GB2271440A - Optimising boiler cleaning - Google Patents

Abstract

The optimum time to activate a boiler cleaning means eg a soot blower 106 is determined from the output of at least two heat flux sensors 100. The optimum time may he considered to be when the first of a number of conditions is satisfied eg when the rate of change of heat flux drops below a threshold thus indicating the onset of sinter or if a local sensor output exceeds a reference level. The threshold level of contaminant build up may be determined from previous heat flux measurements. The sensors may comprise thermocouples (fig. 2, 14) embedded in weld material (fig. 2, 10) in the hollow wall (fig.2, 9) of the boiler so that there is no protrusion from the wall (12, fig. 2). <IMAGE>

Description

IXPROVENENTS IN OR RELATING TO BOILER WALL CLEANING The invention relates to the field of combustion furnaces for fossil fuels, and in particular although not exclusively to the removal of ash and soot from heat transfer surfaces within such furnaces.

Conventional plant or process boilers convert water into steam by the transfer of heat from burning fuel, e.g.

fossil fuel, to the water. The water passes through tubes which form the surface of a combustion chamber in which the fuel is burnt. Transfer of heat from the burning fuel to the tubes is almost entirely by radiation. Heat is absorbed at outer wall surfaces of the tubes and conducts through the material of the tube wall to an inner surface of the tube wall, the inner surface of the tube wall being in contact with the water/steam.

A long standing problem with such boilers is that ash and slag from combustion of the fossil fuel accumulates on the outer surfaces of the tube walls. Since such ash and slag has a low thermal conductivity, heat transfer from the combustion chamber to the inner surfaces of the tube walls is severely reduced.

Such poor heat transfer characteristics can seriously affect the economics of boiler operation. In a typical boiler, even a small percentage loss in efficiency due to ash and slag build-up can cause a loss in efficiency costing thousands of pounds.

The formation and properties of the ash and slag deposits are dependent upon boiler conditions, the mineral content of the fuel, the fuel/air ratio, the impingement of flames on the furnace walls, and variation in ash mineralisation.

Conventionally the slag and ash is removed at periodic intervals from the outer surfaces of the tube walls in the combustion chamber wall. Early removal methods required complete shut down of the boiler and removal of the slag and ash by hand. Later methods included introducing a cleaning fluid e.g. air or steam, through a hand hole in the boiler e.g. by a high pressure hose to remove the slag by hand.

A subsequent method has been to fix a movable cleaning device within a boiler which removes slag during a cleaning cycle conducted periodically. Such cleaning devices are commonly called soot blowers. Modern boilers include several soot blowers which can be operated automatically without shut down of the boiler. However, such soot blowing apparatus has a disadvantage that operation of the soot blowers causes a temporary reduction in steam making capacity due to the cooling effect of the soot blowing agent on the combustion process and tube surfaces. Furthermore, when a boiler is operating in a low steam demand condition, and the boiler firing rate is at a low level, the combustion may be extinguished by a quenching effect of the soot blowing.Thus, conventionally, deciding when to operate the soot blowing apparatus has hitherto been a manual operation, or has been semi-automatic, requiring manual supervision.

Attempts have been made to fully automate such systems. For example in US Patent 3,785,351. A system is described which activates soot blowing apparatus at predetermined fixed intervals. However, this does not result in optimum efficiency of the boiler, since the soot blower is activated regardless of the actual amount of slag and ash built up in the boiler.

The optimum period between cleaning operations is variable, and pre-setting a fixed interval before cleaning operations can lead to insufficient or excessive cleaning.

Premature or too frequent cleaning can result in unnecessary cooling and stressing of the boiler tubes, leading to loss of efficiency and stress cracking of the tubes, whereas insufficient cleaning reduces the operating efficiency of the boiler.

To provide an indication of ash and slag build up on the combustion chamber walls, another prior art system disclosed in US 4,408,568 uses a first heat flux sensor positioned in the combustion chamber such that ash deposits do not build up on it and a second heat flux sensor upon which ash or slag deposits are allowed to build up. A change in output of each heat flux sensor is monitored to determine a change in heat flux due to ash build up on the second heat flux sensor.

Prior art devices for measuring boiler heat flux comprise a sensor of known geometry, e.g. an elongate rod, having one portion exposed to the combustion chamber, and an end which is heat sunk to the wall of the combustion chamber. A temperature gradient along the elongate rod is measured by two or more thermocouples. However, such conventional sensors have the disadvantage that they project outwardly from the boiler wall, and therefore effect the local gas flow along the boiler wall, which in turn effects the amount of slag and ash build up, leading to measurement of a soot or ash build up which is unrepresentative of that on the immediately surrounding boiler wall. The surface projection of the sensor interferes both physically and thermally with the nature and thickness of the ash and slag deposit.

