Molten metal casting liquid level real-time control system and method based on mechanical arm

Abstract

The invention relates to the technical field of liquid level control, in particular to a molten metal casting liquid level real-time control system and method based on a mechanical arm, and the system comprises a mechanical arm positioning module, a liquid level detection module, a motion control module, an anti-disturbance optimization module and a safety interlocking module. According to the method, the relation between liquid level fluctuation and mechanical arm movement is accurately modeled, dynamic changes of the liquid level and the position of the mechanical arm are captured in real time, it is ensured that the influence of the liquid level changes on the precision of the mechanical arm is effectively controlled, the problem that in a traditional method, the liquid level is unstable due to dependence on manual adjustment and hardware limitation is solved, and a liquid level correction data set is provided; the action frequency and peak node distribution of the mechanical arm are optimized, the stability of the liquid level in the pouring process is ensured, the coordination of the liquid level and the action of the mechanical arm is ensured by recognizing and adjusting abnormal distribution caused by liquid level fluctuation in real time, the accurate, quick-response and automatic-adjustment liquid level control method is provided, manual intervention is reduced, and the production efficiency is improved. And the safety and the production efficiency in the pouring process are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of liquid level control, in particular to a molten metal casting liquid level real-time control system and method based on a mechanical arm. BACKGROUND

[0002] The technical field of liquid level control relates to real-time monitoring and adjusting of liquid height in various containers or devices, the core matters of this field include liquid level detection, liquid level signal transmission and liquid level execution control, and is widely used in chemical industry, metallurgy, power, environmental protection and other industries to ensure processing continuity and production safety. The liquid level control of molten metal in metallurgy has particularity due to high temperature and large fluctuation of liquid surface. The traditional liquid level detection method relies on manual observation or uses sensor elements such as float, conductive rod and thermocouple to complete measurement, and the liquid level control is mostly realized by manually operating gates or mechanical baffles for adjusting. The traditional molten metal casting liquid level real-time control system refers to a device and method for real-time monitoring and control of metal liquid surface height during metal liquid casting, and is aimed at the technical matters of maintaining stable liquid level in the casting link. The traditional method mostly uses manual measurement of liquid level and manual adjustment, or uses float rod detection combined with hydraulic valve opening and closing for liquid level control, and some systems also use thermocouple temperature difference to detect liquid level height and drive motor to adjust casting speed.

[0003] The existing technology relies on manual operation or simple sensor technology for liquid level monitoring and control. In the case of large fluctuation of liquid level or high control requirement, this method cannot respond to the dynamic changes of liquid level in real time, resulting in unstable liquid level control. Due to the use of simple sensors such as float and conductive rod, it is difficult to effectively resist the influence of high temperature and strong fluctuation of molten metal, and large measurement errors are easily generated. In the case of rapid change of liquid level, it is difficult to timely adjust the casting speed and control action, and manual intervention is relied on for adjustment. This operation method not only reduces the automation degree of the system, but also increases the risk of human error, and cannot fully guarantee the accuracy and stability of the casting process, thereby affecting the overall production efficiency and safety. SUMMARY

[0004] In order to solve the technical problems existing in the prior art, the present application provides a molten metal casting liquid level real-time control system and method based on a mechanical arm. The technical solution is as follows: On the one hand, a molten metal casting liquid level real-time control system based on a mechanical arm is provided, which comprises: The mechanical arm positioning module obtains the liquid level target value and the initial position distribution curve in the casting pool, extracts the liquid level key parameters, analyzes the influence of liquid level fluctuation on the positioning accuracy of the mechanical arm, sorts out the liquid level and mechanical arm motion relationship model, and obtains the liquid level adaptation table; The liquid level detection module extracts a real-time liquid level value and a target liquid level deviation based on the liquid level adaptation table, identifies a liquid level change trend and a response time, quantifies a liquid level error accumulation effect, induces a liquid level correction weight, and obtains a liquid level correction dataset; The motion control module extracts a mechanical arm action frequency and a peak node within a unit time based on the liquid level correction dataset, performs hierarchical analysis in combination with a liquid level fluctuation characteristic and a mechanical arm action law, and obtains a mechanical arm action level matching table; The anti-disturbance optimization module sorts liquid levels and action demands according to a priority and a proportion, identifies an abnormal distribution relationship between the liquid levels and the mechanical arm actions, adjusts an action sequence through an abnormal node characteristic, and constructs a liquid level driven anti-disturbance control scheme.

[0005] As a further scheme of the present application, the liquid level adaptation table comprises a liquid level interval, a positioning accuracy range, a liquid level compensation factor, and a motion coupling factor, the liquid level correction dataset comprises a liquid level error factor, a response delay index, a change trend weight, and a correction coefficient, the mechanical arm action level matching table comprises a hierarchical action interval, a demand priority level, a fluctuation threshold, and a matching weight, and the liquid level driven anti-disturbance control scheme comprises an abnormal distribution structure, a node adjustment sequence, an abnormal correction factor, and an action allocation factor.

[0006] As a further scheme of the present application, the mechanical arm positioning module comprises: A key parameter extraction submodule extracts liquid level key parameters according to a liquid level target value and an initial position distribution curve, classifies liquid level fluctuation influence factors, and generates a liquid level characteristic distribution table; A positioning accuracy analysis submodule analyzes an influence of liquid level fluctuation on mechanical arm positioning accuracy based on the liquid level characteristic distribution table, identifies an adaptive value of positioning accuracy under different liquid levels, and generates a liquid level and mechanical arm motion relationship model; An adaptation table generation submodule sorts liquid level parameters based on the liquid level and mechanical arm motion relationship model, analyzes a corresponding relationship with a target liquid level value, and generates a liquid level adaptation table.

[0007] As a further scheme of the present application, the liquid level detection module comprises: A liquid level deviation extraction submodule extracts a real-time liquid level value and a target liquid level deviation based on the liquid level adaptation table, records liquid level change trend and response time data, and generates a liquid level deviation data table; An error accumulation quantification submodule quantifies a liquid level error accumulation effect based on the liquid level deviation data table, analyzes a liquid level distribution and demand proportion difference, and generates a liquid level error accumulation effect table; The correction weight induction submodule extracts a liquid level deviation magnitude and a correction frequency based on the liquid level error cumulative effect table, screens a high-frequency error section and labels a deviation direction, induces a liquid level correction weight, and generates a liquid level correction dataset.

[0008] As a further scheme of the present application, the motion control module comprises: The action frequency extraction submodule extracts a mechanical arm action frequency and a peak node within a unit time based on the liquid level correction dataset, classifies mechanical arm action priority data, and generates an action frequency distribution table. The fluctuation characteristic analysis submodule analyzes fluctuation node states based on the action frequency distribution table, in combination with liquid level fluctuation characteristics and mechanical arm action rules, and generates a liquid level fluctuation state table. The hierarchical analysis submodule performs mechanical arm action and fluctuation multidimensional comparison based on the liquid level fluctuation state table, screens action adaptation relationships, and generates a mechanical arm action hierarchical matching table.

