WO2024240751A1 - Method and system for calibrating the parameters of a device for automatically controlling a furnace - Google Patents

Abstract

Disclosed is a method for calibrating the parameters of a device for automatically controlling a furnace, preferably a submerged-combustion furnace, for melting a mixture (1001a) of batch materials comprising mineral waste. The method makes it possible to obtain optimal values for the parameters of the transfer function of the automatic-control device. It is based on: - modelling the thermal behaviour of the furnace (1002) using a non-stationary heat transfer function H(s) with, as input data, the variations as a function of time in the temperature ΔT of the mixture (1001a) of molten batch materials; said variations being measured following modification (3004), for a given and limited period of time, of at least one operating parameter of the furnace (1002) selected from the moisture content in the mixture (1001a) of batch materials, the pull of the furnace, the charging rate of the mixture (1001a) of batch materials, the power of the heating means (2002a-c), and/or the amount of organic compounds or of carbon-containing fuels in the mixture (1001a) of batch materials; - modelling the parameters of the transfer function C(s) of the automatic-control device (2010) applied to the modelled transfer function H(s).

Description

Method and system for calibrating the parameters of a furnace servo device

The present invention relates to a method and a system for calibrating the parameters of a control device of a furnace, preferably with submerged combustion, for the melting of a mixture of raw materials comprising mineral waste. It also relates to a furnace, preferably with submerged combustion, which implements said method or said system.

Technical background

It is common practice to use a submerged combustion furnace (SCF), also called a submerged burner furnace (SBF), for melting a mixture of vitrifiable raw materials. In this type of furnace, combustion means such as oxygen/air-fuel burners are directly immersed in the mixture of vitrifiable raw materials and in the molten glass bath formed. The direct injection, in the form of a flow, of the reactants and gaseous combustion products into the glass bath, followed by their rapid expansion and rise, accelerate the melting and improve the homogeneity of the bath.

US 351,413 B, J. T. WAINWRIGHT 26.10.1886 describes, for example, a method of injecting air and/or fuel such as gas or oil into a bath of molten glass so that said gases circulate and heat through the bath before their combustion takes place under or near the surface of the bath. The bath is thus constantly stirred and mixed, thereby keeping it hot.

US 1,656,828 A, POWELL EDWARD R, 17.01.1928 describes a method and apparatus for manufacturing rock wool in which a mixture of raw materials is introduced into a vertical tank provided with an adjacent inclined combustion chamber. The combustion chamber comprises a burner arranged so that the combustion gases enter and melt the mixture of raw materials at the lower base of the tank. At the base of the tank, an opening allows the flow of molten glass through an air jet for blowing molten glass in the form of fibers.

FR 876569 A, UNION DES VERRERIES MECANIQUES, 10.11.1942 describes a submerged combustion furnace comprising a vertical tank equipped, in its lower part, with at least one submerged burner with a combustion fluid pressure of more than 0.2 atmospheres and a speed greater than 30 meters per second.

WO 2009 091 558 A1, GAS TECHNOLOGY INST [US] 29.07.2009 describes a submerged combustion furnace comprising a tank with a double wall with fluid circulation. The bottom of the furnace is provided with a plurality of submerged burners whose relative spatial arrangement is optimized to improve the thermal homogeneity of the glass bath and reduce the number of unmelted parts.

In addition to the tank suitable for melting the mixture of vitrifiable raw materials, a submerged combustion furnace may comprise one or more other adjacent tanks in communication with the first tank, into which the molten glass flows from the first tank to undergo various treatments such as, in particular, refining.

DE 651 687 C, GLASHUETTE ACHERN A G, 18.10.1937 describes a furnace comprising a vertical and rotating melting tank fed with raw materials through an upper opening using a hopper. The tank comprises, in the center of its lower base, oriented along the axis of revolution of the tank, a submerged burner, and is provided, on the periphery of said lower base, with openings allowing the flow of glass towards a lower refining tank.

GB 1 028 481 A, SELAS CORP OF AMERICA, 04.05.1966 describes a furnace provided with a plurality of melting tanks each comprising, at the centre of their lower base, a plurality of submerged burners. The furnace further comprises a main refining tank into which the molten glass from the melting tanks flows.

In the context of the manufacture of mineral fibres, the mixture of vitrifiable raw materials used in a submerged combustion furnace comprises mineral materials which are sources of oxides, hydroxides and/or carbonates of metals, metalloids, alkalis and/or alkaline earths, the relative proportions of which are adjusted so as to obtain the desired chemical composition of the glass at the end of the melting.

Mineral materials are typically mining materials such as, for example, silica sand, bauxite, dolomite, calcium carbonates, magnesium carbonates and/or sodium carbonate. They may also be mineral co-products of other manufacturing industries.

However, with the aim of reducing the ecological impacts of manufacturing industries, in particular reducing the exploitation of natural resources, energy consumption and greenhouse gas emissions, it is now common practice to substitute all or part of the raw materials in the mixture with so-called recyclable mineral waste.

A first example of recyclable mineral waste is “cullet”, of which two types are distinguished:
- so-called “internal” cullet which may include glass waste from the same manufacturing process or the same glass product production line and generally includes cutting waste, defective products detected and rejected during quality control or during adjustments to product compositions;
- so-called "external" cullet which may include glass waste collected from other processes or glass product manufacturing lines, from consumers with the aim of recycling glass products after their use, for example deconstruction waste, glass bottles, used glazing, etc.

In general, and more for external cullet, cullet is a mixture of glass debris of different colors and different compositions. The surface of this debris can also be covered with organic and/or inorganic layers, layers inherited from the different surface functionalization processes for certain applications of glass products. Finally, cullet can also include a certain number of foreign body debris such as ceramic, earthenware, porcelain, terracotta, plastic, metals, electronic components, etc.

A second example of recyclable mineral waste is mineral fibre waste, whether sized or not, i.e. with or without an organic binder. This waste can come from manufacturing industries, building sites or other works and/or recycling channels.

A third example of recyclable mineral waste is raw materials from biomass recovery channels of plant, animal, bacterial or fungal origin. This material can be used as fuel, but also as a means of adjusting the composition of glass and/or its redox state.

Due to their origin, recyclable mineral waste, particularly external waste, has variable levels of humidity and/or organic matter. When introduced into the furnace, this humidity and organic matter alter the thermal and chemical balance of the furnace and cause sudden variations in the temperature of the glass bath. The refining and redox balance processes of the glass bath are also disrupted. The stationary heating and refining regime of the furnace is then destabilized. Reductions in yield and the generation of non-compliant glass products for the intended applications, particularly for the manufacture of mineral fibers, are likely to occur.

