The device described herein is a new design for a sampler head or sample cutter head 50 for the extraction of representative sample increments from a material stream conveyed on a belt conveyor. The sampler. is of a type generally called a cross-belt sampler, but is also known as a go-belt sampler or a hammer sampler. This type of sampler has been known for many years.

The new design corrects the technical faults that are present in the known designs, especially in regard to the tendency of existing samplers to throw material off the belt that is supposed to be part of the sample collected.

2. Prior Art

Sampling of bulk materials conveyed on a conveyor belt is a critical operation in the transfer of bulk commodities between buyer and seller or within industrial processing operations. It is usual to sample the commodity as it is loaded onto transport (road, rail or ship) and to sample it as it is off-loaded at the buyer's site or at the boundaries of the processing operation. The sample that is analysed to estimate the quality or value of the shipment must be representative of the entire shipment. A sample is said to be representative of the material sampled when the expected size distributions and assays of the sample are equal to the true mean size distribution and assays of the lot of material being sampled. Modern statistical sampling theory provides a means of determining how many sample increments and the total mass of material that must be extracted from the flow of material as it is loaded or off-loaded in order to ensure that the expected value of the mean squared difference (also called the variance) between the properties of the accumulated sample and the true properties of the entire shipment is limited to an acceptable magnitude.

The increments extracted from the primary stream are combined into the representative sample and that sample may be further sub-sampled by other devices to arrive that the final mass of material submitted for physical and chemical analysis. If all the sampling equipment used in taking and processing the sample increments and the final sample is correctly designed, any difference between the properties of the accumulated sample and the true properties of the entire shipment will be randomly distributed with some variance and a statistical expected value of zero. When the statistical expected value of the difference is zero, the sampling equipment is said to be unbiased. Lack of bias in commercial sampling of commodities is usually a contractual requirement and it is essentially mandatory that the sampling equipment used be unbiased.

Devices that extract samples or increments of material from a flowing stream can be shown to be biased or unbiased by carrying out an analysis of the motion of the device and the motion of the particles induced by the motion of the device and by considering the manner in which the device intercepts the stream of material.

A mechanical sampler, in the most general terms, is a device which can move through a stream of moving material and collect and subsequently discharge or divert a volume of that stream. In the case at hand, the cross-belt sampler causes some volume of material to be swept off a conveyor belt. The sampler has two or more cutting edges which define the passage through which the material must pass in order to be collected into the body of the device from which it will be finally collected as an increment or sample. Generally, two of the cutter head edges will be critical in defining what material is collected and what is not. In the case of a cross-belt sampler, the sample or increment of material is collected into the cutter head and the conveyor belt forms one of the confining surfaces of the cutter head. The material collected by the cross- belt sampler is discharged under centrifugal force.

It is also critical that the sampler design be such that once a particle has passed into the cutter head, it does not then pass out of the cutter head and return to the flow from which it was sampled.

As a mechanical sampler moves, the two critical edges of the cutter head sweep out a surface. Both edges may or may not sweep out the same surface. There are three cases of interest in terms of the orientation of the surfaces.

In a first case, if the surfaces are both the same, then the theoretical extent of bias of the sampler can be assessed by determination of the time between the passage of the leading and trailing cutter head edges through an arbitrary particle trajectory in the flow of material past the cutter head. If the time interval between the passage of the leading and the following edge is a constant, for all possible particle trajectories, then the sampler can be deemed to be theoretically unbiased. Cutter heads moving in linear motion (traversing) and meeting the above conditions must have cutter head edges whose projections onto a plane substantially perpendicular to the material flow are straight and parallel. Rotating cutter heads meeting the above condition must have edges whose projections onto a plane perpendicular to the axis of rotation are straight and they or their linear projections meet at the centre of rotation.

