Melting process monitoring system of medium-frequency electric furnace and process guidance method thereof

Abstract

The invention discloses a smelting process monitoring system of a medium-frequency electric furnace and a process guidance method of the smelting process monitoring system. The system comprises an edge acquisition terminal and a cloud platform, the edge acquisition terminal comprises an intermediate frequency voltage transmitter used for acquiring a voltage signal at the output side of the intermediate frequency electric control cabinet and converting the voltage signal into a first analog signal; the intermediate-frequency current transmitter is used for collecting a current signal at the output side of the intermediate-frequency electric control cabinet and converting the current signal into a second analog signal; the analog quantity conversion module is used for converting the first analog signal and the second analog signal into digital signals; the intelligent thermodetector is used for measuring the temperature of the molten metal in real time and sending temperature data; the cloud platform is used for receiving and processing the digital signal and the temperature data and storing high-frequency sampling data; identifying a smelting stage through an LSTM network; power control is optimized on the basis of a Transform model and a GAN model; and integrating an expert knowledge base and providing process guidance. According to the system, comprehensive digitization and real-time accurate sensing of the smelting process are achieved; and the timeliness and scientificity of decision making are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of metallurgical industry automation, and more particularly to a smelting process monitoring system of a medium-frequency electric furnace and a process guidance method thereof. BACKGROUND

[0002] At present, in the traditional medium-frequency induction electric furnace smelting process relying on manual operation, due to the lack of real-time data monitoring and standardized control, there are systematic defects in production management, and the specific problems are as follows:

[0003] 1. Full reliance on worker experience, large randomness of operation:

[0004] Non-quantitative standard: workers manually adjust power according to experience, lacking critical value guidance of voltage, current and power.

[0005] Large operation fluctuation: significant operation difference between workers of different shifts, leading to smelting efficiency and energy consumption fluctuation exceeding ±15%.

[0006] 2. Key data missing, unable to trace the root cause of the problem:

[0007] No temperature record: molten iron temperature relies on naked eye observation (such as "bright yellow ≈ 1500℃"), actual temperature error up to ±50℃, leading to overburning or insufficient smelting.

[0008] No electrical curve: power / voltage / current curve is not recorded, unable to identify abnormalities (such as a 30% sudden drop in current when material is in the shed, but the worker did not respond in time).

[0009] No power analysis: active power is not monitored, a large amount of electric energy is wasted in invalid inductive reactance (measured power factor is only 0.75, lower than the standard value of 0.9).

[0010] 3. Personnel changes lead to management effectiveness collapse:

[0011] Experience dependence: core operation skills are "passed down by word of mouth" by the master, and new workers are prone to misoperation (such as directly full power when starting a cold furnace, causing thermal shock cracks in the furnace lining).

[0012] No standard inheritance: after personnel rotation or resignation, operation specifications are lost, and error operations are "passed down" and spread (such as adding material in advance to increase production, causing short circuit and fire).

[0013] 4. Control layer lag:

[0014] Lack of real-time monitoring data of smelting process, leading to 1-2 minutes of power adjustment after workers find temperature abnormalities, at which time the molten iron has been overburned.

[0015] 5. Quality risk:

[0016] The smelting process has no standardized control, and the batch rejection rate of castings caused by uneven composition is more than 5%.

[0017] There is no system solution in the prior art that can realize millisecond-level data acquisition, multi-stage intelligent identification and closed-loop process guidance. SUMMARY

[0018] Therefore, the present application provides a smelting process monitoring system of an intermediate frequency electric furnace and a process guidance method thereof, which at least partially solve the above technical problems.

[0019] To achieve the above purpose, the present application adopts the following technical solutions:

[0020] In a first aspect, the present application provides a smelting process monitoring system of an intermediate frequency electric furnace, comprising: an edge collection terminal and a cloud platform.

[0021] The edge collection terminal comprises:

[0022] An intermediate frequency voltage transmitter is configured to collect the voltage signal on the output side of the intermediate frequency electric control cabinet and convert it into a first analog signal.

