Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
MEASURING A PROPERTY OF MOLTEN GLASS
Document Type and Number:
WIPO Patent Application WO/2013/034918
Kind Code:
A1
Abstract:
A method for determining a property of molten glass in contact with a refractory section is described. An alternating electrical current is passed through a transmitter coil positioned relative to a surface of the refractory section not in contact with molten glass. An alternating primary magnetic field is produced that passes through the refractory section into the molten glass to induce eddy currents in the molten glass. The eddy currents produce a secondary magnetic field that interacts with a receiver coil. By measuring the interaction between the receiver coil and the secondary magnetic field the property of the molten glass can be determined. The method is particularly useful for determining the thickness of the refractory section, which may be a portion of the bottom of a glassmaking furnace. Apparatus for determining the thickness of a refractory section of a glassmaking furnace is also described.

Inventors:
DENNO RICHARD CHRISTOPHER SOMER (GB)
ZYSKO GRZEGORZ (GB)
PEYTON ANTHONY (GB)
Application Number:
PCT/GB2012/052199
Publication Date:
March 14, 2013
Filing Date:
September 07, 2012
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
PILKINGTON GROUP LTD (GB)
DENNO RICHARD CHRISTOPHER SOMER (GB)
ZYSKO GRZEGORZ (GB)
PEYTON ANTHONY (GB)
International Classes:
G01F23/26; C03B5/16; F27D21/00; G01B7/06; G01N27/06; G01N33/38; G01V3/10
Foreign References:
US20050046419A12005-03-03
US20110068807A12011-03-24
CN2172857Y1994-07-27
JPS5983004A1984-05-14
EP0554895A11993-08-11
US5925159A1999-07-20
US20080252287A12008-10-16
Other References:
See also references of EP 2753901A1
"High Temperature Glass Melt Property Database for Process Modelling", 2005, THE AMERICAN CERAMIC SOCIETY, pages: 183
IEEE TRANSACTIONS ON MEDICUL IMAGING, vol. 22, no. 5, 2003, pages 627 - 635
"High Temperature Glass Melt Property Database for Process Modelling", 2005, THE AMERICAN CERAMIC SOCIETY
Attorney, Agent or Firm:
STANLEY, Andrew Thomas (Intellectual PropertyPilkington European Technical Centre,Hall Lane, Lathom, Ormskirk, Lancashire L40 5UF, GB)
Download PDF:
Claims:
CLAIMS

1. A method for determining a property of molten glass, the molten glass being in contact with a refractory section, the refractory section having a first surface in contact with the molten glass and a second opposing surface not in contact with molten glass, the method comprising the steps

(i) positioning a transmitter coil relative to the second surface;

(ii) positioning a receiver coil in a spaced relationship to the transmitter coil;

(iii) passing an alternating electrical current at a first frequency through the transmitter coil to produce an alternating primary magnetic field such that the primary magnetic field passes through the refractory section and the into the molten glass to induce eddy currents in the molten glass, the eddy currents producing a secondary magnetic field and there being an interaction between the secondary magnetic field and the receiver coil;

(iv) measuring the interaction between the receiver coil and the secondary magnetic field; and

(v) using the measurement of the interaction between the receiver coil and the secondary magnetic field to determine the property of the molten glass.

2. A method according to claim 1, wherein the transmitter coil is opposite the second surface.

3. A method according to claim 1 or claim 2, wherein the property is the distance from the transmitter coil or the receiver coil to the molten glass in contact with the first surface, thereby providing a means for determining the thickness of the refractory section.

4. A method according to claim 1 or claim 2, wherein the property is a topographical representation of the first surface, thereby providing means of determining the shape of the interface between the first surface of the refractory section and the molten glass.

5. A method according to claim 1 or claim 2, wherein the property that is determined is the electrical conductivity of a volume element of the molten glass.

6. A method according to claim 5, wherein upon determining the electrical conductivity of the volume element, the temperature of the molten glass in the volume element is calculated.

7. A method according to claim 6, wherein upon calculating the temperature of the molten glass in the volume element, a temperature dependent property of the molten glass at the calculated temperature in the volume element is calculated.

8. A method according to claim 7, wherein the temperature dependent property is chosen from the list consisting of the molten glass viscosity, the molten glass density and the solubility of a gas in the molten glass.

9. A method according to any preceding claim, wherein the interaction between the receiver coil and the secondary magnetic field induces a voltage in the receiver coil and the induced voltage is used to determine the property of the molten glass.

10. A method according to any of the claims 1 to 8, wherein the interaction between the receiver coil and the secondary magnetic field induces a voltage in the receiver coil, and mutual inductance between the transmitter coil and the receiver coil is used to determine the property of the molten glass.

11. A method according to any preceding claim, wherein the property is determined by comparing the measured interaction between the receiver coil and the secondary magnetic field with a prediction of the interaction between the receiver coil and the secondary magnetic field.

12. A method according to claim 11, wherein the prediction is obtained from a mathematical model of the system comprising the transmitter coil, the receiver coil, the refractory section and the molten glass, the mathematical model simulating the magnetic fields.

13. A method according to claim 12, wherein the property is the distance from the transmitter coil or the receiver coil to the molten glass in contact with the first surface and in the mathematical model of the system, the molten glass has a constant value of electrical conductivity.

14. A method according to claim 12 or claim 13, wherein the prediction is obtained by using an inverse model technique.

15. A method according to any preceding claim, wherein the refractory section is part of a glassmaking furnace, preferably a portion of a wall of a glassmaking furnace or a portion of the bottom of a glassmaking furnace.

16. A method according to any preceding claim, wherein the first frequency is one of a range of frequencies over which measurements of the interaction between the receiver coil and the secondary magnetic field are made.

17. A method according to any preceding claim, wherein the first frequency of the alternating electric current is at least 100kHz, preferably 100kHz to 50MHz.

18. A method according to any preceding claim, wherein the receiver coil is opposite the second surface of the refractory section.

19. A method according to any preceding claim, wherein the transmitter coil is in registration with the receiver coil.

20. A method according to any preceding claim, wherein there is a plurality of transmitter coils, preferably eight, spaced apart from one another.

21. A method according to any preceding claim, wherein there is a plurality of receiver coils, preferably eight, spaced apart from one another.

22. A method according to any preceding claim, wherein the alternating current has a sinusoidal waveform.

23. Apparatus for determining the thickness of a refractory section of a glass making furnace, the refractory section having a first surface in contact with molten glass and a second opposing surface not in contact with molten glass, the apparatus comprising a transmitter coil and a receiver coil, an alternating current power supply in electrical communication with the transmitter coil, and an impedance analyser in electrical communication with the receiver coil, the transmitter coil and the receiver coil being housed in a heat resistant enclosure, the transmitter coil being configured such that upon apply an alternating current through the transmitter coil a primary magnetic field is produced that is transmitted through the refractory section into the molten glass to induce eddy currents in the molten glass, the eddy currents generating a secondary magnetic field that induces a voltage in the receiver coil, and the mutual inductance between the receiver coil and the transmitter coil is measurable with the impedance analyser.