In a second case, the cutter head edges may define a single plane and the plane may be perpendicular to the motion of the cutter head. In this case the time between the passage of the cutter head edges through an arbitrary particle trajectory is zero and the unbiased nature of the cutter head depends on the cutter head intercepting a constant path length of particle trajectory on all particle trajectories. In a case in which the motion of the solids is uniform and parallel to the plane of the cutter head edges, the edges must be parallel so that a swath of constant cross- section is cut from the flow of solids. In any other case, the surfaces formed by the motion of the cutter head edges are complex and the required shapes of the edges for unbiased sampling are correspondingly complex. In the second case above, the cutter head 'slices' a section of the flowing stream out of the flow and captures it as an increment. In the first case above, the cutter head presents an aperture through which the material to be sampled flows and the material moving on any particular particle trajectory is collected for a fixed time period. For cutter head geometries that do not comply with either the first or second set of conditions described, the lack of bias in sampling is determined by both the time of flow into the cutter head and the length of the particle trajectory that is 'sliced' out of the material flow. In what follows these two components will be referred to as the 'flow time' and the 'slice length'. The flow time multiplied by the particle velocity will be denoted as the 'flow path'. The sum of the slice length and the flow path will be denoted as the 'total increment path length'. For mechanically correct sampling, the total increment path length must be constant for all parts of the flow of material being sampled.

Sample cutter heads that do not satisfy the appropriate criterion (constant flow time, constant slice length or constant total increment path length) can be said to be mechanically incorrect and will take a sample increment that is biased if there is any persistent segregation of the particles in the flow.

A further problem arises with cross-belt samplers in that the motion of the head through the load on the belt may impart motion to particles that are outside the sample cutter head and these particles may be swept off the belt along with the particles that properly form the sample increment. The inclusion into the increment of material that should not be part of the increment will cause the sampler to be mechanically incorrect and in the presence of persistent segregation, the sampling will be biased.

Similarly, failure to collect all material that should be collected will bias the sampler. Operators of cross-belt samplers often fear that the head will damage the belt and so adjust the head relative to the belt in such a way as to cause fine material to be left behind on the belt. Long considers this problem, US Patent No. 5,767,421. In addition to providing special belt idlers to ensure that is a section of a cylinder in section normal to the direction of travel of the belt, Long's design discloses displacing the axis of rotation from the axis of the belt towards the discharge side of the belt. This simple change to the geometry of the system ensures that the head of the sampler pushes every more firmly against the belt as it moves across the belt. The extra pressure substantially assists the removal from the belt of all particles that pass into the cutter head.

With particular reference to cross-belt samplers, six design variations can be distinguished, based on the orientation of the plane formed by the leading edges of the cutter head, the orientation of the sides of the cutter head at the leading edges or the flat or curved nature of the side plates of the cutter head.

The first design is one in which the cutter head edges are parallel and form a plane that is parallel to the direction of motion of the particles on the belt. The side plates are flat. This is the original design of the cross-belt sampler and the cutter head sweeps across the belt in a plane perpendicular to the motion of the belt. This design is shown in Figure 1. Such sampler geometry is disclosed by Ford in US Patent No. 5,392,659. The solids trapped between the side plates and the rear plate of the sampler and the surface of the belt are accelerated in a direction normal their motion on the belt and are thrown off the side of the belt into a chute of suitable geometry (not shown). With the objective of sweeping off the belt any particles remaining within the path of the sampler head, there is often an adjustable wiper or scraper or brush fixed to the trailing edge of the sampler; a wiper is usually made of a flexible material and may be adjustable. The use of a brush at the trailing edge of a cross-belt sampler is described in van der Merwe, in US Patent No. 5,115,688.

A second design has side plates that are also flat and parallel, but in this design the plates are set at an angle to the flow so that the motion of the solids relative to the head is substantially parallel to the side plates. Such a device is shown in Figure 2. The axis of rotation is parallel to the centre-line of the belt and the plane of the cutter head edges is parallel to the solids motion and may or may not pass through the axis of rotation of the head. This design with angled side plates is also disclosed in van der Merwe, US Patent No. 5,115,688.