It is therefore common to carry out preliminary treatments of mineral waste to reduce, or even eliminate, the humidity and organic compounds they contain.

WO 0248612 A1, SAINT GOBAIN [FR] 20.06.2002 describes, for example, a method for destroying and/or rendering inert mineral waste, in particular making it possible to obtain a recoverable cullet for the subsequent manufacture of mineral wool. In this method, the mineral waste is introduced into a liquid and/or foamy phase maintained at a temperature of at least 800°C and previously formed in a tank equipped with a submerged burner from a mixture of partly vitrifiable materials.

WO 2006 018 582 A1, SAINT GOBAIN ISOVER [FR] 23.02.2006 describes a process for treating mineral waste, in particular mineral fibre waste, in which pure oxygen or oxygen-enriched air is injected into a mass of materials to be recycled, itself subjected to heating via submerged burners. By burning the organic compounds and melting the mass, it is possible to obtain a cullet that can be used for the subsequent manufacture of glass fibres.

US 4877449 A, INST GAS TECHNOLOGY [US] 10/31/1989 describes a submerged combustion furnace comprising a vertical tank provided, in its upper part, above the glass bath, with a cooled grid on which solid charges are deposited using a hopper. The combustion gases coming from the glass bath pass through the grid and heat the solid charges causing them to melt and the liquid thus formed to flow into the glass bath.

Alternatively, it is possible to introduce glass waste directly, i.e. without prior treatment, into the glass bath of a furnace, in particular a submerged combustion furnace intended for the manufacture of mineral fibres. Generally, a control system, such as a feedback loop, is implemented for at least one operating, functioning or control parameter of the furnace, which makes it possible to compensate for the disturbances to which the furnace is subjected compared to its steady state. Such an approach is made possible by the very low thermal inertia intrinsic to submerged combustion furnaces, in particular those equipped with a double wall with fluid circulation.

EP 2 433 911 A1, JOHNS MANVILLE [US] 28.03.2012 describes a method and a device for recycling a glass wool mattress, in which said mattress is introduced into the glass bath of a submerged combustion furnace at a uniform charging speed. A PID servo device or a predictive control system allows the regulation, at the output, of several operating parameters of the furnace, in particular the speed of the mattress conveyors, from different input signals, such as the temperature of the glass bath, the draft and/or the flow rate of fuel and/or oxidant in the burners.

WO 2022/180 345 A1, SAINT GOBAIN ISOVER [FR] 01.09.2022 describes a method for regulating a submerged combustion furnace supplied with a wet mixture of mineral wools and/or biomass, in which the flow rate of the mixture or the power of the submerged burners is regulated using a PID control device based on a measurement of the humidity level of said mixture.

A major disadvantage of processes involving a prior stage of treatment of recyclable mineral waste is that they require more complex installations or treatments, and therefore more substantial material and financial investments for their implementation.

It is therefore advantageous to favour processes and systems based on direct introduction, i.e. without prior treatment, into the furnace of a mixture of raw materials including mineral waste. The particularity of these processes and systems is that they require the implementation and configuration of control systems and processes, such as feedback loops, or predictive control systems.

However, despite the progress made in the automation of servo systems with regard to their configuration and calibration, these remain non-trivial for furnaces, particularly submerged combustion furnaces, for the fusion of a mixture of raw materials including mineral waste.

The moisture content and organic compound content of mineral waste vary greatly depending on its origin and storage conditions. Thus, during the operating time of a furnace, particularly a submerged combustion furnace, a mixture of raw materials including mineral waste in given proportions is highly likely to show sudden and significant variations in its moisture content and organic compound content.

In addition, the proportion of mineral waste in the mixture is likely to vary depending on the ease or difficulty of supplying raw materials. These variations can also contribute to sudden and significant variations in the humidity rate and organic compound content of the mixture. As explained above, such variations cause significant disturbances in the thermal and chemical balance of the kiln. The kiln leaves its steady state and can switch to unstable regimes.

It is also common practice to add organic mineral fuels based on carbon, generally solid, to the 1001a mixture of raw materials. These fuels constitute an additional source of energy for the furnace and often come from energy recovery channels. The type and quantity of fuel can vary greatly depending on their origin and availability on the market. A typical example of fuel can be coal or petroleum coke.

However, it has been found that automatically configured and calibrated control systems are generally incapable of correcting or compensating, in a rapidly convergent manner and with limited amplitudes of overshoot of the setpoint, the disturbances caused by sudden and intense variations in the humidity rate and the content of organic compounds or carbonaceous fuels in the mixture of raw materials.

There therefore remains a need for a method and a system allowing reliable and efficient calibration of the control devices of a furnace, in particular submerged combustion, allowing a rapidly convergent response in the event of sudden and intense variations in the humidity level and/or the content of organic compounds or carbonaceous fuels of a mixture of raw materials including mineral waste.

Solution to the technical problem

In a first aspect of the invention, there is provided a method for calibrating the parameters of a control device of a furnace, preferably with submerged combustion, for the melting of a mixture of raw materials comprising mineral waste.
- said oven comprises at least one tank equipped with at least one heating means, preferably in the form of at least one submerged burner, and at least one control device configured to regulate the power of said heating means according to a set temperature, T0;
- said tank is suitable for melting a mixture of raw materials;
- said tank comprises at least one temperature measuring device;
- said temperature measuring device is configured for continuous measurement of the temperature of said mixture of molten raw materials, and connected to said servo device;
said method comprises the following steps:
(a) the continuous introduction of a mixture of raw materials of given composition into the tank;
(b) continuous measurement of the temperature of the molten raw material mixture using the temperature measuring device;
(c) stationary heating of the mixture of molten raw materials at a given temperature, Ti;
(d) the modification, for a given and limited period of time, without active regulation by the servo device, of at least one operating parameter of the furnace chosen from the moisture content in the mixture of raw materials, the furnace draw, the charging speed of the mixture of raw materials, the power of the heating means, and/or the quantity of organic compounds or carbonaceous fuels in the mixture of raw materials;
(e) measurement of the temporal variations of the temperature, ΔT, of the molten raw material mixture and of the power ΔP of the heating means;
(f) modeling, using a data processing device, the thermal behavior of the furnace using a non-stationary heat transfer function H(s) with, as input data, the temporal variations of the temperature ΔT of the mixture of molten raw materials and of the at least one operating parameter modified in step (d);
(g) modeling, using a data processing device, the parameters of the transfer function C(s) of the servo device applied to the transfer function H(s) modeled in step (f).

