WO2023208567A1

WO2023208567A1 - Method and system for the end-of-cast control in a blast furnace operation

Info

Publication number: WO2023208567A1
Application number: PCT/EP2023/059184
Authority: WO
Inventors: Colin COMMANDEUR; Gerard Marie LOUWERSE; Jan VAN DER STEL; Joost Christiaan Storm; Rudolf SPRIK; Christiaan Cornelis STOLK
Original assignee: Tata Steel Ijmuiden B.V.
Priority date: 2022-04-26
Filing date: 2023-04-06
Publication date: 2023-11-02

Abstract

The invention relates to a method and system of controlling the end-of-cast of a tapping operation in a metallurgical furnace, preferably in a blast furnace, the method comprising: initiating a tapping operation by tapping liquids from at least one taphole in a metallurgical furnace; receiving acoustic signals from the tapping operation in at least one acoustic receiver located remotely from the taphole; converting the acoustic signals received by the least one acoustic receiver into electrical signals, the electrical signals corresponding to acoustic signals transmitted by the tapping operation and sensed by the acoustic receiver; processing the electrical signals to determine if an acoustic event in relation to the tapping operation has occurred by comparing the electrical signals with pre-determined tapping operation conditions; generating indication data, depending on the comparing; providing an indication based on the indication data; and initiating an action in relation to the tapping operation using the indication data.

Description

METHOD AND SYSTEM FOR THE END-OF-CAST CONTROL IN A BLAST FURNACE OPERATION

FIELD OF THE INVENTION

The invention relates to a system and a method of controlling and optimizing the end-of- cast control in a blast furnace operation when tapping molten iron and slag from at least one taphole in the blast furnace.

BACKGROUND TO THE INVENTION

Most metallurgical furnaces have at least one tapblock or taphole for the purpose of draining molten process material from the furnace. A blast furnace is an example of an industrial metallurgical furnace. The process of draining liquids from a metallurgical furnace is called tapping or casting. The blast furnace process results in liquid iron and slag being produced. These two liquids drip down into the coke-filled hearth of the blast furnace, where they wait to be tapped, or cast, from the furnace. The regular removal of liquids from the hearth is done through the taphole or tapholes. The number of tapholes can range from one to five, depending on the size and output of the furnace. The majority of modern high productivity blast furnaces have two, three or four tapholes. The tapholes are openings in the blast furnace shell with special refractory constructions built into the hearth sidewall. Each taphole has an opening or innerside facing the inside of the blast furnace and an opening or outerside facing the outside of the blast furnace. The tapholes are opened by either drilling through the refractory or by placing a bar in the refractory that is later removed. The holes are closed by forcing a plug of malleable refractory clay into the hole, which quickly hardens to securely seal the hole. A blast furnace will be cast or tapped between 8 and 14 times per day. These casts may last between about 90 and 180 minutes, with the end of the cast indicated by a spraying of liquids and noises caused by gas from the raceway escaping out of the taphole.

The casthouse operation is an extremely important area for the blast furnace. The main objectives of good casthouse operation may be summarized as follows: (i) to remove liquid iron and slag from the furnace at a rate that does not allow the process to be affected by increasing liquid levels in the hearth, (ii) to separate and sample the iron and slag that is cast from the furnace, and (iii) to direct the iron to the ladle and the slag to the slag pot, slag pit or slag granulator. The extraction of liquids from the hearth is crucial for maintaining stable process parameters, and the damaging effects of not casting the furnace occur very fast.

For the safe operation of a blast furnace, sufficient iron and slag needs to be drained to prevent flooding of the blast furnace. When the liquid level becomes too high in the furnace, it will cause operational problems and potentially lead to dangerous situations. On the other hand, tapping too much liquid from the furnace may result in short-circuiting of the hot blast and gas escaping via the taphole to the outside of the furnace. This causes damage to the taphole, releases gas and dust into the environment and causes severe spraying of the cast which is undesired. Especially the environmental issues caused by gas and dust release are important. One may estimate the levels of liquid iron and slag in the furnace to predict when the cast needs to be ended. A heat and mass balance are an option; however such models need frequent fine- tuning to be used in practice. Thermocouple data and electromagnetic field (EMF) sensors can be used to provide feedback to such models. As the liquid level models lack the required accuracy in practice the end of cast is often judged by experienced operators mainly by sight (e.g. increased splatter near the taphole exit and hot blast coming through the taphole) and hearing. At the end of the cast the liquid may produce more and more sparks and may eventually start spraying. For the longevity of the taphole it should be closed preferably before severe spraying occurs. Hence, the chosen moment of closing the taphole depends commonly on the judgement and experience of the operator.

Patent document W02009/039665-A1 discloses systems and methods for acoustic monitoring of a tapblock of a metallurgical furnace, in particular a blast furnace. The method comprising receiving electrical signals from a plurality of acoustic emission sensors along at least one acoustic waveguide that is at least partially received within an outer structure of a tapblock, in particular of a blast furnace. The electrical signals correspond to acoustic signals being transmitted along the acoustic waveguide, preferably at least one of a cooling conduit and a thermal well, and sensed by the acoustic emission sensors, preferably by a plurality of accelerometers. The electrical signals are processed to determine the occurrence of events in relation to an inner structure of the tapblock, particularly the integrity of the refractory lining. The method and system monitor the tapping process, drilling process, and specifically the lancing process, and provide feedback in order to minimize damage to the tapblock or the refractory lining.

