AU2009352124A1

AU2009352124A1 - Method for characterizing the combustion in lines of partitions of a furnace having rotary firing chamber(s)

Info

Publication number: AU2009352124A1
Application number: AU2009352124A
Authority: AU
Inventors: Nicolas Fiot; Pierre Mahieu
Original assignee: Solios Carbone SA
Current assignee: Solios Carbone SA
Priority date: 2009-09-07
Filing date: 2009-09-07
Publication date: 2012-03-08
Anticipated expiration: 2029-09-07
Also published as: CN102597678A; WO2011027042A1; EP2475948A1; RU2012113704A; EP2475948B1; ZA201201211B; RU2500961C1; CA2772693C; AU2009352124B2; CA2772693A1; CN102597678B

Abstract

The invention relates to a method including a series of tests consisting of totally stopping the injection of fuel, one line of partitions (6) after the other, without any activity on the lines of partitions (6) other than that of the test, calculating the variation between the measurements of an image parameter of the total content of unburnt material in the combustion gases before and after totally stopping the injection in each tested line of partitions (6), and identifying any line of partitions (6) as having incomplete combustion if said variation is greater than x% of the initial value of said image parameter at the start of the corresponding test, x% preferably being between 5% and 10%.

Description

1 METHOD FOR CHARACTERIZING THE COMBUSTION IN LINES OF PARTITIONS OF A FURNACE HAVING ROTARY FIRING CHAMBER(S) The invention relates to the field of chamber-type furnaces having 5 rotary firing, known as "ring" furnaces, for baking carbon blocks, particularly carbon anodes and cathodes for the production of aluminum by electrolysis. The invention more particularly relates to a method for characterizing the combustion in lines of partitions of such a chamber-type ring furnace. Ring furnaces for baking anodes are described in particular in the 10 following patent documents: US 4,859,175, WO 91/19147, US 6,339,729, US 6,436,335 and CA 2550880, which may be referred to for further information. Their structure and operation will be partially reviewed here, however, with reference to figures 1 and 2. A schematic plan view of the structure of a ring furnace with open chambers is represented with two fires in the example in 15 figure 1, and a partial perspective cross-sectional view with a cutaway section showing the internal structure of such a furnace is represented in figure 2. The baking furnace (FURN) 1 comprises two parallel bays or spans 1a and lb extending the length of the furnace 1 along the longitudinal axis 20 XX, each comprising a succession of transverse chambers 2 (perpendicular to the axis XX), separated from each other by transverse walls 3. The length of each chamber 2, meaning in the transverse direction of the furnace 1, is constituted of pits 4 open in their upper part to allow loading the carbon blocks to be baked and unloading the cooled baked blocks, and in which are 25 stacked the carbon blocks 5 to be baked, packed in carbonaceous powder, adjacent to and alternating with hollow heating partitions 6, with thin walls, generally separated from each other by transverse spacers 6a. The hollow partitions 6 of one chamber 2 are in the longitudinal extension (parallel to the major axis XX of the furnace 1) of the hollow partitions 6 of the other 30 chambers 2 in the same span 1a or 1b, and these hollow partitions 6 are in communication with one another by means of pods 7 in the upper part of their longitudinal walls, facing longitudinal passages arranged at that level in 2 the transverse walls 3, such that the hollow partitions 6 form longitudinal lines of partitions parallel to the major axis XX of the furnace and within which gases will circulate (combustion air, combustible gases, and combustion gases and fumes) to ensure the anodes 5 are preheated and baked, then 5 cooled. The hollow partitions 6 additionally comprise baffles 8 to prolong and more uniformly distribute the path of the combustion gases or fumes, and these hollow partitions 6 have openings 9 in their upper part, called peep holes, which can be closed with removable covers that are arranged in a crowning block of the furnace 1. 10 The two spans la and lb of the furnace 1 are placed in communication at their longitudinal ends by means of crossovers 10, which transfer the gases from one end of each line of hollow partitions 6 in a span Ia or lb to the corresponding end of the line of hollow partitions 6 in the other span 1b or 1a, so as to form substantially rectangular loops of lines of 15 hollow partitions 6. The operating principle of ring furnaces, also known as advancing fire furnaces, consists of advancing a flame front from one chamber 2 to an adjacent one during a cycle, with each chamber 2 successively undergoing the stages of preheating, forced heating, full fire, then cooling (natural then 20 forced). The anodes 5 are baked by one or more fires or fire groups (two fire groups are represented in figure 1, with one extending in this example across thirteen chambers 2 of span 1 a and the other across thirteen chambers 2 of span 1b), which move cyclically from chamber 2 to chamber 2. Each fire or 25 fire group consists of five successive zones A to E, which are, as represented in figure 1 for the fire for span 1 b, from downstream to upstream relative to the direction of the gas flow in the lines of hollow partitions 6, and in the direction opposite to the cyclic displacements from chamber to chamber: A) A preheating zone consisting of, if referring to the fire for span 1 a 30 and taking into account the direction of rotation of the fires indicated by the arrow at the crossover 10 at the end of the furnace 1 appearing at the top in figure 1: 3 - an exhaust manifold 11 which is equipped, for each hollow partition 6 of the chamber 2 above which this exhaust manifold extends, with a system for measuring and adjusting the flow rate of the combustion gases and fumes per line of hollow partitions 6, this 5 system possibly comprising, in each exhaust pipe 11 a which is integrally attached to the exhaust manifold 11 and opens into said exhaust manifold, and engaged in the opening 9 of one of the respective hollow partitions 6 of this chamber 2, an adjustable shutter pivoted by a shutter actuator, for adjusting the flow rate, as 10 well as a flow meter 12 slightly upstream in the corresponding pipe 11a, a temperature sensor (thermocouple) 13 for measuring the temperature of the combustion gases when they are suctioned out, and - a preheating measurement ramp 15, substantially parallel to the 15 exhaust manifold 11 and upstream from the latter, generally, above the same chamber 2 and fitted with temperature sensors (thermocouples) and pressure sensors to prepare the static negative pressure and the temperature that will prevail in each of the hollow partitions 6 of this chamber 2, in order to be able to 20 display and regulate this negative pressure and this temperature in the preheating zone; B) A heating zone comprising: - several identical burner ramps 16, numbering two or, preferably, three as represented in figure 1; each fitted with fuel (liquid or gas) 25 burners or injectors and temperature sensors (thermocouples), each of the ramps 16 extending above one of the corresponding number of adjacent chambers 2, such that the injectors of each burner ramp 16 engage with the openings 9 of the hollow partitions 6 in order to inject the fuel into them; 30 C) A blowing zone or natural cooling zone, comprising: 4 - a zero point ramp 17, extending above the chamber 2 immediately upstream from the one below the burner ramp 16 furthest upstream, and fitted with pressure sensors to measure the pressure prevailing in each of the hollow partitions 6 of this 5 chamber 2, in order to be able to adjust this pressure as indicated below, and - a blowing ramp 18, fitted with electric fans equipped with a means for adjusting the flow of ambient air blown into each of the hollow partitions 6 of a chamber 2 upstream from the one located under 10 the zero point ramp 17, such that the flow of ambient air blown into these hollow partitions 6 can be regulated to obtain a desired pressure (slight negative or positive pressure) at the zero point ramp 17; D) A forced cooling zone, which extends across three chambers 2 15 upstream from the blowing ramp 18, and which comprises, in this example, two parallel cooling ramps 19, each fitted with electric fans and pipes for blowing ambient air into the hollow partitions 6 of the corresponding chamber 2; and E) A work zone, extending upstream from the cooling ramps 19, to 20 allow the placement and removal of anodes 5 in the furnace, and maintenance of the chambers 2. The furnace 1 is heated by the burner ramps 16. The burner injectors are introduced through the openings 9 into the hollow partitions 6 of the chambers 2 concerned. Upstream from the heating ramps 16 (relative to 25 the direction in which the fire advances and in which the air and combustion gases and fumes circulate in the lines of hollow partitions 6), the blowing ramp 18 and the cooling ramp(s) 19 comprise pipes for blowing combustion air supplied by electric fans, these pipes being connected, via the openings 9, to the hollow partitions 6 of the chambers 2 concerned. The exhaust 30 manifold 11 is arranged downstream from the burner ramps 16, for extracting the combustion gases and fumes (referred to for brevity as "flue gases"), circulating in the lines of hollow partitions 6.