[0009] As a further scheme of the present application, the anti-disturbance optimization module comprises: The priority sorting submodule extracts a liquid level peak position and an action difference threshold based on the mechanical arm action hierarchical matching table, analyzes liquid level contribution degrees under a unit action difference, and generates a liquid level priority sorting table. The abnormal distribution identification submodule extracts a main channel node abnormal change trajectory based on the liquid level priority sorting table, identifies abnormal equilibrium points and deviation directions, and generates an abnormal distribution relationship table. The abnormal adjustment submodule adjusts an action sequence through abnormal node characteristics based on the abnormal distribution relationship table, extracts a node abnormal change frequency and an amplitude sequence, counts a deviation amplitude and a continuous number of abnormal over-limit nodes, divides stable intervals and abnormal transition sections, and generates a liquid level driving anti-disturbance control scheme.

[0010] As a further scheme of the present application, the liquid level peak position and the action difference threshold refer to setting a liquid level peak position according to a difference between monitoring of liquid level changes and actions, and calculating a corresponding action difference threshold. The liquid level priority sorting table is obtained by extracting a liquid level peak position and an action difference threshold through a mechanical arm action hierarchical matching table, and analyzing liquid level contribution degrees under a unit action difference. The main channel node abnormal change trajectory refers to recording a trajectory of abnormal changes of a main channel node through real-time monitoring.

[0011] As a further scheme of the present application, the system further comprises a safety interlocking module. The safety interlocking module monitors the liquid level state and the mechanical arm action execution based on the liquid level driven anti-disturbance control scheme, compares the unmet action demand and the remaining capacity in real time, fills the abnormal gap by adjusting the action allocation sequence, and generates a safety interlocking global optimization execution scheme. The safety interlocking global optimization execution scheme includes residual adjustment parameters, execution sequence configuration, remaining capacity utilization rate, and adjustment completion criterion.

[0012] As a further scheme of the present application, the safety interlocking module comprises: The state monitoring submodule collects liquid level node values and mechanical arm action feedback based on the liquid level driven anti-disturbance control scheme, records the jump time and deviation amplitude, and generates a state monitoring data table. The demand comparison submodule extracts the unmet action corresponding time point based on the state monitoring data table, identifies the remaining capacity and instantaneous gap, matches the target gap and capacity segment, and generates a demand comparison result table. The dynamic adjustment submodule fills the abnormal gap by adjusting the action allocation sequence based on the demand comparison result table, identifies the capacity gap node and response lag segment, updates the action output timing and liquid level curve, and generates a safety interlocking global optimization execution scheme.

[0013] On the other hand, the real-time control method of the molten metal casting liquid level based on the mechanical arm is executed based on the above-mentioned real-time control system of the molten metal casting liquid level based on the mechanical arm, comprising the following steps: S1: According to the target value of the liquid level in the pouring pool and the initial position distribution curve, the liquid level key parameters and the target value are extracted, the liquid level fluctuation influence factor is normalized, the relationship between the liquid level and the mechanical arm movement is matched, and a liquid level adaptation table is generated. S2: Based on the liquid level adaptation table, the target and real-time liquid level deviation value is extracted, the deviation amplitude and response duration is analyzed, the liquid level change trend and response time data is screened, and a liquid level correction data set is generated. S3: Based on the liquid level correction data set, the mechanical arm action frequency and peak node in unit time are extracted, the liquid level response value is associated to identify abnormal transition points and stable recovery points, the echo time and jump boundary in the fluctuation interval are extracted, and a mechanical arm action level matching table is generated. S4: Based on the mechanical arm action level matching table, the high-frequency liquid level priority section and the fluctuation peak position are analyzed, the step change node is identified and the main channel and compensation path are reconstructed, and a liquid level driven anti-disturbance control scheme is constructed. S5: Based on the liquid level driven anti-disturbance control scheme, the unmet action demand parameters and the abnormal state value of the key node are screened, the offset frequency peak value is extracted and the action adjustment logic is corrected, and a safety interlocking global optimization execution scheme is generated.

[0014] The technical scheme provided by the embodiment of the present application has at least the following beneficial effects: By accurately modeling the relationship between the liquid level fluctuation and the mechanical arm movement, the dynamic changes of the liquid level and the mechanical arm position are captured in real time, ensuring that the influence of the liquid level change on the mechanical arm precision is effectively controlled, solving the problem of unstable liquid level control caused by the dependence on manual adjustment and hardware limitations in the traditional method. Based on the correlation between the liquid level and the mechanical arm, a liquid level correction dataset is proposed, and hierarchical analysis is performed in combination with the liquid level fluctuation characteristics and the movement law of the mechanical arm, thereby optimizing the action frequency of the mechanical arm and the distribution of peak nodes, ensuring stable control of the liquid level during pouring. At the same time, by identifying and adjusting the abnormal distribution caused by the liquid level fluctuation in real time, the coordination between the liquid level and the mechanical arm action is ensured, thereby realizing the generation of a global optimization and safety interlocking execution scheme. The scheme provides a highly accurate, rapid response and automatic adjustment liquid level control method, reducing manual intervention and improving the safety and production efficiency during pouring. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0016] Figure 1 is a schematic diagram of the real-time control system of the molten metal pouring liquid level based on the mechanical arm provided by the embodiment of the present application; Figure 2 is a schematic diagram of the system framework of the present application; Figure 3 is a flowchart of the mechanical arm positioning module in the present application; Figure 4 is a flowchart of the liquid level detection module in the present application; Figure 5 is a flowchart of the motion control module in the present application; Figure 6 is a flowchart of the anti-disturbance optimization module in the present application; Figure 7 is a flowchart of the safety interlocking module in the present application; Figure 8 is a flowchart of the real-time control method of the molten metal pouring liquid level based on the mechanical arm provided by the embodiment of the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the present application will be described below in combination with the drawings.

[0018] In the embodiments of the present application, the words such as "for example", "for instance", "such as", etc. are used to indicate an example, an illustration or an example. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the word "example" is intended to present the concept in a specific manner. In addition, in the embodiments of the present application, the meaning expressed by "and / or" can be both, or can be one of the two.

[0019] In the embodiments of the present application, "image" and "picture" can be used interchangeably at times. It should be pointed out that when the distinction is not emphasized, the meanings expressed are consistent. "Of", "corresponding" and "corresponding" can be used interchangeably at times. It should be pointed out that when the distinction is not emphasized, the meanings expressed are consistent.

[0020] In the embodiments of the present application, sometimes the subscript such as W1 can be written in the form of non-subscript such as W1. When the distinction is not emphasized, the meanings expressed are consistent.

[0021] In order to make the technical problems, technical solutions and advantages of the present application clearer, specific embodiments will be described in detail below with reference to the drawings.