There is a need for a method and for a corresponding system to better control the tapping process, and in particular the end-of-cast process from a taphole in a metallurgical furnace, and provide feedback in order to minimize damage to the taphole and minimize the release of gas and dust into the environment.

DESCRIPTION OF THE INVENTION

It is an object of the invention to provide a method of controlling the tapping operation, in particular the end-of-cast, in a blast furnace when tapping molten iron from at least one taphole. It is another object of the invention to provide a system for controlling the end-of-cast of a tapping operation when tapping molten iron from a blast furnace.

These and other objects and further advantages are met or exceeded by the present invention providing a method according to independent claim 1 and an acoustic monitoring system according to independent 10, and with preferred embodiments of the method and the system in the dependent claims and the description.

All features and advantages described herein with respect to the method of the invention may also be applied to the system of the invention and vice versa.

The invention is embodied in a method of controlling a tapping operation from a taphole, in particular the end-of-cast of a tapping operation of a blast furnace 110, the method comprising: initiating a tapping operation by tapping molten iron from at least one taphole 120 in the blast furnace 110; receiving acoustic signals 201 from the tapping operation in at least one acoustic receiver 140 located remotely from the tapblock or taphole 120; converting the acoustic signals 201 received by the least one acoustic receiver 140 into electrical signals 202, the electrical signals 202 corresponding to acoustic signals transmitted by the tapping operation and sensed by the at least one acoustic receiver 140 located remotely from the taphole; processing the electrical signals 202 to determine if an acoustic event in relation to the tapping operation has occurred by comparing the electrical signals 202 with pre-determined tapping operation conditions 203; generating indication data 210, resulting from the comparison or depending on the comparing; providing an indication based on the indication data 210; and initiating an action in relation to the tapping operation using the indication data 210.

The invention is also embodied in an acoustic monitoring system 100 for controlling a tapping operation from a taphole 120, in particular the end-of-cast of a tapping operation, in a blast furnace 110, for implementing to method according to this invention, the system comprising: at least one acoustic receiver 140 located remotely from a tapblock or taphole 120 of the blast furnace 110 and the at least one acoustic receiver 140 being configured to sense acoustic signals 201 from a tapping operation of a blast furnace 110; a monitoring station 150 comprising a processor 412 configured to execute instructions stored on a computer-readable medium in a memory 418, the monitoring station 150 causing the processor 412 to perform processor operations including: receiving the acoustic signals 201 associated with the tapping operation; converting the acoustic signals 201 received by the acoustic receiver 140 into electrical signals 202, the electrical signals 202 corresponding to acoustic signals 201 transmitted by the tapping operation and sensed by the acoustic receiver 140; processing the electrical signals 202 to determine if an acoustic event in relation to the tapping operation has occurred by comparing the electrical signals 202 with predetermined tapping operation conditions 203; generating indication data 210 resulting from the comparison or depending on the comparing; and providing an indication based on the indication data 210.

In accordance with the invention it has been found that during a tapping operation in a blast furnace 110 a wide range of sounds or acoustic emissions by various mechanisms are created. The liquid iron and slag should be drained from a blast furnace 110 just until the point that the slag level will dip or fall below the level at which the taphole is present. The taphole 120 is drilled under a slight angle and is a few centimetres in diameter. The taphole size will increase during the casting period due to erosion. The gas pressure present above the slag is several bars above atmospheric pressure. When the slag level dips for a short period of time below the top of the taphole on the hot side a bubble will be trapped. These gas bubbles will travel up to taphole and escape from the blast furnace 110 on the outside of the taphole 120. As the slag level drops further, the gas bubbles will grow bigger and bigger. Due to the lower pressure on the outside of the blast furnace 110 the gas bubbles will expand and accelerate towards the outlet of the taphole 120 and likely form the cause of the splashing. At the point that gas bubbles start to escape from the blast furnace 110 the taphole 120 is commonly closed shortly thereafter via an action of human operators to prevent a gas break-through. The inventors have found that already several minutes prior to a regular end-of-cast situation, there are a distinct series of sounds or characteristic acoustic signals, e.g., one or more of frequency, pitch, loudness, decibel level, or tone, announcing that an end-of-cast is approaching. By detecting these sounds or acoustic signals 201 at a location remotely or at distance, e.g. at several meters, from the outerside of the tapblock or taphole 120 and separating these sounds from the background noise a reliable method is provided to earlier than before initiating the end of the tapping operation, viz. end-of-casting process, by shutting off the tapping operation and thereby limiting or avoiding the disadvantageous events of splatter, the release of gas and dust. This offers substantial environmental advantages. A further advantage is that it offers a safer working environment for the human operators 130. Furthermore, the longevity of the taphole 120 is increased due to limiting excessive wear of the refractory lining of the taphole or tapblock.