[0023] An intermediate frequency current transmitter is configured to collect the current signal on the output side of the intermediate frequency electric control cabinet and convert it into a second analog signal.

[0024] An analog quantity conversion module is configured to convert the first analog signal and the second analog signal into digital signals.

[0025] An intelligent temperature measuring instrument is configured to measure the temperature of the metal melt in real time and send temperature data.

[0026] The cloud platform is configured to receive and process the digital signals and the temperature data, store high-frequency sampling data, identify the smelting stage through an LSTM network, optimize power control based on a Transformer and a GAN model, and integrate an expert knowledge base and provide process guidance.

[0027] In one embodiment, the intelligent temperature measuring instrument sends the temperature data by wireless radio and aligns the time stamp with the voltage and current data.

[0028] In one embodiment, the cloud platform comprises:

[0029] A stage identification module adopts a DTW and LSTM network, and automatically segments the inverter stop stage, inverter start stage, full-power melting stage, continuous charging stage, continuous heating stage and holding temperature stage in the smelting process based on the time sequence characteristics of voltage, current and power data.

[0030] An anomaly detection module using a Transformer and a GAN network for real-time detection of current oscillation, power mutation or temperature anomaly events during the smelting process, and triggering a shutdown alarm when an anomaly is detected.

[0031] A process optimization module using a knowledge graph and reinforcement learning technology to generate optimization strategies based on historical smelting data, furnace life and material type, and dynamically adjust the start-up parameters or cooling time of the next furnace;

[0032] A visual dashboard based on WebGL technology for real-time three-dimensional dynamic rendering of a three-dimensional thermal map, damage score curve and optimization strategy analysis report of the smelting process.

[0033] In one embodiment, the cloud platform further comprises:

[0034] A digital twin module for building a digital twin of the smelting process and achieving a millisecond-level data acquisition and decision-making closed loop through an edge-cloud collaborative architecture.

[0035] In one embodiment, during the full-power melting stage, a multi-objective optimization model is constructed as follows:

[0036]

[0037] where: η represents energy efficiency, v represents melting rate, w represents electrode wear; P input represents input power; P rated represents rated power; T coolant represents cooling temperature.

[0038] In one embodiment, the cloud platform further comprises:

[0039] An adjustment module for dynamically adjusting the inverter trigger angle to control the power factor when the smelting process is in the full-power melting stage, and automatically recommending the charging strategy and temperature curve according to the smelting stage.

[0040] In one embodiment, the cloud platform further comprises:

[0041] A recommendation module for displaying the temperature stratification in the ladle in real time through an AR interface and recommending the amount of alloy additive when the smelting process is in the holding and tapping stage.

[0042] In a second aspect, the embodiments of the present application also provide a smelting process monitoring process guidance method for an intermediate frequency electric furnace, which uses the smelting process monitoring system for an intermediate frequency electric furnace as described in any one of the embodiments of the first aspect. The method comprises the following steps:

[0043] Collecting voltage analog signals, current analog signals and temperature data network signals through an edge collection terminal;

[0044] The analog signal is converted into a digital signal and uploaded to a cloud platform through a 5G network along with the network signal of temperature data;

[0045] The cloud platform identifies the smelting stage through a DTW and LSTM fusion model;

[0046] Abnormal detection and early warning are performed through a Transformer and GAN model, and power parameters are dynamically adjusted based on a knowledge graph and a reinforcement learning model;

[0047] Process guidance information is generated and fed back to the operation end through a visual board.

[0048] According to the technical solution, compared with the prior art, the present application has the following technical effects:

[0049] 1. The smelting process is fully digitized and accurately perceived in real time. Through the deployment of an edge collection terminal composed of a medium-frequency voltage transmitter, a medium-frequency current transmitter and an intelligent temperature meter, the system can synchronously collect key process parameters such as voltage, current and temperature with millisecond-level precision, completely changing the traditional extensive operation mode relying on manual experience and naked-eye observation, and solving the problems of missing key data and being unable to trace back.

