Vacuum suspension melting system condensate shell overheating early warning method

Abstract

The present application relates to the technical field of metallurgical equipment, and relates to a kind of overheat early warning method of vacuum suspension smelting system condensate shell. Including the establishment of power-time-temperature prediction model, obtain the prediction surface temperature dataset;Real-time acquisition of the surface temperature of crucible / melt, obtain the actual surface temperature dataset, obtain the instantaneous surface temperature difference deviation dataset by the above dataset;Establish cooling liquid temperature rise prediction model, obtain the cooling loop temperature difference deviation dataset;Instantaneous surface temperature difference deviation dataset and cooling loop temperature difference deviation dataset are normalized, and obtain comprehensive instantaneous risk score;And cumulative anomaly index and time weighting are carried out, and obtain cumulative anomaly index;Vacuum suspension smelting system enters corresponding early warning state according to corresponding rule. Thus, the problem that the monitoring means of existing vacuum suspension smelting process is single, response lag, and lacks recognition ability to the coupling state of upper condensate shell and lower overheating is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of metallurgical equipment, in particular to a superheat early warning method for a vacuum suspension melting system crust. BACKGROUND

[0002] The vacuum suspension melting furnace is a kind of equipment that realizes non-contact suspension heating and melting of metal melt by using high-frequency electromagnetic induction force. This kind of equipment has the advantages of high purity, no crucible pollution and fast temperature control response, and is widely used in the preparation field of high-temperature alloy, rare metal and special functional material. However, in the vacuum suspension melting process, due to the exposure of the melt to low pressure environment and the complex effect of electromagnetic force and thermal convection above the melt, an upper crust is easily formed on the surface of the melt. When the upper crust cannot be melted in time, it may block the high-temperature melt below, so that the surface temperature is significantly lower than the actual temperature distribution in the furnace, and even under the condition of continuous heating, the lower melt is locally overheated and the thermal load of the furnace wall is abnormally increased, which may cause damage to the equipment and safety risk.

[0003] In the prior art, the surface temperature distribution of the melt is usually obtained by single-point infrared temperature measurement or thermal imaging monitoring to judge the temperature state of the melting process. However, such scheme mainly has the following defects:

[0004] 1. Insufficient representativeness of surface temperature measurement: due to the high reflectivity and low emissivity of the crust surface, and the existence of high-temperature melt exposed in the cracks of the crust, single-point infrared temperature measurement may misjudge the high temperature in the gap as the overall temperature, and cannot truly reflect the average temperature state of the upper crust.

[0005] 2. Lack of indirect judgment ability of lower melt overheating: the existing surface temperature measurement and thermal imaging monitoring can only observe the temperature change of the outer surface of the melt, and cannot directly obtain the temperature information of the area shielded by the crust; when the upper crust is formed, the temperature continues to rise below but the surface temperature remains unchanged, and the system is difficult to identify the abnormality in time.

[0006] 3. The monitoring system is not sensitive to the change of working condition: the current temperature control is mainly based on instantaneous value or single-parameter closed-loop regulation, without considering the dynamic relationship between input energy (heating power) and temperature change. When the power is continuously input but the surface temperature does not rise as expected, the control system cannot identify the risk signal behind this "temperature rise lag" phenomenon.

[0007] 4. Cooling load abnormality cannot be associated with melting state: although the abnormality of cooling liquid temperature or flow can reflect the change of equipment thermal load, it lacks a comprehensive judgment mechanism with melting power and temperature response, so that the abnormal trend is often discovered late. SUMMARY

[0008] To solve the problems of single information, lagging response and lack of recognition ability for the coupling state of the upper skull and the lower overheating in the monitoring means of the existing vacuum suspension smelting process, the present application provides an overheating early warning method for skull of a vacuum suspension smelting system.

[0009] In the first aspect, the present application provides an overheating early warning method for skull of a vacuum suspension smelting system, the overheating early warning method comprising:

[0010] S1: establishing a power-time-temperature prediction model to obtain a predicted surface temperature data set of the crucible / melt;

[0011] S2: collecting the surface temperature of the crucible / melt in real time to obtain an actual surface temperature data set of the crucible / melt, and obtaining an instantaneous surface temperature difference deviation data set through the predicted surface temperature data set and the actual surface temperature data set;

[0012] S3: establishing a cooling liquid temperature rise prediction model, collecting the heating power and the cooling liquid mass flow in real time to obtain a cooling loop temperature difference deviation data set;

[0013] S4: performing normalization processing on the collected instantaneous surface temperature difference deviation data set and the cooling loop temperature difference deviation data set to obtain a comprehensive instantaneous risk score; and performing accumulation anomaly index and time weighting to obtain an accumulation anomaly index;

[0014] S5: the vacuum suspension smelting system enters a first early warning state according to the comprehensive instantaneous risk score and the accumulation anomaly index through a first rule;

[0015] Or, the vacuum suspension smelting system enters a second early warning state according to the comprehensive instantaneous risk score and the accumulation anomaly index through a second rule;

[0016] Or, the vacuum suspension smelting system enters a third early warning state according to the comprehensive instantaneous risk score and the accumulation anomaly index through a third rule.

