the range of 750-1500°C.

By Yefractory material(s)' is meant materials which will retain their strength at

temperatures above 1000F (538°C).

As used here the term "and/or" means "and" or "or", or both.

As used herein "(s)" following a noun means the plural and/or singular forms of the noun. The term "comprising" as used in this specification means "consisting at least in part of". When interpreting statements in this specification which include that term, the features prefaced by that term in each statement all need to be present but other features can also be present. Related terms such as "comprise" and "comprised" are to be interpreted in the same manner.

BRIEF DESCRIPTION OF THE FIGURES

An exemplary embodiment of the invention will be described with reference to the drawings, in which :

Figure 1 is a tube reinforced with helical windings;

Figure 2 is an exemplary embodiment of the invention;

Figure 3 is another exemplary embodiment of the invention;

Figure 4 is another exemplary embodiment of the invention;

Figure 5 is another exemplary embodiment of the invention;

Figure 6 shows the specifications of each reinforcement sheet used in Figures 2-5;

Figure 7a shows the stress-strain curves for the woven reinforcement sheets used in Figures 2-5;

Figure 7b shows a table of the measured longitudinal stiffness of the stainless steel reinforcement;

Figure 8 shows the results of the periodically interrupted F-42 pressurized pipe test as a plot of diametral and longitudinal strain versus time;

Figures 9 shows mean creep rates in the diametral direction;

Figure 10 shows mean creep rates in the longitudinal direction;

Figure 11 shows a crack running along the length of the control pipe oriented in response to hoop stress;

Figure 12 shows indentations left by the helical winding reinforcement on the pipe surface, and a crack oriented in response to longitudinal stress;

Figure 13 shows measured longitudinal stiffness E i _ong plotted against reinforcement angle, along with curves generated for various E ₂ (in the order of 3 GPa);

Figure 14 shows measured mean tangential, longitudinal and radial creep rates plotted against initial reinforcement angle (solid points) and final angle (unfilled points), with creep rate trends superimposed. Shaded bands cover range of predicted behaviour from initial angle (solid line) to final angle at t = 820 hours or failure (dashed line). Creep strain data from the interrupted F-42°test is superimposed, with x and + symbols, showing tangential and longitudinal creep rates, respectively;

Figure 15 shows measured mean effective creep rates and calculated effective creep rates plotted against reinforcement angle. DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

Referring to the drawings a tube (not shown) reinforced with the invention is shown in Figures 2-5. The tubes may be brass tubes formed of an alloy such as for example, an alloy comprising a nominal composition of 65% by weight copper and 35% by weight zinc, with a copper composition typically in the range of 64-68.5%.

In other examples the tubes may be formed of a corrosion resistant alloy such as for example, an alloy comprising in the range 23-26% by weight chromium, 32-36% by weight nickel, and 0.35-0.4% by weight carbon, and the alloy may also comprise about 1.5% by weight manganese and about 1.5% by weight silicon. The alloy may also optionally comprise other 'micro-alloying' additions. The balance of the alloy comprises iron. In one particular embodiment metal alloy tube is formed of a corrosion resistant alloy comprising approximately 25% chromium, 35% nickel, 0.4% carbon, and 39.6% iron. Typically, the unreinforced metal alloy tube is suitable for use at temperatures above 400°C and typically at temperatures in the range of 750-1100°C and the tube may be suited for use with internal pressures of 45 bars, for example. The tube may be a reformer tube for a catalytic reformer, for example, which contains a catalyst which is contacted by a process gas stream flowing through the reformer tube.

Referring to Figures 2-5, around the exterior of the tube over at least a part of the length of the tube is a layer of reinforcement material 1. The reinforcement layer comprises a refractory material provided in a braided sheath form, comprising a refractory metal such as stainless steel, tungsten, molybdenum, niobium, tantalum, or rhenium, or of alumina which has a greater creep resistance than the tube material. According to the invention, the reinforcement layer 1 is formed by interweaving wire bundles which each comprise a plurality of wires in a diagonally overlapping pattern as shown. Other weaving or braiding patterns may also be used to prepare the reinforcement layer. The braided sheath is then wound around at least a portion of or the entire length of the tube to reduce creep.

As shown, each wire bundle used in the braided sheath comprises a plurality of wires 2 which are of a similar cross section dimension and are substantially arranged side by side so that there is no or minimal gap between adjacent wires. Each wire bundle comprises about for example 2 to 50 wires, or about 5 to 40 wires, or about 5 to 30 wires, or about 8 to 25, or about 8 to 20 wires arranged side by side.