5 The anodes 5 are heated and baked by the combustion of the fuel (gas or liquid) injected in a controlled manner by the burner ramps 16, and to an appreciably equal extent by the combustion of volatile materials (such as polycyclic aromatic hydrocarbons) from pitch released by the anodes 5 in the 5 pits 4 of the chambers 2 in the preheating and heating zones. As these volatile materials released in the pits 4 are for the most part combustible, and are able to flow into the two adjacent hollow partitions 6 through passages arranged in these partitions, they catch fire in these two partitions because of the residual combustion air present among the combustion gases in these 10 hollow partitions 6. Thus the air and combustion gases circulate the length of the lines of hollow partitions 6, and a negative pressure created downstream from the heating zone B, by the exhaust manifold 11 at the downstream end of the preheating zone A, allows controlling the flow rate of combustion gases 15 inside the hollow partitions 6, while the air coming from the cooling zones C and D due to the cooling ramps 19 and the blowing ramp 18 in particular, is preheated in the hollow partitions 6, cooling the anodes 5 baked in the adjacent pits 4 during its path and serving as an oxidizer when it reaches the heating zone B. 20 As the anodes 5 bake, the assembly of ramps 11 to 19 is cyclically advanced (for example every 24 hours) with the associated measurement and recording equipment and instruments, by one chamber 2, each chamber 2 thus successively performing, upstream from the preheating zone A, a function of loading unbaked carbon blocks 5, then, in the preheating zone A, 25 a function of natural preheating by the combustion gases from the fuel and the pitch vapors which leave the pits 4 and enter the hollow partitions 6 because of the negative pressure in the hollow partitions 6 of the chambers 2 in the preheating zone A, then, in the heating zone B or baking zone, a function of heating blocks 5 to about 11 00"C, and lastly, in the cooling zones 30 C and D, a function of cooling the baked blocks 5 with ambient air and, correlatively, preheating this air which constitutes the oxidizer for the furnace 1. The forced cooling zone D is followed, in the direction opposite the 6 direction the fire advances and the combustion fumes circulate, by a zone E for unloading the cooled carbon blocks 5, then possibly loading unbaked carbon blocks into the pits 4. The process of regulating the furnace 1 essentially consists of 5 regulating the temperature and/or pressure of the preheating A, heating B, and blowing or natural cooling C zones of the furnace 1 as a function of predefined setpoint rules. The combustion gases extracted from the fires by the exhaust manifolds 11 are collected in a duct 20, for example a cylindrical duct partially 10 represented in figure 2, with a flue 21 which can be in the shape of a U (see dotted lines in figure 1) or can extend around the furnace, its outlet 22 conveying the exhausted and collected combustion gases to a flue gas treatment center (FGTC) which is not represented because it is not a part of the invention. 15 In order for the anodes (carbon blocks) to achieve their optimum characteristics, and therefore to guarantee that a final baking temperature is reached, the current preference for furnaces of this type is to supply the burner ramps 16 with fuel (liquid or gas fuel) independently of the pressure differential and air flow conditions in the partitions 6, and therefore 20 incomplete combustion can result in a not insubstantial or even a high number of the lines of partitions 6. This may result in high operating costs for the furnace, not only because of the excess consumption of fuel, but also because of fouling in the exhaust pipes and ducts which leads to the retention of unburnt material, representing an increased potential risk of fire 25 and of an improper baking process. There is a general need to improve the continuous optimization of the operation of such furnaces, in order to reduce operating costs and prevent the risk of fire and impacts on the baking process. For this reason, the invention proposes a method or process for characterizing the 30 combustion in lines of partitions of a chamber-type ring furnace for baking carbon blocks, by analyzing the value of at least one parameter indicative of the total content of unburnt material in the combustion gases and residual air 7 issuing from said lines of partitions and collected in an exhaust manifold of said furnace, said furnace comprising a succession of chambers for preheating, heating, natural cooling and forced cooling, arranged serially along the longitudinal axis of the furnace, each chamber consisting of pits in 5 which are arranged the carbon blocks to be baked, adjacent to and alternating with, transversely to said longitudinal axis, hollow heating partitions, in communication with and aligned with the partitions of other chambers, parallel to the longitudinal axis of the furnace, in lines of partitions in which circulate the cooling and combustion air and combustion gases, said 10 exhaust manifold being connected to each of the partitions of the first preheating chamber by one of the respective exhaust pipes, the necessary combustion air being partially injected by a blowing ramp of the natural cooling zone, connected to at least one fan, and partially infiltrating through the lines of partitions due to the negative pressure, and the fuel necessary for 15 baking the carbon blocks being partially injected by at least two burner ramps each respectively extending over one of the at least two chambers adjacent to the heating zone, and each able to inject fuel into each of the partitions of the respective corresponding chamber of the heating zone, the regulation of the furnace combustion essentially comprising a regulation of the 20 temperature and/or pressure of the preheating, heating, and natural cooling zones, per line of partitions, as a function of predefined temperature and/or pressure setpoint rules, said method for characterizing the combustion being characterized in that it comprises at least one step of successive tests of totally stopping the fuel injection, one line of partitions after another, for a 25 sufficient period to allow the measurement of said parameter indicative of the total content of unburnt material in the combustion gases to stabilize, and without ordering any action in the lines of partitions other than for the one concerned by the totally stopped injection test for the duration of the test, the characterization of the combustion being based on calculating the variation 30 between the measurements of said indicative parameter made before and after totally stopping injection in each of the tested lines of partitions, in order to identify one or more lines of partitions in a situation of incomplete combustion, if said variation is greater than x% of the value of said indicative VVP U iAUl /U t A VI/ L A'A A UU7fU )JAVOA 8 parameter at the start of said totally stopped injection test, where x% is preferably between about 5% to 10%, the value of x depending in particular on the number of partitions per chamber, the detection threshold values, and the accuracy of the measurement of said indicative parameter by at least one 5 detector. Thus, by a test in which fuel injection is totally stopped in a line of partitions but only for a sufficient period for the measurement of the indicative parameter to stabilize and without modifying anything in the other lines of partitions, the method of the invention allows identifying a line of partitions 10 operating in a situation of incomplete combustion, and subsequent measures can be taken to optimize the combustion. In order to limit the number of stopped injection tests and allow the system to identify more quickly the partition or partitions in a situation of incomplete combustion, the method of the invention additionally comprises at 15 least one prior step, the step of preselecting the lines of partitions likely to be in a situation of incomplete combustion, in order to limit the number of stopped injection tests, in said step of successive totally stopped injection tests, to only the preselected lines of partitions, and consisting of: calculating, for each line of partitions of row n, a combustion ratio which is equal to the 20 ratio of the amount of combustion air available to the amount of fuel injected into said line of partitions of row n; empirically defining a limit ratio called the stoichiometric ratio based on measurements of said parameter indicative of the content of unburnt material in the combustion gases collected at the outlet from a benchmark line of partitions, representative of the best state of 25 the lines of partitions of the furnace, and such that this stoichiometric ratio corresponds to a measured threshold for said indicative parameter below which the combustion is considered to be incomplete; comparing the combustion ratio of all lines of partitions to the stoichiometric ratio; and considering the combustion as incomplete in any line of partitions of row n for 30 which the corresponding combustion ratio is less than the stoichiometric ratio.