[0017] In some embodiments, in step S1, establishing the power-time-temperature prediction model comprises: collecting historical operation data of the vacuum suspension smelting furnace under different charging amounts, different materials and different process parameters, synchronously processing the historical operation data according to a unified time reference, and removing, smoothing or filtering abnormal values, and then constructing an input sequence reflecting the power-time behavior characteristics according to a preset time window, taking the surface temperature of the crucible / melt at each time point as the model output target, establishing a power-time-temperature mapping model for time series prediction, and completing the fitting of the model parameters by minimizing the error between the predicted temperature and the actual measured temperature;

[0018] The historical operation data includes: heating power changing over time, cumulative energy, heating time length, cooling liquid inlet temperature, cooling liquid outlet temperature, surface temperature of the crucible / melt.

[0019] In some embodiments, in step S1, the heating power and the heating time are collected in real time to obtain the predicted surface temperature data set of the crucible / melt.

[0020] In some embodiments, in step S1, establishing the power-time-temperature prediction model includes: regarding the input coil power as an input heat source of the vacuum suspension smelting system, establishing a theoretical temperature rising relationship of the melt and the surface temperature by calculating the effective absorbed energy per unit time, i.e. ; in the formula, is the real-time input power of the coil, is the electromagnetic coupling efficiency, is the melt mass, and c is the specific heat capacity, is the predicted surface temperature change rate, is the heat dissipation term changing with temperature.

[0021] In some embodiments, in step S1, the heating time is collected in real time to obtain the predicted surface temperature data set of the crucible / melt.

[0022] In some embodiments, in step S3, establishing the cooling liquid temperature rise prediction model includes: regarding the input coil power as an input heat source of the vacuum suspension smelting system, establishing a cooling loop temperature difference deviation relationship of the cooling liquid, i.e. ; in the formula, is the real-time input power of the coil, is the cooling liquid mass flow rate, and cp is the specific heat capacity of the cooling liquid at constant pressure, is the cooling loop temperature difference deviation, is the effective heat transfer proportion parameter.

[0023] In some embodiments, in step S4, the normalization processing includes:

[0024] defining a normalized instantaneous score of the instantaneous surface temperature difference deviation data set, i.e. ; in the formula, is the first reference scale, is the instantaneous surface temperature difference deviation, is the first normalized instantaneous score;

[0025] defining a normalized instantaneous score of the cooling loop temperature difference deviation data set, i.e. ; in the formula, is the second reference scale, is the cooling loop temperature difference deviation, is the second normalized instantaneous score;

[0026] The comprehensive instantaneous risk score is defined as ; wherein, is a first initial recommended value, is a second initial recommended value, is the comprehensive instantaneous risk score.

[0027] In some embodiments, in step S4, the cumulative abnormality indicator is weighted with time comprises:

[0028] The cumulative abnormality indicator is weighted with time, i.e. ; wherein, is a time weighting function, W is a cumulative window length, is the cumulative abnormality indicator;

[0029] The discretization is implemented, i.e. ; wherein, , is a sampling interval.

[0030] In some embodiments, the first rule is: and ; wherein, is a first graded alarm threshold, is a first set threshold;

[0031] The vacuum suspension smelting system enters a first level early warning state, i.e. interface prompts, records events, and recommends manual inspection.

[0032] In some embodiments, the second rule is: ; or, and exceeds a second set threshold; wherein, is a first graded alarm threshold, is a second graded alarm threshold, is an outlet temperature of the cooling liquid, is an inlet temperature of the cooling liquid;

[0033] The vacuum suspension smelting system enters a second level early warning state, i.e. automatically reduces the heating power by a preset proportion or to the next power level, extends the steady state observation time, and simultaneously triggers sound or light alarms and short message or remote notifications, and records snapshots.

[0034] In some embodiments, the third rule is: ; or, instantaneous and continuously rises; wherein, is a third graded alarm threshold, is a critical threshold of the instantaneous change amount of the melt or crucible surface temperature, is an outlet temperature of the cooling liquid, temperature of the cooling liquid;

[0035] The vacuum suspension smelting system enters a three-level early warning state, namely, immediately disconnecting the heating power supply, opening the emergency cooling circuit, recording all historical data and notifying the maintenance and on-duty management personnel.