In the embodiments shown in Figure 2 and 4, each wire bundle comprises 8 individual wires arranged side by side and then braided or woven into a planar sheet. In the embodiment shown in Figure 3, each wire bundle comprises about 11 wires and the embodiment shown in Figure 5 comprises about 15 individual wires. In the examples shown each wire bundle comprises about the same number of wires. This is so that the resulted braided sheath can have a substantially uniform thickness thereby allowing the stress to be evenly distributed in the braided sheath.

In the embodiments shown in Figures 2-5, the wires used in each wire bundle are simply arranged side by side and are not bonded or joined prior to being woven into a planar sheet. Alternatively, the wires may be joined or bonded or twisted along their length before being woven or braided into a planar sheet.

The wire bundles are tightly woven or braided so that they overlap at wire bundle crossings and cover a major or a substantial area of the tube exterior. In some

embodiments the wire bundle crossings extend in a first direction which is substantially parallel to the longitudinal axis of the tube, and/or in a second direction which follows the circumference of the tube. At each wire bundle crossing where two wire bundles intercept, a first wire bundle extends in a first direction which is at an angle Θ relative to the tube axis, and a second wire bundle extends in a second direction which is at an angle of -Θ relative to the tube axis, i.e. the two wire bundles which meet at the crossing are symmetrically arranged about the longitudinal tube axis.

The wires may be of a diameter in the range of about 0.1mm to 1mm, or may be of a diameter of in the range of about 0.12mm to 0.5mm, or about 0.12mm to 0.28mm. The braided sheath may comprise about 2 to 80 wire bundles, or about 2 to 70 wire bundles, or about 2 to 60 wire bundles, or about 10 to 50 wire bundles, or about 20 to 48 wire bundles. The number of wire bundles used in each exemplary embodiment is shown in Figure 6. In the examples shown, at least some wire bundles of the reinforcement layer are provided at a non-zero angle, which is measured relative to the longitudinal axis (0°) of the tube and defined from 0° to 90°, so as to resist creep both in diametral and loop direction. At least some wire bundles extend at an angle of approximately + 2° to 88° to the

longitudinal axis of the tube. In other embodiments, at least some wire bundles extend at an angle of about + 10° to 80°, or about + 20° to 75°, or about + 30° to 70°, or about + 35° to 70°, or about + 40° to 70°, or about + 45° to 70°, or about + 50° to 65°, or about + 50° to 60°, or most preferably at an angle of about + 50° to 55° relative to the longitudinal axis of the tube. The reinforcement layer 1 generally has a cover factor of less than 100% due to its braided nature. Preferably the reinforcement layer has a cover factor or surface area coverage of more than 70%, or preferably more than 80%, 85%, 88%, 90%, 93%, 95%, or 98%.

The strain and stiffness of each reinforcement layer is determined by factors such as number of wires per wire bundle, number of wire bundles used in the braid, the angle between the wire bundles and the axis of the tube, cross section dimension of the wires, material used for the wire, the cross section dimension and length of the filaments in each individual wire, crimp angle, and the overall cover factor of the reinforcement layer 1. These parameters can be adjusted to achieve an optimum performance.

The reinforcement has inherently significantly higher creep resistance than the metal alloy tube, such as 20%, 50%, 100% or higher creep resistance, and possibly 2-3 orders of magnitude higher of creep resistance, at temperatures above about 40% of the absolute melting point of the metal alloy tube. The reinforcement thus assists in inhibiting downward (i.e., longitudinal or axial) creep where the tubes are vertically mounted, and also assists in reducing diametral creep where the tubes are subject to internal pressure during plant operation. The reinforcement thus acts to prolong the effective working life of the tubes and the plant of which the tubes are a part of.

In the embodiments shown there is a single layer of reinforcement, in a braided sheath form. In alternative embodiments the reinforcement may comprise two or more such braided layers.

Tubes of the invention may be used in catalytic reformers in oil refineries, in which the tube may carry a vaporising crude oil and hydrogen mixture at a temperature up to 1000°C and pressure up to 45 bars, or in reformers in hydrogen production, methanol production, ammonia production or ethylene production for example, or in other industries. Tubes of the invention may be used in steam catalytic reformers. In such applications the tubes may exhibit increased creep resistance, higher strength, and/or higher resistance to corrosion such as oxidation, at temperatures of use, relative to the equivalent un- reinforced metal alloy tube. Tubes of the invention may also be used in high temperature heat exchangers, for example in hydrogen production in jet engines, or in solar thermal energy production, for example in solar thermal high temperature collectors.