24. Apparatus according to claim 23, wherein the transmitter coil and the receiver coil are configured such that in use, the receiver coil and transmitter coil are opposite the second surface.

25. Apparatus according to claim 23 or claim 24,wherein there are eight transmitter coils arranged on one surface of a substrate and eight receiver coils arranged on the opposite surface the substrate, and means for connecting each transmitter coil individually to the alternating current power supply such that the mutual inductance between each transmitter coil and receiver coil pair is measurable.

26. Use of an alternating magnetic field to determine the electrical conductivity of molten glass.

27. Use of an alternating magnetic field to determine the thickness of a refractory section in contact with molten glass.

Description:
MEASURING A PROPERTY OF MOLTEN GLASS

The present invention relates to a method of measuring a property of molten glass in contact with a refractory section, in particular the electrical conductivity of molten glass in a glassmaking furnace or to making an indirect measurement of the thickness of the bottom of a glassmaking furnace containing molten glass.

In the commercial production of glass products such as containers and sheets, a refractory walled furnace is typically used to convert the glass making raw materials into molten glass. Such a furnace comprises side walls, a roof, often referred to as the crown, and a bottom. Usually glass making raw materials are fed into one end of the furnace (the feeder end comprising the front wall of the furnace) and heated to a sufficiently high temperature to convert the glassmaking raw materials into molten glass. The molten glass that is produced forms a pool that contacts the sidewalls and the furnace bottom. Molten glass usually flows out of the furnace from the end of the furnace opposite the feeder end. The molten glass may be fed to a float bath or through rollers, for the production of glass sheets, or formed into gobs, that are subsequently blown in the production of glass containers such as bottles.

It is well known that the temperature of the molten glass has an influence on the glass manufacturing process and typically thermocouples embedded into refractory blocks or optical pyrometers are used to monitor the glass temperature. It is often not desirable to place a thermocouple directly into the molten glass as this can affect the quality of the final product.

It is also well known that in glass making furnaces, the refractory materials in contact with molten glass are susceptible to corrosion and that there are many factors that affect the amount of corrosion of refractory material in contact with molten glass. Such factors include the temperature of the molten glass, the composition of the molten glass and the operating temperature of the refractory. Ideally the operating temperature of the refractory should be kept as low as possible, although in practice this may not be possible due to (i) the energy requirements needed to convert the glass making raw materials into molten glass and the subsequent energy needed to refine the molten glass to remove bubbles, and (ii) the amount of heat that is absorbed by the molten glass. For example, when overhead firing is used to heat the molten glass, with for example gas burners, the furnace bottom temperature is influenced by the amount of iron in the molten glass. For the same base glass composition and the same heating conditions, as the iron content of the glass increases, the furnace bottom temperature decreases. This is because the transparency of the glass to infra red radiation decreases as the iron content of the molten glass increases. This makes it difficult for heat to pass through the molten glass and heat the furnace bottom, with the result being that the furnace bottom temperature decreases. A similar effect if observed when the amount of ferrous iron in the molten glass increases, keeping the total amount of iron constant. Conversely, a glass with very low iron does not absorb as much heat and the furnace bottom temperature increases in temperature.

For example, a soda-lime-silicate float glass composition containing low levels of iron expressed as Fe 2 0 3 , for example between 0.001 and 0.1% by weight has a higher furnace bottom temperature than the same glass composition with a higher level of Fe 2 0 3 , for example 0.11% by weight (as is typical for clear float glass), or greater than 0.5% by weight (as is typical for "tinted" float glass). It has been reported in the art that the temperature rise of the furnace bottom is between 100-150°C higher for a glass composition with low iron compared to clear float glass iron levels. This increased temperature may lead to increased corrosion of the refractory materials and it is important to be able to monitor the thickness of the furnace bottom to avoid catastrophic failure of the furnace. In addition, by extending the useable life of the furnace there are cost savings to be made.

Several methods have been proposed for measuring refractory wall thickness of glass making furnaces including the use of ultra sound, radio wave reflection and electrodes implanted in refractory blocks.

Another known technique for determining the thickness of a furnace bottom uses thermal imaging. The temperature of a part of the exterior of the furnace bottom is measured using a thermal imaging camera. By knowing the exterior temperature of the furnace bottom and the thermal conductivity of the block, the thickness of the bottom can be determined. This method has the problem that any structural changes to the bottom, for example cracking or delamination will decrease the accuracy of the measurement. Additionally, it is generally not possible to install such an arrangement to continually monitor the bottom thickness, so errors may arise due to different ambient conditions each time the measurement is made. Options are available for measuring the furnace wall thickness when not in use, i.e. when the furnace is at ambient temperature. However such techniques are not suitable for measuring properties on a glass making furnace that is operational, that is, when the glassmaking furnace contains molten glass. In the technical field of furnaces for molten metal, use of eddy currents induced in the molten metal has been described and used to measure the thickness of the furnace walls. In EP0554895A1 a sensor coil is used to induce an eddy current in molten metal. The eddy current produces a secondary magnetic field that is measured by a sensor coil. The detected signal is used to determine the position of the metal. Using the distance of the primary coil from the wall, it is possible to calculate the thickness of material in between the outer face of the furnace and the detected metal position. This is an indicator of the wall thickness of the furnace. The method disclosed in EP0554895A1 can be used with metal lined furnaces or all- refractory material furnaces.

The eddy current principle described in EP0554895A1 relies on the electrical conductivity (σ) of molten metal to induce an eddy current in the molten metal. The electrical conductivity σ of molten metal is typically 2 x 10 6 S/m and the electrical conductivity of solid metal is typically 1 x 10 6 S/m. For the avoidance of doubt, the unit 'S' refers to the SI unit of electrical conductivity, the Siemen, and S=Q ~1 where Q=ohms. There is relatively little change in the electrical conductivity of metals such as steel with

temperature.

By comparison, the electrical conductivity of solid glass, such as float glass, is in the region of lO "10 - 10 "14 S/m, that is, many orders of magnitude lower than the electrical conductivity of a solid metal such as steel and as such glass is an electrical insulator at room temperature. However unlike steel, the electrical conductivity of molten glass increases by many orders of magnitude to such an extent that the molten glass becomes electrically conductive at higher temperatures.

The electrical resistivity (p) of typical float glass compositions as a function of temperature has been presented in the text book "High Temperature Glass Melt Property Database for Process Modelling", p. 183, Ed. T. P. Seward, Pub. The American Ceramic Society (2005), ISBN 1 -57489-225-7. In this textbook, the electrical resistivity of a number of float glass compositions at temperatures between 950°C and 1450°C are given. In the aforementioned reference, for the float glass composition referred to as FL base, the electrical resistivity is given below in table 1. The electrical conductivity σ is obtained therefrom because σ=ρ _1 .

Table 1.

At 1450°C, it can be seen that the electrical conductivity of molten float glass is many orders of magnitude lower that the electrical conductivity of molten steel (22.42 S/m compared with 2xl0 6 S/m).