[0050] 2. An intelligent analysis architecture of "edge-cloud" cooperation is constructed, and the timeliness and scientificity of decision-making are improved. The edge side is responsible for high-frequency data collection and preliminary conversion, and the cloud platform is responsible for massive data storage and complex algorithm operation. This architecture not only guarantees the real-time nature of data collection, but also utilizes the powerful computing power of the cloud to achieve deep insight into the smelting process. Through the LSTM network to identify the smelting stage, and based on the Transformer and GAN model to optimize power control, the process control is changed from lagging and passive response to leading and active optimization.

[0051] 3. The standardization and intelligent inheritance of process knowledge are realized. By integrating an expert knowledge base and providing real-time process guidance, the system converts the "implicit experience" of experienced operators into "explicit rules" that can be replicated and executed, effectively avoiding the loss of operation specifications and quality fluctuations due to personnel changes, and reducing the over-reliance of production management on individual "masters".

[0052] 4. The production efficiency, product quality and equipment safety are significantly improved. The system ensures that the smelting process always runs under optimal or near-optimal parameters through real-time monitoring and intelligent algorithms, thereby achieving the multi-objective optimization effect of improving energy efficiency, accelerating the melting rate and reducing electrode furnace lining loss, and effectively preventing the occurrence of production accidents such as stock material, overburning and furnace lining perforation through abnormal detection. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute a part of the embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of the provided drawings.

[0054] Figure 1 The device structure schematic diagram of the smelting process monitoring system of the intermediate frequency electric furnace provided by the present application.

[0055] Figure 2 The overall architecture diagram of the smelting process monitoring system of the intermediate frequency electric furnace provided by the present application.

[0056] Figure 3 The data curve diagram of different smelting stages provided by the present application.

[0057] Figure 4 The process guidance method flow chart of the smelting process monitoring system of the intermediate frequency electric furnace provided by the present application.

[0058] Figure 5 The process guidance principle diagram of the smelting process monitoring system of the intermediate frequency electric furnace provided by the present application. DETAILED DESCRIPTION

[0059] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0060] Referring to Figures 1-2 The embodiments of the present application disclose a smelting process monitoring system of an intermediate frequency electric furnace, mainly composed of an edge collection terminal deployed in a factory workshop and a cloud platform deployed in a remote data center. The edge side is responsible for high-frequency data collection and preliminary conversion, and the cloud platform is responsible for massive data storage and complex algorithm operation. This architecture not only guarantees the real-time nature of data collection, but also realizes deep insight into the smelting process by using the powerful computing power of the cloud.

[0061] Among them, the edge collection terminal is responsible for the collection and uploading of raw data, and its specific implementation is as follows:

[0062] Intermediate frequency voltage transmitter: for example, a transmitter with model HUT-3000U3S2-A1 is selected, with a range of ±3000VAC and an output of 4-20mA analog signal. It is installed on the output side of the intermediate frequency electric control cabinet, and the intermediate frequency voltage signal is collected through a special voltage transformer.

[0063] Intermediate frequency current transducer: such as the model R1201-12KA-4-20mA transducer, with a range of ±12000A AC and an output of 4-20mA analog signal. It is installed on the output side of the intermediate frequency electric control cabinet, and the intermediate frequency current signal is collected through the Rogowski coil or current transformer.

[0064] Analog quantity conversion module: such as an 8-channel analog quantity acquisition module with the model WJ188. The 4-20mA signal output by the voltage and current transducer is connected to the module. The module is responsible for converting analog signals to digital quantities, and through its built-in Ethernet interface, it is connected to the factory's 5G private network gateway through an industrial switch.

[0065] Intelligent temperature measuring instrument: such as an infrared temperature measuring gun or an immersion thermocouple with the model KD-T102L, with a range of 500-1800℃. The temperature measuring instrument accurately measures the temperature of the metal melt when the operator measures the temperature, automatically binds the measured temperature data with the current furnace number, and sends the data to the wireless access point in the factory through the built-in wireless communication module (such as LoRa, Wi-Fi or 5G module), and finally transmits the data to the cloud platform through the network. To ensure data consistency, all edge devices (conversion modules, temperature measuring instruments) are connected to a unified NTP time server to accurately align all data timestamps.