The invention uses acoustics, i.e. the characteristics sounds like for example poppingtype sounds, hissing-type sounds or other sounds related to release of energy or gas, which can be registered and processed even when a substantial amount of smoke is present near the taphole and on the shopfloor hindering a human visual assessment or limiting the use of image processing solutions like optical sensors. The use of image processing solutions is also limited because any smoke, dust and sparks will result over time in the fouling of the hardware used, for example a camera to capture or detect optical data. The use of image processing solutions is also impeded by the lids covering a trough or launder configured to deliver the liquids from the taphole to an appropriate vessel during a tapping operation, e.g. to a torpedo car.

The acoustic monitoring system 100 used in the method and system according to this invention is configured to detect acoustic emissions from a variety of sources, more in particular noise or acoustic emissions related to molten metal flow, energy release, gas development and gas flow during a tapping operation of molten iron from a blast furnace 110. The acoustic receiver 140 or a group of acoustic receivers 140 located remotely from the taphole 120 serve as transducers to convert the acoustic signals 201 into corresponding electrical signals 202 that can be processed by the monitoring station 150 comprising a processor 412. The acoustic receiver 140 or acoustic sensor 140 may include one or more microphones to capture sound data, e.g., audible, infrasound and/or ultrasound data, from different portions of the tapping process. Acoustic receiver 140 captures the sounds, e.g. audible, infrasounds and/or ultrasounds, themselves or captures characteristics or other data associated with those sounds. The acoustic receiver 140 may include a variety of different types of receivers or sensors. In a preferred embodiment the acoustic receiver 140 is a microphone located remotely from the taphole. The microphone may be directional, bi-directional, omnidirectional, dynamic, condenser, or other types. In a preferred embodiment the microphone is an omnidirectional microphone. In an embodiment the acoustic receiver 140 comprises at least two sensors positioned remotely from the taphole of the blast furnace. Advantageously a setup of a group of acoustic sensors or acoustic receivers 140 can be used to improve the suppression of background noise.

In an embodiment the acoustic monitoring system 100 has a monitoring station 150 comprising a processor 412 configured to execute instructions stored on a computer-readable medium in a memory 418, the monitoring station 150 causing the processor 412 to perform processor operations including: receiving the acoustic signals 201 associated with the tapping operation; converting the acoustic signals 201 received by the acoustic receiver 140 into electrical signals 202, the electrical signals corresponding to acoustic signals 201 transmitted by the tapping operation and sensed by the acoustic receiver 140; processing the electrical signals 202 to determine if an acoustic event in relation to the tapping operation has occurred by comparing the electrical signals 202 with pre-determined tapping operation conditions 203; generating indication data 210, resulting from the comparison or depending on the comparing; providing an indication based on the indication data 210; and initiating an action in relation to the tapping operation using the indication data 210, in particular shutting off the tapping operation.

The acquisition and processing of the acoustic signal data can be performed using a variety of methods or commercially available systems known to those skilled in the art. In an embodiment of the acoustic monitoring system 100, the monitoring station 150 is a personal computer (PC) or a server-based processing system.

In an embodiment the monitoring station 150 includes a controller 410 that is implemented digitally and is programmable using conventional computer components. The controller 410 comprises a processor 412 executing code or instructions stored on a tangible computer-readable medium in a memory 418, or elsewhere such as portable media or on a server or in the cloud among other media, to cause the controller 410 to receive and process data and to perform actions and/or control components of equipment, e.g., features 160, 170, or 180 such as shown in Fig. 1. The code can be stored on a computer-readable storage medium, for example in the form of a computer program comprising a plurality of instructions executable by one or more processors 412. In an embodiment the computer-readable storage medium is non-transitory. The controller 410 can be any device that may process data and execute code that is a set of instructions to perform actions such as to control industrial equipment. As non-limiting examples, the controller 410 takes the form of a digitally implemented and/or programmable PID controller, a programmable logic controller, a microprocessor, a server, a desktop or laptop personal computer, a laptop personal computer, a handheld computing device, and a mobile device.

Examples of the processor 412 include any desired processing circuitry, an applicationspecific integrated circuit (ASIC), programmable logic, a state machine, or other suitable circuitry. The processor 412 may include one processor or any number of processors. The processor 412 may access code stored in the memory 418. The memory 418 can be any transitory or non-transitory computer-readable medium configured for tangibly embodying code and may include electronic, magnetic, or optical devices. Examples of the memory 418 include random access memory (RAM), read-only memory (ROM), flash memory, compact disc, digital video device, magnetic disk, an ASIC, a configured processor, or other storage device.

In an embodiment the instructions are stored in the memory 418 or in the processor 412 as executable code. The instructions may include processor-specific instructions generated by a compiler and/or an interpreter from code written in any suitable computer-programming language. The instructions may take the form of an application that includes a series of setpoints, parameters, and programmed steps which, when executed by the processor 412, allow the controller 410 to monitor and control various components of the acoustic monitoring system 100. In a preferred embodiment the instructions comprise the processor 412 initiating an action in relation to the tapping operation using the indication data and includes shutting off the tapping operation.

Sounds or acoustic signals 201 received by an acoustic receiver 140 are converted into electrical signals 202 enabling an analyses to determine an acoustic profile or acoustic characteristics such as frequency, pitch, loudness, decibel level, or tone, among others. Using the acoustic receiver data, i.e. acoustic signals 201 , and characteristics of the acoustic receiver data, one or more acoustic profiles may be generated. The acoustic profiles are associated with an event associated with the tapping operation, particularly in relation to an approaching end-of- cast of a tapping operation. In other words, the acoustic profiles correlate certain sounds or characteristics of those sounds, or both, to a certain event associated with the tapping operation.