[0036] To solve the problems of single information, response lag and lack of recognition ability for the coupling state of the upper solidified skull and the lower overheating in the existing monitoring means of the vacuum suspension smelting process, the present application has the following advantages:

[0037] Through the technical scheme of the present application, the expected temperature rise curve of the heating power and the heating time in the smelting process is established, the surface temperature and the cooling liquid temperature collected in real time are combined, the deviation value is calculated and the abnormality is accumulated, when the accumulated deviation exceeds the set threshold value, the early warning is triggered, thereby realizing the early monitoring and prevention and control of the upper solidified skull formation and the lower overheating. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 A flow chart of a solidified skull overheating early warning method of a vacuum suspension smelting system is shown. DETAILED DESCRIPTION

[0039] The present disclosure will now be discussed with reference to a number of example embodiments. It should be appreciated that these embodiments are discussed merely to better enable a person of ordinary skill in the art to better understand and thus implement the present disclosure, and are not intended to limit the scope of the present disclosure in any way.

[0040] As used herein, the term "includes" and its variants are to be read as open-ended terms that mean "including, but not limited to." The term "based on" is to be construed as "based at least in part on." The terms "one embodiment" and "an embodiment" are to be read as "at least one embodiment." The term "another embodiment" is to be read as "at least one other embodiment." The terms "upper," "lower," "left," "right," "front," "rear," "top," "bottom," "inner," "outer," "vertical," "horizontal," "lateral," "longitudinal," and similar terms are used for orientation or positional relationships based on the orientation or position as shown in the drawings. These terms are used merely for purposes of description and are not intended to limit the indicated device, element, or component to a particular position, orientation, or configuration, unless otherwise indicated by the specific context. Also, these terms are used in conjunction with the terms "coupled" and "connected" to describe the position, orientation, and connection of a device, element, or component relative to another device, element, or component. Any of these terms can be used to indicate that two devices, elements, or components are in some way physically or logically connected or coupled together, but these terms do not imply that a particular orientation or position is required. For example, a device, element, or component can be directly connected or coupled to another device, element, or component, or it can be indirectly connected or coupled through an intermediate device, element, or component. For example, a device, element, or component can be connected or coupled to another device, element, or component via an internal connection, an external connection, or an internal and external connection. For those skilled in the art, the specific meaning of these terms in the present application can be understood according to the specific context. In addition, the terms "first," "second," and the like are used primarily to distinguish different devices, elements, or components from one another (the specific type and configuration of which can or can not be the same), and are not necessarily used to indicate or imply a relative importance or quantity of the indicated devices, elements, or components. Unless otherwise specified, the meaning of "a plurality" is two or more.

[0041] Embodiment I

[0042] The present embodiment discloses a superheat early warning method for a vacuum suspension melting system crust, as shown in the accompanying drawings, the superheat early warning method comprises the following steps: Figure 1

[0043] S1: Establish a power-time-temperature prediction model to obtain a predicted surface temperature data set of the crucible / melt;

[0044] S2: Real-time acquisition of the surface temperature of the crucible / melt, obtaining an actual surface temperature data set of the crucible / melt, and obtaining an instantaneous surface temperature difference deviation data set by the predicted surface temperature data set and the actual surface temperature data set;

[0045] S3: Establish a cooling liquid temperature rise prediction model, real-time acquisition of the heating power and the cooling liquid mass flow, and obtain a cooling loop temperature difference deviation data set;​

[0046] S4: normalizing the collected instantaneous surface temperature difference deviation dataset and the cooling loop temperature difference deviation dataset to obtain a comprehensive instantaneous risk score; and performing time weighting on the cumulative anomaly index to obtain a cumulative anomaly index;

[0047] S5: the vacuum suspension melting system enters a first early warning state according to the comprehensive instantaneous risk score and the cumulative anomaly index through a first rule;

[0048] Or, the vacuum suspension melting system enters a second early warning state according to the comprehensive instantaneous risk score and the cumulative anomaly index through a second rule;

[0049] Or, the vacuum suspension melting system enters a third early warning state according to the comprehensive instantaneous risk score and the cumulative anomaly index through a third rule.

[0050] In the embodiment, a superheating early warning method for a vacuum suspension melting system is provided, and the method is applied to a vacuum suspension melting furnace. The superheating early warning method is to establish a power-time-temperature prediction model and a cooling liquid temperature rise prediction model.

[0051] Further, the power-time-temperature prediction model is used to describe the dynamic relationship between the coil heating power input and the surface temperature response in the melting process of the vacuum suspension melting furnace. The establishment and application of the model include an offline modeling stage and an online running stage.

[0052] Specifically, in the offline modeling stage, historical running data of the vacuum suspension melting furnace under different charging amounts, different materials and different process parameters are collected, and the historical running data at least includes: heating power , cumulative energy , heating time t, cooling liquid inlet temperature , cooling liquid outlet temperature , crucible / melt surface temperature , etc. The above data are synchronized according to a unified time reference, and after removing, smoothing or filtering the abnormal values, the data are constructed into an input sequence reflecting the power-time behavior characteristics according to a preset time window. The surface temperature corresponding to each time point is taken as the model output target, a power-time-temperature mapping model for time series prediction is established, and the fitting of the model parameters is completed by minimizing the error between the predicted temperature and the actual measured temperature, so as to obtain the power-time-temperature prediction model. The power-time-temperature prediction model is used to establish a function mapping relationship from the power input sequence to the surface temperature output; the calculation logic is: according to the power, energy and running parameters in the current time and the historical time window, the surface temperature value at the corresponding time is predicted.