[0066] Cloud platform for receiving and processing digital signals and temperature data, storing high-frequency sampling data; identifying smelting stages through LSTM network; and optimizing power control based on Transformer and GAN model; and integrating expert knowledge base and providing process guidance.

[0067] The device structure and connection relationship of the system are shown in Figure 1 By using intermediate frequency voltage transducers and intermediate frequency current transducers, the active power (hereinafter referred to as power) during the smelting process, voltage, and current can be obtained, and it can be determined whether the sintering process is controlled according to the regulations. For example, whether the temperature of 1100℃ is rising too fast, whether the holding power is correct, and whether the temperature is seriously overheated. The temperature data of the intelligent temperature measuring instrument is automatically uploaded to the cloud, and the timestamp can be aligned with the intermediate frequency voltage, intermediate frequency current, and power data for joint comparison and analysis. Finally, the smart casting cloud platform and the device cloud summarize and display all the data, and configure and manage the entire sensing network.

[0068] In specific implementation, the system can transmit data to the smart casting cloud platform through intermediate frequency voltage transducers, Hall current sensors, temperature measuring guns, etc. through wireless networks, providing data support for large model training and inference.

[0069] In one specific embodiment, the cloud platform includes:

[0070] The stage recognition module uses DTW and LSTM networks to automatically segment the smelting process into the inverter stop stage, inverter start stage, full-power melting stage, continuous feeding stage, continued heating stage and heat preservation and soup discharge stage based on the temporal characteristics of voltage, current and power data.

[0071] Data preprocessing: Extract the voltage, current, and power time series data of the current furnace from the TSDB and perform data cleaning and standardization.

[0072] Model inference: Real-time inference is performed using a trained LSTM-DTW fusion model. This model, trained on historical data from over 2,000 furnaces, takes as input data in 500ms time windows and outputs probability distributions for the six melting stages (inverter stop, inverter start, full power melting, continuous feeding, continued heating, and holding for hot water).

[0073] Stage determination: The stage with the highest probability is taken as the current smelting stage, and the result is updated to the digital twin in real time.

[0074] The anomaly detection module uses Transformer and GAN networks to detect current oscillations, power mutations or temperature anomalies in the smelting process in real time, and triggers a shutdown alarm when an anomaly is detected. In this embodiment, the Transformer model is used as an encoder to learn the contextual features of the power and current sequences under normal working conditions. The discriminator in the GAN model is used to calculate the difference score (Anomaly Score) between the current real-time data sequence and the normal sequence. When the difference score exceeds the preset threshold, it is determined to be an anomaly (such as a 30% drop in current may indicate shed material), and a shutdown alarm is immediately sent to the engineer through API calls to corporate WeChat, SMS and other interfaces, and a red flashing prompt is displayed on the visual dashboard.

[0075] The process optimization module uses knowledge graphs and reinforcement learning technology to generate optimization strategies based on historical smelting data, furnace age and material type, and dynamically adjust the startup parameters or cooling time of the next furnace. Construct a knowledge graph with "furnace condition-material-process" as the core. For example, it stores more than 2,000 metallurgical expert rules (IF-THEN format), such as "IF furnace age>200&material=high Cr steel THEN recommends that the target soup temperature be increased by 20°C." Among them, reinforcement learning: in the full-power melting stage, start the multi-objective optimization algorithm based on Q-learning. Its reward function R=α·energy efficiency+β·melting rate-γ·electrode loss, and the action space is the fine-tuning instruction for the inverter trigger angle and frequency. The algorithm satisfies P input ≤1.1P rated and T coolant Find the optimal control strategy under the hard constraint of ≤65℃.

[0076] The optimized strategy (such as the adjusted power setting value and recommended feeding timing) will be sent to the human-machine interface (HMI) at the edge in real time to guide workers' operations or directly executed through the PLC interface.