Processing the acoustic receiver data, i.e., acoustic signals 201 , converted into electrical signals 202 with the set of acoustic profiles including comparing characteristics of the acoustic signals captured relative to the tapping operation to characteristics of acoustic signals captured during previous tapping operations and forming the pre-determined tapping operation conditions 203. An acoustic profile of the set of acoustic profiles includes a machine learning algorithm generated using data associated with multiple instances of a particular event, particularly the end of cast event of a tapping operation, associated with one or more metallurgical furnaces 110 and forming the pre-determined tapping operation conditions 203 or threshold conditions.

In an embodiment the profiles include the identification of one or more ranges of sound characteristics, e.g., a range of frequencies, associated with the event that the profile is associated with. In another embodiment, the profiles identifies patterns of characteristics of the captured sound data to create pre-determined tapping operation conditions 203 or other threshold conditions. These characteristics, and any ranges or patterns associated with those characteristics, can be refined over time using for example neural networks and machine learning. For example, the amount of acoustic receiver data, i.e., acoustic signals 201 , may increase over time using aggregation of acoustic receiver data from multiple tapping operations from one single blast furnace or multiple tapping operations from multiple different but comparable blast furnaces. Preferably outlier acoustic receiver data are filtered out as anomalies. Over time, the profiles associated with an approaching end-of-cast event can be adjusted and refined based on new acoustic receiver data to improve the accuracy of the predetermined tapping operation conditions 203. This refining allows the acoustic receiver data to be used to predict future events more accurately or characteristics of future events, and therefore allow a user to more accurately or effectively prepare for such events.

Furthermore, in an embodiment the monitoring station 150 comprising a processor 412 programmed to automatically act based on a determined likely event in the future. For example, instead of or in addition to notifying a user of an approaching end-of-cast, the system may automatically take a particular action, i.e., shutting off the tapping operation, that the system has been programmed to take when such an event is identified or a certain threshold likelihood that such an event occurs. For example, if a threshold or pre-determined tapping operation condition 203 is set by a user or by a neural networks or machine learning algorithm at for example 95% likelihood for a certain event, and the monitoring station 150 determines that the likelihood of the event occurring during a particular tapping operation is 96%, then the processor 412 in the monitoring station 150 may take action automatically to prevent the event from happening, for example by shutting off the tapping operation.

In an embodiment an acoustic receiver 140 is configured to detect ultrasound signals 201 during a tapping operation, and preferably in a frequency range of about 15 kHz to 60 kHz. It has been found that there is a significant rising trend of the amplitude within this frequency range towards the end of the tapping operation, typically 2 to 4 minutes before the end-of-cast, and therefore may define a reliable threshold that indicates a too high amount of gas and energy is escaping and the cast may soon start to splash.

In an embodiment an acoustic receiver 140 is configured to detect sound signals 201 in a frequency range of about 100 Hz to 2 kHz during a tapping operation. It has been found that there is a significant rising trend of plopping-type sounds within this frequency range towards the end of a tapping operation, typically 2 to 4 minutes before the end-of-cast, and therefore may define a reliable threshold that indicates for example a too high amount of gas and energy is escaping and the cast may soon start to splash. Processor 412 may continuously analyse and process, e.g. a machine learning algorithm, the incoming electrical signals 202 and compare these with pre-determined tapping operation conditions 202. In an embodiment it counts the number of popping-type noise registered in a timeframe and notifies the user when the count has crossed a pre-defined threshold.

In another embodiment the acoustic receiver 140 or group of acoustic receivers are configured to detect sounds signals both in the in a frequency range of about 100 Hz to 2 kHz and in a range of about 15 kHz to 60 kHz. It has been found that the output or emission within both frequency ranges start to increase at the same time, viz. about 2 to 4 minutes before the end-of-cast of a tapping operation in a blast furnace operation.

Having performed the necessary acoustic signal acquisition and processing, the processor 412 of the monitoring station 150 provides an output to the indicator 160 or human machine interface (HMI) and the status display 170.

The indicator 160 provides a highly visible display positioned in proximity to the blast furnace 110 to provide real-time feedback to employees and human operators 130 in proximity to the blast furnace 110. Specifically, the indicator 160 or HMI may provide visual feedback to an operator 130 during tapping. The indicator 160 or HMI is advantageously configured to display feedback to a tapper in real-time and to inform the human operator 130 whether the tapping must be stopped or not, by displaying for example a message on an operator screen. The real-time feedback allows an operator 130 to modify the operator’s actions by timely ending the tapping operation or cast in order to avoid amongst others unnecessary damage to the taphole and the refractory lining therein.

It is understood that the indicator 160 is configured to display any combination of visual information, e.g., lights, text, images, photographs, or animations, and audio information, e.g., horns, buzzers, sirens, pre-recorded dialogue, recorded warning messages, etc.. In another embodiment the indicator 160 is configured to act automatically by shutting off the tapping process. The indicator 160 is advantageously communicatively connected to the monitoring station 150 in a wired connection. Alternatively, the indicator 160 can use a wireless connection, e.g., Bluetooth or WiFi, with the monitoring station 150. In still another embodiment, the indicator 160 can be integrated to the monitoring station 150 such that they share at least one common electronic processor.

