A method and system for fully automatic production of LF refining

Abstract

The invention discloses a method and system for fully automatic production of LF refining, comprising the following modules: an information acquisition module, an argon blowing and stirring module, a smelting state detection module, a desulfurization calculation module, a deoxidation calculation module, a slag making calculation module, a circulation calculation module, a temperature calculation control module, a cleanliness calculation module, an alloying calculation module, a cost analysis module, and a self-learning module. The method and system for metallurgical LF refining control proposed by the invention realize the fully automatic and intelligent production of LF, further optimize the LF production process, thereby reducing costs, improving molten steel quality, and saving manpower, having greater economic and social benefits, forming a set of LF fully automatic intelligent control systems with universal applicability, and having greater practical and popularization significance.

Description

Technical Field

[0001] The invention belongs to the technical field of automatic control of refining outside a furnace, and in particular relates to a method and a system for fully automatic production of LF refining. Background Art

[0002] In current steel enterprises, the LF process can be used to add deoxidizers, slag-making agents, argon blowing and stirring, electrode heating, alloy addition, wire feeding and other operations. LF is responsible for desulfurization, deoxidation, temperature adjustment, composition control, cleanliness control and buffer rhythm, and is an important part of the steel production chain. LF is flexible in operation, the process operation is changeable, the processing time is not fixed, and the smelting state relies on manual judgment, which makes it impossible for LF to achieve fully automatic one-button production, the mode is not fixed, and the smelting process cost is not optimized. At present, there is no universally adaptable LF fully automatic production control system in the world.

[0003] With the continuous maturity of technologies such as artificial intelligence and big data, the intelligentization of the entire steel production line is an inevitable trend. Under this technical background, the fully automated intelligent production of LF is realized, the LF production process is further optimized, and the cost is reduced, the quality of molten steel is improved, and manpower is saved. It has great economic and social benefits, and a set of LF fully automated intelligent control systems with universal applicability is formed, which has great practical and promotion significance. For this reason, the present invention proposes a method and system for fully automatic production of LF refining. Summary of the invention

[0004] The object of the present invention is to provide a method and system for fully automatic production of LF refining to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: a fully automatic production system for LF refining, comprising the following modules: an information acquisition module, an argon blowing and stirring module, a smelting state detection module, a desulfurization calculation module, a deoxidation calculation module, a slag making calculation module, a circulation calculation module, a temperature calculation control module, a cleanliness calculation module, an alloying calculation module, a cost analysis module, and a self-learning module;

[0006] Information acquisition module: cyclically acquires production plan information, target exit time, target molten steel composition and temperature, target molten steel cleanliness, auxiliary material alloy information and equipment information;

[0007] Argon blowing and stirring module: calculate and control the appropriate argon stirring at each stage during the LF smelting process;

[0008] Smelting state visual inspection module: through deep learning visual algorithm, obtain slag state, slag thickness, molten steel stirring state, foaming state, obtain the amount of slag return and its composition and ladle state information;

[0009] Desulfurization calculation module: calculates the slag requirements, basicity, and necessary desulfurization time required for desulfurization;

[0010] Deoxidation calculation module: calculates the real-time oxygen content in molten steel and slag, and calculates the required amount of deoxidizer;

[0011] Slag calculation module: calculates the basicity, oxidizability, viscosity, fluidity, interfacial tension, activity, foaming ability, slag volume, and inclusion capacity of the slag, and calculates the feed combination of the slag according to the existing raw material conditions to achieve precise control of the slag;

[0012] Cyclic calculation module: triggered by events, cyclically calculates and displays the molten steel status information and slag status information within the set period;

[0013] Temperature calculation control module: Based on the principle of energy conservation, it is triggered by cyclic calculation and time, calculates the energy change within the cycle, and calculates and displays the calculated temperature of the molten steel and the heating operation required to reach the target temperature in real time;

[0014] Cleanliness calculation module: Calculates the impact of various factors in the molten steel smelting process on inclusions, and calculates the number, form and type of inclusions in the molten steel based on thermodynamics and kinetics principles and big data algorithms;

[0015] Alloying calculation module: with the goal of cost optimization, calculate the amount of various alloys required to achieve the target molten steel composition;

[0016] Self-learning module: conduct rolling self-learning on the furnaces with better smelting status, correct the self-learning parameters, and improve the performance of each model;

[0017] Cost analysis module: calculates the cost composition of each stage of smelting, alloy recovery rate and electric heating efficiency, and conducts cost optimization analysis on smelting operations, and seeks the optimal cost operation through heuristic algorithms.