[0022] The embodiments of the present application provide a molten metal casting liquid level real-time control system based on a mechanical arm, as shown in Figures 1-2 The schematic diagram of the molten metal casting liquid level real-time control system based on the mechanical arm, the system comprises: The mechanical arm positioning module obtains a target value of a liquid level in a casting pool and an initial position distribution curve, extracts key parameters of the liquid level, analyzes the influence of liquid level fluctuation on the positioning accuracy of the mechanical arm, sorts out a relationship model between the liquid level and the movement of the mechanical arm, and obtains a liquid level adaptation table; The liquid level detection module extracts a real-time liquid level value and a target liquid level deviation based on the liquid level adaptation table, identifies a liquid level change trend and a response time, quantifies a liquid level error accumulation effect, induces a liquid level correction weight, and obtains a liquid level correction data set; The motion control module extracts the action frequency and peak node of the mechanical arm per unit time based on the liquid level correction data set, performs hierarchical analysis in combination with the liquid level fluctuation characteristics and the action law of the mechanical arm, and obtains a mechanical arm action level matching table; The anti-disturbance optimization module sorts the liquid level priority and the action demand ratio according to the mechanical arm action level matching table, identifies the abnormal distribution relationship between the liquid level and the action of the mechanical arm, adjusts the action sequence through the abnormal node characteristics, and constructs a liquid level driven anti-disturbance control scheme; The safety interlocking module is based on a liquid level driven anti-disturbance control scheme, monitors the liquid level state and the mechanical arm action execution, compares the unsatisfied action demand and the remaining capacity in real time, fills in the abnormal gap by adjusting the action allocation sequence, and generates a safety interlocking global optimization execution scheme.

[0023] The liquid level adaptation table includes a liquid level interval, a positioning accuracy range, a liquid level compensation factor, and a motion coupling factor. The liquid level correction data set includes a liquid level error factor, a response delay index, a change trend weight, and a correction coefficient. The mechanical arm action level matching table includes a hierarchical action interval, a demand priority level, a fluctuation threshold, and a matching weight. The liquid level driven anti-disturbance control scheme includes an abnormal distribution structure, a node adjustment sequence, an abnormal correction factor, and an action allocation factor. The safety interlocking global optimization execution scheme includes a residual adjustment parameter, an execution sequence configuration, a remaining capacity utilization rate, and a regulation completion criterion.

[0024] Specifically, as shown in Figure 2 , 3 The mechanical arm positioning module includes: The key parameter extraction submodule extracts liquid level key parameters, classifies liquid level fluctuation influencing factors, and generates a liquid level feature distribution table according to the liquid level target value in the pouring pool and the initial position distribution curve. By obtaining the liquid level target value in the pouring pool, which is set to 50.0 mm, and collecting the data of the initial position distribution curve in the pouring pool, the data is obtained by measuring the liquid level reading of the mechanical arm at different initial positions. For example, when the mechanical arm is at P1 (100 mm, 100 mm, 50 mm), the liquid level sensor displays 49.8 mm; at P2 (120 mm, 100 mm, 50 mm), the liquid level reading is 50.2 mm; at P3 (100 mm, 120 mm, 50 mm), the liquid level is 49.9 mm. Based on the data, the key parameters of the liquid level are first extracted, for example, the average value of all initial liquid level readings is calculated, the average value of the initial liquid level is 49.97 mm, and the maximum value of the liquid level is recorded as 50.2 mm and the minimum value is 49.8 mm, which provides a basis for subsequent fluctuation analysis, identifies and classifies the factors affecting the liquid level, and analyzes the running state of the pouring equipment, finds that the main influencing factors include the instantaneous flow of the feeding pump, the degree of blockage of the discharge port and the material temperature. Specifically, under the initial state, the flow of the feeding pump is 2.0 liters per minute, the degree of blockage of the discharge port is 5%, and the material temperature is 80℃. By associating the key parameters of the liquid level with the influencing factors, a liquid level feature distribution table is generated, which reflects the changes of the liquid level under different initial conditions and the specific situation of being affected by the influencing factors.

[0025] The positioning accuracy analysis submodule analyzes the influence of liquid level fluctuation on the positioning accuracy of the mechanical arm based on the liquid level feature distribution table, identifies the positioning accuracy adaptation value under different liquid level conditions, and generates a liquid level and mechanical arm motion relationship model; Based on the liquid level feature distribution table, which records different initial liquid levels and their corresponding fluctuation influence factor data, for example, when the liquid level is 49.8 mm, the surface fluctuation amplitude is 0.18 mm, the feed flow is 2.0 L / min, and the material temperature is 80℃, the influence of liquid level fluctuation on the positioning accuracy of the mechanical arm is analyzed. The specific process is as follows: first, the calibration positioning accuracy of the mechanical arm under the static liquid surface is obtained, which is set to ±0.05 mm; second, according to the fluctuation amplitude in the liquid level feature distribution table, the actual positioning deviation of the mechanical arm under the dynamic liquid surface is experimentally measured, for example, when the liquid level fluctuation amplitude is 0.18 mm, the actual positioning deviation range of the mechanical arm expands to ±0.15 mm, and when the liquid level fluctuation amplitude is 0.05 mm, the positioning deviation range shrinks to ±0.08 mm. On this basis, the positioning accuracy adaptation value under different liquid level conditions is identified. By setting the division interval of the liquid level fluctuation amplitude, greater than 0.15 mm is defined as high fluctuation, less than or equal to 0.05 mm is defined as low fluctuation, and between 0.05 mm and 0.15 mm is defined as medium fluctuation. For liquid level conditions in the high fluctuation interval, the positioning accuracy adaptation value is set to ±0.20 mm; for liquid level conditions in the medium fluctuation interval, the adaptation value is set to ±0.12 mm; for liquid level conditions in the low fluctuation interval, the adaptation value is set to ±0.07 mm. The adaptation value is obtained through actual testing, for example, in one test, when the liquid level fluctuation amplitude is 0.17 mm, the positioning accuracy adaptation value is adjusted to ±0.20 mm, and the mechanical arm operation success rate is 98%. This test is repeated 20 times to verify the effectiveness of the adaptation value. Finally, the liquid level and mechanical arm motion relationship model is generated according to the adaptation value, which describes how the liquid level state determines the allowed positioning accuracy of the mechanical arm.