VT X1 VU /U I U fA& 9 Thus the identification of the lines of partitions in a situation of incomplete combustion, using the totally stopped injection test, is advantageously preceded by preselecting the lines of partitions likely to be in this situation of incomplete combustion, using calculations of the combustion 5 ratio for each of all the lines of partitions in the furnace, and using said stoichiometric ratio defined empirically based on measurements of the indicative parameter in a benchmark line of partitions chosen as being representative of the best state of the lines of partitions, and lastly by comparing each combustion ratio to the stoichiometric ratio in order to 10 deduce which is or are the combustion line(s) in which combustion can be considered as incomplete. In an advantageous implementation of the method for characterizing the combustion according to the present invention, in said step of preselecting the lines of partitions in a state of incomplete combustion, the 15 combustion ratio (RCcIn) in a line of partitions of row n can be calculated as being proportional to the square root of the static negative suction pressure measured in the preheating zone for said line of partitions, and inversely proportional to the sum of the fuel injection capacities from the injectors of the burner ramps operating on the same line of partitions of row n. 20 In particular, during this preselection step, the combustion ratio for the line of partitions of row n can easily be calculated by applying the following formula: (1) RCei10lx P-P,x( tN IjHR i where P1 and P7 are the pressures measured in the partitions of row n of the 25 chambers respectively in communication with the exhaust manifold and the "zero point" ramp in the natural cooling zone, N is the number of burner ramps, generally equal to 2 or 3, and InjHRi is the total injection capacity in the partition of row n for the injectors of the burner ramp of row i, where i varies from 1 to N.