According to other advantageous embodiments:
- the non-stationary heat transfer function H(s) is modeled using a first-order response transfer function with or without dead time;
- in step (d), at least two parameters, preferably three parameters, are modified sequentially or in parallel;
- in step (d), the moisture content in the raw material mixture is modified so that the variation in the moisture content of said mixture is between 0 and 10%, and/or the furnace draw is modified so that the relative variation in said draw is between 0 and 2000kg/h;
- in step (d), the amount of organic compounds or carbonaceous fuels in the mixture of raw materials is modified so that the relative variation of said amount of organic compounds or carbonaceous fuels is between 0 and 15% by weight, preferably between 0 and 10% by weight;
- in step (d), the power of the heating means is modified so that the relative variation of said power with respect to the initial power is between 0 and 100%, preferably 0 and 50%, or even between 0 and 25%;
- the servo device is a Proportional – Integral – Derivative (PID) regulator;
- in step (g), the furnace transfer function H(s) further takes, as input data, a set of simulated values, I(s), of the variations in the moisture content in the mixture of raw materials, the furnace draw, the charging speed of the mixture of raw materials, the quantity of organic compounds or carbonaceous fuels and/or the value of the set temperature, T0;
- the values of the moisture content in the raw material mixture, the kiln draw, the charging speed of the raw material mixture, and/or the amount of organic compounds or carbonaceous fuels are simulated in the form of a random signal, such as white noise or pink noise.

The method according to the invention can be used for the calibration of a control device of a furnace, preferably with submerged combustion for the melting of a mixture of raw materials including mineral waste.

In other words, the invention also relates to the use of a method according to the invention for calibrating a control device for a furnace, preferably with submerged combustion, for melting a mixture of raw materials comprising mineral waste.

In a second aspect of the invention, there is provided a system for calibrating the parameters of a control device of a furnace, preferably with submerged combustion, for the melting of a mixture of raw materials comprising mineral waste,
- the oven comprises at least one tank equipped with at least one heating means in the form of at least one submerged burner;
- the tank is suitable for melting a mixture of raw materials;
said system comprises:
- at least one temperature measuring device, said one temperature measuring device being configured for continuous measurement of the temperature of said molten raw material mixture;
- at least one servo device configured to regulate the power of said heating means; the parameters of the transfer equation C(s) of said servo device being calibrated using a method according to the invention.

In a third aspect of the invention, there is provided a furnace, in particular a submerged combustion furnace, for melting a mixture of raw materials comprising mineral waste, in which a method according to the first aspect of the invention is implemented.

The furnace, preferably with submerged combustion, for melting a mixture of raw materials comprising mineral waste, said furnace comprises:
- a first tank suitable for melting a mixture of raw materials and equipped with at least one heating means in the form of at least one submerged burner,
- at least one temperature measuring device configured for continuous measurement of the temperature of said molten raw material mixture;
- at least one servo device configured to regulate the power of said heating means and to receive at least one continuous measurement of the temperature using said temperature measuring device; the values of the parameters of the transfer function C(s) of the servo device being fixed from values obtained using a calibration method according to the invention.

According to other advantageous embodiments:
- the oven is such that the heating means is an oxygen/air-fuel submerged burner and the servo device is further configured to regulate the power of said submerged burner by adjusting the flow rate of fuel injected into said burner while maintaining a constant oxygen flow rate to fuel flow rate ratio;

– the furnace is such that the heating means is an oxygen/air-fuel submerged burner and the servo device is further configured to inject oxygen or air at a constant total flow rate of oxygen or air into said burner and into a bubbler, and at a constant oxygen or air flow rate to fuel flow rate ratio in the burner when the power of the submerged burner varies.

In a fourth aspect of the invention there is provided an installation for the manufacture of mineral fibres comprising a submerged combustion furnace according to the fourth aspect of the invention.

Advantages of the invention

A first remarkable advantage of the invention is the obtaining of optimal values for the parameters of the transfer function of the servo device of a submerged combustion furnace. When the furnace is regulated by a servo device thus calibrated, the temperature of the furnace converges rapidly towards the set temperature of the furnace in the event of sudden and intense variations in the set temperature, T ₀ , the moisture content in the mixture 1001a of raw materials, the draw of the furnace, the charging speed of the mixture 1001a of raw materials, and/or the quantity of organic compounds or carbonaceous fuels in the mixture 1001a of raw materials.

A second remarkable advantage is that it is possible to model, or even simulate, different values for the parameters of the transfer function of the servo device and to select those allowing an optimal regulation of the furnace without there being any need to physically implement disturbance tests on the furnace to evaluate said calibration. The adjustment of the servo device therefore requires less stress on the furnace.

is a schematic representation of an example of a glass or rock fiber manufacturing line.

is a schematic sectional representation of a submerged combustion furnace for melting a mixture of raw materials including mineral waste.

is a flowchart of a method according to the first aspect of the invention.

is a graphical representation of the evolution of the temperature of a submerged combustion furnace as a function of time after a sudden decrease in its power.

is an example block diagram of a servo device according to certain embodiments.

is a graphical representation of an example of the temporal evolution of power, expressed as a variation relative to the base power, linked to temporal variations in the moisture content (upper frame) and the quantity of organic compounds or carbonaceous fuels (lower frame).

is a graphical representation of the temporal evolution of power (lower frame) and temperature (solid line, upper frame) of a furnace whose values of the parameters of the transfer function of its servo device are previously adjusted in accordance with the invention.

is a graphical representation of the time evolution of the temperature (upper frame) and the power (middle frame) of a furnace as a function of the variations in the moisture content and the amount of organic matter in the raw material mixture, and the furnace draw.

Detailed description of embodiments

In reference to the , a 1000 glass or rock fiber manufacturing line by the internal centrifugation method generally includes:
- silos 1001 for storing raw materials 1001a, for example mineral compounds and/or cullet;
- a glass or rock melting furnace 1002 for melting the raw materials 1001a;
- a conveyor 1003 for transporting the raw materials 1001a from the silos 1001 to the furnace 1002;
- one or more fiberizing tools 1005a, 1005b, 1005c fed with molten glass or rock 1006;
- an open or closed feed channel 1004 provided with openings located just above each fiberizing tool 1005a, 1005b, 1005c to feed them with glass or molten rock 1006.