[0018] Preferably, the system establishes a LF refining model based on metallurgical principles, deep learning algorithms and big data algorithms, including a desulfurization model, a deoxidation model, a temperature model, an alloying model, a molten steel and slag calculation model, an inclusion model and a cost analysis optimization model, so as to improve the calculation accuracy and applicability of the model on the basis of the calculation conforming to the metallurgical process.

[0019] Preferably, a deep learning visual algorithm is used to establish a smelting state visual detection module to obtain slag state, slag thickness, molten steel stirring state, foaming state, slag return amount and ladle state information; and automatic and accurate acquisition of smelting process state information is achieved.

[0020] A method for fully automatic production of LF refining, comprising the following steps:

[0021] Step S1, obtaining molten steel information. If there is temperature measurement sampling information and slag testing information, the latest temperature measurement sampling information is used as the state information of the molten steel in this furnace. If there is no temperature measurement sampling information, the last sampling information and process event information are used as the state information of the molten steel through model calculation.

[0022] Step S2, obtaining the target exit information of molten steel, obtaining the target composition requirements, temperature requirements and cleanliness requirements of the molten steel, and combining the time factor, calculating the gap between the current molten steel indicators and the exit indicators, and calculating the desulfurization, deoxidation, alloying, heating, inclusion denaturation and floating time;

[0023] Step S3, determining whether the phosphorus element exceeds the allowable upper limit, and if so, issuing a phosphorus component alarm;

[0024] Step S4, determine whether the ladle is in the entry position, determine whether the ladle is successfully uncapped, calculate the amount of slag returned from the ladle according to the machine vision algorithm, and transmit it to the system. According to the ladle number and furnace tracking, the composition of the slag returned is transmitted to the system;

[0025] Step S5, updating the material information and equipment information of the LF smelting process, updating the silo availability, material composition, temperature drop coefficient, electrode transformer data and ladle information;

[0026] Step S6, determining whether deoxidation is required, performing deoxidation calculation, calculating the oxygen level of molten steel and slag, calculating the amount of deoxidizer added, the amount of deoxidation products and the real-time inclusion content, and calculating the number of inclusions in the steel and the floating time;

[0027] Step S7, determining whether desulfurization is required. If desulfurization is required, the slag basicity and inclusion adsorption capacity characteristics are set according to the initial sulfur and target sulfur, and the required slag making material is calculated;

[0028] Step S8, determine whether alloy elements need to be added, calculate the change of each element during the processing time, and correct the change into the element control calculation; with the goal of cost optimization, trigger specific elements according to the smelting process and calculate the amount of alloy added;

[0029] Step S9, determine whether heating is required, perform heating calculation, calculate the temperature loss caused by various factors within the specified exit time, calculate the heating time required for the temperature to reach the target in combination with the electrode efficiency, and allocate the heating time;

[0030] Step S10, calculate the necessary time for heating, the necessary time for inclusion treatment and floating, the necessary time for alloying, the necessary time for deoxidation and desulfurization, remove the overlapping time, calculate the necessary time for standard treatment, if the necessary time is greater than the required time for exit, then issue a processing time alarm, switch to priority setting, process according to the logic of composition, temperature and cleanliness priority, and generate the relationship axis of pre-calculated smelting treatment steps, molten steel and slag state calculation and smelting time;

[0031] Step S11, drive the ladle car to the processing position, lower the furnace cover, and start the processing process;

[0032] Step S12, the smelting state visual inspection module monitors the slag state, slag thickness, slag state, and stirring state, and transmits them to the system for feedback stirring control;

[0033] Step S13, measuring temperature and taking samples, updating the latest information on the composition of molten steel and slag, and repeating S1-S10 until the molten steel meets the requirements;

[0034] Step S14, after the temperature and composition meet the requirements, soft blowing is performed to wait for the calcium treatment opportunity;

[0035] Step S15, after the scheduled time is reached, the furnace cover is raised, the furnace exits the station, detailed furnace cost information is generated, and cost analysis and self-learning parameter optimization are performed.