10 In addition, in the characterization method of the invention, the step of preselecting the lines of partitions in a state of incomplete combustion may advantageously also comprise a step which consists of classifying the lines of partitions in a state of incomplete combustion, ordered from the one in 5 which combustion is the most incomplete to the one in which combustion is the least incomplete, by applying a scoring system to the lines of partitions in which any line of partitions of row n is assigned a score NCcin given by the following formula: (2) NC 1 n = 20 - 10( ) RS 10 In addition, in order to obtain preselection information quickly that is easy to use, the step of classifying the lines of partitions can advantageously consider that, for a line of partitions of row n in a good state, the combustion is complete if NCoin<10, the combustion is incomplete if 10<NC 1 ,<12, and the combustion is very incomplete and therefore critical if NCc,[>12 . 15 To ensure an implementation of this characterization method which is advantageous in the simplicity of the detection means and the signal processing provided by these means, the parameter indicative of the total content of unburnt material in the combustion gases is the carbon monoxide (CO) content, which is measured, for determining said stoichiometric ratio, in 20 the exhaust pipe of said exhaust manifold which is connected to the partition of the benchmark line of partitions in the first preheating chamber, said threshold for this indicative parameter to which the stoichiometric ratio corresponds being approximately 500 ppm of CO measured in said exhaust pipe, which corresponds, in the standard operating conditions for this type of 25 furnace, to a level of 1000 ppm of CO at the point of combustion. Thus, as a carbon monoxide detector may already be present in such prior art furnaces, in the collectors of the exhaust manifold, the method of the invention may be implemented without requiring the installation of a specific detection and/or measurement device, by using measurement data 30 already available because they are provided by the sensors of the detection 11 instrumentation already installed on such furnaces. The method of the invention can therefore be implemented by means of a software module which can be easily and simply integrated with the current programs for operating such furnaces. 5 In addition, the method of the invention can be supplemented by implementing, after the characterization steps which allow identifying and selecting the lines of partitions in a state of incomplete combustion, at least one later combustion optimization step. Advantageously, such a combustion optimization can consist of 10 automatically modifying regulation parameters in the preheating, heating, and/or natural cooling zones of the furnace, in order to balance the stoichiometric ratio RS of combustion air to fuel, for the purpose of restoring a situation of complete combustion, which can be defined simply as the value of said indicative parameter falling below a configurable threshold. 15 Whether this optimization step takes place as stated in the above paragraph or in any other manner, the method of the invention can advantageously be such that, after said optimization step, at least one additional step of characterizing the combustion as defined above, in the lines of partitions not pre-selected, in the manner indicated above, from 20 among the lines of partitions thought to be in a state of incomplete combustion, is activated if at least one combustion optimization step as described above does not restore a situation of complete combustion. Other features and advantages of the invention will be apparent from the following non-limiting description and the accompanying drawings in 25 which: - figures 1 and 2, already described above, are respectively a schematic plan view of the structure of a ring furnace with two fires and open chambers, and a partial perspective cross-sectional view with a cutaway representing the internal structure of such a furnace, 30 - figure 3 is a dual graph representing the evolution in the measured CO (in ppm) and the percentage of residual oxygen in the gases VV I AUI A/ U. t A A A llUU7/U J3A UOA 12 collected at the exhaust pipe, for the same line of partitions, as a function of the total injection capacity in the line of partitions, expressed as a percentage of the installed maximum capacity, according to three different values for the static negative suction 5 pressure measured at the preheating measurement ramp associated with the first preheating chamber of the furnace, - figure 4 is a curve characterizing the combustion in a line of partitions of row n, indicating the measured CO content (in ppm) per line of partitions as a function of the combustion ratio RCcin; 10 - figure 5 is a diagram representing, on the x axis, the combustion scores NCcin in a line of partitions of row n, resulting from applying the combustion classification system according to the invention, while the measured CO content (in ppm) per line of partitions in the corresponding exhaust pipe is represented on the y axis, and 15 - figure 6 is a diagram corresponding to an example of a test of totally stopped fuel injection conducted successively in three lines of partitions a, P, and y, and representing on the y axis the value of the total CO content (in ppm) measured in the exhaust manifold over time (expressed in minutes), and revealing a reduction in the total 20 CO content, measured by the test for the first line of partitions a tested, which is greater than a limit value indicative of a state of incomplete combustion in this line of partitions a. The method of the invention concerns a loop for characterizing the combustion in the lines of partitions 6 of the furnace 1 by analyzing the total 25 carbon monoxide (CO) content, or any other parameter indicative of the content of unburnt material, in the gases collected at the exhaust manifold 11 for a fire of the furnace 1, where this total CO content is measured by the CO analyzer-detector 14 in the collector of the exhaust manifold 11 (see figure 2), and the method for characterizing the combustion in the lines of partitions 30 6 comprises a first step of estimating the combustion quality in each of the lines of partitions 6 and preselecting lines of partitions estimated to be in a state of incomplete combustion, then classifying the lines of partitions using a scoring system which allows selecting lines of partitions considered to be in a VV N. UllI UAJ/ U j X U A I A' AL.UU7/ U-JL UaIri 13 