Reinforced industrial tubes of the invention may be manufactured by centrifugal casting and then placed on a winding machine to wind the one or more braided reinforcement sheets around at least a portion or the entire exterior of the tube. A gas diffusion barrier layer may be applied to the interior of the tube by, for example, thermal spraying, or to the exterior before the reinforcement is applied to the tube. In another form, reinforcement material is first shaped to a tubular form, for example by winding or wrapping about a mandrel, and the reinforcement tube is then placed inside a centrifugal casting mold. A metal alloy tube is then centrifugally cast against the interior of the tube of the reinforcement material. A gas diffusion barrier may then be applied to the interior of the tube.

The tubes may also be manufactured by extrusion and reinforcement winding or wrapping, in which the metal alloy tube is extruded and then placed on for example a winding machine where one or more layers of braided reinforcement material are wound around the tube. The tubes may also be manufactured by co-extrusion, by passing the woven reinforcement material through an extrusion die as the metal alloy tube is extruded, so that the reinforcement is encased within the metal material of the tube wall. Optionally after co-extrusion a further layer or layers of reinforcement may be formed around the exterior of the tube. A gas diffusion barrier layer may be applied to the interior or exterior of the tube before or after the reinforcement material.

The reinforcement layer may be made up of different refractory materials, so a functionally graded composite may result.

For ease of manufacture and testing, brass tubes reinforced with stainless steel wires are tested and the methodology and results of which are discussed under the Experimental section below. In another embodiment, a stainless steel tube reinforced with a tungsten wire braid sheath also achieved satisfactory results. It is preferred to isolate the tungsten wires from the surrounding atmosphere by an oxidation sleeve as oxidation of refractory materials can be a limiting factor.

The tubes shown in the drawings have a circular cross section but in other embodiments the tubes may have an oval or multi-segmented cross-sectional shape. While in describing the tubes, vertical mounting applications thereof have been referred to and the tubes are suitable for use in industrial plant in which the tubes extend horizontally or at an angle between the vertical and horizontal.

The reinforcement may be applied over substantially the full length of a tube such as a reformer tube or over a major part of the length of the reformer tube. Alternatively the reinforcement may be applied over a minor part of the length of the tube, at or towards one end for example and typically an end further along the tube length in the direction of gas flow through the tube in use. The number of layers of the reinforcement may also vary over the length of a tube to provide for optimum performance of the tube under operating temperature and pressure.

Typically for mounting the tube the tube will have flanges or other mechanical mounts at either end thereof. Experimental

The invention is further illustrated by the following description of experimental work.

Figure 1 shows a tube reinforced using a helically wound refractory metal. Such helical windings provide about 5 to 7 times life extension over a monolithic pipe of equal dimensions in an accelerated test of 4000 hours. For further detailed proof of concept, model materials were selected for ease of manufacture and testing. Brass pipes (65% by weight copper and 35% by weight zinc) are reinforced with 304 or 316L austenitic stainless steel. Life extensions of greater than 10 times have been achieved with the embodiments shown in Figures 2-5.

Figure 6 lists the details of the reinforcement layer used in each embodiment. The table lists the reinforcement layer details such as surface area coverage, braid/weaving angle, wire packing fraction, longitudinal stiffness, wire size, number of wire bundles and number of wires per wire bundle. Reinforcement angle is measured relative to the tube axis (0°) and defined from 0° < Θ < 90°. Braid naming convention is selected based on the wire coarseness: C denote "coarse wires" and F denote "fine wires".

While the helically wound reinforcement serves as an extremely effective means of restricting diametral expansion, it has low longitudinal stiffness and thus very limited ability to control longitudinal expansion. By altering the reinforcement angle, the woven reinforcement sacrifices some of this diametral stiffness, but gains the ability to more effectively constrain the pipe in the longitudinal direction, particularly due to its

interlocking nature. Longitudinal stiffness measurement

The reinforced tubes are pressurized and creep rupture tested at 400°C and pressurized to 2MPa. Tensile testing is performed on the reinforced tubes in an 810 Material Test System using customized grips to hold the ends of the reinforcement layer at the diameter of the brass pipe. The samples consisted of a 140mm length of 304 SS braided sheath pulled over a 100mm length of brass tube. With the tube left free in the middle of the braided sheath, the excess braided sheath is clamped at either end and elongated at 3mm/min to produce a stress-strain curve. Strain is measured from crosshead displacement.