Specific embodiments of the present invention aim to alleviate and improve upon some of the above mentioned prior art and its associated problems.

According to one aspect of the present invention there is provided a boiler cleaning management apparatus comprising: a plurality of heat flux sensors; reading means for reading signals of said plurality of sensors and converting said readings into computer readable signals; and computer means for receiving said computer readable signals and determining therefrom an optimum time to activate a boiler cleaning means.

Preferably, said optimum time is determined as a time when a slag layer on a combustion chamber wall of said boiler begins to sinter.

Said apparatus may further comprise said boiler cleaning means.

Preferably, said computer is arranged to activate said boiler cleaning means automatically without the need for human intervention.

According to a second aspect of the present invention, there is provided a method of determining an optimum time to perform a boiler cleaning operation, in which data from one or a plurality of sensors is characterised using a first characterisation method to produce a first characterisation of said data over a first time period, and data from one or a plurality of said sensors is characterised using a second characterisation method to produce a second characterisation of said data over a second time period, wherein said second time period is substantially greater than said first time period.

The invention includes a method of determining an optimum time to clean a boiler, in which a decision on whether to activate cleaning is made based on sensor data received in a time period not exceeding 1 minute.

The invention includes a method of determining an optimum time to clean a boiler, in which a decision on whether to activate a cleaning operation is made based on sensor data received in a time period of more than 1 minute but preferably not exceeding 20 minutes.

The invention includes a method of determining an optimum time to clean a boiler, in which a heat flux measurement corresponding to a locality of the boiler is compared with a variable alarm level.

Preferably, said alarm level is dependent upon a plurality of previous heat flux measurements.

Preferably, said alarm level is periodically variable at intervals of greater than 1 minute.

According to a third aspect of the present invention, there is provided a first method of determining an optimum time to clean a boiler, said first method comprising: (i) obtaining in a first time period a plurality of sensor data readings corresponding to data of one or a plurality of sensor devices; (ii) analyzing said data readings to obtain one or more characterisations of said data; (iii) comparing a rate of change value of one or more said characterisations to one or more correspondiong reference values; (iv) generating an affirmative signal when a said rate of change becomes less than or equal to a said corresponding reference value.

Preferably, a said characterisation comprises a fourier summation.

Preferably, a said reference value is dependent on two or more successive sensor data readings.

Preferably, said sensor data readings comprise data corresponding to one or more readings of heat flux.

Preferably, said first period has a duration in the range 0.1 to 60 seconds.

According to a fourth aspect of the present invention, there is provided a second method of determining an optimum time to clean a boiler, said second method comprising: (a) obtaining over a predefined time period a first plurality of sensor data readings; (b) calculating an average of said first plurality of readings; (c) obtaining at least one further plurality of sensor data readings over at least one further pre defined time period; (d) calculating an average of said further plurality of readings; (e) determining a difference value between said averages of said first and further pluralities of readings; and (f) generating an affirmative signal dependent on a said difference value.

Preferably, said method further comprises: (g) repeating steps (a) to (f) a number B of times to obtain a number B of difference values; and (h) determining an average of said B difference values.

Preferably, said predefined time period is greater than 1 minute.

According to a fifth aspect of the present invention, there is provideds a third method of determining an optimum time to clean a boiler, said third method comprising: determining an alarm level dependent on two or more sensor data readings of one or more sensor devices and a highest and lowest possible operational loading of said boiler; obtaining local sensor data corresponding to a local area of said boiler from said one or more sensors and comparing said sensor data with said alarm level; and generating an affirmative signal if said local sensor data crosses a threshold determined by said alarm level.

The invention includes a method of determining an optimum time to clean a boiler comprising: monitoring the boiler using any two or more said methods according to any of the above methods, to generate an affirmative signal from any one or more of said methods; and generating an optimum time signal in response to the first affirmative signal received from any of said methods.

According to a sixth aspect of the present invention, there is provided a method of manufacturing a heat flux sensor in a furnace combustion chamber wall comprising the steps: (i) indenting a portion of the wall to produce an indent: (ii) filling said indent with a heat conductive in fill material; (iii) forming in said heat conductive material a channel to a first depth, said channel having an open upper end; (iv) forming in said heat conductive material a recess adjacent to said channel, said recess being of a second depth which is shallower than said first depth; (v) inserting a sealing member into said open upper portion of the channel and said recess, such that a lower portion of said channel lies unfilled by said sealing member; (vi) sealing said sealing member into said upper portion of the channel such that said channel remains sealed at the upper portion thereof;; and (vii) profiling said in-fill material and said sealing member to form an unbroken outer surface.

Preferably, said wall comprises a hollow metal tube.

Preferably, said indented wall portion is indented to protrude into a central bore of the tube to a depth in the range 2mm to 9mm.