In reference to the , a submerged combustion furnace 1002 generally comprises at least one refractory melting tank 2001 provided at its base with a series of submerged burners 2002a-c of the oxidant-fuel type, for example oxygen-gas, air-fuel or oxygen-fuel. The furnace 1002 is supplied with a mixture 1001a of raw materials by a screw conveyor 1003 via an opening provided on its side wall. The opening may be submerged or emerged. The submerged burners 2002a-c ensure the melting of the mixture 1001a and the stirring of the cast iron 2004.

The furnace 1002 may comprise a second tank 2005, for example a refining tank, into which the cast iron 2004 flows via a groove 2006 provided for this purpose. The second tank 2005 may be provided with a plurality of flame burners 2007 arranged above the surface of the cast iron 2004, with immersed or non-immersed electrodes and with a means 2008 for supplying a refining, oxidizing and/or reducing agent. The means 2088 for supplying an oxidizing or reducing agent may also be arranged in the first tank 2001.

The stay of the 2004 cast iron in the second tank continues its thermal and chemical homogenization and allows the adjustment of its redox state in accordance with the specifications. At the end of this stay, the 2004 cast iron constitutes the molten glass or rock 1006 which is then conveyed to forming tools, such as fiberizing tools 1005a-c via channel 1004, glass aggregate manufacturing tools, or even molding tools.

The melting tank 2001 has a temperature measuring device 2009, for example a thermocouple 2007, immersed or not, configured for continuous measurement of the temperature, T, of said mixture 1001a of molten raw materials, i.e. the cast iron 2004. The temperature measuring device is generally connected to a control device 2010, such as a Proportional-Integral-Derivative (PID) controller, allowing the regulation of the furnace according to a set temperature, T ₀ . The set temperature, T ₀ , can be a fixed value or a time profile.

The servo device 2010 is connected to the controllers (not shown) of the submerged burners and adjusts their power so that the temperature, T, of the cast iron 2004 reaches the set temperature, T ₀ . In the case of oxygen-fuel burners, the power of the burners is adjusted by varying the flow rates of oxygen, fuel and/or the ratio of these two flow rates.

The servo device 2010 may further be configured to control the quantity or flow rate of oxidizing and reducing agent conveyed via the supply means 2008. This control may be exercised via a connection to the control device (not shown) of the supply means 2008. For example, the supply means 2008 is a conveyor, the control device may vary the conveying speed according to a setpoint value provided by the servo device 2010.

In reference to the and 3, in a first aspect of the invention, there is provided a method 3000 for calibrating the parameters of a servo device 2010 of a furnace 1002, preferably with submerged combustion, for the melting of a mixture 1001a of raw materials comprising mineral waste,
- said oven 1002 comprises at least one tank 2001 equipped with at least one heating means 2002a-c, preferably in the form of at least one submerged burner, and at least one control device 2010 configured to regulate the power of said heating means 2002a-c;
- said tank 2001 is suitable for the fusion 2004 of a mixture 1001a of raw materials;
- said tank 2001 comprises at least one temperature measuring device 2009;
- said temperature measuring device 2009 is configured for continuous measurement of the temperature of said mixture 1001a of molten raw materials 2004, and connected to said servo device 2010;
said method 3000 comprises the following steps:
(a) the continuous introduction 3001 of a mixture 1001a of raw materials of given composition into the tank 2001;
(b) continuous measurement 3002 of the temperature of the mixture 1001a of molten raw materials 2004 using the temperature measuring device 2009;
(c) the stationary heating 3003 of the mixture 1001a of molten raw materials 2004 according to a given temperature, T _i ;
(d) the modification 3004, for a given and limited period of time, without active regulation by the servo device 2010, of at least one operating parameter of the furnace 1002 chosen from the moisture content in the mixture 1001a of raw materials, the furnace draw, the charging speed of the mixture 1001a of raw materials, the power of the heating means 2002a-c, and/or the quantity of organic compounds or carbonaceous fuels in the mixture 1001a of raw materials;
(e) measurement 3005 of the temporal variations of the temperature, ΔT, of the mixture of raw materials 1001a in fusion 2004;
(f) modeling 3006, using a data processing device, of the thermal behavior of the furnace 1002 using a non-stationary heat transfer function H(s) with, as input data, the temporal variations of the temperature, ΔT, of the mixture 1001a of molten raw materials and of the at least one operating parameter modified in step (d);
(g) modeling 3007, using a data processing device, of the parameters of the transfer function C(s) of the servo device 2010 applied to the transfer function H(s) modeled in step (f).

The 2009 temperature measuring device may be a thermocouple or a pyrometer.

For the purposes of the invention, “carbonaceous fuels” means any type of organic mineral fuel based on carbon, preferably solid, which can be added to the mixture 1001a of raw materials. An example of a carbonaceous fuel may be coal or petroleum coke.

Modeling steps (f) and (g) are typically performed using a data processing device. An example of a device may be a device configured to automatically perform sequences of arithmetic or logical operations to perform tasks or actions. Such a device, typically referred to as a computer, may include one or more central processing units (CPUs) and at least one control device adapted to perform such operations.

The device may also include other electronic components such as input/output interfaces, non-volatile or volatile storage devices, and communication buses for transferring data between components within the device. One of the input/output devices may be a user interface for human-machine interaction, such as a graphical user interface for displaying human-understandable information.

The data processing device may advantageously comprise one or more graphics processing units (GPUs) whose parallel structure makes them more efficient than central processing units in performing complex calculations.

Step (g) of modeling the thermal behavior of the furnace 1002 using a non-stationary heat transfer function H(s) makes it possible to obtain a digital model of the furnace 1002 from which it is possible to model the parameters of the transfer function C(s) of the control device 2010 without it being necessary to physically intervene on the furnace 1002 to implement this modeling. In other words, the transfer function H(s) provides a model of the furnace 1002 on which the control device 2010 can be applied, via its transfer function C(s), in order, in step (g), to determine the values of the parameters of said device for optimal regulation of the furnace 1002.

By means of the method according to the first aspect of the invention, it is therefore possible to model, or even simulate, different values for the parameters of the transfer function C(s) and to select those allowing optimal regulation of the furnace 1002 without there being any need to physically implement tests on the furnace 1002 to evaluate it. A remarkable advantage is that the furnace 1002 is considerably less used for the adjustment of the servo device 2010.

In step (f), the thermal behavior of the furnace 1002 is modeled using a non-stationary heat transfer function H(s).