Converting measured load into stress is achieved through knowledge of the wire diameter and the number of wires present in the braided sheath. This sectional area is projected onto the direction of the applied load by dividing by cos6, wherein Θ is the angle between the wire bundles and the longitudinal axis of the tube. Pressurized pipe tests

The tube samples were placed in a horizontal tube furnace at 400 °C, pressurized to 2MPa with argon and allowed to creep until rupture, or 820 hours had elapsed. Internal pressure and temperature are logged every 60 seconds during all tests, and maintained stability to within ±0.2 MPa and ± 1 °C. A sustained pressure drop indicated pipe rupture, in which case the test was ended. Unlike conventional uni-axial creep tests and some more sophisticated multi-axial creep tests where strains may be measured in- situ, the tubes reinforced with a braided sheath made in-situ diametral strain

measurement of the pipe problematic. Longitudinal strain was measured over a 50mm gauge length scribed across the mid-span of the tube, in the middle of the hot zone of the furnace. At least 5 measurements are made using digital callipers and the readings are then averaged . Measurements of the inside and outside diameter were made on a length of pipe sectioned from the midspan and positioned vertically in a 3-jaw chuck. A Giddings and Lewis Discovery D12 Cordax Series coordinate-measuring machine (CMM) with an 0 8mm probe is used to take 12 diametral coordinate measurements to 1 m resolution. These diameter measurements allowed diametral, as well as radial strain to be determined.

In order to help understanding of strain behaviour during testing, an F-42 sample was interrupted periodically and the reinforcement was removed to enable strain

measurements on the pipe. This method of periodic interruption to take strain

measurements has been used previously with some success on unreinforced metallic creep rupture specimens under internal pressure. To quantify the effect of this periodic interruption, two additional tests of F-42 samples were performed without interruption. In this case, as well as further tests, final maximum strains in the diametral, longitudinal and radial directions are determined upon failure.

Three unreinforced control pipes are tested to rupture, along with a single helical wrap- reinforced pipe. C-65 and F-42 reinforced pipes are both tested twice, whereas C-36 and F- 50 reinforced pipes are tested once only. All reinforced tests are ended if rupture had not occurred after a 10 times life extension over the average control test (approximately 820 hours).

Results and Discussion

Longitudinal stiffness measurement

The stress-strain curves for the woven braid reinforcements are shown in Figure 7a; Figure 7b shows a table of the measured longitudinal stiffness of stainless steel reinforcement. Two vertices were identified on each curve. The region before the first vertex was attributed to a period of the tightened weave, and is not used for further analysis involving longitudinal stiffness of the braided sheaths. The slope of the linear region between the two vertices, where the braided sheath is fully engaged on the pipe, was denoted E | _0ng and measured over a minimum of 100 data points well away from the vertices. Figure 8 shows the results of the periodically interrupted F-42 pressurized pipe test as a plot of diametral and longitudinal strain versus time.

Pressurized pipe tests

The average life of three creep rupture control tests is 83 hours, which is significantly lower than the predicted life of 130.1 hours. The creep life prediction was made based on data for a slightly different alloy composition (70-30 rather than 65-35) and did not account for microstructural influence such as grain size. However this inconsistency is of little concern, as the purpose of these unreinforced control tests was to provide a creep rupture life baseline to which reinforced cases could be compared. A post-mortem maximum diametral expansion of 12.85% was observed at the midspan of the ruptured unreinforced pipe, with a corresponding 0.3% longitudinal expansion over the gauge length. The results of the periodically interrupted F-42 pressurized pipe test are shown in Figure 8 as a plot of diametral and longitudinal strain versus time. The average uninterrupted F-42 ° sample strained 12.4% in the diametral direction and -6.14% longitudinally. Similar to typical creep response, there is a region of primary creep, followed by secondary, or steady-state, creep. As expected, the absence of tertiary creep is consistent with the fact that the sample has not ruptured.

Figure 9 and 10 show mean creep rates in the diametral and longitudinal directions, respectively, for reach reinforcement type. Error bars denote variation in repeated tests. All of the reinforced pipes reached a 10 times life extension over the control pipe without rupturing, at which point testing was ended due to practicality. The helically wound reinforcement denoted Wrap-86° failed after 647 hours. The data shows that diametral creep decreases with increasing reinforcement angle, and that longitudinal creep increases with increasing reinforcement angle. Both diametral and longitudinal creep appear to be minimized between the 50° and 65° samples. The braid architecture parameters such as wire diameter, number of strands and number of wires per strand appear to have little effect on the observed trends, reinforcement angle appears to be dominant.