Preferably, said indented wall portion is indented to protrude into said central bore such that no tangent to an inner surface of said indented portion makes an angle of greater than 9 degrees to a main longitudinal axis of said tube.

Preferably, said channel is formed to be substantially annular.

Preferably, said sealing member is formed of a substantially annular ring.

Preferably, said sealing member is formed of a same material type as said in-fill material.

Preferably, said in-fill material complies to British standard 309S92L.

Preferably, said sealing member is sealed into said upper portion of the channel by welding.

Preferably, said unbroken outer surface is formed to be a gas tight surface.

According to yet another aspect of the present invention, there is provided a heat flux measuring device for a boiler, comprising an elongate wall of a first thickness, said wall being provided with an indented portion and a sensor portion which closely fits the indented portion to locally thicken said wall to a second thickness, wherein an outer surface of said sensor portion and an immediately surrounding outer surface portion of the tube wall presents a substantially continuous unbroken surface.

Preferably, said sensor portion comprises two or more heat sensitive sensors buried under said unbroken surface.

Preferably, said wall is a substantially cylindrical boiler tube wall.

Preferably, the outer surface of the sensor portion matches the outer surface of the immediately surrounding wall outer surface.

The invention includes a heat flux measuring device made of a material which complies to British Standard 309S92L.

For a better understanding of the invention, and to show how the same may be carried into effect, reference will now be made, by way of example, to the accompanying diagrammatic drawings, in which: Figure 1 shows a boiler incorporating a boiler cleaning management apparatus according to a first specific embodiment of the present invention; Figure 2 shows a schematic waveform of an output of a single heat flux sensor with time; Figure 3 shows a control routine for a boiler cleaning management apparatus according to specific method of the present invention; Figure 4 shows a cross-sectional view of a tubular sensor unit for a boiler combustion chamber wall according to a second specific embodiment of the present invention; Figure 5 shows a partial cross-section through an annular stud sensor of the second specific embodiment, in a partially manufactured state; ; Figure 6 shows a plan view of the stud sensor of Figure 5; Figure 7 shows a cut away view through a side wall of the sensor unit of Figure 4, showing the stud sensor of Figures 5 and 6 in a direction parallel to a main axis of the sensor tube; Figure 8 shows a side view of the stud sensor of Figures 5 to 7 in a partially completed state of manufacture; and Figure 9 shows a cross sectional view of an annular stud sensor and surrounding portion in a completed manufactured state.

A first specific embodiment according to the present invention will now be described.

Referring to Figure 1 of the accompany drawings, a boiler cleaning management apparatus comprises a plurality of sensors 100 distributed around a combustion chamber wall 101 of a boiler 102, a plurality of Data Acquisition Units (DAU's) 103 mounted locally to one or a group of sensors 100, and a computer 104.

The combustion chamber wall 101 is conventionally a series of hollow metal pipes welded together in the form of a cylindrical chamber. Combustion takes place within the chamber 102 and water/steam passing through the pipes 101 is heated by energy radiated through the metal side walls of the boiler pipes.

The sensors are distributed around the boiler combustion chamber so as to provide heat flux data representative of the heat flux in the combustion chamber wall near to each sensor, and in specific local regions of the combustion chamber wall. The sensors may be similar to those described with reference to Figures 4 to 9 herein, or may be of a conventional type. Although only one DAU 103 as shown in Figure 1, each sensor 100 will be connected to a corresponding DAU, as indicated by the dotted lines.

The DAU's are connected to a computer 104 which can be situated remotely in a control room. The computer has a monitor and means for manual interface 105 to allow an operator to key in data, change programming parameters, and generally monitor the performance of the boiler.

The computer 104 can be connected directly or via an interface system to activate a plurality of soot blowers 106. Although only two soot blowers are shown in Figure 1, a plurality of such soot blowers may be provided at locations around the combustion chamber to clean local areas of the combustion chamber wall. The computer can selectively activate each soot blower individually, or can activate a combination of soot blowers together to clean the whole combustion chamber in a global clening operation, or a cleaning operation of a locality of the combustion chamber wall. The soot blowers are of a conventional type.

In use the apparatus operates generally as follows.

Each sensor measures a local slag level at the position of that sensor inside the combustion chamber, and produces a heat flux signal dependent upon the particular local slag level. The heat flux signals from the sensors are read by the DAU's which are mounted near to the sensors on the outside of the boiler. The heat flux signals are converted into computer readable data at the DAU's and are then transmitted to the computer, which is preferably remote from the boiler.

The computer operates a programmed set of variable algorithms to characterise the heat flux data and to determine therefrom an optimum time to activate one or more soot blowers. The optimum time to remove the slag is when the slag layer is in an optimum condition in which it starts to sinter, i.e. when it begins to change from a particulate phase to an interconnected phase. Each DAU receives a sensor signal from one or a plurality of sensors.