In the field of molten glass, eddy currents have been used to determine the weight of a molten glass gob produced in a container furnace, as described in US 5,925,159. As a gob of molten glass is dropped through a magnetic field, the perturbation to the magnetic field is detected and converted into a weight measurement. In this way, the gob weight could be accurately controlled. In contrast to the present problem, the molten glass gob is not contained in a refractory vessel and would not be considered relevant to the problem of determining the wall thickness of a refractory vessel containing molten glass.

It is known from US2008/0252287Al to use a Lorentz force anemometer to determine properties of electrically conductive moving substances. The Lorentz force anemometer comprises a magnetic system with at least two magnetic poles (NORTH and SOUTH) to create a primary field. The poles are arranged at opposite sides of the cross section of a substance being inspected. This arrangement ensures that the primary field or lines of force running between the NORTH and SOUTH in the air gap pass through the entire cross section of the substance and thus create the precondition for a complete determination of the spatial distribution of flow velocity and electrical conductance. Such systems do not work well where flow velocities and/or electrical conductivity are low.

Surprisingly the present inventors have discovered that an alternating magnetic field is able to penetrate the refractory bottom of an operational glass making furnace and generate eddy currents in molten glass contained in the furnace. The magnetic field components created by these induced eddy currents can then be sensed using one or more receiver coils.

Accordingly the present invention provides from a first aspect a method for determining a property of molten glass, the molten glass being in contact with a refractory section, the refractory section having a first surface in contact with the molten glass and a second opposing surface not in contact with the molten glass, the method comprising the steps (i) positioning a transmitter coil relative to the second surface; (ii) positioning a receiver coil in a spaced relationship to the transmitter coil; (iii) passing an alternating electrical current at a first frequency through the transmitter coil to produce an alternating primary magnetic field such that the primary magnetic field passes through the refractory section and the into the molten glass to induce eddy currents in the molten glass, the eddy currents producing a secondary magnetic field and there being an interaction between the secondary magnetic field and the receiver coil; (iv) measuring the interaction between the receiver coil and the secondary magnetic field; and (v) using the measurement of the interaction between the secondary magnetic field and the receiver coil to determine the property of the molten glass.

Upon the discovery that molten glass may be used as an electrically conductive liquid in which eddy currents may be induced by passing an alternating magnetic field through the molten glass, there are a number of properties of the molten glass that may be determined using this concept. Some properties are related directly to the molten glass itself, whereas some properties are obtainable once the properties of the molten glass are known, for example the position of the molten glass in contact with the first surface provides a means of determining the thickness of the refractory section.

Preferably in step (i) the transmitter coil is positioned to be opposite the second surface. This is advantageous because more lines of magnetic flux are able to penetrate the refractory section than when the transmitter coil is positioned beyond the perimeter of the second surface, for example when the transmitter coil is positioned towards one side of the refractory section.

Preferably the property is the distance from the transmitter coil or the receiver coil to the molten glass in contact with the first surface, thereby providing a means for determining the thickness of the refractory section. The thickness of the refractory section may be determined using the known position of transmitter coil or receiver coil relative to the second surface of the refractory section and the determined distance from the transmitter coil or receiver coil of the molten glass in contact with the first surface of the refractory section.

Preferably the property is a topographical representation of the first surface. This provides a means of determining the shape of the interface between the first surface of the refractory section and the molten glass. This is useful for determining whether there has been localised corrosion of the refractory section due to contact of the first surface with the hot molten glass.

Preferably the property that is determined is the electrical conductivity of a volume element of the molten glass. Given that the electrical conductivity of molten glass for a given composition is a function of the molten glass temperature, see for example table 1 , once the electrical conductivity of a volume element of the molten glass has been determined, the temperature of the molten glass in the volume element may be calculated. When the temperature Tv of a volume element of the molten glass has been calculated it is possible to calculate a temperature dependent property of the molten glass at the calculated temperature Tv of the molten glass in the volume element. Preferably the temperature dependent property is chosen from the list consisting of the molten glass viscosity, the molten glass density and the solubility of a gas in the molten glass. Examples of such a gas are oxygen, nitrogen, carbon dioxide, carbon monoxide or sulphur dioxide. It is possible to calculate flow velocity and visualisations of any of the preceding properties using the method according to the first aspect of the present invention.

Preferably the interaction between the receiver coil and the secondary magnetic field induces a voltage in the receiver coil, and the induced voltage is used to determine the property of the molten glass. Preferably the interaction between the receiver coil and the secondary magnetic field induces a voltage in the receiver coil, and the mutual inductance between the transmitter coil and the receiver coil is used to determine the property of the molten glass. The measured mutual inductance has a real part and an imaginary part. The magnitude and the phase of the mutual inductance may be calculated from the real and imaginary parts. Preferably the property of the molten glass is determined using the imaginary part of the measured mutual inductance. Preferably the property of the molten glass is determined using the real part of the measured mutual inductance. Suitably both the real part of the measured mutual inductance and the imaginary part of the measured mutual inductance are used to determine the property of the molten glass.

Preferably the property is determined by comparing the measured interaction between the receiver coil and the secondary magnetic field with a prediction of the interaction between the receiver coil and the secondary magnetic field. Preferably the prediction is obtained from a mathematical model of the system which comprises the transmitter coil, the receiver coil, the refractory section, the molten glass and any other nearby objects that may affect the magnetic field, the mathematical model simulating the magnetic fields. Preferably when the property being determined is the distance from the transmitter coil or the receiver coil to the molten glass in contact with the first surface, the electrical conductivity of the molten glass in the mathematical model has a constant value. Preferably the prediction is obtained by using an inverse model technique. If the electrical conductivity of the molten glass in the mathematical model is not assigned a constant value, the electrical conductivity distribution may be used as a variable in the solution of an inverse model. It is possible to assign the electrical conductivity distribution of the molten glass with a predicted set of values using another model that predicts the temperature distribution of the molten glass based on known boundary conditions, and to then use the known temperature dependence of the electrical conductivity. The refractory section may be in the vicinity of a metal member such that the alternating primary magnetic field induces eddy currents in the metal member. This makes measurement of the secondary magnetic field due to eddy currents in the molten glass more difficult because the secondary magnetic field due to eddy currents in the metal member may be stronger than the secondary magnetic field due to eddy currents in the molten glass. Preferably the refractory section is part of a glassmaking furnace. Preferably the refractory section is a portion of a wall of the glassmaking furnace or a portion of the bottom of the glassmaking furnace. For a newly built glassmaking furnace, the furnace bottom may be between 30cm and 70cm thick. Over the campaign life of the glassmaking furnace, the thickness of the refractory bottom may corrode, in which case the furnace bottom may be less thick than the furnace bottom at the start of the campaign life. Usually refractory corrosion occurs gradually, although it is possible that the furnace bottom suffers localised corrosion, in which case a portion of the furnace bottom is thinner than other portions of the bottom of the glassmaking furnace.

Suitably the refractory section is less than 70cm thick. Suitably the refractory section is more than 20cm thick. Suitably the refractory section is between 20cm and 70cm thick.

In an operational glassmaking furnace, the molten glass may have a depth of less than 2m, typically about lm. It will be readily apparent to a person skilled in the art that as the depth of molten glass in the glassmaking furnace approaches zero, the ability to induce eddy currents in the molten glass diminishes.