According to an exemplary embodiment, the thermal balance of the furnace 1002 may be modeled using the following equation:
where T is the temperature of the furnace 1002 including the cast iron 2004, t is the time, M is the mass of the furnace 1002 including the cast iron 2004 it contains, c _p is the specific heat capacity of the furnace including the cast iron 2004, P(t) is the power of the heating means at time t, φ is the power required to melt the mixture 1001a of raw materials at a given charging speed, and I(t) = δ(t) + Ω(t) is the instantaneous variation in power linked to the temporal variations in the content, δ(t), in humidity δ and in the quantity Ω(t) of organic compounds or carbonaceous fuels in the mixture 1001a of raw materials.

The power I(t) can also include the temporal variations of the draw ψ(t) of the furnace and/or the speed v(t) of charging of the mixture 1001a of raw materials.

Generally, in the case of convective heat transfers, the power φ varies proportionally to the difference between the temperature T(t) of the furnace 1002 including the cast iron 2004 at time t and a so-called fictitious boundary layer temperature T _p . The boundary layer temperature T _p can be interpreted as a temperature representative of the temperature of the cast iron 2004 near the walls of the furnace 1002 and the mixture 1001a of unmelted raw materials. The power φ can then be estimated using the following relationship:
where β and T _p are constants unknown a priori.

Setting x(t) = T(t) – _T0 and u(t)=P(t) – P0, with _T0 the set temperature and _P0 the base power to reach the set temperature _T0 without disturbance, i.e. with I(t) = 0, a non-stationary heat transfer function H(s) in the Laplace domain can be:

with

And

In the initial state, that is to say before any disturbance of the oven 1002, the temperature, T(t), of the oven 1002 follows the set temperature, T ₀ , i.e. T(t) = T ₀ for the base power, P ₀ :

And

The quantities M, c _p , T _p and β are generally not known and depend on the structure of the furnace 1002, its constituent materials, the chemical nature of the cast iron 2004 and its quantity. β also depends on the charging speed of the mixture 1001a of raw materials.

According to the first aspect of the invention, the modeling of the thermal behavior of the furnace 1002 is modeled using the transfer function H(s) with, as input data, the temporal variations of the temperature ΔT of the mixture 1001a of molten raw materials and of the operating parameter modified in step (d). The temporal variations of the temperature ΔT of the mixture 1001a of molten raw materials and of the operating parameter modified in step (d) can be interpreted as the consequences of a disturbance introduced by the modification, in step (d), of at least one operating parameter of the furnace. The use of this disturbance makes it possible to calculate the parameters of the transfer function H(s), in particular the constant γ and the temperature T _p .

In some embodiments, the non-steady heat transfer function H(s) is modeled using a first-order response transfer function with or without dead time.

Generally, submerged combustion furnaces have a certain inertia and when one of their operating parameters is suddenly modified, in the form of a pulse, the response of said furnace 1002 is not immediate. The furnace 1002 shows a delay in response to the disturbance.

According to an exemplary embodiment, for a power impulse response, for example following a sudden change in power in the form of a unit step U(s)=∆P/s, and a zero change, ΔI= Δδ + ΔΩ = 0, in the moisture content and the amount of organic compounds, the response can be written in the form:

Where ΔP = P ₁ – P ₀ is the variation, in the form of a unit step, from the initial power P ₀ to the power P ₁ different from P ₀ .

In the time domain, this function has the following expression, for t > 0:

The values of the parameters γ and T _p can be obtained by carrying out a function adjustment f(t) on the temporal variations of the temperature ΔT of the mixture 1001a of molten raw materials and of the power ΔP of the heating means 2002a-c.

In step (d), for a given and limited period of time, without active regulation by the servo device 2010, at least one operating parameter of the furnace 1002 is modified, chosen from the moisture content in the mixture 1001a of raw materials, the furnace draw, the charging speed of the mixture 1001a of raw materials, the power of the heating means 2002a-c, and/or the quantity of organic compounds or carbon fuels in the mixture 1001a of raw materials.

The number and nature of the operating parameters to be modified depends on the composition of the mixture of raw materials including mineral waste and the precision required for the calibration of the control parameters.

According to certain embodiments, in step (d), at least two, or even at least three, operating parameters are modified sequentially or in parallel. The modification of at least two, or even three parameters, is generally sufficient for representative modeling of the thermal behavior of a furnace, in particular with submerged combustion and, ultimately, precise modeling of the parameters of the transfer function of the control device 2010 for effective regulation.

As discussed previously, the moisture content and organic or carbonaceous fuel content of mineral wastes can vary greatly depending on their origin and storage conditions. During the operating time of a submerged combustion furnace, a raw material mixture including mineral wastes in given proportions can show sudden and significant variations in its moisture content and organic or carbonaceous fuel content.

Thus, according to certain embodiments, in step (d), the moisture content in the mixture 1001a of raw materials is modified so that the variation in the moisture content of said mixture is between 0 and 10%, and/or the furnace draw is modified so that the relative variation in said draw is between 0 and 2000 kg/h.

These variation intervals in furnace output and moisture content of raw material mixtures including mineral waste allow effective calibration of furnace control devices, particularly submerged combustion furnaces.

According to certain embodiments, the oven 1002 may comprise at least one means 2011 for measuring the humidity of said mixture 1001a of raw materials. This means may be any type of humidity sensor suitable for measuring the humidity of the mixture 1001a of raw materials. It may be arranged at the conveyor 1003 just before loading, or further upstream, in the storage silos 1001. The humidity measuring means 2011 allows precise measurement of the variation in the humidity content of the mixture 1001a in step (d).

According to other embodiments, in step (d), the amount of organic compounds in the mixture 1001a of raw materials is modified so that the relative variation of said amount of organic compounds or carbonaceous fuels is between 0 and 15% by weight, preferably between 0 and 10% by weight.

These intervals of variation in the quantity of organic compounds allow modeling of the transfer function H(s) of the furnace 1002 and, then, of the transfer function C(s) of the control device 2010 which, when the latter is ultimately implemented in a device 2010, results in rapid convergence of the temperature of the furnace 1002 to the set temperature in the event of sudden and intense variations in the content of organic compounds or carbon fuels in the mixture of raw materials comprising mineral waste.

According to other embodiments, complementary or not, in step (d), the power of the heating means (2002a-c) is modified so that the relative variation of said power compared to the initial power is between 0 and 100%, preferably 0 and 50%, or even between 0 and 25%.