state of incomplete combustion, and defined as a function of the ratio of the combustion air to the fuel available in each line of partitions 6 and a stoichiometric ratio RS defined empirically by measurements in a benchmark line of partitions 6, representative of the best state of the lines of partitions in 5 the furnace. This first step in the method for characterizing the combustion allows preselecting lines of partitions 6 which are estimated to be in a state of incomplete combustion if their combustion ratio RC, which is the ratio of the combustion air to the fuel available for each line of partitions 6 considered, is 10 less than the stoichiometric ratio RS presented above. This step of preselecting lines of partitions believed to be in a state of incomplete combustion is immediately followed by a step of selecting lines of partitions 6 considered to be in at state of incomplete combustion, by classifying them using a system of scoring the quality of combustion in the 15 lines of partitions, based, as already mentioned, on the principle of the stoichiometry of the ratio of the amount of combustion air to the amount of fuel available in each line of partitions. In fact, the maximum amount of fuel that can be injected at a given moment into a line of partitions 6 depends on the flow rate of the air in this 20 line of partitions, or the level of static negative pressure measured in this line of partitions at the same time. Below the stoichiometric ratio, the combustion is incomplete, and a portion of the fuels present in the line of partitions no longer burns completely, giving rise to the formation of carbon monoxide (CO). 25 This threshold phenomenon is better understood by considering figure 3, which shows three continuous curves representing the CO content measured in ppm by a CO analyzer 14 in the exhaust pipe 11 a (see figure 2) for a given line of partitions, as a function of the amount of fuel injected, expressed as the total injection capacity into said line of partitions and 30 evaluated as a percentage of the installed maximum capacity. The three continuous curves for the CO measurements are each established for one of three different static negative suction pressures in the line of partitions VV J UtlfU IU't A t LIAAUl'f.U /U.JAUJtYM 14 considered and respectively correspond to three dotted curves indicating the percentage of residual oxygen in the gases collected in the exhaust pipe 11 a of the exhaust manifold 11 considered, these three different static negative pressures being measured by the preheating ramp 15, at the first preheating 5 chamber 2. Thus the curves 23, 24 and 25 for the CO content (in ppm) measured at said exhaust pipe 11a while varying the total injection capacity from 10% to about 30% of the installed maximum capacity, with a static negative suction pressure of -140 Pa, -120 Pa, and -70 Pa, respectively 10 correspond to the dotted curves 26, 27 and 28 indicating the corresponding variation (continuous reduction) in the percentage of residual oxygen, as indicated on the y axis to the right in figure 3, for the same respective negative suction pressures. One will note that for a total injection capacity into a line of 15 partitions 6 of between 10% and 15% of the installed maximum capacity, the curves 23, 24 and 25 for the CO measured at the exhaust pipe 11 a of said line of partitions 6 differ little from each other, and indicate low CO contents (substantially less than 500 ppm), corresponding to a combustion considered to be complete, while for total injection capacity values greater than 15% of 20 the installed maximum capacity, the three curves 23, 24 and 25 for the measured CO diverge from each other with slopes that are first progressively increasing and then substantially constant, but to a greater extent when the absolute value of the negative suction pressure is low. In addition, for a total injection capacity per line of partitions that is greater than about 25% of the 25 installed maximum capacity, the three curves 23, 24 and 25 for the measured CO give results above 1000 ppm, which corresponds to a combustion that is increasingly incomplete with lower absolute values of the negative suction pressure. Simultaneously, the curves 26, 27 and 28 indicating the variation in the percentage of residual oxygen are decreasing, with a negative slope that 30 is substantially constant and differing little between curves. Based on this observation, a combustion ratio RC 1 n is defined for each line of partitions 6 of row n, which gives the ratio of the amount of fuel 15 injected into said line of partitions of row n to the amount of combustion air available in this same line of partitions of row n. The amount of combustion air available in the line of partitions of row n corresponds to the flow rate of the air in this line of partitions of row n, which can be estimated by calculating 5 the square root of the static negative suction pressure in this line of partitions of row n, measured in the preheating zone A by the preheating measurement ramp 15 (see figure 1). The amount of fuel injected into the same line of partitions of row n can be directly obtained by adding the capacities of the injectors which are 10 operating on this same line of partitions. Thus the formula (1) expressing the combustion ratio or relation for this line of partitions of row n, which is RCcln, can be as follows: (1) RCen= 1 0 x VP- P'J x N N SInjHRi where P1 and P7 are pressures measured in the line of partitions of row n at 15 the chambers 2 respectively in communication with the exhaust manifold 11 for P1, in the preheating zone A, and with the zero point ramp 17 in the natural cooling zone C, and where N is the number of burner ramps 16, generally equal to 2 or 3, and lnjHRi is the sum of the injection capacities of the injectors for the burner ramp 16 of row i where i varies from 1 to N (2 or 20 3) in the line of partitions of row n. In addition, one will note that each burner ramp 16 generally contains two injectors per partition 6 of the same corresponding chamber 2, such that if N = 3, as in the example in figure 1 (with three burner ramps 16), a line of partitions of row n is supplied with fuel by six injectors. Thus the combustion ratio RCcln in a line of partitions of row 25 n is proportional to the square root of the static negative suction pressure measured in the preheating zone A for this line of partitions 6, and inversely proportional to the sum of the fuel injection capacities of the injectors for the burner ramps 1 operating on this same line of partitions 6 of row n.