[0036] Preferably, a cost analysis optimization model is established based on big data algorithms and heuristic algorithms, and the optimal cost combination operation plan is globally searched through the analysis of smelting micro-operation costs and the overall cost analysis of furnaces, combined with the simulation of smelting procedures.

[0037] Compared with the prior art, the beneficial effects of the present invention are: the method and system for metallurgical LF refining control proposed in the present invention realize the fully automated intelligent production of LF, further optimize the LF production process, thereby reducing costs, improving molten steel quality, and saving manpower, with greater economic and social benefits, forming a set of LF fully automated intelligent control systems with universal applicability, which has great practical and promotional significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of the system structure of the present invention;

[0039] Figure 2 It is the work flow chart of the present invention. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] See also Figure 1 to Figure 2 , the present invention provides a technical solution: a LF refining fully automatic production system, including the following modules: information acquisition module, argon blowing and stirring module, smelting state detection module, desulfurization calculation module, deoxidation calculation module, slag calculation module, circulation calculation module, temperature calculation control module, cleanliness calculation module, alloying calculation module, cost analysis module, self-learning module;

[0042] Information acquisition module: cyclically acquires production plan information, target exit time, target molten steel composition and temperature, target molten steel cleanliness, auxiliary material alloy information and equipment information;

[0043] Argon blowing and stirring module: calculate and control the appropriate argon stirring at each stage during the LF smelting process;

[0044] Smelting state visual inspection module: through deep learning visual algorithm, obtain slag state, slag thickness, molten steel stirring state, foaming state, obtain the amount of slag return and its composition and ladle state information;

[0045] Desulfurization calculation module: calculates the slag requirements, basicity, and necessary desulfurization time required for desulfurization;

[0046] Deoxidation calculation module: calculates the real-time oxygen content in molten steel and slag, and calculates the required amount of deoxidizer;

[0047] Slag calculation module: calculates the basicity, oxidizability, viscosity, fluidity, interfacial tension, activity, foaming ability, slag volume, and inclusion capacity of the slag, and calculates the feed combination of the slag according to the existing raw material conditions to achieve precise control of the slag;

[0048] Cyclic calculation module: triggered by events, cyclically calculates and displays the molten steel status information and slag status information within the set period;

[0049] Temperature calculation control module: Based on the principle of energy conservation, it is triggered by cyclic calculation and time, calculates the energy change within the cycle, and calculates and displays the calculated temperature of the molten steel and the heating operation required to reach the target temperature in real time;

[0050] Cleanliness calculation module: Calculates the impact of various factors in the molten steel smelting process on inclusions, and calculates the number, form and type of inclusions in the molten steel based on thermodynamics and kinetics principles and big data algorithms;

[0051] Alloying calculation module: with the goal of cost optimization, calculate the amount of various alloys required to achieve the target molten steel composition;

[0052] Self-learning module: conduct rolling self-learning on the furnaces with better smelting status, correct the self-learning parameters, and improve the performance of each model;

[0053] Cost analysis module: calculates the cost composition of each stage of smelting, alloy recovery rate and electric heating efficiency, and conducts cost optimization analysis on smelting operations, and seeks the optimal cost operation through heuristic algorithms.

[0054] In this embodiment, preferably, the system establishes a LF refining model based on metallurgical principles, deep learning algorithms and big data algorithms, including a desulfurization model, a deoxidation model, a temperature model, an alloying model, a molten steel and slag calculation model, an inclusion model and a cost analysis optimization model, so as to improve the calculation accuracy and applicability of the model on the basis of the calculation conforming to the metallurgical process.