[0077] The visual dashboard, using WebGL technology to achieve 3D dynamic rendering, is used to display real-time 3D thermal maps of the melting process, damage score curves, and optimization strategy analysis reports. Implementation is developed using a WebGL-based graphics engine (such as Three.js). Based on data from temperature sensors and model inference, the temperature field distribution is dynamically rendered on the 3D model of the electric furnace. The algorithm-based estimated cumulative damage curve for the furnace lining is displayed in real time, predicting the remaining life. After each furnace is completed, an analysis report is automatically generated, including energy consumption, duration, abnormal events, and optimization suggestions.

[0078] The cloud platform also includes a digital twin module for building a digital twin of the smelting process. This module uses an edge-cloud collaborative architecture to achieve millisecond-level data acquisition and closed-loop decision-making. A time series database (TSDB) can be used to store 10ms sampling data to build a digital twin of the smelting process.

[0079] The adjustment module dynamically adjusts the inverter trigger angle to control the power factor during the full-power melting phase of the smelting process. It also automatically recommends a charging strategy and temperature profile based on the melting phase. If abnormal power factor fluctuations are detected during the full-power melting phase, the cloud platform uses the knowledge graph to match similar cases and, combined with current lining thickness sensor data, dynamically adjusts the inverter trigger angle to keep power factor fluctuations within a controllable range.

[0080] The recommendation module is used to display the temperature stratification in the ladle in real time through the AR interface when the smelting process is in the insulation and soup discharge stage, and recommend the amount of alloy additives.

[0081] This invention achieves a closed loop between millisecond-level data acquisition and cloud-based intelligent decision-making by building an edge-cloud collaborative architecture. A multi-stage adaptive recognition algorithm is developed, accurately classifying smelting stages through a fusion of dynamic time warping (DTW) and LSTM models. Furthermore, a large-scale model-based process knowledge graph is constructed to enable cross-domain mapping between equipment operating parameters and metallurgical process parameters.

[0082] At different smelting stages, refer to Figure 3 As shown, the vertical axis represents the three curves, namely power (kW), inverter current (A), and voltage (V), and the horizontal axis represents time. Six specific smelting stages are shown. The cloud platform algorithm layer makes the following decisions:

[0083] 1. Inverter stop phase (preparation period)

[0084] The time series database (TSDB) is used to store 10 ms level sampling data, and the digital twin of the smelting process is constructed.

[0085] Data characteristics: voltage drop to zero, current exponential decay; cooling water temperature / PLC state signal activation;

[0086] Process guidance:

[0087] Through the knowledge graph matching historical maintenance records, the best shutdown time is pushed (such as: the furnace age > 200 times, the cooling time is extended by 15%). A three-dimensional thermal map is generated to guide the next furnace loading distribution.

[0088] 2. Inverse starting phase (about 30-60 seconds)

[0089] Data characteristics: voltage / current presents pulse oscillation (frequency 2-5 kHz). The power factor quickly rises from 0.

[0090] Process guidance:

[0091] Dynamic adjustment of IGBT trigger sequence to reduce the peak value of starting current by 20%-30%. The best soft start curve is predicted through digital twin simulation.

[0092] 3. Full power melting stage

[0093] Data characteristics: power is stable at rated value ± 5%. Harmonic content THD < 8%.

[0094] Cloud platform algorithm implementation:

[0095] (1) Multi-objective optimization model

[0096]

[0097] Where: η represents energy efficiency, v represents melting rate, and w represents electrode wear; P input represents input power; P rated represents rated power; T coolant represents cooling temperature.

[0098] Input power P input must not exceed 1.1 times the rated power P rated . This is a safety margin factor. It allows for a slight overload operation for a short time (for example, during startup or acceleration of melting), but cannot exceed this limit. The purpose of the constraint: protect power electronic devices: prevent inverter, thyristor, etc. Large power devices from burning out due to overcurrent. Protect the electric furnace coil: excessive power will generate huge electromagnetic force and heat, which may damage the insulation of the induction coil. Ensure grid stability: avoid excessive impact on the plant's power grid.