16 Figure 4 shows a crosshatched and curved zone 29 for this line of partitions 6 of row n, which corresponds to the envelope of the different measurement points for the CO measured in ppm at the corresponding exhaust pipe 11a as a function of the variation in the corresponding 5 combustion ratio RCcIn. The threshold value of RC below which combustion is considered to be incomplete, meaning the value of said stoichiometric ratio RS, is defined empirically by observing the value of the CO in a line of partitions representative of the best state of the partitions of the furnace. Beyond a value of 1000 ppm of undiluted CO, which approximately 10 corresponds to a value of 500 ppm measured at the CO detector 14 in the exhaust pipe 11a (figure 2) when taking into account the dilution in the furnace 1, the combustion is considered to be incomplete. In figure 4, the incomplete combustion threshold is therefore indicated at 500 ppm of measured CO, which corresponds to a value for the 15 stoichiometric ratio RS of about 6, at the intersection of the crosshatched zone 29 of the envelope of CO measurement points and the incomplete combustion threshold of 500 ppm. Thus a preselection of lines of partitions 6 likely to be in a situation of incomplete combustion is made, keeping in mind that the CO content, 20 chosen in this exemplary embodiment as a parameter indicative of the total content of unburnt material in the combustion gases, is measured, to determine the stoichiometric ratio RS, in the one of the exhaust pipes 11 a of the exhaust manifold 11 which is connected to the one of the partitions 6 which is located at the intersection of the benchmark line of partitions and the 25 first preheating chamber 2, the threshold of the CO content to which the stoichiometric ratio RS corresponds being about 500 ppm of CO measured at this exhaust pipe 11a, which corresponds, under the standard operating conditions for this type of furnace 1, to a level of 1000 ppm of CO at the point of combustion. 30 From the calculation of the combustion ratio RCcIn, one also can deduce, at least for the lines of partitions 6 believed to be in a state of incomplete combustion from comparing their combustion ratio RCcln with the 17 stoichiometric ratio RS, but preferably for all lines of partitions 6 of the furnace 1, a score which allows classifying the lines of partitions in decreasing order from the one having the most incomplete combustion to the one having the least incomplete combustion, or even the most complete if all 5 lines of partitions are given scores, for example by a scoring system from 0 to 20 defined such that for values above 10, the stoichiometric limit is exceeded and the combustion is considered to be incomplete in the corresponding line of partitions. As an example, to classify the preselected lines of partitions as 10 being in a state of incomplete combustion in the manner described above, these lines of partitions are ordered from the one in which combustion is the most incomplete to the one in which combustion is the least incomplete, by applying the scoring system for the lines of partitions in which any line of partitions 6 of row n is assigned a classification score NCcIn given by the 15 following formula (2): (2) NCe = 20-10( " RS where RCcIn and RS are the ratios previously defined, which are respectively the combustion ratio in the partition of row n and the stoichiometric ratio. Once the lines of partitions are scored from 0 to 20 according to 20 their RCcnl/RS ratio, if the combustion score NCcln is less than 10 the combustion is considered to be complete, while if this combustion score NCcln is between 10 and 12 the combustion is incomplete, and is very incomplete and therefore critical if the score NCcin is greater than 12. An example of such a scoring is represented in figure 5, in which 25 the scores NCcIn are indicated by dots on a continuous curve passing through three crosshatched rectangular zones, one of which 30 extends between the scores 0 and 10 along the x axis and between 0 and the incomplete combustion threshold of 500 ppm of measured CO for the lines of partitions in a state of complete combustion, with a second zone 31 30 extending along the x axis between the scores 10 and 12 and on the y axis 18 between the values of 500 and 1000 ppm of measured CO, for one or more line(s) of partitions in a state of incomplete combustion, and lastly with a third zone 32 for scores above 12 on the x axis and a measured CO greater than 1000 ppm on the y axis, for any line of partitions having highly incomplete 5 and therefore critical combustion. With such a scoring system, the lines of partitions are considered to be in a state of incomplete combustion if they have a score above 10, and these selected lines of partitions are then each subjected to a step of identifying the lines of partitions in a state of incomplete combustion, using a 10 test in which fuel injection is totally stopped for a given period in the selected lines of partitions, in succession, starting with the one having the highest score and conducting the test successively on the lines of partitions by decreasing order of their combustion scores. Figure 6 schematically represents the steps in the totally stopped 15 fuel injection test in three successive lines of partitions of row a, P, and y having progressively lower combustion scores NC. In figure 6, the y axis shows the total CO content measured in ppm by the CO detector 14 in the collector for the exhaust manifold 11 (see figure 2), while the x axis indicates the time in minutes. The curve 33 represents the evolution over time of the 20 total CO content measured in the collector for the exhaust manifold 11. At time t1, the order is issued to totally stop the supply of fuel to the injectors of the burner ramps 16 operating on the line of partitions a, in an almost instantaneous cutoff, from an initial value (for the totally stopped injection test) of the fuel injection rate to a zero rate, which corresponds to the left 25 downward arrow in the "a" rectangle, symbolizing the injector fuel supply control for this line of partitions a during this totally stopped injection test. The injection is stopped for a time interval t1 t2 sufficient for the measurement of the CO content to stabilize before the time t2 at the end of the injection total cutoff. The curve 33 of the CO content indicates a drop to a stabilized value 30 of, for example, 500 ppm during the interval t1 t2, such that it is possible to measure the value ACO corresponding to the difference between the initial value of the CO content at time t1 and the final value at time t2 due to this 19 interruption in the fuel supply. Then, at time t2, the fuel supply to this line of partitions a is returned to its initial value, as symbolized by the upward arrow on the right side of the "a" rectangle in figure 6. Then an interval of time t2 t3 passes, of a duration slightly greater than or substantially equal to the interval 5 ti t2, about 2 minutes, with the same totally stopped fuel injection test starting at time t3 in the line of partitions of row fl, keeping in mind that during the execution of a totally stopped injection test in a particular line of partitions, no change is ordered in the baking process in the all other lines of partitions. The duration of the second test in the line of partitions /, 10 corresponding to the interval t3 t4, is the same as the duration ti t2, and the curve 33 of the CO content, which returned after the end of the test on the line of partitions a to a normal level, only shows, as a result of the test on the line of partitions fl, a limited decrease in the CO content measured after injection was totally stopped in the line of partitions # during the interval t3 t4. 15 The same is true for the third totally stopped injection test, conducted on the line of partitions y for the interval of time t5 t6, for the same duration of about 2 min as the durations of the other tests t1 t2 and t3 t4, such that the measurement of the CO content during each test can stabilize, after the fuel injection is no longer cut off, during the interval of time separating two 20 successive tests. For each test, the reduction in the CO content which results, ACO, is compared to a percentage X of the initial value of the CO content at the start of this test, COi, and if ACO is greater than X% of COi, as is the case for the line of partitions a, the line of partitions a is identified as being in a state 25 of incomplete combustion, which is not the case for the lines of partitions # and y illustrated by the curve 33 in figure 6. The totally stopped fuel injection test is therefore conducted, one line of partitions after another, on the lines of partitions previously selected by their combustion score NC. It is essential that no action be ordered on lines 30 of partitions 6 other than the one involved in the totally stopped injection test, for the entire duration of the test, in order to avoid interfering with the characterization of the combustion. This characterization depends on 20 calculating the variation in the CO content measured between the initial moment of the test and the final moment, with the measurements of CO content always being total measurements. The sudden downward slope followed by the climb in the curve 33 in figure 6 therefore expresses quite 5 well the impact of totally stopping the fuel injection in the line of partitions a on the CO content in the collector of the manifold 11, which therefore takes into account the exhaust gases extracted from all lines of partitions in the furnace. As for the threshold of X% of the value of the COi content at the 10 start of each totally stopped injection test, this value of X depends in particular on the number of partitions 6 per chamber 2 in the furnace, as well as on the accuracy of the measurement and the limit detection values for the CO detector 14 more particularly. In general, X% is chosen to be within a range of 5% to 10%. Typically for a furnace 1 with nine partitions 6 per 15 chamber 2, a characterization system which makes use of the method of the invention must be able to detect at least one partition of row n from among nine partitions 6 where combustion tends to become incomplete. If we consider the flows circulating in each line of partitions, and therefore in each partition, to be equivalent, the drop in the CO content after stopping the 20 injection of fuel into the partition of row n will be at least equal to ACOn = 500 ppm/9 = 56 ppm, due to dilution, which is about X = 10% of the CO content measured at the collector of the exhaust manifold 11, where this content is equal to at least 500 ppm. After having thus selected the lines of partitions considered to be in 25 a state of incomplete combustion, using the stoichiometric ratio RS, the combustion ratios RC for the lines of partitions, the comparison of combustion ratios to the stoichiometric ratio, and the assignment of combustion scores NC to the lines of partitions, and after the lines of partitions in a state of incomplete combustion are identified by the totally 30 stopped fuel injection test, at least one later step, the combustion optimization step, can be applied.