[0055] In this embodiment, preferably, a deep learning visual algorithm is used to establish a smelting state visual detection module to obtain slag state, slag thickness, molten steel stirring state, foaming state, slag return amount and ladle state information; and automatic and accurate acquisition of smelting process state information is achieved.

[0056] A method for fully automatic production of LF refining, comprising the following steps:

[0057] Step S1, hardware and environment settings: install two high-definition cameras on the LF platform, one is aimed at the observation hole of the LF furnace, and the viewing angle and imaging range are the same as the workers' viewing angle; the other is aimed at the steel car exit position to monitor and calculate the slag return amount, ladle number and ladle status; configure the server system operating environment, improve the automatic execution logic of the first-level program, and match and connect the system with the LF first-level program.

[0058] Step S2, based on the convolutional neural network and image monitoring algorithm, establish a smelting state visual detection module to obtain the slag state, slag thickness, molten steel stirring state, foaming state, slag return amount and ladle state information; realize automatic and accurate acquisition of smelting process state information.

[0059] Step S3, establishing a LF refining model based on metallurgical principles, deep learning algorithms and big data algorithms, including a desulfurization model, a deoxidation model, a temperature model, an alloying model, a molten steel and slag calculation model, an inclusion model and a cost analysis optimization model.

[0060] Step S4, based on the big data algorithm and heuristic algorithm, a cost analysis optimization model is established. Through the analysis of the smelting micro-operation cost and the overall cost analysis of the furnace, combined with the simulation of the smelting procedure, the model can globally search for the optimal cost combination operation plan.

[0061] Step S5, establish the relationship axis between the pre-calculated smelting treatment steps, the calculation of the state of molten steel and slag, and the smelting time, calculate the necessary time for heating, the necessary time for inclusion treatment and floating, the necessary time for alloying, the necessary time for deoxidation and desulfurization, remove the overlapping time, and calculate the necessary time for standard processing. If the necessary time is greater than the required time for exit, a processing time alarm is issued, and the processing is switched to priority setting and processed according to the priority logic of composition, temperature, and cleanliness.

[0062] Step S6, obtain the molten steel status information. If there is temperature measurement sampling information and slag testing information, the latest temperature measurement sampling information is used as the status information of the molten steel in this furnace. If not, the last sampling information and process event information are used as the status information of the molten steel through model calculation.

[0063] Step S7, obtain the target exit information of molten steel, obtain the target composition requirements, temperature requirements and cleanliness requirements of the molten steel, and combine the time factor to calculate the gap between the current molten steel indicators and the exit indicators, and calculate the time for desulfurization, deoxidation, alloying, heating, inclusion denaturation and floating.

[0064] Step S8, determining whether the phosphorus element exceeds the allowable upper limit, and if so, issuing a phosphorus component alarm;

[0065] Step S9, determine whether the ladle is in the entry position, determine whether the ladle cover is successfully removed, calculate the amount of slag returned from the ladle according to the machine vision algorithm, and transmit it to the system. According to the ladle number and furnace tracking, the composition of the slag returned is transmitted to the system.

[0066] Step S10, updating the material information and equipment information of the LF smelting process, updating the silo availability, material composition, temperature drop coefficient, electrode transformer data and ladle information, etc.

[0067] Step S11, determine whether deoxidation is needed, perform deoxidation calculation, calculate the oxygen level of molten steel and slag, calculate the amount of deoxidizer added, the amount of deoxidation product and the real-time inclusion content, calculate the number of inclusions in the steel and the floating time, and perform related control operations.

[0068] Step S12, determining whether desulfurization is required, if desulfurization is required, then setting the slag basicity and inclusion adsorption capacity and other characteristics according to the initial sulfur and target sulfur, calculating the required slag making material, and performing relevant control operations.