[0026] The adaptation table generation submodule generates a liquid level adaptation table based on the liquid level and mechanical arm motion relationship model, sorts the liquid level parameters, and analyzes the corresponding relationship with the target liquid level value. Based on the liquid level and mechanical arm motion relationship model, for example, the model specifies that when the liquid level fluctuation amplitude is less than 0.05 millimeters, the positioning accuracy adaptation value of the mechanical arm is ±0.07 millimeters, when the fluctuation amplitude is between 0.05 millimeters and 0.15 millimeters, the adaptation value is ±0.12 millimeters, and when the fluctuation amplitude is greater than 0.15 millimeters, the adaptation value is ±0.20 millimeters, the liquid level parameters are sorted, and the specific operation is to select a typical range of pouring pool liquid level from 48.0 millimeters to 52.0 millimeters, with a step of 0.1 millimeter, to generate a series of discrete liquid level points, for example, 48.0 millimeters, 48.1 millimeters, 48.2 millimeters, and so on to 52.0 millimeters, and the liquid level points are a total of 41, then the corresponding relationship between each liquid level parameter and the target liquid level value is analyzed, while considering the fluctuation amplitude information in the liquid level feature distribution table, for example, for the liquid level point 49.0 millimeters, by calling the liquid level and mechanical arm motion relationship model, and combining the fluctuation amplitude (for example, 0.16 millimeters) of the liquid level under the liquid level feature distribution table query, the positioning accuracy adaptation value is calculated to be ±0.20 millimeters, for the liquid level point 50.0 millimeters (i.e. the target liquid level value), the fluctuation amplitude is low (for example, 0.03 millimeters), and the positioning accuracy adaptation value is calculated to be ±0.07 millimeters, the corresponding relationship establishes a mapping that associates each liquid level value with the positioning accuracy that the mechanical arm should adopt under the liquid level condition, for example, in actual operation, if the real-time liquid level detection is 49.0 millimeters, the corresponding positioning accuracy adaptation value of ±0.20 millimeters is immediately extracted or calculated from the model, and finally, a liquid level adaptation table is generated, which lists each liquid level parameter and its corresponding positioning accuracy adaptation value in detail.

[0027] Specifically, as shown in Figure 2 , 4 The liquid level detection module includes: The liquid level deviation extraction submodule extracts the real-time liquid level value and the target liquid level deviation based on the liquid level adaptation table, records the liquid level change trend and response time data, and generates a liquid level deviation data table. Based on the liquid level adaptation table, the table stores the corresponding mechanical arm positioning accuracy adaptation value of each liquid level value, for example, when the liquid level is 50.0 millimeters, the adaptation accuracy is ±0.07 millimeters, when the liquid level is 49.5 millimeters, the adaptation accuracy is ±0.12 millimeters, the real-time liquid level value is obtained, for example, the real-time liquid level data in the pouring pool is collected every 0.1 second through the liquid level sensor, at a certain time T0, the real-time liquid level value is 49.7 millimeters, at this time the target liquid level value is set to 50.0 millimeters, the real-time liquid level value and the target liquid level deviation are extracted, the difference between the real-time liquid level value and the target liquid level value is calculated, for example, at T0, the liquid level deviation = 49.7 millimeters-50.0 millimeters =-0.3 millimeters, the liquid level change trend and response time data are recorded, specifically, after collecting the deviation-0.3 millimeters at T0, 49.5 millimeters is collected at subsequent time T1 (T0+1 second), the deviation is-0.5 millimeters, 49.2 millimeters is collected at T2 (T0+2 seconds), the deviation is-0.8 millimeters, by comparing the continuous deviation values, it is judged that the liquid level is in a continuous downward trend, and the time interval of each deviation update is recorded as the response time, for example, the response time from T0 to T1 is 1 second, and the response time from T1 to T2 is 1 second, the data reflects the dynamic process of liquid level change, for example, if the liquid level deviation deviates downward for three times in a row, and the deviation amplitude of each time is greater than the last time, it is determined that the trend is "significant downward", finally, a liquid level deviation data table is generated, which details each sampling time (for example, T0, T1, T2), the corresponding real-time liquid level value (49.7, 49.5, 49.2), the target liquid level value (50.0), the liquid level deviation (-0.3, -0.5, -0.8), the liquid level change trend (for example, "down", "down", "significant down") and the response time (for example, 1 second, 1 second).

[0028] The error accumulation quantification sub-module quantifies the liquid level error accumulation effect based on the liquid level deviation data table, analyzes the difference between the liquid level distribution and the demand ratio, and generates a liquid level error accumulation effect table. Based on the liquid level deviation data table, which includes time-series liquid level deviation, trends, and response times (e.g., at continuous sampling time points T0, T1, and T2, the liquid level deviations are -0.3 mm, -0.5 mm, and -0.8 mm, respectively, with a response time of 1 second for each), the cumulative effect of liquid level error is quantified. Specifically, a cumulative time window is set, such as 10 seconds. At the current moment, the arithmetic mean of the absolute values ​​of all liquid level deviations over the past 10 seconds is calculated as the cumulative error within that time window. The difference between the liquid level distribution and the demand ratio is analyzed by querying the liquid level adaptation table to obtain the positioning accuracy adaptation value corresponding to the current real-time liquid level. For example, if the current real-time liquid level is 49.2 mm, the liquid level adaptation table is queried. The table shows that the positioning accuracy adaptation value is ±0.12 mm. The current liquid level deviation (e.g., -0.8 mm) is compared with this adaptation accuracy value to calculate the required proportional difference. This difference is defined as the ratio of the absolute value of the current liquid level deviation to the positioning accuracy adaptation value. For example, the required proportional difference = |-0.8 mm| / 0.12 mm = 6.67. A ratio greater than 1 indicates that the current liquid level deviation has exceeded the allowable range of the adaptation accuracy and needs attention. The larger the ratio, the more severe the deviation. For example, when the required proportional difference is greater than 5, it indicates that the error accumulation effect has reached a significant level. Finally, a liquid level error accumulation effect table is generated, which records the liquid level deviation, error accumulation, and required proportional difference corresponding to each sampling time point.

[0029] The correction weight summarization submodule extracts the magnitude and correction frequency of liquid level deviation based on the liquid level error cumulative effect table, filters high-frequency error segments and marks the deviation direction, summarizes the liquid level correction weight, and generates a liquid level correction dataset. Based on the liquid level error accumulation effect table, the liquid level deviation, error accumulation and demand proportion difference at each sampling time are recorded. By setting the medium deviation threshold of 0.4 mm and the large deviation threshold of 0.7 mm, the absolute value of the real-time liquid level deviation is divided into three categories: low (<0.4 mm), medium (0.4-0.7 mm) and high (>0.7 mm). At the same time, the number of times that the absolute value of the liquid level deviation exceeds the medium deviation threshold in a rolling time window (such as 30 seconds) is counted, which is used to calculate the correction frequency. For example, at T0, T1 and T2, the liquid level deviation is -0.3 mm, -0.5 mm and -0.8 mm respectively, T0 is low deviation, T1 is medium deviation, and T2 is high deviation, and if there are 5 times of deviation exceeding 0.4 mm in 30 seconds, the correction frequency is 5 times, by identifying the section with significantly higher correction frequency than the average level (for example, more than 5 times in 30 seconds), it is marked as a high-frequency error section, and the deviation direction (liquid level too high or too low) is determined. For example, if the liquid level deviation is continuously negative and the correction frequency reaches 8 times in the continuous time period from T0 to T10 seconds, this section is identified as a high-frequency error section of low liquid level, and the deviation direction is negative. Based on the deviation magnitude and correction frequency, the liquid level correction weight is calculated, the initial weight is set to 0.1, the weight is multiplied by 1.5 when the deviation magnitude is medium, and multiplied by 2.0 when the magnitude is high. If the correction frequency is significantly high in the high-frequency error section, multiply by a coefficient of 1.2. Through this method, a liquid level correction data set is generated, recording the start and end time of the high-frequency error section, the deviation magnitude, the correction frequency, the deviation direction and the correction weight.