[0099] In addition, the cooling water temperature Tcoolant Must be below 65 degrees Celsius. This is a critical value. Too high a temperature will cause a series of problems:

[0100] Risk of scaling: When the water temperature exceeds 60-70°C, the dissolved calcium and magnesium minerals in the water are more likely to precipitate and form scale on the pipes and inner walls, greatly reducing the heat exchange efficiency, forming a vicious cycle, and ultimately leading to overheating damage.

[0101] Cavitation phenomenon: Water is more likely to vaporize at high temperatures, producing bubbles that break in the pump (cavitation), damaging the water pump impeller.

[0102] Insulation aging: The insulation material of the coil will accelerate aging and embrittlement if it works at high temperature for a long time, leading to insulation failure and short circuit. Ensure cooling efficiency: Ensure that the cooling system can continuously and effectively cool the key components.

[0103] Must be below 65 degrees Celsius, the purpose of the constraint is to prevent equipment damage: avoid serious accidents such as coil insulation damage, power electronic component scrap due to overheating. Maintain system stability: prevent unexpected downtime due to cooling failure.

[0104] These two constraints are the "safety red line" that the system's intelligent decision cannot cross, and are the core safety logic that must be built in after the automation system replaces manual experience.

[0105] (2) Process guidance:

[0106] Adjust the frequency of the inverter every 10 seconds (0.5-1kHz range) to suppress specific harmonics.

[0107] Automatically trigger the feeding signal when the metal liquid level reaches the critical value.

[0108] 4. Continuous feeding stage

[0109] Data characteristics: Power shows periodic fluctuations (amplitude about 15%). Acoustic emission signal appears 300-500Hz characteristic frequency.

[0110] Process guidance:

[0111] Automatically match the best feeding interval according to the material type (scrap steel / alloy) (error <3 seconds). Monitor the material pile shape through millimeter wave radar and dynamically adjust the feeding machine swing speed.

[0112] 5. Continue to heat up stage

[0113] Data characteristics:

[0114] dT / dt gradually decreases to 5-8℃ / min. Impedance spectrum appears capacitive characteristics.

[0115] Process guidance:

[0116] When the furnace bottom / furnace wall temperature difference is detected to be > 50℃, the electromagnetic stirring is automatically triggered. The target temperature curve is adjusted according to the alloy composition (such as increasing the holding slope for Cr-containing steel grades).

[0117] 6. Holding stage for discharging soup

[0118] Data characteristics: power drops to 30%-40% of rated value. Characteristic eddy current appears on the surface of molten steel.

[0119] Process guidance:

[0120] The amount of alloying additive is recommended according to the predicted deoxidation degree (accurate to ±0.5 kg). The temperature stratification in the ladle is displayed in real time through the AR interface.

[0121] The present application designs a closed-loop system process structure: collect data -> fusion analysis (evaluate the accuracy of power setting and process stability / constant temperature) -> calibrate the holding power setting according to the analysis result -> solidify the optimal setting value to form a standardized SOP -> guide production -> continue data collection.

[0122] Taking "full-power melting stage detects abnormal fluctuation of power factor" as an example, the operation process of the system is as follows:

[0123] (1) The current transducer detects current waveform distortion, and the THD value rises.

[0124] (2) The Transformer model of the anomaly detection module identifies that the mode is similar to the historical case of "furnace lining thinning leading to electromagnetic coupling change" in the knowledge base.

[0125] (3) The process optimization module is immediately started, and the knowledge graph matches the coping strategy: fine-tune the inverter frequency to suppress specific harmonics.

[0126] (4) The optimization instruction is issued to the electric control system through the MQTT protocol, and the adjustment is executed.

[0127] (5) After adjustment, the power factor is restored to stability, and the system records the event and the processing scheme to the knowledge base for subsequent optimization.