[0053] Specifically, in the online running phase, the power-time-temperature prediction model takes the real-time collected heating power (sampling frequency , initial sampling frequency ), current heating time and related operating parameters as inputs, calculates the predicted surface temperature at the corresponding time point according to the trained model, and continuously outputs the predicted temperature rise curve over time to obtain the predicted surface temperature dataset of the crucible / melt, which is used as the temperature reference for the current smelting condition.

[0054] Meanwhile, the surface temperature of the crucible / melt is collected in real time, which can be obtained by the ROI mean value or the area array average, and the sampling frequency (initial sampling frequency ) is used to obtain the actual surface temperature rise curve of the crucible / melt, that is, the actual surface temperature dataset of the crucible / melt, and the instantaneous surface temperature difference deviation curve is obtained by comparing the predicted surface temperature dataset with the actual surface temperature dataset, and the instantaneous surface temperature difference deviation dataset is obtained. In this application, the instantaneous surface temperature difference deviation (unit: ℃) is positive when the surface temperature of the crucible / melt is cooler than expected.

[0055] To adapt to different charging amounts, material qualities and furnace conditions, the above machine learning model can also use incremental learning or sliding window retraining to update the parameters regularly, so that it can continuously absorb the latest working condition data during online operation, thereby maintaining the prediction accuracy and model stability. The prediction model established by this data-driven method can automatically depict the complex nonlinear relationship between heating power, time and temperature, and provide a reliable temperature reference curve for the instantaneous deviation calculation and early warning determination of the application.

[0056] Further, to obtain the expected temperature change of the cooling loop, the application further establishes a cooling liquid temperature rise prediction model in addition to the power-time model to improve the accuracy of overheating judgment. The model is based on the combination of the energy conservation simplification principle and the experimental calibration mechanism, and the temperature change of the cooling liquid is predicted by inputting the power, the cooling liquid flow and the inlet temperature.

[0057] In the actual smelting process, the coil input power is not all transmitted to the melt, and part of the heat is dissipated to the cooling liquid channel through the furnace wall, the coil and the structural member. To avoid complex solution of heat dissipation, the application uses an "effective heat transfer proportion parameter" for simplified description. This parameter is calibrated by multiple sets of smelting experimental data in steady state and quasi-steady state, and can reflect the average heat transfer efficiency under different furnace conditions, melt materials, load amounts and coil states. The calibrated can be set in segments according to the power section, frequency or smelting stage, thereby forming an adaptive parameter set.

[0058] Based on this, the cooling liquid temperature rise prediction model is established, including: taking the input coil power as the input heat source of the vacuum suspension smelting system, establishing the cooling loop temperature difference deviation relationship of the cooling liquid, that is, the expected temperature rise of the cooling liquid can be expressed as: ;

[0059] Wherein, is the real-time input power of the coil, is the mass flow rate of the cooling liquid, and cp is the constant pressure specific heat capacity of the cooling liquid. Then the predicted temperature of the cooling liquid outlet is obtained: ; wherein, is the inlet temperature of the cooling liquid, is the cooling loop temperature difference deviation.

[0060] Therefore, according to the above formula, the heating power, the mass flow rate of the cooling liquid, and the actual temperature of the cooling liquid outlet (sampling frequency , initial sampling frequency ) are collected in real time to obtain the cooling loop temperature difference deviation data set.

[0061] In order to further improve the accuracy of the cooling liquid temperature rise prediction model under dynamic working conditions, the data-driven method can be used to correct the above or . By collecting a large amount of historical smelting working condition data, a regression model of and input variables is constructed. The regression model can use polynomial fitting, random forest regression, gradient boosting regression or other commonly used data training methods in the art, and the optimal model parameters are determined by cross-validation. The data-driven model can implicitly compensate for factors such as furnace wall heat dissipation, environmental changes, coil state aging, etc., so that the expected temperature of the cooling liquid remains highly consistent under varying conditions.

[0062] Through the above method, the reliable expected temperature curve of the cooling liquid outlet can be obtained without complex solving of electromagnetic coupling efficiency or structural heat dissipation, which provides a stable reference for subsequent deviation calculation, crust identification and overheating warning.

[0063] Further, the collected instantaneous surface temperature difference deviation data set and the cooling loop temperature difference deviation data set need to be normalized to obtain a comprehensive instantaneous risk score; and the cumulative anomaly index and time weighting are performed to obtain a cumulative anomaly index, which is used to reflect the comprehensive of deviation duration and amplitude.