In addition to the indicator 160, information from the monitoring station 150 can be sent to a status display 170. The status display 170 may display the same information displayed by the indicator 160 or HMI, or it may a different set of information. In addition, the status display 170 may be physically located in close proximity to the blast furnace 110, or alternatively the status display 170 may be located in a remote location, such as a control room or supervisor’s office. The status display 170 may take the same form as the indicator 160 or it may be of a different form. For example, the status display may comprise a computer monitor, an analogue meter, a digital display, an auditory alarm, a television monitor, or any other appropriate display apparatus. While the embodiment of the acoustic monitoring system 100 is shown comprising both an indicator 160 and a status display 170, it is understood that the acoustic monitoring system 100 could be configured to operate without the indicator 160 and/or the status display 170 or that the functions of both the indicator and the status display 170 could be combined into a single element.

In an embodiment the monitoring station 150 is connected to a network 180 such that it is in communication with user stations 190. The network 180 can be open or a closed network. The network 180 can be a wired or wireless network. The user stations 190 connected to the network may be PCs or any similar device. Once connected to the network 180, output information from the monitoring station 150 may be accessed from, or stored in, user stations 190 that can be at remote locations. The information displayed on the user stations 190 may be the same information displayed by the indicator 160 and the status display 170, or the user stations 190 may be configured to display a different set of information. In addition to displaying the real-time information output by the monitoring station 150, the user stations 190 may also be configured to access any stored signal data or acoustic event information contained with the monitoring station 150. Having access to stored data enables an operator working at a user station 190 to compare real-time acoustic emissions data to previously recorded acoustic emissions data, e.g. an acoustic profile or acoustic characteristic. Such a comparison allows an operator to trend the acoustic emissions information over an extended period thereby enabling an operator to track changes in the acoustic emissions.

In an embodiment the steps of processing the electrical signals by comparing the electrical signals 202 with pre-determined tapping operation conditions 203, the generating of indication data 210, and the providing of an indication are performed repetitively as electrical signals 202 are received, thereby providing the indication in real-time, in response to the electrical signals 202.

DESCRIPTION OF THE FIGURES

The invention will also be explained by means of the following, non-limiting figures of embodiments according to the invention. In the claims, the drawings and in the description, like reference numerals are used to indicate like elements, functions or features as between the claims, the drawings and the described embodiments. Fig. 1 shows a block diagram of an acoustic monitoring system used in the method for controlling the tapping operation in a blast furnace, more in particular for controlling the end-of- cast of a tapping operation.

Fig. 2 shows a block diagram illustration of an embodiment of monitoring station 150 as shown in Fig. 1.

Fig. 3 shows an example of the amplitude of the electrical signal versus time of during a tapping operation after low-pass filtering to remove most of any hissing-type noise on a shopfloor near a taphole or tapblock of a blast furnace.

Fig. 4 shows as comparison of neural network versus logistic regression in the last few minutes of a tapping operation at a blast furnace at Tata Steel in IJmuiden, NL.

Fig. 5 shows an example of the steps in calculating the output count for a neural net approach.

Referring to Fig. 1, there is shown a block diagram of a real-time acoustic monitoring system 100 for monitoring and controlling a tapping operation, more in particular the end-of- cast of a tapping operation, used in relation to a blast furnace operation where molten iron and slag are drained through taphole 120. The acoustic monitoring system 100 used in the method and in the system according to this invention are each configured to detect acoustic emissions from a variety of sources during a tapping operation from taphole 120 of a blast furnace 110. The acoustic receiver 140 is located remotely, e.g. in a range of 3 to 15 meters, from the taphole 120 and serves as a transducer to convert the acoustic signals into corresponding electrical signals 202 (not shown) that can be processed by the monitoring station 150 comprising a controller 410 having a processor 412, as shown in Fig. 2, which can execute code stored on a tangible computer-readable medium in a memory 418, as shown in Fig. 2, or elsewhere such as portable media, on a server or in the cloud among other media, to cause the controller 410 to receive and process data and to perform actions and/or control components of equipment such as an indicator 160, a status display 170 or a network 180 in turn connected to user station(s) 190. The monitoring station 150 causing the processor 412 to perform processor operations including processing the electrical signals 201 to determine if an acoustic event in relation to the tapping operation has occurred by comparing the electrical signals with pre-determined tapping operation conditions 203 (not shown), generating indication data, and depending on the comparing providing an indication based on the indication data, and next initiating an action in relation to the tapping operation using the indication data, in particular shutting off the tapping operation. And Fig. 2 shows a block diagram illustration of an embodiment of monitoring station 150 shown in Fig. 1. The acoustic signals 201 are received by at least one acoustic receiver 140 converting the acoustic signals 201 received by the least one acoustic receiver 140 into electrical signals 202 (not shown), the electrical signals corresponding to acoustic signals transmitted by the tapping operation and sensed by the acoustic receiver 140. Conversion of the acoustic signals 201 into electrical signals 202 can be done for example using an analogue-to- digital (A/D) converter. The A/D converter is configured to receive analogue acoustic emission signals and convert them into corresponding electrical signals 201 that are communicated to the processor 410. The A/D converter can be any commercially available A/D converter known to those skilled in the art. Also, the A/D converter may be single-channel for processing the acoustic signals from a single acoustic receiver 140, or, the A/D converter may be multi-channel for processing the acoustic signals 201 from a plurality of acoustic receivers 140 located remotely from the taphole 120. The A/D converter can be received within the monitoring station 150, or may be integral to the acoustic receiver 140 as shown in Fig. 2, or it may be a self- contained device located remotely from, but communicably linked with, the monitoring station 150. The monitoring station 150 comprising a controller 410 having a processor 412 which can execute code or instructions stored on a tangible computer-readable medium in a memory 418 to cause the controller 410 to process the electrical signals to determine if an acoustic event in relation to the tapping operation has occurred by comparing the electrical signals with predetermined tapping operation conditions, generating indication data and depending on the comparing, and next providing an indication on control components of equipment such as an indicator 160, a status display 170 or a network 180. The processor 412 may initiate an action in relation to the tapping operation using the indication data, in particular shutting off the tapping operation.