For the avoidance of doubt, in a preferred embodiment the present invention provides a method for determining the thickness of a refractory section, the refractory section having a first surface in contact with molten glass and a second opposing surface not in contact with molten glass, the method comprising the steps (i) positioning a transmitter coil to face the second surface; (ii) positioning a receiver coil in a spaced relationship to the transmitter coil; (iii) passing an alternating electrical current at a first frequency through the transmitter coil to produce an alternating primary magnetic field, such that the primary magnetic field passes through the refractory section and the into the molten glass to induce eddy currents in the molten glass, the eddy currents producing a secondary magnetic field and the secondary magnetic field inducing a voltage in the receiver coil; (iv) measuring the mutual inductance between the transmitter coil and receiver coil; and (v) using the measured mutual inductance to determine the thickness of the refractory section. When the refractory section is part of a glassmaking furnace, there is a particular difficulty in measuring the interaction of the secondary magnetic field with the receiver coil because of the presence of supporting steelwork that is used in the furnace superstructure. Since steel has a much higher electrical conductivity than molten glass, eddy currents may be more readily induced in the supporting steelwork by the primary magnetic field than in the molten glass at certain frequencies. For low frequencies, eddy current generation in the surrounding steelwork creates a large strength secondary magnetic field which causes a problem in being able to measure the lower strength secondary magnetic field produced by eddy currents generated in the molten glass. Furthermore the ferromagnetic properties of the steel also affect the magnetic fields.

One way to reduce the effects of the steelwork in a glassmaking furnace is to use at least two different frequencies of primary magnetic field by using two different frequencies of alternating electric current. The first frequency should be chosen to minimise the effects of the steelwork, and this first frequency is chosen such that the penetration depth of the primary magnetic field into the steelwork is of the order of a few um. The second frequency is chosen such that the effect of the molten glass is low, and the observed signal is primarily due to the steelwork. The signal due to the steelwork can then be used to remove the signal from the steelwork at the higher frequency. Measurements can then be made at each frequency and the effects of the steelwork removed by suitably normalising the two sets of measurements.

The optimum frequencies can be determined by sweeping across a range of excitation frequencies. This allows the response at each frequency to be determined and to then use a single frequency to perform measurements. The single frequency can be chosen such that effects due to the steelwork are minimised. Alternatively all of the frequency response data can be used.

Other embodiments have other preferable features.

Preferably the first frequency is one of a range of frequencies over which

measurements of the interaction between the receiver coil and the secondary magnetic field are made.

Preferably the first frequency of the alternating electric current is at least 100kHz.

Preferably the first frequency of the alternating electric current is in the range of lOOkHz to 50MHz. Preferably the first frequency of the alternating current is in the range of 500kHz to

10MHz, more preferably 500kHz to 5MHz, even more preferably 700kHz to 3MHz, most preferably 900kHz to 2MHz.

Suitably the transmitter coil is rectangular, preferably square, in outline.

Suitably the receiver coil is rectangular, preferably square, in outline. Preferably at step (ii) the receiver coil is positioned to be opposite the second surface of the refractory section.

Preferably the transmitter coil and the receiver coil are both opposite the second surface of the refractory section. Preferably the transmitter coil faces the second surface of the refractory section.

Preferably the receiver coil faces the second surface section.

Preferably both the receiver coil and the transmitter coil face the second surface of the refractory section.

Preferably the transmitter coil has circular symmetry about a transmitter coil axis. Preferably the transmitter coil axis is at an angle of 90° to the second surface of refractory section when measurements of the interaction between the receiver coil and secondary magnetic field are made.

Preferably the receiver coil has circular symmetry about a receiver coil axis.

Preferably the receiver coil axis is at an angle of 90° to the second surface of the refractory section when measurements of the interaction between the receiver coil and secondary magnetic field are made.

Preferably there is a plurality of transmitter coils spaced apart from one another.

Preferably there is a plurality of receiver coils spaced apart from one another.

The plurality of transmitter coils and/or receiver coils may be in a linear array. Preferably the plurality of transmitter coils is in registration with the plurality of receiver coils.

Preferably there are eight transmitter coils. Preferably the eight transmitter coils are in a linear array.

Preferably there are eight receiver coils. Preferably the eight receiver coils are in a linear array.

Preferably the mutual inductance is measured between each combination of pairs of transmitter coil and receiver coil. The transmitter coil may be on a first substrate and the receiver coil may be on a second substrate. Preferably the transmitter coil and the receiver coil are on the same substrate. Preferably the transmitter coil is on a first major surface of the substrate and the receiver coil is on the opposing second major surface of the substrate.

It will be readily apparent to a person skilled in the art that when the transmitter coil and the receiver coil are in mechanical communication, for example by being on the same substrate, step (i) and step (ii) of the method of the first aspect of the invention are carried out at the same time.

Preferably the alternating current has a sinusoidal waveform.

Preferably the alternating current has a periodic waveform, suitably a saw tooth waveform or a square wave waveform.

The refractory section may comprise a portion that is not in contact with molten glass, for example a roof portion of a glassmaking furnace. When the method is used in this way, the property of the molten glass that is determined is the glass level.

Preferably the glass has a velocity less than 30mm/s, more preferably less than 20mm/s, even more preferably less than 15mm/s, even more preferably less than lOmm/s.

It will be readily apparent to one skilled in the art that the molten glass must be at a sufficiently high temperature such that the molten glass is electrically conductive so that eddy currents may be induced in the molten glass by an alternating magnetic field. Preferably the molten glass has a temperature between 700°C and 1600°C, more preferably between 1000°C and 1600°C.

The molten glass may have a soda- lime-silicate composition, such as is used for making windows or containers, or a borosilicate composition. Glass for windows is usually made using the float process and usually has a soda-lime-silicate composition. For low levels of iron oxide, typically around 0.1% by weight Fe 2 0 3 , such a float glass composition is known as clear float glass. Clear float glass has a composition as defined in BS EN 572-1 and BS EN 572-2 (2004). When such a composition is manufactured with lower levels of iron oxide being present, such a composition is usually known as low iron float glass, and typically has an iron oxide level of <0.1% by weight, usually 0.001-0.1% by weight Fe 2 0 3 . The present invention also provides from a second aspect an apparatus for

determining the thickness of a refractory section of a glass making furnace, the refractory section having a first surface in contact with molten glass and a second opposing surface not in contact with molten glass, the apparatus comprising a transmitter coil and a receiver coil, an alternating current power supply in electrical communication with the transmitter coil, and an impedance analyser in electrical communication with the receiver coil, the transmitter coil and the receiver coil being housed in a heat resistant enclosure, the transmitter coil being configured such that upon applying an alternating current through the transmitter coil a primary magnetic field is produced that is transmitted through the refractory section into the molten glass to induce eddy currents in the molten glass, the eddy currents generating a secondary magnetic field that induces a voltage in the receiver coil, such that the mutual inductance between the receiver coil and the transmitter coil is measurable with the impedance analyser.