[0030] Specifically, as shown in Figure 2 、 5 , the motion control module includes: The action frequency extraction sub-module extracts the mechanical arm action frequency and peak node in unit time based on the liquid level correction data set, classifies the mechanical arm action priority data, and generates an action frequency distribution table. Based on the liquid level correction data set, the data set contains correction weight, deviation direction and other information of different error sections, for example, in the time period T0 to T10, it is identified as a high frequency error section, the deviation direction is negative, and the correction weight is 0.24, the action frequency and peak node of the mechanical arm in unit time are extracted, specifically, set the unit time as 5 seconds, real-time monitor and count all the actions of the mechanical arm in each 5 second time window as the action frequency, for example, in T0 to T5 seconds, the mechanical arm performs 3 actions (1 times of feeding port adjustment, 2 times of detection), the action frequency is 3, at the same time, define the peak node as the time when the action frequency exceeds the preset threshold (for example, 4 times every 5 seconds), for example, if in T5 to T10 seconds, the mechanical arm performs 5 actions (2 times of feeding port adjustment, 2 times of stirring, 1 time of detection), then T10 is marked as the peak node, then classify the mechanical arm action priority data, specifically, according to the influence degree of the mechanical arm action on the liquid level stability, the action is divided into different priorities, for example, the action directly affecting the liquid level adjustment (such as feeding port adjustment, discharge port adjustment) priority is set to high (priority value 3), the auxiliary adjustment action (such as stirring, oscillation) priority is set to medium (priority value 2), the monitoring action (such as liquid level detection, visual detection) priority is set to low (priority value 1), then, combined with the deviation direction and correction weight in the liquid level correction data set, the priority of the related action is dynamically adjusted, for example, if the liquid level correction data set shows that the liquid level is continuously low and the correction weight is high (0.24), the priority of the feeding port adjustment action is temporarily increased from 3 to 4, and the priority of the discharge port adjustment action is appropriately reduced, finally, the action frequency distribution table is generated, which records the start and end time of each unit time window, the total action frequency, the peak node identification, and the priority value and execution times of each specific action (such as feeding port adjustment, stirring, detection), the action frequency distribution table provides quantitative data of the mechanical arm action for subsequent fluctuation characteristic analysis.

[0031] The fluctuation characteristic analysis submodule analyzes the fluctuation node state based on the action frequency distribution table, combined with the liquid level fluctuation characteristics and the mechanical arm action rule, to generate a liquid level fluctuation state table. Based on the action frequency distribution table, which records the action frequency, peak node and action priority of the mechanical arm in different time periods, for example, in T0 to T5 seconds, the action frequency is 3, the feed inlet adjustment priority is 3, at the same time, combined with the liquid level correction data set, which provides information such as liquid level deviation magnitude, correction frequency and correction weight, for example, in T0 to T5 seconds, the liquid level deviation is-0.5mm, the correction frequency is high, and the correction weight is 0.24, combined with the liquid level fluctuation characteristics and the mechanical arm action law, the specific analysis is as follows: by comparing the execution of specific actions in the action frequency distribution table with the liquid level fluctuation state indicated in the liquid level correction data set, for example, if the liquid level correction data shows that the liquid level continues to drop and the deviation magnitude is medium (-0.5mm) in T0 to T5 seconds, but the action frequency distribution table shows that the execution frequency of the feed inlet adjustment action is lower than the preset reference adjustment frequency (for example, at least once every 5 seconds), and the frequency of the detection action is too high (for example, 2 times every 5 seconds), which indicates that the action of the mechanical arm cannot effectively respond to the liquid level drop trend, then, analyze the fluctuation node state, define the fluctuation node state by comprehensively judging the above comparison results, for example, if the liquid level deviation continues to increase and the key adjustment action is insufficient, the fluctuation node state is marked as "obvious liquid level instability trend, action response lag", if the liquid level deviation decreases and the key adjustment action is sufficient, it is marked as "liquid level tends to be stable, action response is good", for example, in T0 to T5 seconds, the liquid level deviation changes from-0.3mm to-0.5mm, and the feed inlet adjustment action is executed only once, the node state is judged as "obvious liquid level instability trend, action response lag", the judgment is based on the condition that the liquid level deviation rate (for example, -0.2mm / s) exceeds the safety threshold (for example, -0.1mm / s), finally, generate a liquid level fluctuation state table, which details each analysis time period (for example, T0 to T5 seconds), the main fluctuation characteristics of the liquid level in this time period (for example, "continuous decline, medium deviation level"), the main action law of the mechanical arm (for example, "frequent detection action, insufficient feed adjustment") and the fluctuation node state judged accordingly.

[0032] The hierarchical analysis submodule performs mechanical arm action and fluctuation multi-dimensional comparison based on the liquid level fluctuation state table, filters action adaptation relationship, and generates a mechanical arm action level matching table. Based on the liquid level fluctuation state table, which records the liquid level fluctuation characteristics, mechanical arm action law and fluctuation node state in different time periods, for example, from T0 to T5 seconds, the fluctuation node state is "obvious liquid level instability trend, action response lag", according to the fluctuation node state, the actual action sequence of the mechanical arm in this time period (such as feed inlet adjustment, detection) is extracted from the action frequency distribution table, and compared with the predefined ideal action sequence, for the "obvious liquid level instability trend, action response lag" state, the ideal action sequence should include the feed inlet rapid adjustment with priority 4 and the discharge outlet fine adjustment with priority 3, by quantifying the matching degree of the actual action sequence and the ideal action sequence, the matching degree percentage is calculated, the matching degree calculation method is to sum the priority of each action in the ideal action sequence, sum the priority value of the matched action in the actual action sequence, and then calculate the matching degree. For example, the total priority of the ideal sequence is 7, the priority of the actual executed feed inlet adjustment is 3, and the priority of the detection action is 1, then the matching degree is 42.8%. Set the matching degree below 60% as not suitable, 60% to 80% as partially suitable, and above 80% as completely suitable. By analyzing the historical running data, it is found that when the matching degree is below 60%, 85% of the cases will further deviate from the target value. Finally, a mechanical arm action level matching table is generated, which records each fluctuation node state, actual action sequence, ideal action sequence and action adaptation relationship.