Fig. 4 shows as comparison of the pop count obtained by a neural network versus logistic regression as set out above in the last few minutes of a tapping operation at a blast furnace at Tata Steel in IJmuiden, NL. In this example at about 02:30 there is a significant increase in the number of pops and providing a characteristic acoustic profile. Having this characteristic acoustic profile the formulation of pre-determined tapping operation conditions 203 (not shown) for use in the method and system according to this invention for comparing against real-time acoustic signals 201 converted into electrical signals 202 in a processor 412 can be used. Based on the comparing indication data 210 are generated and an indication is based on the indication data 210. This allows for an automated and reliable detection of an approaching end- of-cast event such that it will be feasible to consistently and timely close the taphole from a blast furnace and achieving the various advantages set out in this description.

And Fig. 5 shows an embodiment of the basic steps in calculating the output count for a neural net approach. In this embodiment the count number is a 10 second interval, but it will be apparent to the skilled person that other suitable time intervals can be used. The “output” count can be compared against a pre-determined tapping operation condition, the comparing creating an indication data based on which an action can be initiated such as shutting off the tapping operation in a blast furnace.

The invention will now be illustrated also with reference to non-limiting embodiments according to the invention.

EXAMPLE 1 .

In Fig. 3 an example is shown of the amplitude of the electronic signal versus time during a regular and complete tapping operation of about 160 minutes from a taphole of a blast furnace at Tata Steel in IJmuiden, NL. A Dodotronic UltraMic 384k microphone had been positioned at about 10 meters away from the taphole to prevent damage from the heat radiation emitted from the molten iron. The microphone was able to detect sound signals in a frequency range of 100 Hz to 2 kHz. A sampling rate of 192 kSPS was used. The sound data have been low-pass filtered to remove most of the background noise. It has been found that most energy bursts of the popping-type sounds occurred below 1 kHz.

The start of the tapping operation or cast is clearly marked by the drilling of the taphole. During the cast several noises are audible and recorded, such as intercom speech and movement of tools on the shopfloor. At this particular blast furnace a spray screen is activated at the closing of the taphole to prevent dust from escaping. The spray and mist produce a loud hissing-type sound clearly marking the end-of-cast. Throughout the tapping operation there are pops detected which are mostly false positives. However, when approaching the end-of-cast, e.g. about 2 to 4 minutes before a regular end, a characteristic acoustic profile is emitted which can be received and processed. Having this characteristic acoustic profile from multiple tapping operation from a furnace allows, e.g. using an algorithm via machine learning, the formulation of pre-determined tapping operation conditions 203 for use in the method and system according to this invention for comparing against real-time acoustic signals 201 converted into electrical signals 202 in a processor 412. On the basis of the comparison indication data 210 are generated and an indication is based on these indication data 210. This allows an automated and reliable detection of an approaching end-of-cast event in a tapping operation such that it will be feasible to consistently and timely close the taphole 120 of a blast furnace 110 and achieving the various advantages set out in this description.

EXAMPLE 2.

In an embodiment of the method and the system of the invention logistic regression is applied to separate characteristic acoustic profile from the background noise. The algorithm aims to count the number of pop events that happen in a given amount of time. The clearly audible pops create a short peak which is above the noise level. The first step is to select the peaks in the signal. To relieve the computational burden the 20 highest peaks are processed. To reduce the number of coefficients necessary to classify the algorithm calculates the Mel- Frequency Cepstral Coefficients (MFCC). To do this a Fourier spectrum is calculated. From the results the logarithm of the power spectrum is computed, and the bins of the power spectrum are redistributed according to the Mel-Scale. Finally a Discrete Cosine Transform is used to calculate the coefficients, which are the amplitudes of the resulting spectrum. To further reduce the required amount of coefficients Principle Components Analysis can be used. By this technique the basis is determined on which the MFC coefficient vectors can be projected. The samples collected around each peak in the signal can now be classified using the logistic regression coefficients. A positive identification of a pop sound is found when the output of the logistic regression is a “1” and the event is counted. A value of 5 counts per time frame, e.g. 5 counts per 10 seconds, can be used as a pre-determined tapping operation condition 203 against which a comparison is to be made. If the count within a time frame is for example higher than five the cast is found to convey gas bubbles indicating the approaching end of the cast. At this time the casting or tapping operation needs to be stopped either by action of a human operator following an audio or a visual indication or an action may be taken automatically.