Preferably the transmitter coil and the receiver coil are arranged to be opposite the second surface in use.

Preferably the transmitter coil and the receiver coil are arranged to be face the second surface in use.

In a preferred embodiment the apparatus has eight transmitter coils arranged on one surface of a substrate and eight receiver coils arranged on the opposite surface of the substrate. Preferably the apparatus comprises means for connecting each transmitter coil individually to the alternating current power supply such that the mutual inductance between each transmitter coil and receiver coil pair is measurable.

In another embodiment the alternating current power supply is operable at a frequency of at least 100kHz, preferably in a frequency range of more than 100kHz, preferably in the range of 500kHz to 10MHz, more preferably 500kHz to 5MHz, even more preferably

700kHz to 3MHz, most preferably 900kHz to 2MHz. Suitably the alternating current power supply generates a sinusoidal output.

The present invention also provides from a third aspect use of an alternating magnetic field to determine the electrical conductivity of molten glass. The present invention provides from a fourth aspect use of an alternating magnetic field to determine the thickness of a refractory section in contact with molten glass. The invention will now be described with reference to the following figures (not to scale) in which:

Figure 1 shows a cross section through the melt end of a typical glassmaking furnace;

Figure 2 shows a perspective view of a portion of the furnace bottom of a typical glassmaking furnace;

Figure 3 shows a schematic diagram of a sensor used to make the measurements in accordance with the first aspect of the present invention;

Figure 4 shows a schematic diagram of the apparatus used to make the measurements in accordance with the first aspect of the present invention;

Figure 5 shows a perspective view of part of the sensor prior to being positioned below a furnace bottom;

Figure 6 shows a cross section of the sensor head located below a furnace bottom in a position ready to make measurements;

Figure 7 shows the variation of the comparison function at two different frequencies with thickness of refractory bottom used in the modelled data (measured data normalised using first reference measurements);

Figure 8 shows the variation of the comparison function at two different frequencies with thickness of refractory bottom used in the modelled data (measured data normalised using the second reference measurements);

Figure 9 shows the variation of the comparison function at two different frequencies with thickness of refractory bottom used in the modelled data (measured data normalised using the first reference measurements) when the measurements are repeated;

Figure 10 shows the variation of the comparison function at two different frequencies with thickness of refractory bottom used in the modelled data with reduced lift-off of the sensor (measured data normalised using the first reference measurements);

Figure 11 shows the variation of the comparison function at two different frequencies with thickness of refractory bottom used in the modelled data with modelled glass electrical conductivity of lOS/m (measured data normalised using the first reference measurements); Figure 12 shows the variation of the comparison function at two different frequencies with thickness of refractory bottom used in the modelled data with modelled glass electrical conductivity of lOS/m (measured data normalised using the second reference measurements); and Figure 13 shows the real part of the measured mutual inductance in response to a change in flow of the molten glass in the glassmaking furnace.

Figure 1 shows a schematic vertical cross section of the melter end of a typical float glass making furnace. The furnace 1 has a left sidewall 3 and a right sidewall 5, a crown 7 and a bottom 9. The furnace has a width 11 of about 11m. The length of the melter section is typically 20m, with a refiner and conditioner section adding typically another 10m to the overall furnace length. The sidewalls, crown and bottom are made of suitable refractory materials well know to a person skilled in the art. Inside the glassmaking furnace there is a pool of molten glass 13 at a depth of about lm. The thickness of the refractory bottom 9 is typically 0.3m to 0.7m. Typical glass velocity in such a melter is l-3mm/s, with a maximum glass velocity of 5-1 Omm/s.

The furnace 1 is supported above ground level by a steelwork superstructure. Dwarf walls 15 are located in steel channels 17 and the furnace sits on top of the dwarf walls. The steel channels are themselves resting on refractory brickwalls 19, and each refractory brickwall 19 is supported on a steel beam 21. This type of glassmaking furnace construction is well known in the art.

The cut-out region on the figure shown in refractory brick wall 19 and steel beam 21 is used for illustration purposes so that the spacing 23 between adjacent dwarf walls, which is typically 40cm, can be seen.

Figure 2 shows in more detail a perspective view of a refractory section 9' of the furnace bottom 9. The refractory section 9' comprises AZS refractory tiles 25 on top of clay flux refractory blocks 27. The lower surface 29 of the clay flux refractory block 27 is the exterior facing surface of the furnace bottom and is not in contact with molten glass when the furnace is operational. In the art, this surface 29 is often known as the "cold surface".

Opposite the cold surface 29 is surface 31 of the AZS refractory tile. The surface 31 is in contact with molten glass when the furnace is operational. The clay flux refractory block sits on dwarf walls 15. The dwarf walls 15 are made of a suitable refractory material and are located in steel U-shaped beams 17. The steel U-shaped beams 17 sit on top of brick walls 19 and the brick walls 19 sit on top of steel I-shaped beams 21. The beams 17 and 21 are part of the steelwork superstructure supporting the glassmaking furnace above ground level. The steel U-shaped beams 17 typically have a width 41 between 10cm and 20cm, usually 16cm. The separation 43 of adjacent beams 17 is typically between 30cm and 50cm, usually 40cm. The height 51 of the steel U-shaped beam is about 8cm.

The separation 45 of adjacent steel I-shaped beams 21 is typically less than 100cm, usually about 80cm. Each steel I-shaped beam 21 has a height 47 of about 80cm.

The brickwall 19 typically has a height 49 of about 20cm.

The overall thickness 53 of the furnace bottom prior to the start of the campaign life is usually between 30cm and 70cm thick. When operational, the height of glass relative to surface 31 is typically less than 2m, usually about lm.

As will be readily apparent from figure 2, apparatus suitable for implementing the method according to the first aspect of the present invention has limited access to the cold surface 29 due to the surrounding steelwork.

Figure 3 shows a cross section of the sensor used to make measurements on a glassmaking furnace of the type shown in figure 1. The sensor 60 comprises an array of coils in a planar arrangement on a printed circuit board (PCB) 62. On the upper surface of the PCB there is an upper array 64 of transmitter coils and on the opposite lower surface of the PCB there is a lower array 66 of receiver coils. The upper array 64 consists of eight copper coils 68, 70, 72, 74, 76, 78, 80 and 82 formed by etching the upper surface of the PCB. The lower array 66 consists of eight copper coils 84, 86, 88, 90, 92, 94, 96 and 98 formed by etching the lower surface of the PCB. The two arrays 64, 66 are in registration, such that transmitter coil 68, 70, 72, 74,76, 78, 80 and 82 is directly above the respective receiver coil 84, 86, 88, 90, 92, 94, 96 and 98.

Each coil is essentially square in outline and has two turns. Each coil has a side length of about 70mm. The separation between adjacent transmitter or receiver coils is about 8mm.

Each transmitter coil is configured to be connectable to a source of alternating electric current. It is possible to supply an alternating current to each transmitter coil individually or in combinations thereof, although it is preferred to supply an alternating electric current to one transmitter coil at a time to make measurements.