[0064] Specifically, the instantaneous surface temperature difference deviation is normalized to reduce the dimension effect and facilitate threshold setting: wherein, the normalized instantaneous score of the instantaneous surface temperature difference deviation data set is defined, that is, ; where, is a first reference scale, representing the magnitude of "significant deviation", is the instantaneous surface temperature difference deviation, is a first normalized instantaneous score.

[0065] Specifically, the cooling loop temperature difference deviation is normalized to reduce dimensionality effects and facilitate threshold setting: where the normalized instantaneous score of the cooling loop temperature difference deviation dataset is defined as ; where, is a second reference scale, representing the magnitude of "significant deviation", is the cooling loop temperature difference deviation, is a second normalized instantaneous score.

[0066] Specifically, is the comprehensive risk score (dimensionless), calculated from the normalized deviation and weights, i.e. the comprehensive instantaneous risk score can be expressed as: ; where, is the deviation weight (dimensionless, or set empirically), is the first initial recommended value, is the second initial recommended value, is the comprehensive instantaneous risk score. In this application, is used as the initial recommendation (surface deviation first), but can be optimized based on historical data.

[0067] Specifically, the cumulative anomaly indicator is weighted by time:

[0068] To avoid false positives caused by transient jitter, a weighted accumulation (sliding window integration) is used: ; where, is the time weight function (can be constant 1, or an exponential decay more biased towards the near term) to emphasize recent deviations; W is the accumulation window length, with window length = 120 s as the initial value, adjustable, e.g. 60-300 s.

[0069] Discretization is achieved:

[0070] ;

[0071] where , is the sampling interval.

[0072] In this embodiment, the cumulative deviation indicator D cum(t) The system samples the surface temperature deviation at a preset sampling period, and the deviation is accumulated step by step at each discrete time, so that the deviation evolution process in continuous time is converted into a cumulative anomaly index in the form of a discrete time sequence. Through this discretization processing mode, the calculation process of the cumulative deviation can be consistent with the sampling period of the actual measurement data, which is convenient for real-time implementation in the control system, effectively suppresses the instantaneous noise interference, highlights the continuous accumulation characteristics of the abnormal deviation, and provides stable and reliable judgment basis for subsequent threshold judgment and interlocking protection.

[0073] Further, the vacuum suspension melting system of the vacuum suspension melting furnace defines three levels in the skull process, namely a first warning state, a second warning state and a third warning state.

[0074] Specifically, the condition of the first warning state is: And ; wherein, is the first grading alarm threshold, is the first set threshold.

[0075] Its action is: interface prompt, record events, recommend manual inspection; not automatic intervention.

[0076] Its purpose is: to find the deviation trend early, and manually check whether it is transient or interference.

[0077] Specifically, the condition of the second warning state is: ; or, And exceeds the second set threshold; wherein, is the first grading alarm threshold, is the second grading alarm threshold, is the outlet temperature of the cooling liquid, is the inlet temperature of the cooling liquid. In this application, the second set threshold is the acceptable change rate in the normal melting process. When the cooling water temperature difference rises faster than the acceptable change rate in the normal melting process, it is determined that there is an abnormal heat release trend.

[0078] Its action is: automatically reduce the heating power by a preset proportion (recommended reduction range 10%-30%, or reduce to the next power level), prolong the steady-state observation time, and at the same time trigger sound / light alarm and short message / remote notification, record snapshots (thermal image, power, cooling temperature), etc.

[0079] Its purpose is: to reduce heat input as soon as possible to alleviate the risk under suspected internal heat accumulation.

[0080] Specifically, the condition of the third warning state is: ; or, instantaneous and continuously rising; wherein, is the third hierarchical alarm threshold, is the critical threshold of melt or crucible surface temperature instantaneous change, i.e. the risk boundary of abnormal rise of surface temperature per unit time.

[0081] Its action is: start interlocking protection process - immediately disconnect the heating power, open the emergency cooling circuit (if safety allows), record all historical data and notify maintenance and on-duty management personnel. The system enters the "review pending" mode and needs to be manually reset to restore.

[0082] Its purpose is: to ensure personnel and equipment safety, and to avoid major accidents.

[0083] Wherein, each threshold needs to be optimized on the pilot data according to the ROC curve (hit rate vs. false alarm rate).

[0084] In this embodiment, debouncing and hysteresis processing is also needed: 1) time debouncing for triggering: requires at least to continuously meet the triggering condition before confirming the alarm, to avoid short-term peak triggering; 2) hysteresis for alarm cancellation: the cancellation condition needs to be scored to drop below and last for a period of time to prevent repeated tripping of the alarm.

[0085] In this embodiment, the linkage control action is defined as:

[0086] The action of the secondary warning state (automatic power reduction): reduce to wherein ; and record the execution reason and timestamp. If the power is reduced to the safe range within , it is restored to normal; otherwise, it enters the third warning state processing.