As an illustrative example, the represents the evolution of the temperature (dotted line) of a submerged combustion furnace as a function of time after the introduction of 5.7% by mass of coke in a mixture 1001a of raw materials followed by a sudden decrease in the power of the furnace from 287kW to 173kW. When the coke was introduced, due to the energy released by its combustion, the temperature of the furnace 1002, initially at T ₀ = 1175°C, increased to reach 1288°C. On the , time t=0 corresponds to the moment when the power is abruptly reduced, in the form of a unit step, from 287kW to 173kW. The evolution of the temperature is represented by the dot figures.

The response f(t) previously described for an impulse response can be written with P ₀ = 287kW, P ₁ = 173kW and T ₀ = 1175°C:

To determine the values of the parameters T _p and γ, the response f(t) can be fitted to the data from the using a least squares method with T _p and γ as fitting parameters. The optimal fit, illustrated in the by the solid line, allows to obtain the following values for the parameters T _p and γ:

The transfer function H(s) used to model the thermal behavior in this example can therefore be written as:

And the quantity Mc _p , representing the thermal capacity of the furnace 1002 independent of its mass, can be calculated:

In step (g), the parameters of the transfer function C(s) of the servo device 2010 are modeled by applying it to the transfer function H(s) of the oven 1002, obtained in step (f).

Once the thermal behavior of the furnace 1002 has been modeled using the transfer function H(s), it is possible to model the parameters of the transfer function C(s) of the control device 2010 applied to the transfer function H(s) of the furnace.

This modeling can in particular be implemented by simulation in order to determine ex-situ the optimal values of the parameters of the transfer function C(s) for efficient regulation of the furnace 1002. Numerical calculation software such as Matlab or Scilab provide functions and algorithms adapted to this type of simulation.

In reference to the , an example of modeling may consist of an iterative optimization loop in which the parameters of the transfer function C(s) of the servo device are adjusted until the setpoint power P calculated by said function C(s) from a difference ε between the temperature T of the furnace 1002 and a setpoint temperature T ₀ for said furnace 1002 allows the furnace 1002 to reach said setpoint temperature T ₀ when said power P increased by the power I(s) is provided, as input data, to the transfer function H(s) of the furnace 1002. I(s) = δ(s) + Ω(s) is the instantaneous variation in power linked to the temporal variations in the content, δ(s), in humidity δ and in the quantity Ω(s) of organic compounds or carbonaceous fuels in the mixture 1001a of raw materials.

During the execution of the iterative execution loop, the parameters of the transfer function C(s) of the servo device 2010 can be adjusted manually, using an adjustment method such as the Ziegler–Nichols method, the Cohen-Coon method, the Åström–Hägglund method, or automatically using a numerical optimization method.

With the transfer function H(s) of the example of the , the diagram of the can, for example, be transcribed using the following formula:

With ε(t) = - x(t). From where:

According to preferred embodiments, the servo device 2010 is a Proportional-Integral-Derivative (PID) controller.

The transfer function C(s) of the 2010 servo device can be written in the Laplace domain:
or Kp, Ki and Kd are respectively the proportional, integral, and derivative parameters or gains.

According to another writing, the transfer function C(s) can have the following form:
where τ _i and τ _d are the integration and derivative time constants respectively and, τ _i = Kp/Ki = KpTi and τ _d =Kd/Kp = Td/Kp.

It has been found that in most industrial furnaces, the temperature measurements, T, of the 1001a mixture of 2004 molten raw materials can be affected by significant noise. In order to avoid control divergences, it may be advantageous to neglect the drift term of the transfer function C(s) without there being a significant impact on the control performance of the servo device.

The function C(s) can then be written:

Or again

The transfer function C(s) is used to calculate, in real time, a gain for correcting the setpoint of the power regulators of the heating means 2002a-c in the form of at least one submerged burner. The unit of the gain depends on the type of regulator used. For example, in the case of a submerged burner, it may correspond to a percentage of the fuel flow rate, for example, a gas volume flow rate, injected into the burner. The flow rate value may be linked to a burner power level according to a linear relationship.

From the previous example, the output setpoint value of the 2010 servo device can be written according to the following relation:

The set power, P, at the input of the oven 1002 can be written:

Modeling the parameters, Kp and τ _i of the transfer function C(s) of the 2010 servo device can then consist of adjusting the values which minimize the temporal temperature variations in a simulation of the system in regulation.

This adjustment can be performed manually, using an adjustment method such as the Ziegler–Nichols method, the Cohen-Coon method, the Åström–Hägglund method, or automatically using a numerical optimization method.

For the example of the and a measurement frequency α of the temperature measurement device 2009 fixed greater than γ, i.e. α = 1 Hz, it is a question of adjusting by iteration the values of the parameters, Kp and τ _i so as to obtain x(t) close to zero in time. The advantage is that it is possible to evaluate the stability, over time, of the furnace during regulation when it is subjected to so-called usual disturbances for applications in the glassmaking field. Another advantage is the possibility of taking into account, during modeling, the sampling frequency of the temperature measurement device 2009.

According to certain embodiments, in step (g), the transfer function H(s) of the furnace 1002 further takes, as input data, a set of simulated values, I(s), of the variations in the moisture content in the mixture 1001a of raw materials, the furnace draw, the charging speed of the mixture 1001a of raw materials, the quantity of organic compounds or carbonaceous fuels and/or the value of the set temperature, T ₀ .

In some preferred embodiments, the values of the moisture content in the raw material mixture 1001a, the kiln draw, the charging rate of the raw material mixture 1001a, and/or the amount of organic compounds or carbonaceous fuels are simulated as a random signal, such as white noise or pink noise.

According to an example of modeling the parameters, Kp and τ _i of the transfer function C(s) of the servo device for the example of and 5, with reference to the , the instantaneous variations, expressed as relative variations δ(t) and Ω(t), of the power I(t) = δ(t) + Ω(t) linked to the temporal variations of the moisture content (upper frame) and of the quantity Ω of organic compounds or carbonaceous fuels (lower frame) are first generated in the form of pink noise. The values of the parameters Kp and τ _i are then manually adjusted by iteration so as to obtain a temperature close to the set temperature over time.

On the , the corresponding variations of power (lower frame) and temperature (solid line, upper frame) of the furnace 1002 are represented, obtained with the following adjusted values of the parameters Kp and τ _i :

There shows that these parameter values Kp and τ _i allow effective control of the temperature T (continuous line, upper frame) of the oven 1002 to the set temperature T ₀ (discontinuous line, upper frame) over the set of instantaneous variations of the power I(s) of the .