21 Such a step can consist of modifying, preferably automatically, regulation parameters in at least one of the natural cooling C, heating B, and preheating A zones, in order to balance as much as possible the combustion ratios to the stoichiometric ratio of combustion air to fuel, to restore a 5 situation of complete combustion in as large a number as possible of lines of partitions, it being possible to define this restoration of a situation of complete combustion as the passage, of the measured value for the CO content or of the value of at least one other parameter indicative of the total content of unburnt material in the combustion gases, to below a configurable threshold. 10 But, if the combustion optimization step or steps as generally presented above do not allow or have not allowed restoring a situation of complete combustion for all the lines of partitions in the furnace 1, then the method of the invention proposes at least one additional step for characterizing the combustion, which takes place by applying the totally 15 stopped injection test to those lines of partitions which were not preselected in accordance with the method of the invention from among the lines of partitions thought to be in a state of incomplete combustion solely due to the fact that their combustion ratio RC was calculated to be less than the stoichiometric ratio RS. In addition, this additional characterization step 20 allows identifying partitions in which the stoichiometric conditions are satisfactory, having a combustion score NC of less than 10 in the scoring system example described above, but in which the physical conditions are generating combustion problems due to the fact that the partitions are deformed, restricted, or more or less plugged. 25

Claims

1. Method for characterizing the combustion in lines of partitions of a chamber-type ring furnace for baking carbon blocks (5), by analyzing 5 the value of at least one parameter indicative of the total content of unburnt material in the combustion gases and residual air issuing from said lines of partitions (6) and collected in an exhaust manifold (11) of said furnace (1), said furnace (1) comprising a succession of chambers (2) for preheating, heating, natural cooling, and forced 10 cooling, arranged serially along the longitudinal axis (XX) of the furnace (1), each chamber (2) consisting of pits (4) in which are arranged the carbon blocks (5) to be baked, adjacent to and alternating with, transversely to said longitudinal axis (XX), hollow heating partitions (6), in communication with and aligned with the 15 partitions (6) of other chambers (2), parallel to the longitudinal axis (XX) of the furnace (1), in lines of partitions (6) in which circulate the cooling air and combustion air and combustion gases, said exhaust manifold (11) being connected to each of the partitions (6) of the first preheating chamber (2) by one of the respective exhaust pipes (11 a), 20 the necessary combustion air being partially injected by a blowing ramp (18) of the natural cooling zone (C), connected to at least one fan, and partially infiltrating through the lines of partitions (6) due to the negative pressure, and the fuel necessary for baking the carbon blocks (5) being partially injected by at least two burner (B) ramps (16) 25 each respectively extending over one of the at least two chambers (2) adjacent to the heating zone, and each able to inject fuel into each of the partitions (6) of the corresponding chamber (2) of the heating zone (B), the regulation of the furnace (1) combustion essentially comprising a regulation of the temperature and/or pressure of the 30 preheating (A), heating (B), and natural cooling (C) zones, per line of partitions (6), as a function of predefined temperature and/or pressure setpoint rules, said method for characterizing the combustion being 23 characterized in that it comprises at least one step of successive tests of totally stopping the fuel injection, one line of partitions (6) after another , for a sufficient period to allow the measurement of said parameter indicative of the total content of unburnt material in the 5 combustion gases to stabilize, and without ordering any action in the lines of partitions (6) other than for the one concerned by the totally stopped injection test for the duration of this test, the characterization of the combustion being based on calculating the variation between the measurements of said indicative parameter made before and after 10 totally stopping injection in each of the tested lines of partitions (6), in order to identify one or more lines of partitions (6) in a situation of incomplete combustion, if said variation is greater than x% of the value of said indicative parameter at the start of said totally stopped injection test, x% being preferably between about 5% to 10%, the value of x 15 depending in particular on the number of partitions (6) per chamber (2), the detection threshold values, and the accuracy of the measurement of said indicative parameter by at least one detector.