Detailed operation is as follows. Each DAU fast scans all sensors connected thereto over a fixed read time period, for example 10 seconds. During this time period a heat flux data signal from each connected sensor is read by the DAU. Preferably three signals per sensor are read each second. Thus, for an example of four sensors connected to a single DAU and scanned for a 10 second period each sensor is read three times per second and 120 individual sensor readings will be taken by the DAU.

Referring to Figure 2 of the accompanying drawings, an output signal wave form u (x, y, t) for one sensor is shown over a time period t. In Figure 2, the horizontal axis (t) denotes time and the upright axis denotes the amplitude vector of the output of the sensor. The DAU samples the sensor input signal wave form at integer positions nl, n2,... n30 to produce a set of individual sensor readings u (xl, Y1, tl ) at n1; U (X2, Y2, t2 ) at n2; etc.

where xl is the time elapsed between the start of the current read time period and the instant at which reading nt is taken, x2 is the time elapsed between the start of the current read time period and the instant at which reading n2 is taken and t is the time component of the data under investigation.

After a further period has elapsed, the DAU repeats a scan over a second fixed read time period (which is preferably the same duration as the first fixed read time period), and reads sensor data at further positions n31 to nw.

The DAU continues to repeat scans over further read time periods. Preferably, a 60 second interval elapses between successive data scans.

The heat flux signals are converted in to computer readable heat flux data by the DAU corresponding to the particular local time dependent slag levels around each sensor by the DAU, and fed to the computer as an on-going process.

The apparatus uses three methods to determine from the heat flux data the optimum time to initiate a cleaning operation. This optimum time is when the slag layer is in an optimum condition of starting to sinter.

Firstly, the computer applies a set of changing algorithms to the heat flux data acquired from the sensors, to determine when the slag layer is in the optimum condition.

For example, to determine the optimum condition for slag or ash which is locally starting to sinter around a single sensor, the output from that sensor over a ten second reading period comprising signals corresponding to heat flux, read at nl, n2, n3 - n30 by the DAU every 0.333 seconds over the first read time period, is examined.

This data is then analyzed using a Fourier periodic as follows.

Where:

T is defined as the mean time component and is given by:

where u'(x,y,t) dt is the fluctuating component of the waveform shown in Figure 2, otherwise called the total wave velocity component, and yl, is the modulus of the difference between the magnitude of the sensor output at time tl minus the magnitude of the sensor output at the start of the read time period.

By solving for T in a conventional manner, a value for f (t) can be obtained by entering an initial estimate for a value of T in equation 1 and iterating using equation 2 until a convergence of 0.01 or less is obtained.

The resultant value of f(t) represents a measure of how much "flame flicker" the heat flux sensor is seeing through the developing slag layer.

When the surface of the slag layer sinters and begins to become molten, the rate of change of f(t) changes rapidly, indicating the optimum condition of onset of sintering has been reached. The optimum time is selected as being when the rate of change of f(t) is less than the reciprocal of a Maximum Heat Flux Fraction (MHFF), ie when

where MHFF is the Maximum Heat Flux Fraction defined as MHFF = The Maximum Heat Flux in the period - another Maximum Heat Flux in the period which is no more than 10% greater that a previous Maximum Heat Flux in a previous period.

An affirmative signal is issued by the computer when the above rate of change condition is achieved.

The above first method indicates when the slag local to one sensor has reached its optimum condition. The computer may take the respective values of f(t) for each sensor to gain an overall picture of which slag areas have reached the optimum condition, and then activate one or more sootblowers to effect local or global cleaning of the combustion chamber as appropriate.

However, although analyzing the rate of change of heat flux is important in determining the onset of sintering of the slag, and therefore the optimum time for soot blowing, where a local area of combustion chamber is subjected to a rapid build up of slag, f(t) occasionally does not change in magnitude significantly enough to use the above mentioned first method reliably.

Therefore, to detect rapid slag build up, a second method is used to determine an optimum time to activate the soot blowing apparatus for local cleaning.

The second method monitors the flux data of one or more sensors by determining an average instantaneous gradient (AIG) between successive average heat flux gradients over a plurality B of differences in such gradients.

Referring to Figure 3 of the accompany drawings, a flow chart for implementing the second method using the computer 104 of Figure 1 is shown.

In figure 3: n is the integer value of a counter, denoting a measurement at nl, n2, ... etc. as described above.

B is the integer value of a counter B.

AR is an average value of heat flux data values from a sensor over a twelve minute period.

IG is an "instantaneous gradient" calculated as

where (1) IG is the instantaneous gradient of sensor data at a first interval between readings at positions nl and n2 in figure 2, (1)AR is an average value of heat flux of the twelve readings prior to and including a reading at position nl, and (2)AR is an average value of heat flux of the twelve sensor readings prior to and including a reading at position n2.