[0069] Step S13, determine whether alloy elements need to be added, calculate the change of each element during the processing time, and correct the change into the element control calculation. With the goal of cost optimization, trigger specific elements according to the smelting process, calculate the amount of alloy added, and perform related control operations.

[0070] Step S14, determine whether heating is needed, perform heating calculation, calculate the temperature loss caused by various factors within the specified exit time, calculate the heating time required for the temperature to reach the target in combination with the electrode efficiency, allocate the heating time, and perform related control operations.

[0071] Step S15, calculate the necessary time for heating, the necessary time for inclusion treatment and floating, the necessary time for alloying, the necessary time for deoxidation and desulfurization, remove the overlapping time, and calculate the necessary time for standard processing. If the necessary time is greater than the required time for exit, a processing time alarm is issued, and the process is switched to priority setting, and processing is performed according to the composition, temperature, and cleanliness priority logic to generate the pre-calculated smelting processing steps, molten steel and slag state calculation and smelting time relationship axis.

[0072] Step S16, drive the ladle car to the processing position, lower the furnace cover, and start the processing process.

[0073] Step S17, the smelting state visual inspection module monitors the slag state, slag thickness, slag state, and stirring state, and transmits them to the system for feedback stirring control.

[0074] Step S18, measure temperature and take samples, update the latest information on the composition of molten steel and slag, and repeat S16-S15 until the molten steel meets the requirements.

[0075] Step S19: After the temperature and composition meet the requirements, soft blowing is performed to wait for the calcium treatment opportunity.

[0076] Step S20, after the scheduled time is reached, the furnace cover is raised, the furnace exits the station, detailed furnace cost information is generated, and cost analysis and self-learning parameter optimization are performed.

[0077] In this embodiment, preferably, a cost analysis optimization model is established based on big data algorithms and heuristic algorithms, and the optimal cost combination operation plan is globally searched through the analysis of smelting micro-operation costs and the overall cost analysis of furnaces, combined with the simulation of smelting procedures.