[0128] Through the collection of smelting process data (voltage, current, temperature), the dynamic cycle of directly converting real-time monitoring data and evaluation conclusions into operation standards (SOP) is realized. The process quality control is no longer controlled by experience, but through dynamic analysis of the change trend and distribution state of the real-time voltage, current, power and temperature data curves, the best holding power setting value and its constant temperature effect are "quantized" and "locked", and the judgment of whether the production process meets the process quality requirements is made, so as to realize the effective supervision of workers' execution process based on data. Based on the analysis conclusion, the connection of heating smelting, holding, out of soup, transportation and other process links is dynamically adjusted, and then the optimal practice scheme is solidified as a repeatable SOP. This closed-loop structure ensures that the standardization comes from the operation data of the production process itself, and is continuously verified and optimized, which can effectively promote the continuous self-evolution of the whole smelting production process.

[0129] Based on the same inventive concept, the present application also provides a smelting process monitoring process guidance method of an intermediate frequency electric furnace, using the smelting process monitoring system of the intermediate frequency electric furnace as described in the above embodiment, referring to Figure 4 The method comprises the following steps:

[0130] S1, collecting voltage analog signals, current analog signals and temperature data network signals through an edge collection terminal;

[0131] S2, converting the analog signals into digital signals, and uploading them to a cloud platform through a 5G network together with the temperature data network signals;

[0132] S3, the cloud platform identifies the smelting stage through a DTW and LSTM fusion model;

[0133] S4, performing abnormality detection and early warning through a Transformer and GAN model, and dynamically adjusting power parameters based on a knowledge graph and a reinforcement learning model;

[0134] S5, generating process guidance information and feeding it back to the operation end through a visual board.

[0135] Referring to Figure 5 The present application realizes the value leap from "data" to "action" and realizes the guidance of the production process:

[0136] Guiding the transfer of the soup bag: the system automatically calculates the optimal out-of-soup time according to the progress of the identified "holding out-of-soup stage" and the subsequent production plan, and generates instructions to guide the crane or transfer car to schedule the ladle, optimize the production rhythm and improve the rationality of the plan arrangement.

[0137] Guidance on filling: In the "continuous feeding stage", the system matches the optimal feeding time, feeding amount and feeding type (such as scrap alloy ratio) according to the current furnace temperature, melting and furnace lining condition through knowledge graph, and gives clear prompt on HMI interface, optimizes production process, standardizes operation process, and avoids workers operating by feeling.

[0138] Guidance on heat preservation: In the late "continuous heating stage" and "heat preservation stage", the system accurately calculates and automatically sets the required heat preservation power according to the target temperature of molten steel, tapping plan and real-time temperature, avoids energy waste or too low temperature, and truly adjusts the heat preservation power according to the field beat, reduces energy consumption.

[0139] Guidance on temperature measurement: The system not only receives temperature measurement data, but also intelligently guides the temperature measurement operation itself. It can actively remind workers to measure temperature when the model inferred temperature reaches the key process point, and compare and calibrate the measured value with the predicted value to realize accurate temperature control and improve material quality.

[0140] Data closed loop: The effect data generated by all these guidance actions is collected by the system (back to S1) for verifying the correctness of the guidance strategy and continuously optimizing the algorithm model, forming an evolutionary closed loop that never stops.

[0141] The melting process monitoring and process guidance method of the intermediate frequency electric furnace provided by the application realizes fine production management: the vague "teacher's experience" is converted into accurate and executable data instructions, realizing the standardization and normalization of production process. In addition, through optimizing the transfer beat, accurately controlling the heat preservation and melting power, the energy consumption and waiting time are significantly reduced, and the equipment utilization and production efficiency are improved. In addition, accurate temperature control and standardized feeding guidance reduce the occurrence of quality problems such as composition segregation and overburning from the source, greatly improve product consistency and reduce scrap rate. The system fixes expert knowledge in the algorithm and knowledge base, so that new workers can complete high-quality operation by following the system guidance, effectively solving the technical gap problem caused by personnel flow. In addition, based on data-driven predictive maintenance and abnormal early warning, management changes from "after-the-fact remediation" to "pre-emptive prevention", greatly improving production safety and planning.

[0142] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same and similar parts between various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the related parts can be referred to the method part description.