A suitable algorithm has the following set up:

1. Divide signal in 10 second timeframes.

2. Calculate standard deviation and select the 20 highest peaks above the threshold of 4 times the standard deviation.

3. For each sample 40 Mel-Frequency Cepstral Coefficients (MFCC), often used in audible speech recognition, are calculated.

4. Project the 40 sample long vector on a 5 vector long vector on the principle component basis using the projection matrix published in the next section. 5. Use the coefficients [-0.0914078, 0.02138893, 0.060807, -0.01008841 , -0.02824872] and intercept at 0.25078529 for a logistic regression projection. If the output is “0” the sample is classified as being ‘background noise’, if the output is “1” the sample is classified as a ‘pop’.

6. Count the amount of samples that are classified as being a ‘pop’ within the time frame.

7. The resulting signal can be thresholded at a count of 5 to signal that the cast or tapping operation is about to finish .

The projection matrix to project the 40 sample long MFCC vector on the principle component basis: [0.80210237, -0.52115703, -0.24291631 , 0.01995926, 0.00371678, 0.01965893, 0.01706565, -0.00171448, 0.03226421 , 0.03612797, 0.05389932, 0.04069661 , 0.02362917, 0.0123075, -0.02098245, -0.02052759, 0.03457675, 0.06383888, 0.07627412, 0.04022767, 0.01565359, -0.01184862, -0.01060456, -0.00088371 , -0.00874608, -0.00902498, -0.01763532, 0.00627398, 0.00531473, -0.00679931 , 0.01882923, 0.02477598, 0.01949214, 0.0168795, -0.01310087, -0.01700504, -0.00899117, 0.0152285, 0.0036556, -0.0154393 ]

[-4.07555479e-01, -3.85597254e-01, -5.49116698e-01 , -4.49668172e-01 , -3.55860022e-01, -1 .10105606e-01 , 6.33272888e-03, 5.83319013e-02, 1.27535794e-01, 2.60840498e-02, -5.13929462e-02, -5.84585243e-02, -4.96893987e-02, 8.33153828e-03, 2.19873244e-03, -2.81853767e-02, 1.28942922e-02, 3.23586169e-02, 4.07478074e-02, - 1.15098939e-04, -3.07906787e-02, -3.52971829e-02, - 1 .10992023e-02, -5.23862415e-03,

-1.78274480e-02, -1.28437644e-02, 3.63439639e-02, 2.78496507e-02, -1.94257527e-02, -2.38081790e-02, -2.87862096e-03, 5.08055596e-03, -2.43322420e-02, -5.82008008e-03, -9.77713420e-03, 1 .19749080e-02, 3.32397493e-02, 2.80876996e-02, -5.83254745e-03, -3.70086980e-02]

[-0.31315522, -0.71371375, 0.43066771 , 0.13053489, 0.23366434, -0.1548165, 0.04948481 , -0.0913591 , -0.16222547, -0.05560699, 0.07426838, -0.02094337, -0.0293647, -0.00253364, 0.04884371 , -0.02582137, 0.02230385, -0.0399499, -0.11012333, -0.08852921 , 0.12724104, 0.02948631 , -0.10271297, 0.00212985, 0.05302279, 0.02832348, -0.03260013, -0.01162168, 0.01069588, 0.01955243, -0.00773127, 0.04253882, -0.02178194, -0.01123247, 0.01953555, 0.01875349, -0.01855403, -0.05299779, -0.00672264, 0.00182723]

[-0.14457695, -0.02625074, -0.41290966, 0.74748692, -0.15978448, -0.05378281 , -0.16173336,

-0.1639823, -0.21483835, -0.01802007, -0.09477204, 0.09127816, -0.16227089, 0.03877404,

-0.07482942, -0.07099707, -0.1006323, 0.06933576, -0.04212719, -0.02136363, -0.01899352, -0.04015522, -0.12650468, -0.03907935, -0.02535965, -0.02201294, -0.08113235, -0.00863723,

-0.06764013, -0.07159995, -0.0332903, -0.01546401 , -0.05438275, -0.01384316, -0.02137852,

-0.04306937, -0.02909614, 0.02145389, -0.00468139, -0.0560697 ]

[-0.19904325, -0.04532474, -0.23417324, 0.21746819, 0.30127149, 0.41711145, 0.35192794, 0.28415594, 0.2496828, 0.07739281 , 0.147812, 0.20780272, 0.03562743, 0.147372, 0.07054035, 0.10103601 , 0.13762848, 0.12598257, 0.09454977, 0.11107981 , 0.18955043, 0.11107494, -0.01795131 , 0.09240546, 0.10326522, 0.06899996, 0.01020826, 0.07362531 , 0.07450185, 0.05692723, 0.07816742, 0.17741358, 0.11428467, 0.10604641 , 0.05473969, 0.02843151 , 0.03179489, -0.00228008, 0.07506158, 0.00345292]

To setup the algorithm the recording of several tapping periods can be examined:

1. Divide signal in 10 second timeframes, or any other reasonable timeframe.

2. In each segment calculate the standard deviation of the signal.

3. Select the up to 20 highest peaks.

4. Select a sample of 1200 samples around the peak, starting 200 samples before the peak

5. If an ‘pop’-type sound is clearly heard in the recording label the sample as a pop, otherwise label the sample as ‘noise’.