The average instantaneous gradient AIG is an average of a number B of instantaneous gradients, each as calculated above.

An affirmative signal is generated by the computer when n is greater than or equal to 3, and the value of AIG is positive and greater than 1.5, but the value of (n)AR is less than 165.

A third method of determining the optimum time is used, in case the above two methods fail to determine the onset of slag or ash sintering.

In the third method the optimum soot blow time is calculated as a set of linear load related points. A local heat flux, as measured by local sensors, is compared with an alarm level over a preset time period. When the alarm level is crossed, an affirmative signal is generated. The alarm level is dependent on a Load Factor of the aforementioned boiler and the maximum heat flux fraction (MHFF) by the following relationship.

Alarm level = load factor x (MHFF X 0.23) The Load Factor is calculated by linear interpolation between the highest operational load factor (HL) the boiler can achieve and the lowest operational load factor (HLL) of the boiler at which soot blowers would be used conventionally. The highest load factor (HL) is accorded the value 1 and the lowest load factor (HLL) is accorded the value 0. Intermediate values between the highest and lowest load factors are determined by linear interpolation between 1 and 0.

The Maximum Heat Flux Fraction (MHFF) is determined as follows: Heat flux is continuously monitored by sampling the output of a sensor over a third time period, eg 12 minutes. The MHFF is then calculated as mentioned with reference to the first method.

The MHFF is variable and can be reset upwardly or downwardly during boiler operation as follows: To set MHFF upwards The total average Heat Flux over the third period is calculated and denoted as TA. If TA is greater than the amount MHFF, then the following criteria are applied: 1. If TA is more than 10% greater than the current MHFF, the current MHFF is reset to 10% of TA.

2. If TA is not more than 10% greater than the current NHFF, then MHFF is reset to equal TA.

To set MHFF downwards A 24 hour average of TA is taken and called 24TA.

The following criteria are then applied to set TA downwards.

1. If MHFF > (24TA x 0.94) then a new value of MHFF is set as the current value of MHFF x 0.95.

2. If MHFF > (24TA x 0.9) then a new value of MHFF is set at 0.9 x current MHFF.

3. If MHFF > than (24TA x 0.5) then a new value of MHFF set at 0.87 x current MHFF.

4. If MHFF > than (24TA x 0.75) then a new value of MHFF set at 0.82 x current MHFF.

5. If none of the above apply, then MHFF is incremented to the current value of MHFF minus 0.75 x the current value of MHFF.

When an affirmative signal is generated by any of the above methods, this indicates that an optimum time to perform either a local or global cleaning operation of the boiler has been reached, and an appropriate cleaning operation is activated. The optimum time for a particular cleaning operation to occur is determined as the time at which the first such signal is generated.

The aforementioned specific embodiment may have an advantage of using a plurality of different methods to calculate an optimum time to perform a soot blowing operation. The embodiment may therefore be capable of updating a characterising algorithm used to determine an optimum soot blowing time on the basis of past collective data and more recently updated data relating to the build up of slag or ash in the boiler. Furthermore the apparatus may be capable of coping with short term changes in slag or ash build up and updating any algorithms used to calculate the heat flux data and determine an optimum soot blowing time in response to the short term changes.

The specific embodiment may have an advantage that by defining the optimum time for soot blowing as when the slag surface begins to sinter, the values of indeterminate boiler variables such as coal chemical composition, ash sintering temperature, ash thickness, local gas temperature, mill group operations, and furnace exit gas temperature, which hitherto were monitored in conventional systems to determine when to operate a soot blower, and which were impossible to monitor on the operational time scales necessary to accurately determine the optimum soot blowing time, need not now be monitored.

The specific embodiment may have a further advantage that local areas of combustion chamber wall may be monitored independently of other local areas of the combustion chamber wall and cleaned independently thereof.

Additionally, the embodiment may be capable of monitoring longer term global changes in combustion conditions in the boiler as a whole, as well as short term local changes.

A second specific embodiment according to the present invention will now be described. Referring to Figure 4 of the accompanying drawings, a tubular sensor for a boiler comprises a first cylindrical metal tube 1 having welded thereto a second metal tube 2 supported by brackets 3 welded to the first and second tubes. The cylindrical tube 1 is suitably a piece of conventional boiler tube, adapted according to a method described hereunder.

The sensor tube has a sensor portion 4.

Extending within the side wall 5 of the tube 1, are two channels extending circumferentially around the tube and transversely to the length of the tube. The channels contain connection wires 6 and 7 respectively, to a pair of thermocouple (not shown in Figure 4) at the sensor portion 4.

The sensor portion 4 comprises an inner side wall portion 8 of the tube is indented so as to bulge inwardly into a central bore of the tube 9, the inner portion being of the same thickness as the side wall 5 of the tube 1.