[0033] Specifically, as shown in Figure 2 、 6 The anti-disturbance optimization module includes: The priority sorting submodule extracts the liquid level peak position and action difference threshold based on the mechanical arm action level matching table, analyzes the liquid level contribution degree under unit action difference, and generates a liquid level priority sorting table; The liquid level peak position and action difference threshold refer to the difference between the monitoring of the liquid level change and the action, the peak position of the liquid level is set, and the corresponding action difference threshold is calculated; The liquid level priority sorting table is obtained by extracting the liquid level peak position and action difference threshold from the mechanical arm action level matching table, and analyzing the liquid level contribution degree under unit action difference; Based on the mechanical arm action level matching table, the table records the liquid level fluctuation state, the actual and ideal action sequence and the action adaptation relationship. For example, for the fluctuation state with obvious liquid level instability trend, the adaptation relationship is not adapted. The liquid level peak position and action difference threshold are extracted. Specifically, from the mechanical arm action level matching table, the non-adapted fluctuation state is identified, for example, the state is that the liquid level drops from 49.7 mm to 49.2 mm within T0 to T5 seconds, then the liquid level peak position is 49.2 mm (the lowest point), and then the action difference is calculated, that is, the sum of the priorities of the actions in the ideal action sequence that are not actually executed or insufficiently executed, for example, the ideal action is the priority 4 feeding port rapid adjustment and the priority 3 discharging port fine adjustment, and the actual action is only the priority 3 feeding port adjustment, then the action difference is 4 (from the non-executed feeding port rapid adjustment part) + 3 (from the non-executed discharging port fine adjustment), that is, 7, and the action difference threshold is set to 5, which means that the action difference higher than the threshold is considered as significant action missing. The liquid level contribution degree under unit action difference is analyzed, specifically, by correlating the action difference with the deviation of the liquid level peak position (for example, the deviation of 49.2 mm and the target 50.0 mm is -0.8 mm), the influence degree of a unit action difference on the liquid level deviation is calculated, for example, if the action difference of 7 leads to a liquid level deviation of -0.8 mm, then the liquid level contribution degree under unit action difference is 0.8 mm / 7=0.114 mm / unit action difference, which means that each missing 1 priority unit action execution leads to an additional deviation of 0.114 mm of the liquid level. The contribution degree value is obtained based on historical data regression analysis, for example, by fitting the action difference and the final liquid level deviation of 100 groups of non-adapted cases, the fitting coefficient is 0.114, and finally, the liquid level priority sorting table is generated, which sorts the actions that need to be executed in priority according to the size of the liquid level contribution degree under unit action difference.

[0034] The abnormal distribution identification submodule extracts the main channel node abnormal change trajectory based on the liquid level priority sorting table, identifies the abnormal equilibrium point and the deviation direction, and generates the abnormal distribution relationship table. The main channel node abnormal change trajectory refers to recording the trajectory of the abnormal change of the main channel node through the real-time monitoring main channel node. Based on the liquid level priority ranking table, which indicates the priority and contribution of different actions to the stability of the liquid level, for example, the feed inlet rapid adjustment has a high liquid level contribution, the abnormal change trajectory of the main channel node is extracted, specifically, the data of the main channel node of the pouring pool is continuously monitored, such as real-time liquid level sensor reading (millimeters), feed pump flow (liters / minute) and discharge valve opening (percentage), an abnormal change rate threshold of 0.1 millimeter / second is set, when the change rate of the real-time liquid level value exceeds this threshold for 3 seconds, the sequence of the liquid level value and the time point in this period are recorded, forming an abnormal change trajectory, for example, from T10 to T15, the liquid level continuously decreases from 49.5 mm to 48.8 mm, the change rate reaches 0.14 mm / s, this sequence of data constitutes an abnormal change trajectory, then, the abnormal equilibrium point and the deviation direction are identified, specifically, when the liquid level change rate is less than the stable threshold of 0.01 mm / s for 5 seconds, but the absolute value of the deviation of the liquid level value from the target liquid level (50.0 mm) is greater than the deviation threshold of 0.5 mm, the liquid level value is identified as an abnormal equilibrium point, and according to its relative height from the target liquid level, the deviation direction is determined, for example, at T20, the liquid level is stable at 49.0 mm, the change rate is 0.005 mm / s, then 49.0 mm is identified as an abnormal equilibrium point, and the deviation direction is lower than the target, finally, an abnormal distribution relationship table is generated, which records each identified abnormal change trajectory (including start time, end time, liquid level sequence), abnormal equilibrium point (if any) and its deviation direction.

[0035] The abnormal adjustment submodule generates a liquid level driving anti-disturbance control scheme based on the abnormal distribution relationship table, adjusts the action sequence through the abnormal node characteristics, extracts the node abnormal change frequency and amplitude sequence, counts the offset amplitude and duration of the abnormal over-limit node, divides the stable interval and the abnormal transition section, and generates a liquid level driving anti-disturbance control scheme. Based on the abnormal distribution relationship table, which records the abnormal change trajectory of the liquid level, the abnormal equilibrium point and the offset direction, for example, there is a trajectory of continuous decline of the liquid level, and an abnormal equilibrium point below the target is formed at 49.0 millimeters, the action sequence is adjusted according to the abnormal node characteristics in the abnormal distribution relationship table (for example, the liquid level is continuously low and has reached the abnormal equilibrium point), combined with the liquid level priority ranking table, the high-priority correction action (for example, increasing the feed rate) is identified and inserted into the front end of the current mechanical arm action sequence, for example, if the current action sequence is "detection-stirring", after identifying that the liquid level is low, it is adjusted to "increase the feed rate-detection-stirring", the node abnormal change frequency and amplitude sequence is extracted, specifically, in a rolling time window (for example, 60 seconds), the number of times that the liquid level deviation absolute value exceeds the deviation threshold of 0.5 millimeters is counted as the abnormal change frequency, for example, in the past 60 seconds, the liquid level has deviated more than 0.5 millimeters for 5 times, then the frequency is 5, at the same time, the maximum deviation amplitude of each abnormal change is recorded to form the amplitude sequence, for example, [-0.8 millimeters, -0.9 millimeters, -0.7 millimeters], the offset amplitude and continuous time interval number of abnormal overrun nodes are counted, specifically, when the liquid level deviation absolute value exceeds the overrun threshold of 1.0 millimeter, the time is marked as an abnormal overrun node, the offset amplitude (for example, -1.2 millimeters) at this time is recorded, and the continuous time interval number is counted, for example, if the liquid level deviation exceeds -1.0 millimeters for 3 consecutive samplings (1 second each time), then the continuous number is 3, on this basis, the stable interval and the abnormal transition section are divided, specifically, the time interval in which the liquid level deviation absolute value is less than 0.2 millimeters for 5 seconds and the liquid level change rate is less than 0.01 millimeters / second is defined as the stable interval, and the time interval from the first detection of abnormal change to the entry into the stable interval is defined as the abnormal transition section, for example, the time from the start of the liquid level decline to the final stabilization near the target value is 20 seconds, then the 20 seconds is the abnormal transition section, finally, the liquid level driving anti-disturbance control scheme is generated, which includes the adjusted mechanical arm action sequence, the execution parameters of each action (for example, the feed pump flow is increased by 0.5 L / min), and the detailed conditions for triggering the adjustment.