In the case of the helical winding and C-65 reinforcement, a substantial reduction in average diametral creep strain compared to the control pipe was observed, while the average longitudinal strain rate increased. A comparison of the periodically interrupted F- 42 strain measurements to those from its uninterrupted counterpart shows that regularly interrupting the test and removing the reinforcement layer to take strain measurements has little overall effect on the final strain. At the end of the interrupted F-42 test, 13.38% diametral strain and -8.19% longitudinal strain was recorded .

Comparatively, the average uninterrupted F-42 test strained 12.39% in the diametral direction and -6.14% longitudinally. Furthermore, this interrupted test reveals that creep rate is approximately linear, with only a small region of primary creep. The large region of steady, linear creep suggests there is no clear engagement point where creep is

completely halted in one direction thereafter, but merely that the presence of the reinforcement slows down creep dramatically.

Figure 11 shows a crack running along the length of the control pipe at a point of maximum diametral expansion, suggesting rupture due to diametral creep as expected.

Figure 12 shows indentations left by the helical winding reinforcement on the pipe surface. The helical winding reinforced pipe obtained a life extension of approximately 7.8 times over the control pipe before rupturing due to longitudinal creep, with a crack running around the circumference. This shift in the primary mode of failure between the control and wrap-reinforced pipe from diametral to longitudinal creep affirmed that diametral creep had effectively been arrested by the reinforcement, and identified longitudinal creep as the new failure mode. On the other hand, all braided sheath reinforced pipes completed a 10 times life extension without failure, upon which time the test was halted intentionally.

Stiffness analysis

The stiffness properties of the braided sheaths are strongly influenced by the angle between the wire bundles and the longitudinal axis of the tube. As a simple model, the braided sheath geometry may be considered as a laminate of 2 lamina oriented at ±θ. This permits classical lamination theory (CLT) to be used to determine the stiffness

characteristics in the global coordinate system. The primary differences between the ideal laminate considered by CLT and a braided sheath reinforced tubular structure are the lack of matrix material in the braided sheath to bond the fibres together and fix the reinforcement angle and also the cylindrical geometry of the braided reinforcement. However, stiffness perpendicular to the fibre direction is still nonzero in the case of the braided reinforcement due to frictional interactions between wire bundles and the overlapping, interlocking nature of the basket weave structure. As the radius of the pipe is much greater than the thickness of the braid, the planar

approximation can be used.

Figure 13 shows measured longitudinal stiffness E i _ong plotted against reinforcement angle, along with curves for E _x generated for various E ₂ (in the order of 3 GPa) better describe low Θ behaviour, but are less accurate than the lower E ₂ values as Θ tends to 90°. No single curve generated from the CLT model entirely captures the overall stiffness reinforcement angle relationship seen when testing the braided sheath reinforcement. However, a nonlinear least squares fit shows that an E ₂ value of 2.77 GPa is optimal, and the general trend observed in from the E i _ong measurements in Figure 12 is not entirely dissimilar to what would be expected in an ideal 2-ply laminate.

The curve fitting exercise in Figure 13 suggests that the frictional interaction responsible for E ₂ varies as a function of reinforcement angle. This variable internal friction is related to the woven material contact area, undulation through the thickness of the braided sheath, and the braided sheath-pipe contact area. Creep strain rate analysis

The reinforcement braided sheaths are altered and arranged such that multiaxial strain rates due to creep are minimized. In a simple analytical model with idealized contact conditions (no slip at the interface), the portion of the restorative force F _r acting at Θ degrees to the tube axis opposing the hoop stress is F _r sine, and the longitudinal portion is F _r cos6. It can be shown that for thin walled pressure vessels (i.e. when t/r < 0.05, as defined in Section 2.1) and orienting Θ to allow the force components to balance gives Θ = tan ^_1 ^2 and a predicted optimal reinforcement angle of Θ = 54.7°. The present results will show that the actual optimal reinforcement angle varies considerably from this prediction.