[0143] The foregoing description of the disclosed embodiments enables a person skilled in the art to make or use the application. Modifications of these embodiments will occur to persons of skill in the art, and that the appended claims are intended to cover all such modifications that do not depart from the true spirit and scope of the application. Therefore, the application is not limited to the embodiments shown but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A system for monitoring a smelting process in an intermediate frequency electric furnace, characterized in that The system comprises: an edge acquisition terminal and a cloud platform; wherein the edge acquisition terminal comprises: a medium-frequency voltage transmitter for acquiring the voltage signal on the output side of the medium-frequency electric control cabinet and converting it into a first analog signal; a medium-frequency current transmitter for acquiring the current signal on the output side of the medium-frequency electric control cabinet and converting it into a second analog signal; an analog-digital conversion module for converting the first analog signal and the second analog signal into digital signals; an intelligent temperature measuring instrument for measuring the temperature of the molten metal in real time and sending the temperature data; the cloud platform is configured to receive and process the digital signals and the temperature data, store high-frequency sampling data, identify the smelting stage through an LSTM network, optimize the power control based on a Transformer and a GAN model, and integrate an expert knowledge base to provide process guidance.

2. A system for monitoring the smelting process of an intermediate frequency electric furnace as claimed in claim 1, wherein The intelligent temperature measuring instrument sends the temperature data through a wireless radio and aligns the time stamp with the voltage and current data.

3. A system for monitoring the smelting process of an intermediate frequency electric furnace as claimed in claim 1, wherein, The cloud platform comprises: a stage identification module that adopts a DTW and LSTM network to automatically segment the reverse conversion stop stage, the reverse conversion start stage, the full-power melting stage, the continuous charging stage, the continuous temperature rising stage and the holding and tapping stage in the smelting process based on the time sequence characteristics of the voltage, current and power data; an anomaly detection module that adopts a Transformer and GAN network to detect current oscillation, power mutation or temperature anomaly events in real time during the smelting process and trigger a shutdown alarm when an anomaly is detected; a process optimization module that adopts a knowledge graph and reinforcement learning technology to generate an optimization strategy according to historical smelting data, furnace age and material type, and dynamically adjust the start-up parameters or cooling time of the next furnace; a visual dashboard that realizes three-dimensional dynamic rendering based on WebGL technology to display a three-dimensional thermal map, a damage score curve and an optimization strategy analysis report of the smelting process in real time.

4. A system for monitoring the smelting process of an intermediate frequency electric furnace as claimed in claim 1, wherein, The cloud platform further comprises: a digital twin module for constructing a digital twin of the smelting process and realizing a millisecond-level data acquisition and decision-making closed loop through an edge-cloud collaborative architecture.

5. A system for monitoring the smelting process of an intermediate frequency electric furnace as claimed in claim 3, wherein, The cloud platform further comprises: wherein: η represents energy efficiency, v represents melting rate, w represents electrode wear; P input represents input power; P rated represents rated power; T coolant represents cooling temperature.

6. A system for monitoring the smelting process of an intermediate frequency electric furnace as claimed in claim 5, wherein, an adjustment module for dynamically adjusting the trigger angle of the inverter to control the power factor when the smelting process is in the full-power melting stage, and automatically recommending a charging strategy and a temperature curve according to the smelting stage. The cloud platform further comprises:

7. A system for monitoring the smelting process of an intermediate frequency electric furnace as claimed in claim 3, wherein, a recommendation module for displaying the temperature stratification in the ladle in real time through an AR interface and recommending the amount of alloying additive when the smelting process is in the holding and tapping stage. A method for monitoring the smelting process of a medium-frequency electric furnace according to any one of claims 1-7, the method comprising the following steps:

8. A process control method for monitoring a smelting process in an intermediate frequency electric furnace, characterized by, acquiring the voltage analog signal, the current analog signal and the temperature data network signal through the edge acquisition terminal; converting the analog signals into digital signals and uploading them to the cloud platform through a 5G network together with the temperature data network signal; identifying the smelting stage through a DTW and LSTM fusion model in the cloud platform; detecting and warning through a Transformer and GAN model; and dynamically adjusting the power parameters based on a knowledge graph and reinforcement learning model. ​ Generate process guidance information and feed it back to the operator through a visual dashboard.

Images