A third design of the cross-belt sampler may have plates that are flat and parallel and angled as in the second design, but the plane of the cutter head edges may not be parallel to the axis of rotation of the cutter head or the axis of the belt. Such a design is shown in Figure 3. The projection of the cutter head edges into a plane perpendicular to the motion of the solids is not a line as in the second case described herein, rather it is a portion of a rectangle.

A fourth design of the cross-belt sampler is disclosed by Lyman (PCT Application PCT/AU2007/000847) (International Publication WO 2007/147201 A1), in which the cutter head (50) edges (9) are oriented such that, when orthographically projected into a plane perpendicular to the motion of the solids, they or their extensions meet at the axis of rotation (1) of the sampler head (50). The plane of the leading edges (9) of the cutter head (50) is not perpendicular to the motion of the solids. Such a design is shown in Figure 4.

NB: The disclosure of WO 2007/147201 A1 is incorporated into the present Specification by way of reference.)

Lyman also introduces the concept of shaping the side plates (3, 4) of the cutter head (50) so that the vector of the velocity of the solids outside the cutter head (50) relative to the cutter head (50) "is everywhere parallel to the side plates". He bases his determination of the shape of the side plates (3, 4) on Equation (1) in WO 2007/147201 A1 , which determines the angle between the undisturbed velocity vector of the solids outside the cutter head (50) relative to the cutter head (50) and the axis of rotation (1) of the cutter head (50).

Lyman states that "to determine the shape of the cutter head (50) in general circumstances, it is necessary to consider a series of planes parallel to the axis of rotation (1) of the cutter head (50) and perpendicular to a line bisecting the opening of the cutter head (50). Within one such plane, Equation (1) can be used to construct a differential equation the solution to which gives the locus of the intersection of the outer surface of the side plate (3, 4) with the plane considered. This equation can be solved explicitly to provide the data needed to manufacture the device".

We show below that, while Lyman seeks to find a shape of the side plates (3, 4) of the cutter head (50) so that the vector of the velocity of the solids outside the cutter head (50) relative to the cutter head (50) "is everywhere parallel to the side plates" (3, 4), he bases the shape on a relationship that subsequent experimentation has found not to be optimal to generate a shape with such properties.

The last design variation on the cross-belt sampler is one described in ISO 13909-2-2001 'Hard coal and Coke - Mechanical sampling - Part 2- Coal - Sampling from moving streams' to have a sample cutter head of a geometry as shown in Figure 1 , which rotates on an axis as indicated in Figure 1 , but which is carried on a moveable trolley above the belt, the trolley being put into motion parallel to the motion of the belt before the sampling device contacts the belt or the solids on the belt. The trolley moves at the same speed as the belt. Such a sampler is capable of collecting the material between its parallel side plates and the swath of the device on the belt is at right angles to the motion of the belt. This device is unbiased but has the complication of requiring a transport system that must be controlled to move at the speed of the belt and will require special support of the belt over a substantial length. The five embodiments of the cross-belt sampler described above constitute the "state of the art" (i.e. PRIOR ART) in cross-belt sampler design.

GENERAL DESCRIPTION OF THE INVENTION

Bias arises in sampling systems as a result of the sampling device not collecting equal proportions of all parts of the material flow. For example, on a conveyor belt, the persistent jostling of the load on the belt due to the motion of the belts over the idlers usually leads to the percolation of finer particles to the region close to the belt surface and the movement of the coarse particles towards the surface of the load. Therefore, when the material is sampled by a device, if more material is taken from the material near the belt than from the material near the surface, the size distribution of the sample will be biased towards the fines and the size distribution will be biased. If the assays of the particles vary with size, as is the usual case, the sample will also be biased with respect to any analyte of interest.

The design principle for a cross-belt sampler disclosed herein corrects the faults in known designs and the new design can be shown to produce devices that are theoretically unbiased.

The new design rules are based on a shape of the side plates of the body such that the velocity of a particle relative to the cutter head and adjacent to the outer surface the cutter head is everywhere parallel to the side of the cutter head at that point on the cutter head. In contrast to Lyman's disclosure, the new design optimally achieves the objective of making the velocity of a particle relative to the cutter head and adjacent to the outer surface the cutter head everywhere parallel to the side of the cutter head.