Each receiver coil is configured to be connectable to a suitable measurement device. Each receiver coil may be used individually or in combinations thereof. Preferably each receiver coil is used individually such that when making measurements, one transmitter coil is connected to a source of alternating electrical current and one receiver coil is in electrical communication with a suitable measuring device.

The PCB 62 with transmitter and receiver coils thereon is mounted in a heat resistant housing 100. The preferred material for the housing should be lightweight and be able to withstand the above ambient temperatures experienced beneath a glassmaking furnace, as the temperature of the furnace bottom 9 can be as high as 700°C. Suitable refractory blanket material may be used to cover the housing 100 for added temperature resistance. The sensor may be enclosed in a pre-cast housing made of temperature resistant material, such as a castable refractory. The housing 100 may have means of providing a cooling fluid such as compressed air onto the PCB to ensure stable operating temperatures of the sensor. The housing 100 may also contain other electronic components such as amplifiers. The housing may contain a peltier cooler to cool the coils and/or other electronic components contained therein.

The sensor 60 is used as follows to make measurements. An alternating electric current is supplied to one of the transmitter coils, for example coil 68. No electric current is supplied to the other transmitter coils. The alternating electric current flowing through the coil 68 produces a primary alternating magnetic field. The alternating magnetic field is able to interact with surrounding materials and the influence thereon can be measured using the receiver coils. It has been found that the primary alternating magnetic field is able to penetrate through the entire glassmaking furnace bottom and produce eddy currents in the molten glass above. The eddy currents produce a secondary alternating magnetic field that can be measured using the receiver coils because the secondary alternating magnetic field interacts with the receiver coils. The interaction induces an electromotive force (emf) or voltage in the respective receiver coil. To measure the effect of the alternating primary magnetic field with the molten glass, the mutual inductance between the transmitter coil being supplied with alternating electric current and each receiver coil that has interacted with the secondary magnetic field is measured. For example, when coil 68 is used as the transmitter coil, the mutual inductance between the coil combinations given in table 2 may be used to determine the interaction of the secondary magnetic field and the surroundings.

It has been found that a better signal to noise ratio can be achieved by not using the receiver coil directly below the transmitter coil being supplied with alternating electric current. In addition, for at least the same reason, it has been found that the receiver coil or coils directly adjacent to the chosen transmitter coil is preferably not used. With reference to table 2 above, this means that preferably only combinations 3-8 were used to make measurements when transmitter coil 68 was supplied with an alternating electric current, in order to improve the signal to noise ratio.

Table 2.

It has been found that by having a plurality of transmitter and receiver coils and to use them in pairs as described above, it is possible to obtain a depth profile of the property being measured because as the separation of the transmitter and receiver coils increases, the depth of penetration of the primary magnetic field into the molten glass increases. This means that molten glass at different distances from the first surface of the refractory section may be inspected depending upon which pair of coils is used. In addition, the interaction of the primary magnetic field with the receiver coil is reduced as the transmitter coil and receiver coil are spaced further apart. Figure 4 shows a schematic of the data collection apparatus used with sensor 60. The sensor is in electrical communication with a multiplexer 102 via suitable cabling 104. The multiplexer 102 is in electrical communication with alternating current power supply 106.

The multiplexer 102 allows alternating electric current to be supplied separately to each transmitter coil in the upper array 64. The multiplexer also allows each receiver coil in the lower array 66 to be used separately to make measurements, as described previously.

The multiplexer 102 is in electrical communication with computer 108 via suitable cabling 110. The computer controls the operation of the multiplexer to select the various transmitter/receiver coil combinations.

The power supply 106 is able to produce a sinusoidal excitation current up to a frequency of 32MHz. For frequencies below 10MHz, the maximum amplitude of the voltage and current is 3V and 60mA respectively. For frequencies above 10MHz, the maximum amplitude of the voltage and current is IV and 20mA respectively. A power amplifier was used to increase the current supplied to each transmitter coil to around 0.62A RMS at a frequency of lMHz.

The multiplexer 102 is also in electrical communication with data acquisition unit 112. The data acquisition unit is also in electrical communication with computer 108 and may be controlled by the computer and may transfer data to the computer for storage.

The data acquisition unit 112 is a Solartron SI 1260 A impedance analyser and is used to measure the mutual inductance for each transmitter/receiver coil combination. The data acquisition unit measures both the real part of the mutual inductance and the imaginary part of the mutual inductance for each particular transmitter/receiver coil combination. The magnitude and phase of the mutual inductance can be calculated directly from the real and imaginary parts.

The multiplexer, power supply, data acquisition unit and computer may be located in a temperature controlled cabinet 114 to provide a controlled environment when

measurements are taken.

Figure 5 shows a schematic of part of the sensor 60 before being positioned adjacent the bottom of the glassmaking furnace, that is adjacent the cold surface. The eight transmitter coils in the upper array 64 are identified and are opposite the cold surface 29. As shown in figure 3, there is a lower array of receiver coils on the opposite side of the PCB. The sensor is moved in the direction of arrow 116 and positioned adjacent the cold surface 29 in order to make measurements. The PCB 62 is opposite the cold surface 29 and parallel thereto such that the transmitter coils and receiver coils are facing the cold surface 29. Figure 6 shows a cross section of a portion of a glassmaking furnace with the sensor

60 adjacent the cold surface 29 of the furnace bottom 9. As can be seen from figure 6, the transmitter and receiver coils are opposite the cold surface 29. In this arrangement, the PCB 62 is substantially parallel to the cold surface 29 such that the transmitter and receiver coils face the cold surface 29. It is possible for the PCB 62 not to be parallel to the cold surface 29 when measurements are made.

Measurements are made as follows and with reference to figure 6. Initially a first reference set of readings are made for each particular combination of transmitter/receiver coil. The sensor is located away from any steelwork to reduce the effects of the eddy current generation in the steel by the primary alternating magnetic field. The first reference set of measurements were made using a sinusoidal alternating electric current at a frequency of 1MHz and then repeated at a frequency of 1.58MHz.

Next, the sensor 60 was placed adjacent to the cold surface of the glassmaking furnace bottom, as shown in figure 6. Suitably a mechanical jack arrangement may be used to move the sensor into position and to keep the sensor in position whilst measurements are made.

As described previously, only one of the transmitter coils (for example transmitter coil 68) is supplied with alternating electric current to produce an alternating primary magnetic field. The lines of flux 200 penetrate the entire thickness of the refractory bottom 9 and penetrate the molten glass 13. The primary alternating field generates eddy currents 202 in the molten glass, which generates a secondary alternating magnetic field (not shown). The secondary alternating magnetic field cuts the receiver coils in the lower array 66 generating an electromotive force (emf) in each receiver coil. Measurements are then made of the mutual inductance between the transmitter coil that is used to generate the primary magnetic field and each of the individual receiver coils, see for example table 2 indicating one particular set of transmitter/receiver coil combinations. The magnitude and the phase of the mutual inductance measured by the Solartron impedance analyser are stored in the computer for further analysis to determine the thickness of the refractory bottom of the glassmaking furnace. The real part of the mutual inductance and the imaginary part of the mutual inductance may also be stored.