2. Method according to claim 1, characterized in that it additionally comprises at least one prior step, called the step of preselecting the 20 lines of partitions (6) likely to be in a situation of incomplete combustion, which allows limiting the number of stopped injection tests, in said step of successive totally stopped fuel injection tests, to only the preselected lines of partitions (6), and consisting of: calculating, for each line of partitions (6) of row n, a combustion ratio 25 (RCoin) which is equal to the ratio of the amount of combustion air available to the amount of fuel injected into said line of partitions (6) of row n; empirically defining a limit ratio called the stoichiometric ratio (RS) based on measurements of said parameter indicative of the content of unburnt material in the combustion gases collected at the 30 outlet from a benchmark line of partitions (6), representative of the best state of the lines of partitions (6) of the furnace, and such that this stoichiometric ratio (RS) corresponds to a measured threshold for said indicative parameter below which the combustion is considered to be 24 incomplete; comparing the combustion ratio (RCin) of all the lines of partitions (6) to the stoichiometric ratio (RS); and considering the combustion as incomplete in any line of partitions (6) of row n for which the corresponding combustion ratio (RCc.,) is less than the 5 stoichiometric ratio (RS).

3. Method according to claim 2, characterized in that, in said step of preselecting the lines of partitions (6) in a state of incomplete combustion, the combustion ratio (RCin) in a line of partitions (6) of row n is calculated as being proportional to the square root of the 10 static negative suction pressure measured in the preheating zone (A) for said line of partitions (6), and inversely proportional to the sum of the fuel injection capacities from the injectors of the burner ramps (16) operating on the same line of partitions (6) of row n.

4. Method according to claim 3, characterized in that, during said 15 preselection step, the combustion ratio for the line of partitions (6) of row n is calculated by applying the following formula: (1) RCcin=10 x P -P,I x ( N N SInjHRi where P 1 and Pr are the pressures measured in the partitions (6) of row n of the chambers (2) respectively in communication with the 20 exhaust manifold (11) and a "zero point" ramp (17) in the natural cooling zone (C), N is the number of burner ramps (16), generally equal to 2 or 3, and lnjHRi is the total injection capacity in the partition of row n for the injectors of the burner ramp (16) of row i, where i varies from 1 to N. 25

5. Method according to any one of claims 2 to 4, characterized in that said step of preselecting the lines of partitions (6) in a state of incomplete combustion also comprises a step which consists of classifying the lines of partitions (6) in a state of incomplete combustion, ordered from the one in which combustion is the most 25 incomplete to the one in which combustion is the least incomplete, by applying a scoring system to the lines of partitions (6) in which each line of partitions (6) of row n is assigned a classification score NCi, given by the following formula: RC (2) NCe = 20 - 10 ( " ) 5 RS

6. Method according to claim 5, characterized in that the step of classifying the lines of partitions (6) is carried out by considering that, for a line of partitions (6) of row n in a good state, the combustion is complete if NCoin<10, the combustion is incomplete if 10<NCoIn<12, 10 and the combustion is very incomplete and therefore critical if NCin>1 2.

7. Method according to any one of claims 1 to 6, characterized in that the carbon monoxide (CO) content is chosen as the parameter indicative of the total content of unburnt material in the combustion 15 gases, said CO content being measured, for determining said stoichiometric ratio, in the exhaust pipe (11 a) of said exhaust manifold (11) which is connected to the partition (6) of the benchmark line of partitions (6) in the first preheating chamber (2), said threshold for this indicative parameter to which the stoichiometric ratio (RS) 20 corresponds being approximately 500 ppm of CO measured in said exhaust pipe (11a), which corresponds, in the standard operating conditions for this type of furnace, to a level of 1000 ppm of CO at the point of combustion.

8. Method according to any one of claims 1 to 7, characterized in that, 25 after the characterization steps which allow identifying and selecting the lines of partitions (6) in a state of incomplete combustion, at least one later combustion optimization step is applied.

9. Method according to claim 8, characterized in that the combustion optimization consists of automatically modifying regulation parameters 30 in the preheating (A), heating (B), and/or natural cooling (C) zones of 26 the furnace (1), in order to balance the stoichiometric ratio (RS) of combustion air to fuel, for the purpose of restoring a situation of complete combustion, defined as the value of said indicative parameter falling below a configurable threshold. 5

10. Method according to either of claims 8 or 9, characterized in that, after said optimization step, at least one additional step of characterizing the combustion according to claim 1, in the lines of partitions (6) not pre-selected according to claim 2 from among the lines of partitions (6) thought to be in a state of incomplete combustion, is activated if the 10 combustion optimization step according to claim 8 or 9 did not restore a situation of complete combustion.