[0036] Specifically, as shown in Figure 2 , 7 , the safety interlocking module includes: The state monitoring submodule acquires the liquid level node value and the mechanical arm action feedback based on the liquid level driving anti-disturbance control scheme, records the jump time and the deviation amplitude, and generates a state monitoring data table. Based on a level-driven anti-disturbance control scheme, this scheme specifies the sequence of actions and parameters that the robotic arm should execute under different abnormal conditions. For example, when the liquid level is below 49.0 mm, the action of "increasing the feed rate by 0.5 L / min" should be executed immediately. The scheme collects liquid level node values ​​and robotic arm action feedback. Specifically, the real-time liquid level value is collected every 0.1 seconds via a liquid level sensor (e.g., 49.0 mm). Simultaneously, action feedback information sent by the robotic arm controller is received. For example, at time T0, the action of increasing the feed pump flow rate from 2.0 L / min to 2.5 L / min has been completed. The jump time and deviation amplitude are recorded. Specifically, when the instantaneous rate of change of the real-time liquid level value (e.g., greater than 0.5 L / min) is recorded... When the liquid level sensor reading changes significantly (in millimeters per second) or the robotic arm's motion state, it is recorded as a jump moment, and the deviation of the liquid level from the target value (50.0 mm) is calculated. For example, at time T1 (T0+5 seconds), the liquid level sensor reading rapidly rises from 49.0 mm to 49.2 mm, which is recorded as a jump moment, and the deviation is 49.2 mm - 50.0 mm = -0.8 mm. This jump is caused by the robotic arm increasing the feeding rate. Finally, a status monitoring data table is generated, which records in detail each sampling moment, the corresponding liquid level node value, the robotic arm motion feedback (including the motion name and actual parameters), the jump moment identifier (e.g., a Boolean value indicating whether it is a jump moment), and the deviation magnitude.

[0037] The demand comparison submodule extracts the time points corresponding to unmet actions based on the status monitoring data table, identifies the remaining capacity and instantaneous gaps, matches the target gaps with the capacity segments, and generates a demand comparison result table. Based on the state monitoring data table, which contains real-time liquid level values, mechanical arm action feedback, jump time, and deviation amplitude, etc., for example, at T0 time, the liquid level is 49.0 mm, and the mechanical arm performs the action of "increasing the feed rate by 0.5 L / min", the unsatisfied action time point is extracted, specifically, the action required to be performed by the liquid level driven disturbance rejection control scheme for the current liquid level state (for example, the liquid level is lower than 49.0 mm) is compared with the action actually performed by the mechanical arm in the state monitoring data table, if it is required to perform "increase the feed rate" at T0 time but actually delayed to T1 time, then the T0 to T1 time period is identified as an unsatisfied action time point, the remaining capacity and instantaneous gap are identified, specifically, the CPU occupancy rate, motor load and other data of the mechanical arm at the current time are monitored, the number of additional actions that can be executed by the mechanical arm in unit time is calculated as the remaining capacity, for example, the current remaining capacity of the mechanical arm is 1 additional medium priority action per second, at the same time, the action required by the liquid level driven disturbance rejection control scheme but not executed or insufficiently executed is quantified as the instantaneous gap, for example, if the feed rate is required to increase by 0.5 L / min, but actually only increases by 0.3 L / min, then the instantaneous gap is 0.2 L / min, the target gap is matched with the capacity paragraph, specifically, the instantaneous gap in a period of time is accumulated to form the target gap (for example, 0.2 L / min for 5 seconds, then the target gap is 1.0 L / min), and the target gap is matched with the capacity paragraph in the subsequent time period, for example, the feed rate gap of 1.0 L / min is matched to the capacity paragraph of 2 additional feed rate increasing actions of the mechanical arm in the next 10 seconds, a demand comparison result table is generated, which records each unsatisfied action time point, identified remaining capacity, instantaneous gap, target gap and capacity paragraph matching result.

[0038] The dynamic adjustment sub-module fills the abnormal gap by adjusting the action allocation order based on the demand comparison result table, identifies the capacity gap node and the response lag section, updates the action output timing and the liquid level curve, and generates a safety interlocking global optimization execution scheme; Based on the demand comparison result table, the table identifies unfulfilled actions, transient gaps, target gaps, and matched capacity segments, for example, there is a 0.2 L / min feed rate transient gap from T0 to T5 seconds, and there is an available capacity segment from T10 to T15 seconds, fill the abnormal gap by adjusting the action allocation order, specifically, according to the transient gap and the capacity segment identified in the demand comparison result table, evaluate whether the lower priority action currently being executed (for example, probing, stirring) can be paused or delayed to release resources to fill the transient gap, for example, if it is found that the feed gap from T0 to T5 seconds causes the liquid level to continue to drop, and the mechanical arm is executing 2 probing actions with priority 1 from T5 to T10 seconds, the probing action is delayed to T15 seconds, and a supplementary feed adjustment action with priority 4 is arranged during T5 to T10 seconds to make up for the previous lack, identify the capacity gap node and the response lag segment, specifically, when the demand comparison result table shows that a certain key action (for example, emergency stop) cannot be immediately executed due to insufficient remaining capacity of the mechanical arm (for example, CPU load 100%), the time point is marked as the capacity gap node, record the time difference between the time when the liquid level driven anti-disturbance control scheme requires a certain action (for example, requires to increase the feed rate at T0) and the time when the action is actually executed and produces measurable liquid level response (for example, the liquid level starts to rise at T10), as the response lag segment, for example, if the feed rate increasing action is required at T0, but the liquid level starts to respond at T10, the response lag segment is 10 seconds, update the action output timing and the liquid level curve, specifically, according to the above adjustment, a new mechanical arm action output timing is generated, which contains the re-allocated and sorted actions, for example, T5-T10 seconds execute the supplementary feed adjustment, T15-T20 seconds execute the probe, and based on the new action output timing and the liquid level and mechanical arm motion relationship model, predict and draw the updated liquid level curve, for example, the new liquid level curve is predicted to stabilize at 50.0 mm at T20 seconds, generate a safety interlock global optimization execution scheme, which is a complete execution scheme containing all dynamically adjusted mechanical arm action output timings, predicted liquid level curves, and specific strategies for dealing with various capacity gap nodes and response lag segments.

[0039] Please refer to Figure 8 , the molten metal casting liquid level real-time control method based on the mechanical arm is executed based on the above-mentioned molten metal casting liquid level real-time control system based on the mechanical arm, comprising the following steps: S1: According to the liquid level target value in the pouring pool and the initial position distribution curve, extract the liquid level key parameters and the target value, normalize the liquid level fluctuation influence factor, match the liquid level and the mechanical arm motion relationship, and generate a liquid level adaptation table; S2: Based on the liquid level adaptation table, extract the target and real-time liquid level deviation value, analyze the offset amplitude and response duration, filter the liquid level change trend and response time data, and generate a liquid level correction data set; S3: Based on the liquid level correction data set, extract the mechanical arm action frequency and peak node within a unit time, associate the liquid level response value to identify abnormal transition points and stable recovery points, extract the echo time and jump boundary within the fluctuation interval, and generate a mechanical arm action level matching table; S4: Based on the mechanical arm action level matching table, analyze the high-frequency liquid level priority section and fluctuation peak position, identify the step change node and reconstruct the main channel and compensation path, and build a liquid level driving anti-disturbance control scheme; S5: Based on the liquid level driving anti-disturbance control scheme, filter the action demand parameters and key node abnormal state values that are not met, extract the offset frequency peak value and correct the action adjustment logic, and generate a safety interlocking global optimization execution scheme.