[0087] The action of the third warning state (emergency) is executed in the following order:

[0088] 1) automatically send a shutdown command to the inverter / power supply;

[0089] 2) open the standby cooling circuit or increase the flow;

[0090] 3) shut down suspicious operations related to the safety of operating personnel;

[0091] 4) send multi-channel alarms (local sound and light, SMS, email, SCADA alarm);

[0092] 5) save at least the last 30 minutes of logs, raw high-frequency data snapshots and generate event reports.

[0093] All actions must be consistent with the factory safety management / PLC interlock rules, and the permissions and levels of automatic actions can be set in system configuration (for example, during the trial operation period, turn off the automatic action of the third-level warning state, and only keep the action of the second-level warning state).

[0094] In this embodiment, logs, evidence preservation, and traceability:

[0095] Each warning event needs to be recorded and saved: trigger timestamp, all sensor original time series (at least covering 30 minutes before and after the event), thermal image frames (key frames), decision parameter values (, , , , ), execution actions and execution results, and manual handling records. The logs are saved for the company's specified period (recommended to be no less than 3 years), and support export.

[0096] In this embodiment, safety and abnormal handling:

[0097] When a sensor appears abnormal (offline, out of range, excessive signal noise), the system immediately enters the "reduced trust" mode: reduces the automatic action permission, prompts manual intervention, and compensates with redundant sensors / estimation methods (for example, using coolant and power as alternative signals). The system needs to provide "manual reset" and requires confirmation by two people or a supervisor before the third interlock can be lifted, ensuring safety and reliability.

[0098] Working principle: the working principle of the present application is to establish a corresponding relationship model of power input and heating time in the vacuum suspension smelting process, to predict the temperature rising trend of the crucible surface under normal working conditions, and to compare and analyze the deviation by using the real-time measured surface temperature and the cooling liquid temperature, so as to indirectly infer the thermal state of the upper solid shell and the lower melt. After the smelting starts, the system calculates the theoretical variation curve of the surface temperature according to the set power-time model; in normal conditions, the surface temperature should show a stable rising trend with the power accumulation and heating time. When the actually collected surface temperature is lower than the expected temperature rising curve, and the cooling liquid outlet temperature continues to rise, it indicates that the input heat is not effectively conducted to the surface layer, which may be caused by the formation of solid shell above or overheating of the melt below, resulting in blocked heat transfer. The system calculates the dynamic deviation of the surface temperature and the expected temperature, and the cooling liquid temperature and the theoretical temperature difference, and integrates the deviation on the time axis to reflect the abnormal persistence. This cumulative deviation index can effectively suppress the interference of instantaneous noise or short-term fluctuations, ensuring stable and reliable warning judgment. When the cumulative deviation exceeds the set threshold, the system automatically triggers a warning signal to remind the operator to take intervention measures such as reducing power, prolonging heating time or checking the cooling system, to prevent safety risks such as local overheating of the lower part due to the upper solid shell, crucible ablation or explosion. Through this method, without directly measuring the melt temperature in the furnace, the coupling relationship between power input and surface response can be used to realize non-contact judgment of internal thermal abnormalities, which has the advantages of simple structure, rapid response, safety and reliability, and is especially suitable for thermal safety intelligent warning in high-temperature closed environment such as vacuum suspension smelting furnace.

[0099] Embodiment two

[0100] The present embodiment discloses a method for overheating warning of solid shell in a vacuum suspension smelting system, which is different from embodiment one in that a power-time-temperature prediction model is used.

[0101] Specifically, the power-time-temperature prediction model can be established by mechanism modeling method. Based on the principle of energy conservation, the input electromagnetic power is regarded as the main heat source of the system, and the effective absorbed energy per unit time is calculated to establish the theoretical temperature rising model of the melt and the surface temperature. Specifically, under the conditions of known coil power , electromagnetic coupling efficiency , melt mass , and specific heat capacity c, the temperature change with time is described by the formula, i.e. ;

[0102] wherein, is the heat dissipation term varying with temperature, including radiation heat dissipation, conduction to the crucible and convection loss, etc. The heat dissipation term can be established as a corresponding function according to the Stefan-Boltzmann radiation law, heat conduction equation and cooling liquid temperature, so that the model not only meets the thermodynamic law but also reflects the loss characteristics of the actual smelting structure.

[0103] By numerically solving the above differential equation, the theoretically expected curve between power input and melt temperature can be obtained. For different charging amounts, materials and geometric structures, model parameters such as density, specific heat, radiation coefficient, electromagnetic absorption efficiency, etc. can be adjusted to obtain adaptive temperature rise prediction results. The mechanism model can provide stable and reliable temperature prediction without a large amount of historical data, and is another feasible implementation way of the power-time model described in the present application.