In order to validate the performance of a calibrated servo device with the values of the parameters Kp and τ _i thus adjusted, these were implemented in the servo device of an industrial furnace presenting the same transfer function H(s) as that of the example of and 5. The industrial furnace, operating under real and industrial conditions, was then subjected to sudden variations in the set temperature, _T0 , its draw, the humidity content and the quantity of organic materials or carbonaceous fuels, such as coke, contained in the mixture of raw materials.

The behavior of the oven thus subjected to such variations is shown in the . In the upper frame, the temporal evolution of the temperature (upper frame, solid line) of the furnace and the set temperature T ₀ (dotted line) is shown. In the central frame, the temporal evolution of the furnace power is shown. In the lower frame, the sudden changes in the draft (dashed line) of the furnace, the moisture content (solid line) and the amount of organic matter or carbonaceous fuels (dotted line) of the raw material mixture are shown.

There , upper frame, shows a rapid convergence of the oven temperature (solid line) to the set temperature T ₀ (dotted line) regardless of the nature and simultaneity of the variations (lower frame).

A sudden change in the moisture content δ (solid line) causes a small disturbance in the power (central frame) and the temperature (solid line) of the furnace deviates little from the set temperature T ₀ (dotted line). Sudden changes in the draft ψ (lower frame dashed line) of the furnace after 5 h and shortly before 15 h cause sudden changes in power (central frame), and the temperature (solid line), after an initial drift, converges very quickly towards the set temperature T ₀ (dotted line). The sudden change in the quantity Ω of organic matter or carbonaceous fuels (lower frame dotted line) around 17 h causes frequent changes in power that are quickly compensated.

In addition, the servo-control device is able to compensate for sudden and simultaneous variations, after 07:30 and 19:00, of the set temperature T ₀ (upper frame, dotted line), of the draft (lower frame, dashed line) of the furnace, of the humidity content (lower frame, solid line) and of the quantity of organic materials or carbonaceous fuels (lower frame, dotted line).

The trials of the demonstrate that the present invention allows efficient and robust calibration of a control device for a submerged combustion furnace. The control device 2010 allows rapid compensation for sudden and intense variations in temperature, T ₀ , of the setpoint (variations of 40°C in the figure), of the moisture content in the mixture 1001a of raw materials, the furnace draw, the charging speed of the mixture 1001a of raw materials, and/or the quantity of organic compounds in the mixture 1001a of raw materials. The temperature of the furnace 1002 converges rapidly to the setpoint temperature with very limited amplitudes of overshooting of the setpoint over time regardless of the nature and simultaneity of the variations.

All embodiments of the first aspect of the invention are combinable.

The method according to the first aspect of the invention can advantageously be used for the calibration of a control device 2010 of a furnace 1002, preferably with submerged combustion, for the melting of a mixture 1001a of raw materials comprising mineral waste.

According to a second aspect of the invention, with reference to the and 3, a system is provided for calibrating the parameters of a servo device 2010 of a furnace 1002, preferably with submerged combustion, for the melting of a mixture 1001a of raw materials comprising mineral waste,
- the oven 1002 comprises at least one tank 2001 equipped with at least one heating means 2002a-c in the form of at least one submerged burner;
- the 2001 tank is adapted to the 2004 fusion of a 1001a mixture of raw materials;
said system comprises:
- at least one temperature measuring device 2009, said one temperature measuring device 2009 being configured for the continuous measurement of the temperature of said mixture 1001a of molten raw materials 2004;
- at least one servo device 2010 configured to regulate the power of said heating means 2002a-c; the parameters of the transfer equation C(s) of said servo device 2010 being calibrated using a method according to any one of the embodiments of the first aspect of the invention.

According to a third aspect of the invention, there is provided a furnace 1002, preferably with submerged combustion, for the melting 2004 of a mixture 1001a of raw materials comprising mineral waste, said furnace 1002 comprises:
- a first tank 2001 adapted to the melting 2004 of a mixture of raw materials and equipped with at least one heating means 2002a-c in the form of at least one submerged burner,
- at least one temperature measuring device 2009 for the continuous measurement of the temperature of said mixture 1001a of molten raw materials 2004;
- at least one servo device 2010 configured to regulate the power of said heating means 2002a-c and to receive at least one continuous measurement of the temperature using said one temperature measuring device 2009; the values of the parameters of the transfer function C(s) of the servo device 2010 being fixed from values obtained using a calibration method according to any one of the embodiments of the first aspect of the invention.

The calibration method for determining the values of the parameters of the transfer function C(s) of the servo device 2010 can be implemented using a system according to the second aspect of the invention.

As with the first aspect of the invention, the oven 1002 may comprise at least one means 2011 for measuring the humidity of said mixture 1001a of raw materials.

According to certain embodiments, the heating means 2002a-c is an oxygen/air-fuel submerged burner and the servo device 2010 is further configured to regulate the power of said submerged burner by adjusting the flow rate of fuel injected into said burner while maintaining a constant oxygen flow rate to fuel flow rate ratio.

According to some alternative embodiments, the heating means 2002a-c is a submerged oxygen/air-fuel burner and the servo device 2010 is further configured to inject oxygen or air at a constant total flow rate of oxygen or air into said burner and into a bubbler, and at a constant oxygen or air flow rate to fuel flow rate ratio into the burner as the power of the submerged burner varies.

In a fourth aspect of the invention, there is provided an installation for the manufacture of mineral fibers, such as a glass or rock fiber manufacturing line, comprising a furnace 1002 according to any one of the embodiments of the third aspect of the invention.

The present invention, in all its aspects, can be implemented, without being limited thereto, in many processes and lines for manufacturing glass products such as, for example, glass wool, rock wool, textile glass fibers, flat glass or hollow glass.

List of references Patent literature

US 351 413 B, J. T. WAINWRIGHT 10/26/1886.

US 1 656 828 A, POWELL EDWARD R, 01/17/1928.

FR 876569 A, UNION OF MECHANICAL GLASSWORKS, 10.11.1942.

WO 2009 091 558 A1, GAS TECHNOLOGY INST [US] 07.29.2009.

DE 651 687 C, GLASHUETTE ACHERN A G, 10/18/1937.

GB 1 028 481 A, SELAS CORP OF AMERICA, 04.05.1966.

WO 0248612 A1, SAINT GOBAIN [FR] 06.20.2002.

WO 2006 018 582 A1, SAINT GOBAIN ISOVER [FR] 02.23.2006.

US 4877449 A, INST GAS TECHNOLOGY [US] 10/31/1989.

EP 2 433 911 A1, JOHNS MANVILLE [US] 03.28.2012.

WO 2022 180 345 A1, SAINT GOBAIN ISOVER [FR] 01.09.2022.