Mean creep rates ε were calculated and plotted against reinforcement angle and stiffness measurements and predictions in Figure 14. As Θ is defined from 0° =¾ Θ =¾ 90° due to symmetry, it would be expected that the ε-θ curve be symmetrical about 90°. At Θ = 90°, the reinforcement is oriented entirely in the diametral direction. Therefore it would be expected that there be minimum ε dia and maximum ε i _on g for Θ = 90°, and vice versa at Θ = 0°. A sinusoidal relationship fulfils these requirements, and relates well to the force imposed by the reinforcement onto the tube, which can be decomposed into its effect in the hoop and longitudinal directions by the sine/cosine of the reinforcement angle. With these conditions on symmetry and the location of maxima and minima fixed, amplitude and vertical offset of the sine wave were varied and compared against measured mean creep rates using a nonlinear least squares fit.

From Figure 14 it can be seen that the measured longitudinal creep rate varies with braid angle and is close to zero at approximately Θ = 54°.

Referring to Figure 14, the measured diametral (tangential) creep rate varies with weaving or braiding angle and is zero for Θ > 55°. Diametral contraction is driven by pipe elongation, but any diametral contraction would result in the reinforcement disengaging. Without the influence of this reinforcement, internal pressure forces dictate that diametral creep is dominant, and the pipe would begin to creep diametrally until the reinforcement is once again engaged, at which point longitudinal creep becomes the path of least resistance. From the strain rate plots in Figure 14, an ideal reinforcement angle can be predicted where multiaxial creep rates are minimized. To assist in determining at what point reinforcement angle best restricts multiaxial creep, it is useful to consider the effective strain rate. Figure 15 shows measured mean radial creep rates and predicted radial creep rates plotted against reinforcement angle. A minimum effective strain rate is predicted at a

reinforcement angle of 54.7°, suggesting that creep is most effectively constrained at this point. The predicted effective creep rate and the experimental effective creep rate are in good agreement as shown in Figure 15.

While there is clear minimum in the effective strain rate curve in Figure 15, this point is still nonzero. This suggests that there is no weaving angle where overall creep strain is completely negated, and that altering the stiffness of the reinforcement layer through engineering the architecture may not be the only element to controlling the creep response. For example, it is feasible that altering the surface roughness of the pipe at the reinforcement pipe interface will improve the reinforcement's ability to restrict creep in the longitudinal direction by increasing the effective E ₂ value which manifests as an internal friction in the braid.

Conclusions

Experimental methods have been developed to manufacture hybrid, reinforced pipes with a range of reinforcement architectures. Methods have been developed to subject the hybrid pipes to creep conditions, i.e., elevated temperatures and internal pressure, and to evaluate their response in terms of strain and rupture life.

Compared to an unreinforced pipe of the same dimensions, 7.7 times creep life extension was achieved by restricting the diametral creep strain with a helically-wound reinforcement layer constructed of a material not prone to creep at the service temperature.

As expected, unreinforced pipes fail through cracking in the longitudinal direction, due to hoop stress. When diametral creep is restricted by a helical reinforcement architecture, longitudinal creep is accelerated and the failure mode becomes cracking in the hoop direction due to longitudinal creep caused by longitudinal stress.

When creep is restricted by a braided sheath reinforcement, creep life extensions greater than 10 times are achieved, and rupture is not been observed within that time frame.

The creep strains in both hoop and longitudinal direction can be manipulated by exploiting the anisotropic nature of the braided sheath reinforcement layer, particularly the coupling between its behaviour in the longitudinal and diametral directions. A simple analytical model predicts an optimal reinforcement angle of 54.7° to minimize both hoop and longitudinal creep strain rates. An empirically based model supports that a braid angle of approximately 54.5° ± 1.5° is optimal to minimize the effective multiaxial creep rate of a hybrid pipe under internal pressure.

The tangential/longitudinal stress ratio influences the optimum angle for a given pipe geometry. Therefore, designing an effective reinforcement layer for a hybrid pipe requires analysis of the specific application. In some industrial applications, resultant stresses are not solely the result of internal pressure. It is useful to consider how a change in stress state affects the optimal reinforcement angle. For example, in an ethylene cracking furnace, internal pressure is sufficiently low that creep of ethylene pyrolysis tubes is driven primarily by self weight. Using nominal tube dimensions of an 80mm outer diameter, 8mm wall thickness and an 20m length, as well as a typical internal pressure of 0.5MPa gives pressure stresses of at = 1.8-2.3MPa and σζ = 0.9MPa as well as a maximum longitudinal self weight stress of osw = 1.5MPa. Superimposing longitudinal stresses gives a tangential/longitudinal stress ratio of 0.74-0.95, with a corresponding neutral angle of 41- 44°.