Preferably the indented portion extends over no more than half of the circumference of the tube.

The sensor portion 4 further comprises a region of infill material 10, e.g. metal welded to the indented portion 8 and having a smooth cylindrical outer surface 11, which matches substantially an outer surface 12 of the tube 1.

The infill material 10 also contains a pill shaped portion 14 buried immediately under the surface 11, into which two (or more) thermocouples are placed, spaced apart from each other. One thermocouple is placed near the outer surface 11, and the other thermocouple is placed radially nearer the centre of the tube 1.

Referring to Figure 5 of the accompanying drawings, a cross-section of the sensor portion 4 is shown, cut along the length of the tube 1, and in a partial state of manufacture.

The sensor portion 4 initially comprises the infill material 10. This material is built up by welding the infill material into the indented portion 8 of the tube side wall. The infill material is built up by successive welds to a solid mass which protrudes beyond the cylindrical surface of the tube as projected over the indent region.

The infill material is then machined, e.g. by turning or grinding the tube, such that an outer surface 11 of the infill material is of a substantially cylindrical shape, having an unbroken surface.

A sensor stud 20 is then formed in the infill material. The sensor stud is of an annular shape in plan view, comprising an annular channel of a first depth. A preferred depth is around 5.7mm at the deepest part of the channel from the surface 11.

The indentation is achieved by dimpling the tube, for example by using controlled hydraulic pressure or a hydrualic press.

The infill material is built up in the indentation using a spiral welding process.

The sensor stud is accurately machined into the infill material by electrical discharge machining. This method ensures that the walls of the original tube are untouched, and the mechanical integrity of the tube unimpaired by the welding process.

An annular recess 22 is then formed adjacent an upper portion of the channel 21.

A first set of holes 23 are provided at a first distance from the surface 11 within the sensor stud. A second set of holes 24 are provided at a second distance from the surface 11 within the stud.

The first and second holes house thermocouple heads (not shown) which are preferably packed with conventional heat conductive paste. The positioning of the second set of holes further away from the surface 11 than the first set of holes, allows a radial temperature gradient from the surface 11 down to the bottom of the stud to be measured by the thermocouple. The channels 21 isolate the sensor stud from temperature gradients in a direction along the length of the tube (horizontally in Figure 2) in use. By providing the first and second holes substantially in the centre of the stud, abberations in the radial thermal gradient down the height of stud, which can occur near the edges of the channel 21 are avoided.

Referring to Figures 6 to 8, Figure 6 shows a plan view of the stud 20, channel 21 and recess 22 after electrical discharge machining, but before formation of the holes 23 and 24.

Figure 7 shows a cross-sectional view of the stud sensor 20 in a direction viewed along a length of the tube, and in a partial state of manufacture.

Circumferential channels 26, 27 are shown for containing connection wires to the thermocouple to fit in the first and second holes 23, 24.

Figure 8 shows a view of the sensor portion 4 of the tube from the side and showing one channel 26, the outer surface 11, the stud 20, and upper and lower holes 23 and 24 respectively. Figures 4 and 5 relate to the tube in the centre portion in the same, unfinished, state of manufacture.

Referring to Figure 9 of the accompanying drawings, a completed sensor portion 4 is shown in partial crosssectional view.

Thermocouples 6, 7 respectively are inserted into the holes 23, 24 and sealed therein by welding over the channel 26, 27 with weld material 28, after first having sealed the annular channel 21 by a sealing ring 29, also welded into place.

Welding of the material 28 and ring 29 is accomplished by an automatic welding process and the ring 29 and weld material 28 is profiled to be smooth and continuous with the surface 11 and the rest of the cylinder, to present a smooth, interrupted, gas tight surface, sealing in the thermocouple 6 and 7 into a buried pill shape portion 14 of the sensor stud 20.

Providing a sensor tube having a buried thermocouple pair behind a cylindrical surface 11 which is unbroken and gas tight may provide an advantage that thermocouple life can be prolonged since gases and flames cannot enter into cavities formed in trenches 26 and 27 respectively nor at the thermocouple heads in holes 23 and 24.

Furthermore, since the sensor portion 4 is of substantially the same shape as the surrounding boiler tube, the sensor portion does not effect the gas flow or deposition process on the boiler tube and therefore, ash and soot build up on the sensor portion is truly representative of the ash and soot built up on adjacent boiler tubes in the immediate vicinity of the sensor portion.

Because the infill material 10, the material of the ring 29, the weld material 28, all of which are profiled to present the surface 11 are all of the same material, there are no local aberrations in heat flux due to different thermal conductivities of weld materials between component parts on the surface 11.

The infill material and material of the ring 29 and weld 28 must have good thermal characteristics, good weld integrity, and good resistance to erosion and corrosion.