In addition, it is true that regardless of the manner in which the plane of the cutter head edges is oriented with respect to the cutter head itself, the shape of the side plates will ensure that the increment taken is mechanically correct. This new geometrical rule for design of the cross-belt sampler has not been realised previously and therefore the range of designs possible is entirely novel. The range of designs possible accommodates a range of belt curvatures, widths of sampler openings and ratios of the speed of the sampler cutter head at the surface of the belt to the speed of the belt.

The requirements for a sample increment collected by a cross-belt sampler to be theoretically unbiased have been discussed above. The total increment path length principle for mechanically correct sampling has not previously been disclosed. The technically correct application of the concept that the undisturbed particle velocity outside the sample cutter head, relative to the cutter head, must be tangent to the surface of the cutter head results in a design that ensures that the total increment path length is constant for cutter head opening designs in which the cutter head edges are coplanar and a sampler design that minimises the inclusion into the increment of particles which fall outside the correct sample increment definition.

This design criterion for a cross-belt sampler has not been quantified and is therefore novel. In particular, it has not been previously recognised that shaping the cutter head side plates to provide a tangential relative solids velocity also ensures the compliance with the total increment path length principle.

BRIEF DESCRIPTION OF THE DRAWINGS

To enable the present invention to be fully understood, preferred embodiments will now be described with reference to the accompanying drawings, in which:

Figures 1 , 2 and 3 are respective isometric views of examples of the PRIOR ART cross-belt samplers hereinbefore described;

Figure 4 is an isometric view of the PRIOR ART cross-belt sampler of Lyman disclosed in WO 2007/147201 A1; Figure 5 is a graph of the projection of the edges of a cutter head, into a plane perpendicular to the motion of the solids on the conveyor belt, in a preferred embodiment of the cross-belt sampler cutter head of the present invention;

Figure 6 is an elevational view of the cross-belt sampler cutter head in a plane perpendicular to the axis of rotation;

Figure 7 is a top plan view of the cross-belt sampler cutter head;

Figure 8 is a rear view of the cross-belt sampler cutter head; and

Figure 9 is an isometric view showing the motion of the cross-belt sampler cutter head relative to the motion of the solids on the conveyor belt.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

The cross-belt sampler cutter head 50, of the present invention, to be hereinafter described in more detail, may be used with the belt conveyor (60) disclosed in WO 2007/147201 A1 , being substituted for cutter head (50) and employing the control system (22) described with reference to Figure 12 of WO 2007/147201 Al

Consider first the general problem of finding a shape for the side plates 3, 4 of a cutter head 50 such that the velocity of a particle relative to the cutter head 50 and just outside the cutter head 50 is everywhere parallel to the cutter head 50 side plates 3, 4.

Let a coordinate origin be defined at a convenient point on the axis of rotation 1 of the cutter head 50 with the z axis pointing in the direction of motion of the solids which is parallel to the shaft of the cutter head 50. The x axis will be taken to form a horizontal plane with the z axis. The y axis is then taken as positive in the vertically downward (towards the belt) direction. Unit vectors in the x, y and z directions can be defined by i, j, k . Cylindrical coordinates will also be used defining a radial coordinate as and an angular displacement as

θ = tan ^"1 [ ^l (3)

The undisturbed motion of the particles on the belt is then described by a vector of magnitude v _B pointing in the positive z direction or by v _B \ . The motion of the cutter head 50 is a pure counter clockwise rotation about the z axis when viewed from the origin and looking in the direction of increasing z. The velocity of the cutter head 50 is most conveniently written is cylindrical coordinates; the velocity of a point on the cutter head 50 is v _h = rωe _θ (4) where ω is the angular velocity of the cutter head 50 in radians per second and e _θ is a unit vector in the direction of increasing θ . The extreme points of the cutter head 50 are taken to lie on the surface of a cylinder defined by

X ² +y ² = /? ² (5)