[0078] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for fully automatic production of LF refining, which adopts a fully automatic production system for LF refining, the system comprising the following modules: information acquisition module, argon blowing and stirring module, smelting state visual detection module, desulfurization calculation module, deoxidation calculation module, slag making calculation module, cycle calculation module, temperature calculation control module, cleanliness calculation module, alloying calculation module, cost analysis module, self-learning module; information acquisition module: cyclically acquires production plan information, target exit time, target molten steel composition and temperature, target molten steel cleanliness, auxiliary material alloy information and equipment information; argon blowing and stirring module: calculates and controls suitable argon stirring at each stage in the LF smelting process; smelting state visual detection module: acquires slag state, slag thickness, molten steel stirring state, foaming state, slag return amount and its composition and ladle state information through deep learning visual algorithm; desulfurization calculation module: calculates slag requirements, basicity and necessary desulfurization time required for desulfurization; Deoxidation calculation module: calculates the real-time oxygen content in molten steel and slag, and calculates the required amount of deoxidizer; slag calculation module: calculates the basicity, oxidizability, viscosity, fluidity, interfacial tension, activity, foaming ability, slag volume, and inclusion capacity of the slag, and calculates the feeding combination of slag according to the existing raw material conditions to achieve precise control of slag; cycle calculation module: triggered by events, cyclically calculates and displays the molten steel status information and slag status information within the set period; Temperature calculation control module: Based on the principle of energy conservation, it is triggered by cyclic calculation and time, calculates the energy change within the cycle, and calculates and displays the calculated temperature of the molten steel and the heating operation required to reach the target temperature in real time; Cleanliness calculation module: Calculates the impact of various factors in the molten steel smelting process on inclusions, and calculates the number, form and type of inclusions in the molten steel based on the principles of thermodynamics and kinetics and big data algorithms; Alloying calculation module: takes cost optimization as the goal, and calculates the amount of various alloys required to achieve the target molten steel composition; self-learning module: performs rolling self-learning on furnaces with good smelting conditions, corrects self-learning parameters, and improves the performance of each model; cost analysis module: calculates the cost composition of each stage of smelting, alloy recovery rate and electric heating efficiency, and performs cost optimization analysis on smelting operations, and seeks the optimal cost operation through heuristic algorithms; The system establishes a LF refining model based on metallurgical principles, deep learning algorithms and big data algorithms, including a desulfurization model, a deoxidation model, a temperature model, an alloying model, a molten steel and slag calculation model, an inclusion model and a cost analysis optimization model, and improves the calculation accuracy and applicability of the model on the basis of the calculation conforming to the metallurgical process; the method is characterized in that: Step S1, obtaining molten steel information. If there is temperature measurement sampling information and slag testing information, the latest temperature measurement sampling information is used as the state information of the molten steel in this furnace. If there is no temperature measurement sampling information, the last sampling information and process event information are used as the state information of the molten steel through model calculation. Step S2, obtaining the target exit information of molten steel, obtaining the target composition requirements, temperature requirements and cleanliness requirements of the molten steel, and combining the time factor, calculating the gap between the current molten steel indicators and the exit indicators, and calculating the desulfurization, deoxidation, alloying, heating, inclusion denaturation and floating time; Step S3, determining whether the phosphorus element exceeds the allowable upper limit, and if so, issuing a phosphorus component alarm; Step S4, determine whether the ladle is in the entry position, determine whether the ladle is successfully uncapped, calculate the amount of slag returned from the ladle according to the machine vision algorithm, and transmit it to the system. According to the ladle number and furnace tracking, the composition of the slag returned is transmitted to the system; Step S5, updating the material information and equipment information of the LF smelting process, updating the silo availability, material composition, temperature drop coefficient, electrode transformer data and ladle information; Step S6, determining whether deoxidation is required, performing deoxidation calculation, calculating the oxygen level of molten steel and slag, calculating the amount of deoxidizer added, the amount of deoxidation products and the real-time inclusion content, and calculating the number of inclusions in the steel and the floating time; Step S7, determining whether desulfurization is required. If desulfurization is required, the slag basicity and inclusion adsorption capacity characteristics are set according to the initial sulfur and target sulfur, and the required slag making material is calculated; Step S8, determine whether alloy elements need to be added, calculate the change of each element during the processing time, and correct the change into the element control calculation; with the goal of cost optimization, trigger specific elements according to the smelting process and calculate the amount of alloy added; Step S9, determine whether heating is required, perform heating calculation, calculate the temperature loss caused by various factors within the specified exit time, calculate the heating time required for the temperature to reach the target in combination with the electrode efficiency, and allocate the heating time; Step S10, calculate the necessary time for heating, the necessary time for inclusion treatment and floating, the necessary time for alloying, the necessary time for deoxidation and desulfurization, remove the overlapping time, calculate the necessary time for standard treatment, if the necessary time is greater than the required time for exit, then issue a processing time alarm, switch to priority setting, process according to the logic of composition, temperature and cleanliness priority, and generate the relationship axis of pre-calculated smelting treatment steps, molten steel and slag state calculation and smelting time; Step S11, drive the ladle car to the processing position, lower the furnace cover, and start the processing process; Step S12, the smelting state visual inspection module monitors the slag state, slag thickness, slag state, and stirring state, and transmits them to the system for feedback stirring control; Step S13, measuring temperature and taking samples, updating the latest information on the composition of molten steel and slag, and repeating S1-S10 until the molten steel meets the requirements; Step S14, after the temperature and composition meet the requirements, soft blowing is performed to wait for the calcium treatment opportunity; Step S15, after the scheduled time is reached, the furnace cover is raised, the furnace exits the station, detailed furnace cost information is generated, and cost analysis and self-learning parameter optimization are performed.

2. The method for fully automatic production of LF refining according to claim 1, characterized in that: Based on big data algorithms and heuristic algorithms, a cost analysis optimization model is established. Through the analysis of smelting micro-operation costs and the overall cost analysis of furnaces, combined with the simulation of smelting procedures, a global search for the optimal cost combination operation plan is carried out.

Images