Claims (15)

Method (3000) for calibrating the parameters of a servo device (2010) of a furnace (1002), preferably with submerged combustion, for melting a mixture (1001a) of raw materials comprising mineral waste,
- said oven (1002) comprises at least one tank (2001) equipped with at least one heating means (2002a-c), preferably in the form of at least one submerged burner, and at least one control device (2010) configured to regulate the power of said heating means (2002a-c) according to a set temperature, T ₀ ;
- said tank (2001) is suitable for the melting (2004) of a mixture (1001a) of raw materials;
- said tank (2001) comprises at least one temperature measuring device (2009);
- said temperature measuring device (2009) is configured for continuous measurement of the temperature of said mixture (1001a) of molten raw materials (2004), and connected to said servo device (2010);
said method (3000) comprises the following steps:
(a) the continuous introduction (3001) of a mixture (1001a) of raw materials of given composition into the tank (2001);
(b) continuously measuring (3002) the temperature of the mixture (1001a) of molten raw materials (2004) using the temperature measuring device (2009);
(c) the stationary heating (3003) of the mixture (1001a) of molten raw materials (2004) according to a given temperature, T _i ;
(d) the modification (3004), for a given and limited period of time, without active regulation by the servo device (2010), of at least one operating parameter of the furnace (1002) chosen from the moisture content in the mixture (1001a) of raw materials, the furnace draw, the charging speed of the mixture (1001a) of raw materials, the power of the heating means (2002a-c), and/or the quantity of organic compounds or carbonaceous fuels in the mixture (1001a) of raw materials;
(e) measuring (3005) the temporal variations of the temperature, ΔT, of the mixture of raw materials (1001a) in fusion (2004) and of the power ΔP of the heating means (2002a-c);
(f) modeling (3006), using a data processing device, the thermal behavior of the furnace (1002) using a non-stationary heat transfer function H(s) with, as input data, the temporal variations of the temperature ΔT of the mixture (1001a) of molten raw materials and of the at least one operating parameter modified in step (d);
(g) modeling (3007), using a data processing device, the parameters of the transfer function C(s) of the servo device (2010) applied to the transfer function H(s) modeled in step (f). The method (3000) of claim 1, wherein the non-stationary heat transfer function H(s) is modeled using a first-order response transfer function with or without dead time. Method (3000) according to any one of claims 1 to 2, such that, in step (d), at least two parameters, preferably three parameters, are modified sequentially or in parallel. Method (3000) according to any one of claims 1 to 3, such that, in step (d), the moisture content in the mixture (1001a) of raw materials is modified so that the variation in the moisture content of said mixture is between 0 and 10%, and/or the furnace draw is modified so that the relative variation in said draw is between 0 and 2000kg/h. Method (3000) according to any one of claims 1 to 4, such that, in step (d), the amount of organic compounds or carbonaceous fuels in the mixture (1001a) of raw materials is modified so that the relative variation of said amount of organic compounds or carbonaceous fuels is between 0 and 15% by weight, preferably between 0 and 10% by weight. Method (3000) according to any one of claims 1 to 5, such that, in step (d), the power of the heating means (2002a-c) is modified so that the relative variation of said power with respect to the initial power is between 0 and 100%, preferably 0 and 50%, or even between 0 and 25%. Method (3000) according to any one of claims 1 to 6, such that the servo device (2010) is a Proportional – Integral – Derivative (PID) regulator. Method (3000) according to any one of claims 1 to 7, such that, in step (g), the transfer function H(s) of the furnace (1002) further takes, as input data, a set of simulated values, I(s), of the variations in the moisture content in the mixture (1001a) of raw materials, the furnace draw, the charging speed of the mixture (1001a) of raw materials, the quantity of organic compounds or carbonaceous fuels and/or the value of the set temperature, T ₀ . Method (3000) according to claim 8, such that the values of the moisture content in the mixture (1001a) of raw materials, the kiln draw, the charging speed of the mixture (1001a) of raw materials, and/or the amount of organic compounds or carbonaceous fuels are simulated in the form of a random signal, such as white noise or pink noise. Use of a method (3000) according to any one of claims 1 to 9 for calibrating a device (2010) for controlling a furnace (1002), preferably with submerged combustion, for melting a mixture (1001a) of raw materials comprising mineral waste. System for calibrating the parameters of a servo device (2010) of a furnace (1002), preferably with submerged combustion, for the melting of a mixture (1001a) of raw materials comprising mineral waste,
- the oven (1002) comprises at least one tank (2001) equipped with at least one heating means (2002a-c) in the form of at least one submerged burner;
- the tank (2001) is suitable for the melting (2004) of a mixture (1001a) of raw materials;
said system comprises:
- at least one temperature measuring device (2009), said one temperature measuring device (2009) being configured for the continuous measurement of the temperature of said mixture (1001a) of molten raw materials (2004);
- at least one servo device (2010) configured to regulate the power of said heating means (2002a-c); the parameters of the transfer equation C(s) of said servo device (2010) being calibrated using a method (3000) according to any one of claims 1 to 9. Furnace (1002), preferably with submerged combustion, for the melting (2004) of a mixture (1001a) of raw materials comprising mineral waste, said furnace (1002) comprises:
- a first tank (2001) suitable for the melting (2004) of a mixture of raw materials and equipped with at least one heating means (2002a-c) in the form of at least one submerged burner,
- at least one temperature measuring device (2009) configured for continuous measurement of the temperature of said mixture (1001a) of molten raw materials (2004);
- at least one servo device (2010) configured to regulate the power of said heating means (2002a-c) and to receive at least one continuous measurement of the temperature using said temperature measuring device (2009); the values of the parameters of the transfer function C(s) of the servo device (2010) being fixed from values obtained using a calibration method (3000) according to any one of claims 1 to 9. Oven (1002) according to claim 12, such that the heating means (2002a-c) is an oxygen/air-fuel submerged burner and the servo device is further configured to regulate the power of said submerged burner by adjusting the flow rate of fuel injected into said burner while maintaining a constant oxygen flow rate to fuel flow rate ratio. Oven according to one of Claims 12, such that the heating means (2002a-c) is an oxygen/air-fuel submerged burner and the servo device (2010) is further configured to inject oxygen or air according to a constant total flow rate of oxygen or air into said burner and into a bubbler, and according to a constant oxygen or air flow rate to fuel flow rate ratio in the burner when the power of the submerged burner varies. Installation for the manufacture of mineral fibres comprising a submerged combustion furnace (1002) according to any one of claims 12 to 14.