Preferably a material to British Standard 309S92L is used as the infill material and for the ring 29 and weld 28.

The use of material to British Standard 309S92L may improve the reliability of a tubular sensor due to its improved resistance to erosion and corrosion, whilst still having a weld integrity which allows manufacture of a buried stud sensor, and having substantially linear temperature versus conductivity characteristics.

Furthermore, sealing of the thermocouple into the body of the sensor portion 4 may protect the thermocouple from the highly corrosive furnace gases, thus prolonging their life, in addition to the improved corrosion resistance of the aforesaid material 309S92L.

Whilst formation of a stud sensor herein as been described by discharge machining, an annular channel may be accurately formed by a chemical etching process.

Similarly an annular recess may be so formed.

The terms "Heat flux" and "Heat flux sensor" as used in this specification will be understood by the man skilled in the art.

The reader's attention is directed to all papers and documents which are filed concurrently with or previous to this specification in connection with this application and which are open to public inspection with this specification, and the contents of all such papers and documents are incorporated herein by reference.

All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and/or all of the steps of any method or process so disclosed, may be combined in any combination, except combinations where at least some of such features and/or steps are mutually exclusive.

Each feature disclosed in this specification (including any accompanying claims, abstract and drawings), may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise. Thus, unless expressly stated otherwise, each feature disclosed is one example only of a generic series of equivalent or similar features.

The invention is not restricted to the details of the foregoing embodiment(s). The invention extends to any novel one, or any novel combination, of the features disclosed in this specification (including any accompanying claims, abstract and drawings), or to any novel one, or any novel combination, of the steps of any method or process so disclosed.

Claims (16)

1. A boiler cleaning management apparatus comprising: a plurality of heat flux sensors; reading means for reading signals of said plurality of sensors and converting said readings into computer readable signals; and computer means for receiving said computer readable signals and determining therefrom an optimum time to activate a boiler cleaning means.

2. A boiler cleaning management apparatus according to Claim 1, wherein said optimum time is determined as a time when a slag layer on a combustion chamber wall of said boiler begins to sinter.

3. A boiler cleaning management apparatus according to Claim 1 or 2, further comprising said boiler cleaning means.

4. A boiler cleaning management apparatus according to Claim 1, 2 or 3, wherein said computer is arranged to activate said boiler cleaning means automatically without the need for human intervention.

5. A boiler cleaning management apparatus substantially as herein described with reference to Figures 1 to 4 of the accompanying drawings.

6. A boiler having a boiler cleaning management apparatus according to any one of Claims 1 to 5.

7. A method of determining an optimum time to perform a boiler cleaning operation, in which data from one or a plurality of sensors is characterised using a first characterisation method to produce a first characterisation of said data over a first time period, and data from one or a plurality of said sensors is characterised using a second characterisation method to produce a second characterisation of said data over a second time period, wherein said second time period is substantially greater than said first time period.

8. A method of determining an optimum time to clean a boiler, in which a decision on whether to activate cleaning is made based on sensor data received in a time period not exceeding 1 minute.

9. A method of determining an optimum time to clean a boiler, in which a decision on whether to activate a cleaning operation is made based on sensor data received in a time period of more than 1 minute but preferably not exceeding 20 minutes.

10. A method of determining an optimum time to clean a boiler, in which a heat flux measurement corresponding to a locality of the boiler is compared with a variable alarm level.

11. A method according to Claim 10, wherein said alarm level is dependent upon a plurality of previous heat flux measurements.

12. A method according to Claim 10 or 11, in which said alarm level is periodically variable at intervals of greater than 1 minute.

13. A first method of determining an optimum time to clean a boiler, said first method comprising: (i) obtaining in a first time period a plurality of sensor data readings corresponding to data of one or a plurality of sensor devices; (ii) analyzing said data readings to obtain one or more characterisations of said data; (iii) comparing a rate of change value of one or more said characterisations to one or more correspondiong reference values; (iv) generating an affirmative signal when a said rate of change becomes less than or equal to a said corresponding reference value.

14. A method according to Claim 13, in which a said characterisation comprises a fourier summation.

15. A method according to Claims 13 or 14, wherein a said reference value is dependent on two or more successive sensor data readings.

16. A method of determining when a layer of ash on a combustion chamber wall begins to substantially sinter, said method comprising the steps of; (i) providing a heat flux sensor on said combustion chamber wall to provide a signal dependent on a local heat flux.

(ii) monitoring said signal over a first predetermined time period to provide data corresponding to a first plurality of heat flux measurement values over said first time period.

(iii) determining a value for a mathematical derivative of said first plurality of values; (iv) repeating stages two and three over one or more successive time periods to provide data corresponding to a respective one or more successive plurality heat flux values and determining a respective one or more mathematical derivatives; and (v) monitoring said mathematical derivative values and producing an activation signal if a rate of change between two or more said derivative values exceeds a predetermined value.