Measurements are made with the sensor 60 in this configuration at the same frequencies as used for the first reference measurements, which for the above example were lMHz and 1.58MHz.

The electric current is disconnected from the transmitter coil and the next transmitter coil in the upper array 64 is then connected to the power supply by operating the multiplexer.

The measurements of mutual inductance between this transmitter coil and each of the receiver coils are then measured at each frequency. This is repeated until each of the transmitter coils have been used to produce the primary magnetic field, and the respective measurements of mutual inductance have been made.

Finally the sensor is moved away from the cold surface of the furnace bottom and a second set of reference measurements are taken. Again, frequencies of lMHz and 1.58MHz were used to make the second set of reference measurements.

The measured data when the sensor is adjacent the cold surface of the furnace bottom is then normalised with respect to the first reference measurement data at each respective frequency. The measured data when the sensor is adjacent the cold surface of the furnace bottom is also normalised with respect to the second reference data at each respective frequency.

The data collected consists of the measured magnitude and phase of the mutual inductance at each frequency of primary magnetic field for each combination of

transmitter/receiver coil in the upper and lower coil arrays 64, 66.

Ideally there should be no change in the values (known as "drift") between the first and second reference measurements. Any drift can be compensated by taking the first and second reference measurements.

The steelwork 17, 21 associated with the construction of the glassmaking furnace 1 as shown in figure 1 is able to produce eddy currents in the presence of an alternating magnetic field such as the primary alternating magnetic field. As such, the eddy currents generated in the steelwork also produce an alternating secondary magnetic field that can induce an emf in the receiver coils. Furthermore the ferromagnetic properties of the steel also affect the magnetic fields. This makes measurement of the secondary magnetic field associated solely with the eddy currents in the molten glass difficult.

One way to overcome this problem is to use a range of frequencies for the alternating current applied to the transmitter coils and the measure the mutual inductance of each transmitter/receiver combination at each frequency in the scan range.

This can be a time consuming process and generates a large amount of data for subsequent analysis.

Another way to overcome this problem is to use a frequency for the alternating current that is supplied to the transmitter coil that produces a larger effect in the molten glass than in the surrounding steelwork. This has been achieved by utilising the skin effect of the steelwork, wherein the depth of eddy current generation decreases as the frequency of the primary alternating magnetic field increases. It has been found that by using a frequency for the alternating electric current in the range of 100kHz - 5MHz, the effects of eddy current generation in the steelwork can be minimised by ensuring the skin depth in the steelwork is several μιη, for example less than ΙΟΟμιη, preferably less than ΙΟμιη.

In the above described method, the frequency of the alternating current was lMHz and 1.58MHz. These frequencies were found to provide a suitably strong secondary magnetic field from the eddy currents induced in molten glass whilst the electromagnetic effects of the steelwork may be predicted in a mathematical model.

The data stored in the computer comprises a list of each transmitter/receiver combination with the corresponding measurement of the magnitude and phase of the mutual inductance. In the absence of steelwork, it is possible to use an analytical solution to determine the thickness of the refractory bottom. However in the presence of the surrounding steelwork in a glassmaking furnace, the data is processed as follows to determine the thickness of the refractory bottom.

A mathematical model of the sensor in the measuring position is constructed using commercially available software based on the finite element method (FEM) for solving electromagnetic fields. Examples of such software are MAXWELL vl2 from Ansoft Corporation and COMSOL Multiphysics v4 from COMSOL Inc. The FEM software is able to take into account the dimensions of the various refractory features (dwarf wall etc) and surrounding steelwork. In addition material properties can be changed and appropriate boundary conditions set for calculations. In the simulations the molten glass electrical conductivity was assumed to be a constant value of 8 S/m. This simplifying assumption speeds up calculations and makes entering the parameters in the FEM software more straightforward. The simulation could be refined by using a temperature dependent electrical conductivity for the molten glass, and the temperature of the molten glass could be obtained from a 3-dimensional flow model of the glassmaking furnace, such as can be carried out using Ansys FLUENT software. Changes to the electrical conductivity of the molten glass used in the simulations are discussed with reference to figure 1 1.

Using the FEM model and the known input parameters, the predicted mutual inductance (magnitude and phase) of each pair of transmitter/receiver coil combination can be determined.

One way to determine the thickness of the refractory bottom is to use an inverse modelling technique. The measured mutual inductance data for each transmitter/receiver coil combination can be used as criteria in the inverse model and the values of the FEM forward model adapted to obtain the best fit with the measured data. Such techniques are time consuming and for the present determination of refractory bottom thickness, a simplified approach was used.

The FEM model program was used to simulate the response of the system (the mutual inductance of the various transmitter/receiver pair combinations) for different furnace bottom thicknesses. It is then possible to compare each of the model predictions of the mutual inductance of the various transmitter/receiver pair combinations with the respective measured values. The thickness of the refractory bottom is then determined as that bottom thickness used in the simulation that best matches the measured mutual inductance data.

A quality factor Cp can be calculated using all the measured mutual inductance data and the calculated data as follows: where P represents a lookup table entry, & 1 is a set of measured values, (f represents a lookup table of computed values, n corresponds to all the transmitter/receiver coil combinations and/ represents the frequency at which the measurements were made (in the above example lMHz and 1.58MHz).

Using this approach, the bottom thickness of an operating float glass furnace containing molten float glass was determined and the results shown in figure 7. In figure 7 axis 210 is the value of the quality factor Cp and axis 212 is the refractory thickness in centimetres (cm) of the furnace bottom in the simulation.

Figure 7 shows the difference between quality factor Cp for each thickness of refractory bottom used in the respective model using the measured data normalised using the first reference data set at a frequency of either lMHz (line 214, square points) and 1.58MHz (line 216, diamond points). As described above, a number of forward models were simulated using the FEM model software, each model varying the furnace bottom thickness. The minimum of the quality factor Cp is indicative of the actual thickness of the refractory bottom of the glass making furnace. Ideally at the actual bottom thickness, the quality factory Cp should be zero because the measured data and the predicted data correspond exactly.

However due to modelling uncertainties and noise on the measured mutual inductance, the minimum was not at Cp = 0 and the position of the minimum used instead to indicate the thickness of the furnace bottom.

Figure 8 shows the difference between quality factor Cp for each thickness of refractory bottom in the simulated data using the measured data normalised to the second reference set at a frequency of lMHz (line 218) and 1.58MHz (line 220). As in figure 7, axis 210 represents the quality factor Cp and axis 212 represents the thickness in cm of the furnace bottom in the simulation. The position of the minimum of Cp is very similar to that in figure 7.

The measurement procedure described above was repeated 90 minutes later at the same position on the cold surface of the furnace bottom. That is, a first reference set of data was obtained at lMHz and 1.58MHz, measurement data with the sensor adjacent the cold surface at lMHz and 1.58MHz and a second reference set of data at lMHz and 1.58MHz. This data was used to generate a quality factor for this data. The results are shown in figure 9.