6. Calculate from each sample 40 Mel-Frequency cepstral coefficients.

7. Use Principle Component Analysis to find the 5 most import components.

8. Use logistic regression to obtain 5 regression coefficients.

While good results are achieved using the numbers presented above, it will be evident for the skilled person that further improved results can be obtained using more MFC Coefficients and/or using more principle components.

Claims

1. A method of controlling a tapping operation, in particular an end-of-cast of a tapping operation, of molten iron from at least one taphole (120) of a blast furnace (110), the method comprising: initiating a tapping operation by tapping liquids from at least one taphole (120) in the blast furnace (110); receiving acoustic signals (201) from the tapping operation in at least one acoustic receiver (140) located remotely from the taphole (120); converting the acoustic signals (201) received by the least one acoustic receiver (140) into electrical signals (202), the electrical signals corresponding to acoustic signals transmitted by the tapping operation and sensed by the at least one acoustic receiver; processing the electrical signals (202) to determine if an acoustic event in relation to the tapping operation has occurred by comparing the electrical signals with predetermined tapping operation conditions (203); generating indication data (210), depending on the comparing; providing an indication based on the indication data (210); and initiating an action in relation to the tapping operation using the indication data (210).

2. The method according to claim 1 , wherein the initiating of an action in relation to the tapping operation using the indication data (210) includes shutting off the tapping operation.

3. The method according to claim 1 or 2, wherein the indication comprises at least one of an audio indication and a visual indication.

4. The method according to any one of claims 1 to 3, wherein the acoustic receiver (140) is a microphone, preferably an omnidirectional microphone.

5. The method according to any one of claims 1 to 4, wherein the received acoustic signals (201) are in a frequency range of 15 kHz to 60 kHz.

6. The method according to any one of claims 1 to 5, wherein the received acoustic signals (201) are in a frequency range of 100 Hz to 2 kHz.

7. The method according to any one of claims 1 to 4, wherein the received acoustic signals

(201) are in a frequency range of 100 Hz to 2 kHz and 15 kHz to 60 kHz.

8. The method according to any one of claims 1 to 7, wherein the processing is performed in a monitoring station (150) comprising a processor (412) configured to execute instructions stored on a computer-readable medium in a memory (418), the monitoring station (150) causing the processor (412) to perform processor operations including: receiving the acoustic signals (201) associated with the tapping operation; converting the acoustic signals (201) received by the acoustic receiver (140) into electrical signals (202), the electrical signals corresponding to acoustic signals transmitted by the tapping operation and sensed by the acoustic receiver; processing the electrical signals (202) to determine if an acoustic event in relation to the tapping operation has occurred by comparing the electrical signals (202) with pre-determined tapping operation conditions (203); generating indication data (210), depending on the comparing; providing an indication based on the indication data (210); and initiating an action in relation to the tapping operation using the indication data (210), preferably shutting off the tapping operation.

9. The method according to any one of claims 1 to 8, wherein the steps of processing the electrical signals (202) by comparing the electrical signals (202) with pre-determined tapping operation conditions (203), the generating of the indication data (210) and the providing of an indication are performed repetitively as electrical signals (202) are received, thereby providing the indication in real-time, in response to the electrical signals

(202).

10. An acoustic monitoring system (100) for controlling the tapping operation from a taphole, in particular the end-of-cast of a tapping operation, in a blast furnace (110), the system comprising: at least one acoustic receiver (140) located remotely from a taphole (120) of the blast furnace (110) and configured to sense acoustic signals (201) from a tapping operation of the blast furnace (110); a monitoring station (150) comprising a processor (412) configured to execute instructions stored on a computer-readable medium in a memory (418), the monitoring station (150) causing the processor (412) to perform processor operations including: receiving the acoustic signals (201) associated with the tapping operation; converting the acoustic signals (201) received by the acoustic receiver (140) into electrical signals (202), the electrical signals (202) corresponding to acoustic signals (201) transmitted by the tapping operation and sensed by the acoustic receiver (140); processing the electrical signals (202) to determine if an acoustic event in relation to the tapping operation has occurred by comparing the electrical signals (202) with pre-determined tapping operation conditions; generating indication data (210), depending on the comparing; and providing an indication based on the indication data (210).

11. The system according to claim 10, wherein the system (100) further comprising the processor (412) initiating an action in relation to the tapping operation using the indication data (210).

12. The system according to claim 11 , wherein the system (100) further comprising the processor (412) initiating an action in relation to the tapping operation using the indication data (210) includes shutting off the tapping operation.

13. The system according to any one of claims 10 to 12, wherein the system (100) further comprising the processor (412) to display the output data comprising indication data (210) on an indicator (160) displaying visual information and/or audio information.

14. The system according to any one of claims 10 to 13, wherein at least one acoustic receiver (140) is configured to receive acoustic signals in a frequency range of 15 kHz to 60 kHz and/or of 100 Hz to 2 kHz.