[0104] Specifically, the heating time determines the overall temperature rise process of the melt, and the present application uses the actual heating time and power to jointly construct the expected temperature rise curve. If the surface temperature does not correspondingly rise in sufficient heating time, it means that the heat transfer link is abnormal or the crust has not been melted. This criterion can avoid false judgments caused by instantaneous power fluctuations and improve the stability of temperature rise judgment.

[0105] Specifically, the surface temperature is the key signal to judge the state of the crust. The present application collects the surface temperature in real time and compares it with the model prediction value. When the actual temperature is lower than the expected temperature rise for a long time, it indicates that the surface layer is blocked or the internal temperature stratification is intensified; if the cooling liquid temperature difference also rises, it is more reliable to indicate that the lower part is overheated. The surface temperature is the most direct temperature abnormality discrimination basis.

[0106] Specificly, the melt height reflects the shape change of the melt under the action of electromagnetic force. The present application identifies whether a crust is formed by the height change trend: if the height fluctuation is weakened, fixed at an abnormal position or not matched with the power change, it usually indicates that the surface is solidified or the internal heating is uneven. The height signal can effectively supplement the temperature criterion and improve the abnormal recognition ability.

[0107] Specifically, the thermal image provides a two-dimensional temperature distribution of the surface, and the present application uses the uniformity of the temperature field, the crack bright spot, and the local cold area to identify the thickness of the crust and the internal temperature deviation. When the overall temperature is low but the local crack is high, it is usually that the internal high temperature is shielded; if the temperature field is unevenly intensified, it is usually related to the growth of the crust. The thermal image enhances the spatial recognition ability of the system to abnormal thermal state.

[0108] In summary, by the above scheme, an expected temperature rise model based on power-time relationship is constructed, and a thermal equivalent model, an energy accumulation model or an empirical fitting model is used to establish the theoretical temperature rise trend of the melt surface under normal working conditions, so that the system can judge whether the "input energy-surface temperature response" matches, laying a foundation for subsequent abnormal recognition.

[0109] Through the above scheme, the surface temperature deviation and the cooling liquid temperature difference deviation The double-parameter judgment mechanism breaks through the limitation of single temperature monitoring by simultaneously monitoring surface temperature rise anomaly (surface is colder than expected) and cooling load anomaly (cooling liquid temperature difference is higher), and realizes the identification of comprehensive anomalies such as thermal plugging, crust capping and lower overheating.

[0110] Through the above scheme, the weighted cumulative deviation index is introduced As the core anomaly measure, the instantaneous deviation is integrated and time-weighted by a sliding window and a weight function to realize the comprehensive evaluation of "deviation amplitude + deviation duration", avoid noise interference, and be more consistent with the physical characteristics of thermal anomaly accumulation.

[0111] Through the above scheme, the comprehensive risk score and early warning grading logic are constructed, the normalized deviation index, deviation weight and change rate are combined to construct a quantifiable risk score, and three alarm levels of first-level prompt, second-level automatic intervention and third-level interlock protection are set, effectively reducing the probability of dangerous escalation.

[0112] Through the above scheme, the early warning action closed loop is established, including power reduction, delayed observation and emergency shutdown interlock, the system not only gives an alarm, but also automatically executes power reduction, enhanced cooling or power-off operation, so that the early warning has substantial protection effect, and the safety and reliability of the vacuum smelting equipment is significantly improved.

[0113] Through the above scheme, time de-bouncing, hysteresis judgment, data consistency verification and other robustness enhancement strategies are adopted to ensure that the system can still work stably in high-frequency electromagnetic environment, temperature measurement noise and cooling system fluctuation, and improve the reliability of actual industrial application.

[0114] For those skilled in the art, it is obvious that the present application is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, the scope of the present application is defined by the appended claims rather than the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application.

[0115] In addition, it should be understood that although the present specification is described in terms of embodiments, each embodiment does not contain only one independent technical solution, and the description manner of the specification is only for clarity, those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other implementation manners that those skilled in the art can understand.