The equation of the exterior surface of one of the side plates 3, 4 of the cutter head 50 may be written as g = z -h(r,θ) = 0 (6)

The vector normal to such a surface is dh . 1 dh - .- Vg = e _r e _θ +k dr ^r r dθ ^θ where e _r is a unit vector pointing in the direction of increasing r. The velocity of the solids relative to the rotating cutter head 50 is written as u = v _B k + rωe _θ (J)

The condition that the surface of the side plate be everywhere parallel to the relative velocity vector is written as

Vg u = 0 (8) which results in

v _B - ω— = 0 (9)

^B dθ and this can be integrated immediately to give

h = ^ ^B. ø + f(r) (10) ω where f(r) is an arbitrary function of partial integration. We then have

The plane of the cutter head 50 edges is defined by z ~ x\anβ for any y (12)

The intersections of this plane with a cylinder of radius R (axial) are defined by the solution to

x Y ² - +U y I/ ² - - R P ² ₍₁₃₎ z = xtanβ and one additional condition. Let this condition be x ^ ±wcosβ (14) where w is a fixed quantity. Then To force the surface to pass through one of these points, we have

^ = tan# (16) or

Taking the positive value first,

(18)

Z = M/ sin/?

This solution corresponds to the downstream plate. Putting these values into the equation for the surface, we find

For the upstream plate, we have

and

The above equations are sufficient to define the outer surfaces of the two side plates of the cutter head 50. For example, the upstream surface is defined by

It is critical to note that at any distance r from the centre of rotation, as the angular displacement of the head changes by an amount Δθ/ω in a time Δt , the surface moves a distance Δz/v _B in the same time interval. This demonstrates that there is no component of velocity normal to the surface of the cutter head 50. The next technical aspect of the design that requires consideration is the shape of the intersection of the cutter head 50 surfaces with the plane of the cutter head 50 edges. Using the general solution Equation (22) and the Equation for the plane of the cutter head 50 edges Equation (12), it is possible to show that the total increment path length is and this value is independent of position on the head. It is also to be noted that the value of β can take on any value and it need not be related to the velocity of the head relative to the belt. For the extreme values of β of π/2 and 0, the total increment path is evaluated to be 2w and

2% _S j _n ^{1 /} wcos/T/ j _wn j _cn corresponds to the increment path for a ω I conventional hammer sampler of aperture 2w (see Figure 2) and a Vezin- type rotating sampler.

The shape of the cutter head 50 edges is illustrated in Figure 5.

This proposed design for the cross-belt sampler simultaneously provides the condition for the mechanically correct sampling of the material on the conveyor belt while at the same time satisfying the demand to have the relative motion of the solids tangential to the exterior faces of the cutter head 50.

In practice, the volume of the cutter head 50 will be truncated at some convenient distance from the axis of rotation. It is convenient to truncate on a plane of constant y value. Similarly a rear closure between the two side plates is provided to contain the sample between the side plates until the material is discharged over the side of the belt. The closure can be made in such a way as to avoid acute angled corners into which cohesive solids might pack and stick inside the sampler head. The overall size of the cutter head 50 is related to the mass flow on the belt, the particle size of the solids, the practical range of curvature of the belt and the ratio of the head speed to belt speed. The minimum width of the cutter head 50 opening in a plane perpendicular to the motion of the solids into the cutter head 50 should be 3 times the nominal dimension of the largest particles in the material to be sampled, if the particles exceed 10 mm in size.

The required cutter head 50 volume must be considered for each application and the head designed according to the novel principle disclosed herein that leads to design that is mechanically correct and optimal in relation to minimising the removal of particles from the belt that are not collected between the plates.