Figure 9 shows the variation of the quality factor for each simulation at the respective furnace bottom thickness. Axis 210 represents the quality factor Cp and axis 212 represents the thickness in cm of the furnace bottom in the simulation. Line 222 is for a primary magnetic field at a frequency of lMHz and line 224 is for a primary magnetic field at a frequency of 1.58MHz. The measured data used to construct figure 9 was normalised using the first reference set of data. As can be seen from figure 9, the position of the minimum is similar to the position obtained from figures 7 and 8. The measurement procedure was carried out again except the sensor housing was adapted such that the planar coil array was positioned about 2.5cm closer to the cold surface of the furnace bottom. No adjustment was made to the mathematical models generated using the FEM software for each thickness of refractory bottom. Given that the planar coil array was positioned closer to the exterior surface of the furnace bottom, this would be expected to result in an apparent reduced thickness of the refractory bottom given that the planar coil array is now closer to the molten glass. The results are shown in figure 10.

In figure 10 the measurement data when the planar coil was positioned 2.5cm closer to the molten glass was normalised using the first reference set of data. Again the axis 210 represents the value of the quality factor and the axis 212 represents the thickness of the refractory bottom used in the respective simulation. Line 226 is for a primary magnetic field at a frequency of lMHz and line 228 is for a primary magnetic field at a frequency of 1.58MHz. The position of the minimum of the quality factor Cp indicates that the refractory bottom thickness is about 3 cm less than before the sensor housing was modified (see results in figures 7, 8 and 9). This result indicates that technique has at least centimetre accuracy. To assess one of the assumption made in the mathematical model, the electrical conductivity of the molten glass in the model was changed from a constant value of 8 S/m to a constant value of 10 S/m. Use of a uniform electrical conductivity clearly is inaccurate due to the glass not having a uniform temperature distribution at the measurement position (as determined by flow model calculations, such as can be carried out using FLUENT software). Assuming a uniform electrical conductivity for the glass speeds up the FEM simulations and simplifies data entry.

The results obtained when the molten glass has a uniform electrical conductivity of 10 S/m in the simulations are shown in figures 1 1 and 12. In each of these figures axis 210 represents the quality factor Cp and the axis 212 represents the thickness of the refractory furnace bottom in cm used in the respective FEM simulation. In figure 12 the measurement data was normalised using the first reference set of data and figure 12 was generated using the measurement data normalised using the second reference measurement data. In figure 11 line 230 is for a primary magnetic field at a frequency of lMHz and line 232 is for a primary magnetic field at a frequency of 1.58MHz. In figure 12 line 234 is for a primary magnetic field at a frequency of lMHz and line 236 is for a primary magnetic field at a frequency of 1.58MHz. As figures 11 and 12 show, there is little difference between the predicted bottom thickness using a slightly different electrical conductivity for the molten glass in the FEM simulations.

In summary, the results shown in figures 7-12 indicate that the thickness of the refractory bottom of the glass making furnace is about 32cm.

To compare the results obtained using the above method, the same furnace underwent a repair a few months later whereby the furnace was allowed to cool to ambient temperatures. A core drill sample was taken from the same location at the furnace bottom where the sensor had been positioned to make measurements. From the core drill sample, which included refractory and solid glass, the refractory thickness at that point of the furnace bottom was measured to be 32cm. This measurement compared very well with the results obtained using the eddy current technique described above.

It will be readily apparent that the sensor may be used to make online measurements of the furnace bottom over the campaign life of the furnace. Many sensors may be used a various positions on the furnace bottom to generate a thickness profile map of the furnace bottom thickness.

The present inventors have used the novel and inventive concept of inducing eddy currents in molten glass to measure the thickness of a refractory section. The thickness of the refractory is measured indirectly by essentially making a direct measurement of the position of the molten glass with respect to the receiver coils and/or transmitter coils.

Upon realising that a magnetic field is able to penetrate the thickness of refractory material in a glassmaking furnace, it is possible to use the same concept of inducing eddy currents in molten glass to measure other properties of the glass, not only the position of the molten glass with respect to the transmitter and/or receiver coils. Another useful

measurement is the electrical conductivity of the molten glass. It is known in the medical field to measure the magnetic fields of induced currents as a means to image the electrical conductivity of human tissues, see for example IEEE Transactions on Medical Imaging, vol.22, No.5, p. 627 -635 (2003). Similar principles can be extended to molten glass.

Given that the electrical conductivity of molten glass is a function of temperature, see for example table 1, it is possible to use the measured electrical conductivity of molten glass to indicate glass temperature. To illustrate this, the sensor described in relation to the description of figures 3 and 4 was located adjacent the cold surface of the bottom of an operational float glass furnace for a period of six days. The sensor was located downstream from the furnace waist. One of the transmitter coils was supplied with an alternating current at 400kHz over the course of the measurement period to produce a primary alternating magnetic field that penetrated the furnace bottom and induced eddy currents in the molten glass. One of the receiver coils was used to measure the effect of the secondary magnetic field produced by the eddy currents. The Solartron impedance analyser was used to measure the magnitude and phase of the mutual inductance (obtainable from the real and imaginary parts of the mutual inductance) between the transmitter coil and the receiver coil. The real part Re of the mutual inductance represents changes in the molten glass electrical

conductivity.

The results over a six hour period of this trial are shown in figure 13.

The results in figure 13 show the measured Re (real component) of the mutual inductance when the water cooled pipe located in the glassmaking furnace waist was lifted, with all the other furnace parameters remaining the same. It is well known to one skilled in the art that when a water cooled pipe located in the waist of a glassmaking furnace is raised, this affects the flow of molten glass in the furnace. As the water cooled pipe is lifted, more glass flows into the downstream end of the glassmaking furnace, and for constant operating conditions, this has the effect of cooling the downstream end of the glassmaking furnace. This effect can be modelled using flow simulation software such as FLUENT and the result of such a simulation using proprietary flow simulation software to indicate glass temperatures is shown on figure 13 as line 306.

On figure 13, axis 300 represents time in hours, axis 302 represents the predicted glass temperature in °C calculated using glass flow simulation software, axis 304 represents the real part Re (in mQ) of the measured mutual inductance between the transmitter coil and the receiver coil. Line 308 represents the change in the real part of the measured mutual inductance over time due to the water cooled pipe being removed from the furnace waist. As figure 13 shows, the measured response is similar to the predicted response to lifting the water cooled pipe in the furnace waist. This provides a means of remotely measuring the molten glass temperature, by measuring the molten glass electrical

conductivity. Once the electrical conductivity has been measured, and the temperature determined therefrom, it is possible to calculate other temperature dependent properties of the molten glass, for example viscosity or the solubility of various gases in the molten glass or density variation. This provides a means of calculating flow velocity and visualisations of any of the preceding properties. The temperature dependence of a number of glass properties are given in the text book "High Temperature Glass Melt Property Database for Process

Modelling", Ed. T. P. Seward, Pub. The American Ceramic Society (2005), ISBN 1-57489- 225-7. For example viscosity at pages 141-146, gas diffusivity at pages 71-73, solubility of S03 at pages 75-94, gas solubility at pages 47-64 and density at pages 125-126.




 
Previous Patent: ANALYTICS

Next Patent: DRILL STRING TUBULAR COMPONENT