Claims

1. A method for overheat warning of a skull in a vacuum suspension melting system, characterized in that, The overheating early warning method comprises: S1: establishing a power-time-temperature prediction model to obtain a predicted surface temperature data set of the crucible / melt; S2: collecting the surface temperature of the crucible / melt in real time to obtain an actual surface temperature data set of the crucible / melt, and obtaining an instantaneous surface temperature difference deviation data set through the predicted surface temperature data set and the actual surface temperature data set; S3: establishing a cooling liquid temperature rise prediction model, collecting the heating power and the cooling liquid mass flow in real time to obtain a cooling loop temperature difference deviation data set; S4: performing normalization processing on the collected instantaneous surface temperature difference deviation data set and the cooling loop temperature difference deviation data set to obtain a comprehensive instantaneous risk score; and performing time weighting on the cumulative anomaly index to obtain the cumulative anomaly index; S5: the vacuum suspension melting system enters a first early warning state according to the comprehensive instantaneous risk score and the cumulative anomaly index through a first rule; Or, the vacuum suspension melting system enters a second early warning state according to the comprehensive instantaneous risk score and the cumulative anomaly index through a second rule; Or, the vacuum suspension melting system enters a third early warning state according to the comprehensive instantaneous risk score and the cumulative anomaly index through a third rule; In step S4, the normalization processing comprises: a normalized instantaneous score defining a set of instantaneous surface temperature difference deviation data, i.e. ; wherein is a first reference scale, is an instantaneous surface temperature difference deviation, is a first normalized instantaneous score; defining a normalized instantaneous score of a cooling circuit temperature difference deviation data set, i.e. ; wherein is a second reference scale, is a cooling circuit temperature difference deviation, is a second normalized instantaneous score; defining a composite instantaneous risk score, i.e. ; wherein, is a first initial recommended value, is a second initial recommended value, is the composite instantaneous risk score; In step S4, the time weighting on the cumulative anomaly index comprises: The cumulative with weight is adopted, that is ; in the formula, is a time weight function, W is a length of a cumulative window, is a cumulative anomaly index; Discretization is achieved, i.e. ; wherein , is the sampling interval; The first rule is: and ; wherein, is a first hierarchical alarm threshold, is a first set threshold; The vacuum suspension melting system entering the first early warning state is that: the interface prompts, records the event, and recommends manual inspection; The second rule is that: ; or, and exceeds a second set threshold; wherein, is a first hierarchical alarm threshold, is a second hierarchical alarm threshold, is an outlet temperature of the coolant, is an inlet temperature of the coolant; The vacuum suspension melting system entering the second early warning state is that: the heating power is automatically reduced by a preset proportion or to the next power level, the steady-state observation time is extended, and the sound or light alarm and the short message or remote notification are triggered at the same time, and the snapshot is recorded; The third rule is: ; or, instantaneous and continuously rising; wherein, is a third hierarchical alarm threshold, is a critical threshold for the instantaneous change in the temperature of the melt or crucible surface, is the outlet temperature of the cooling liquid, is the inlet temperature of the cooling liquid; The vacuum suspension melting system entering the third early warning state is that: the heating power is immediately disconnected, the emergency cooling loop is turned on, all historical data are recorded, and the maintenance and on-duty management personnel are notified.

2. The method of claim 1, wherein the method is characterized by: In step S1, establishing the power-time-temperature prediction model comprises: collecting historical running data of the vacuum suspension melting furnace under different charging amounts, different materials and different process parameters, synchronously processing the historical running data according to a unified time reference, removing, smoothing or filtering abnormal values, constructing an input sequence reflecting the power-time behavior characteristics according to a preset time window, taking the surface temperature of the crucible / melt corresponding to each time point as the model output target, establishing a power-time-temperature mapping model for time series prediction, and completing the fitting of the model parameters by minimizing the error between the predicted temperature and the actual measured temperature; The historical running data comprises: the heating power, the cumulative energy, the heating time, the cooling liquid inlet temperature, the cooling liquid outlet temperature, and the surface temperature of the crucible / melt changing over time.

3. The method of claim 2, wherein the method further comprises: determining the temperature of the molten metal in the crucible; and determining the temperature of the molten metal in the vacuum chamber. In step S1, the heating power and the heating time are collected in real time to obtain the predicted surface temperature data set of the crucible / melt.

4. The method of claim 1, wherein the method further comprises: determining a temperature of the molten metal in the crucible; and determining a temperature of the molten metal in the vacuum chamber. In step S1, the power-time-temperature prediction model is established, which includes: taking the input coil power as the input heat source of the vacuum suspension melting system, establishing the theoretical temperature rising relationship of the melt and the surface temperature by calculating the effective absorbed energy per unit time, i.e. ; wherein, is the real-time input power of the coil, is the electromagnetic coupling efficiency, is the melt mass, c is the specific heat capacity, is the predicted surface temperature change rate, is the heat dissipation term varying with the temperature.

5. The method of claim 4, wherein the method further comprises: determining the temperature of the solidified shell of the vacuum arc remelting system; and comparing the temperature of the solidified shell to a predetermined temperature threshold. In step S1, the heating time is collected in real time to obtain the predicted surface temperature data set of the crucible / melt.

6. The method of claim 1, wherein the method further comprises: determining a temperature of the molten metal in the crucible; and determining a temperature of the molten metal in the vacuum chamber. In step S3, the cooling liquid temperature rise prediction model is established by taking the input coil power as the input heat source of the vacuum suspension smelting system, and establishing a cooling loop temperature difference deviation relationship of the cooling liquid, that is ; wherein, is the real-time input power of the coil, is the mass flow rate of the cooling liquid, cp is the specific heat capacity of the cooling liquid at constant pressure, is the cooling loop temperature difference deviation, is the effective heat transfer proportion parameter.

Images