16. A method according to any one of Claims 13 to 15, wherein said sensor data readings comprise data corresponding to one or more readings of heat flux.

17. A method according to any one of claims 13 to 16, wherein said frist period has a duration in the range 0.1 to 60 seconds.

18. A second method of determining an optimum time to clean a boiler, said second method comprising: (a) obtaining over a predefined time period a first plurality of sensor data readings; (b) calculating an average of said first plurality of readings; (c) obtaining at least one further plurality of sensor data readings over at least one further pre defined time period; (d) calculating an average of said further plurality of readings; (e) determining a difference value between said averages of said first and further pluralities of readings; and (f) generating an affirmative signal dependent on a said difference value.

19. A method according to Claim 18, further comprising: (g) repeating steps (a) to (f) a number B of times to obtain a number B of difference values; and (h) determining an average of said B difference values.

20. A method according to Claim 18 or 19, wherein said predefined time period is greater than 1 minute.

21. A third method of determining an optimum time to clean a boiler, said third method comprising: determining an alarm level dependent on two or more sensor data readings of one or more sensor devices and a highest and lowest possible operational loading of said boiler; obtaining local sensor data corresponding to a local area of said boiler from said one or more sensors and comparing said sensor data with said alarm level; and generating an affirmative signal if said local sensor data crosses a threshold determined by said alarm level.

22. A method of determining an optimum time to clean a boiler comprising: monitoring the boiler using any two or more said methods according to any of Claims 7 to 21, to generate an affirmative signal from any one or more of said methods; and generating an optimum time signal in response to the first affirmative signal received from any of said methods.

23. A method of determining an optimum time to clean a boiler substantially as herein described with reference to Figures 1, 2 and/or 3 of the accompanying drawings.

24. A method of manufacturing a heat flux sensor in a furnace combustion chamber wall comprising the steps: (i) indenting a portion of the wall to produce an indent: (ii) filling said indent with a heat conductive in fill material; (iii) forming in said heat conductive material a channel to a first depth, said channel having an open upper end; (iv) forming in said heat conductive material a recess adjacent to said channel, said recess being of a second depth which is shallower than said first depth; (v) inserting a sealing member into said open upper portion of the channel and said recess, such that a lower portion of said channel lies unfilled by said sealing member; (vi) sealing said sealing member into said upper portion of the channel such that said channel remains sealed at the upper portion thereof; and (vii) profiling said in-fill material and said sealing member to form an unbroken outer surface.

25. A method according to Claim 24, wherein said wall comprises a hollow metal tube.

26. A method according to Claim 24 or 25, wherein said indented wall portion is indented to protrude into a central bore of the tube to a depth in the range 2mm to 9mm.

27. A method according to claim 24,25 or 26, wherein said indented wall portion is indented to protrude into said central bore such that no tangent to an inner surface of said indented portion makes an angle of greater than 9 degrees to a main longitudinal axis of said tube.

28. A method according to any one of Claims 24 to 27, wherein said channel is formed to be substantially annular.

29. A method according to any one of Claims 24 to 28, wherein said sealing member is formed of a substantially annular ring.

30. A method according to any one of Claims 24 to 29, wherein said sealing member is formed of a same material type as said in-fill material.

31. A method according to any one of Claims 24 to 30, wherein said in-fill material complies to British standard 309S92L.

32. A method according to any one of Claims 24 to 31, wherein said sealing member is sealed into said upper portion of the channel by welding.

33. A method according to any one of Claims 24 to 33, wherein said unbroken outer surface is formed to be a gas tight surface.

34. A heat flux measuring device for a boiler, comprising an elongate wall of a first thickness, said wall being provided with an indented portion and a sensor portion which closely fits the indented portion to locally thicken said wall to a second thickness, wherein an outer surface of said sensor portion and an immediately surrounding outer surface portion of the tube wall presents a substantially continuous unbroken surface.

35. A device according to Claim 34, wherein said sensor portion comprises two or more heat sensitive sensors buried under said unbroken surface.

36. A device according to Claim 34 or 35, wherein said wall is a substantially cylindrical boiler tube wall.

37. A device according to Claim 34, 35, or 36, wherein the outer surface of the sensor portion matches the outer surface of the immediately surrounding wall outer surface.

38. A heat flux measuring device made of a material which complies to British Standard 309S92L.

xp A method of determining when to activate a cleaning operation of a boiler, said method comprising the steps of; (i) generating a signal dependent upon heat flux absorbed by a portion of a wall of the boiler; (ii) converting said signal in to a measurement data dependent on said absorbed heat flux; (iii) performing a mathematical function on said measurement data; and (iv) activating a boiler cleaning apparatus in response to a signal generated by comparing a result of said mathematical function with a reference data value.