Figure 6 is an elevational view of the cutter head 50 in a position such that the axis of rotation of the device is perpendicular to the plane of the drawing. The cutter head 50 is attached to the shaft 1 by suitable arms 2 fixed to the shaft and to the body of the cutter head 50. The opening of the cutter head 50 into which the sample increment is collected is defined by curved edges 8 of side plates 3 and 4 and top plate 7 and the belt surface. Plates 3 and 4 are curved plates shaped according to the design principles described above that lead to tangential motion of the solids past the head and for an arbitrary plane of the cutter head 50 edges, meet the constant increment path length criterion. The bottom of the cutter head 50 where it contacts the belt is open. The rear of the cutter head 50 is formed by plates 5 and 6 (see Figure 7). While plate 5 is oriented parallel to a plane passing through the axis or rotation with the objective of assisting a full discharge of the solids from the sample cutter head 50, plate 6 truncates the sampler cutter head 50 volume so as to eliminate what would otherwise be a corner with acute angles into which solids might compact. Figure 7 provides a top view and Figure 8 provides a rear view of the cutter head 50 and. The top view is intended to provide an impression of how the shape of the cutter head 50 changes with distance from the axis of rotation in order to ensure that the motion of the solids past the exterior surface of the cutter head 50 is everywhere parallel to the surface.

Figure 9 shows a segment of the conveyor belt surface 9 over which the cutter head 50 must move and the sense of rotation of the head relative to the belt 10 and the motion of the belt and solids 11. Vector 12 indicates direction of entry of the solids into the cutter head 50, relative to the cutter head 50.

The belt must be contoured or shaped to a partially cylindrical shape so that the cutter head 50 remains in close contact with the belt as it moves over the belt surface. Similarly, this contact must be sufficiently positive to ensure that the cutter head 50 sweeps the material that is properly part of the sample increment from the belt without leaving particles behind.

Note that in Figure 6, the segment 5 of the rear of the cutter head 50 is arranged so that its surface falls in a plane that passes through the axis of rotation. While other orientations of the plate may be used, or a curved surface provided inside the cutter head 50, the orientation shown will ensure that the centrifugal force on the solids inside the cutter head 50 is parallel to the plate, leading to a maximum shearing force between the solids and the plate which will assist discharge of the solids from the interior of the cutter head 50 once it loses contact with the belt surface.

The cutter head 50 is driven in the same manner as existing cross-belt samplers. The usual means of driving the cutter head 50 is by a synchronous motor acting through a gear box of suitable gear ratio to achieve the design cutter head 50 speed at the surface of the belt. The drive system is equipped with suitable controls so that the cutter head 50 will carry out a single rotation per increment and will be accelerated to constant speed from its starting position in an inverted position above the belt prior to contact with the belt surface and will be brought to a stop as it returns to the starting position. The critical issue in relation to the drive is that the head move at a constant speed (±5% relative) as it passes through the load of material on the belt. This is achieved by suitable choice of motor drive controls (vector control variable speed drives, inverter drives, servo motor drives, or other speed-controllable drive systems as current technology provides.).

The cutter head 50 may be constructed of a variety of materials that provide sufficient rigidity to retain the shape of the cutter head 50 as it moves through the load of material on the belt. The material(s) of construction may be chosen to provide resistance to wear by the material to be sampled and due to impact between the material being sampled and the cutter head 50. The material(s) and mode of construction of the cutter head 50 may be chosen to reduce the moment of inertia of the cutter head 50 about the axis of rotation, as this reduction reduces the demand for mechanical strength in the mounting frame and for the power needed to accelerate and decelerate the cutter head 50 during its motion. In the simplest case, where the material to be sampled is of relatively low density and not abrasive, common metal construction may prove effective. In more extreme duties, where the material sampled is dense and substantially more abrasive, and/or there is a desire to reduce the moment of inertia of the sampler head, the cutter head 50 may be constructed of composite materials of high wear and impact resistance formed over an internal skeleton of suitably rigid design.

UTILITY OF THE INVENTION

The invention disclosed herein can be employed to sample any material conveyed on a conveyor belt as long as the cutter head 50 design is suited to the effective particle size of the material and the physical circumstances of the installation. The device is particularly suited to the sampling of granular solids.

VARIATIONS AND MODIFICATIONS

Various changes and modifications may be made to the embodiments described and illustrated without departing from the present invention.