[0040] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any skilled person in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A real-time control system for molten metal pouring level based on a robot arm, characterized by, The system comprises: The mechanical arm positioning module obtains a liquid level target value in the pouring pool and an initial position distribution curve, extracts liquid level key parameters, analyzes the influence of liquid level fluctuation on the positioning accuracy of the mechanical arm, sorts out a liquid level and mechanical arm motion relationship model, and obtains a liquid level adaptation table; The liquid level detection module extracts a real-time liquid level value and a target liquid level deviation based on the liquid level adaptation table, identifies liquid level change trends and response times, quantifies liquid level error accumulation effects, induces liquid level correction weights, and obtains a liquid level correction data set; The motion control module extracts mechanical arm action frequencies and peak nodes per unit time based on the liquid level correction data set, performs hierarchical analysis in combination with liquid level fluctuation characteristics and mechanical arm action rules, and obtains a mechanical arm action hierarchical matching table; The anti-disturbance optimization module sorts liquid levels and action requirements according to priorities and proportioning ratios based on the mechanical arm action hierarchical matching table, identifies abnormal distribution relationships between the liquid levels and the mechanical arm actions, adjusts action sequences through abnormal node characteristics, and constructs a liquid level-driven anti-disturbance control scheme.

2. The robotic arm based molten metal pouring level real-time control system of claim 1, wherein: The liquid level adaptation table comprises a liquid level interval, a positioning accuracy range, a liquid level compensation factor, and a motion coupling factor, the liquid level correction data set comprises a liquid level error factor, a response delay index, a change trend weight, and a correction coefficient, the mechanical arm action hierarchical matching table comprises a hierarchical action interval, a demand priority level, a fluctuation threshold, and a matching weight, and the liquid level-driven anti-disturbance control scheme comprises an abnormal distribution structure, a node adjustment sequence, an abnormal correction factor, and an action allocation factor.

3. The robotic arm based molten metal pouring level real-time control system of claim 1, wherein: The mechanical arm positioning module comprises: A key parameter extraction submodule extracts liquid level key parameters from a liquid level target value in the pouring pool and an initial position distribution curve, classifies liquid level fluctuation influence factors, and generates a liquid level characteristic distribution table; A positioning accuracy analysis submodule analyzes the influence of liquid level fluctuation on the positioning accuracy of the mechanical arm based on the liquid level characteristic distribution table, identifies positioning accuracy adaptation values under differentiated liquid level conditions, and generates a liquid level and mechanical arm motion relationship model; An adaptation table generation submodule sorts out liquid level parameters based on the liquid level and mechanical arm motion relationship model, analyzes corresponding relationships with the target liquid level value, and generates a liquid level adaptation table.

4. The robotic arm based molten metal pouring level real-time control system of claim 3, wherein: The liquid level detection module comprises: A liquid level deviation extraction submodule extracts a real-time liquid level value and a target liquid level deviation based on the liquid level adaptation table, records liquid level change trends and response time data, and generates a liquid level deviation data table; An error accumulation quantification submodule quantifies liquid level error accumulation effects based on the liquid level deviation data table, analyzes liquid level distribution and demand proportion differences, and generates a liquid level error accumulation effect table; A correction weight induction submodule extracts liquid level deviation magnitudes and correction frequencies based on the liquid level error accumulation effect table, screens high-frequency error sections and labels deviation directions, induces liquid level correction weights, and generates a liquid level correction data set.

5. The robotic arm based molten metal pouring level real-time control system of claim 4, wherein: The motion control module comprises: An action frequency extraction submodule extracts mechanical arm action frequencies and peak nodes per unit time based on the liquid level correction data set, classifies mechanical arm action priority data, and generates an action frequency distribution table; An action sequence adjustment submodule adjusts action sequences based on the action frequency distribution table, identifies abnormal action sequences, and generates an action sequence adjustment table; and An action allocation submodule extracts action allocation factors based on the action sequence adjustment table, identifies abnormal action allocation factors, and generates an action allocation table. The fluctuation characteristic analysis submodule analyzes fluctuation node states based on the action frequency distribution table, in combination with liquid level fluctuation characteristics and mechanical arm action rules, and generates a liquid level fluctuation state table; The hierarchical analysis submodule performs mechanical arm action and fluctuation multidimensional comparison based on the liquid level fluctuation state table, screens action adaptation relationships, and generates a mechanical arm action hierarchical matching table.

6. The robotic arm based molten metal pouring level real-time control system of claim 5, wherein: The anti-disturbance optimization module includes: The priority sorting submodule extracts liquid level peak position and action difference threshold based on the mechanical arm action hierarchical matching table, analyzes liquid level contribution degree under unit action difference, and generates a liquid level priority sorting table; The abnormal distribution identification submodule extracts main channel node abnormal change trajectories based on the liquid level priority sorting table, identifies abnormal equilibrium points and offset directions, and generates an abnormal distribution relationship table; The abnormal adjustment submodule adjusts action order based on abnormal node characteristics, extracts node abnormal change frequency and amplitude sequence, counts offset amplitude and duration of abnormal over-limit nodes, divides stable intervals and abnormal transition sections, and generates a liquid level driving anti-disturbance control scheme.

7. The robotic arm based molten metal pouring level real-time control system of claim 6, wherein: The liquid level peak position and action difference threshold refer to setting the peak position of the liquid level according to the difference between the monitoring of the liquid level change and the action, and calculating the corresponding action difference threshold; The liquid level priority sorting table is obtained by extracting the liquid level peak position and the action difference threshold from the mechanical arm action hierarchical matching table, and analyzing the liquid level contribution degree under unit action difference; The main channel node abnormal change trajectory refers to recording the trajectory of the abnormal change of the main channel node through real-time monitoring.

8. The robotic arm based molten metal pouring level real-time control system of claim 1, wherein: The system also includes a safety interlocking module: The safety interlocking module monitors liquid level states and mechanical arm action execution based on the liquid level driving anti-disturbance control scheme, compares unmet action requirements and remaining capacity in real time, fills in abnormal gaps by adjusting action allocation order, and generates a safety interlocking global optimization execution scheme; The safety interlocking global optimization execution scheme includes residual error adjustment parameters, execution order configuration, remaining capacity utilization rate, and adjustment completion criteria.

9. The robotic arm based molten metal pouring level real-time control system of claim 8, wherein: The safety interlocking module includes: The state monitoring submodule collects liquid level node values and mechanical arm action feedback based on the liquid level driving anti-disturbance control scheme, records jump time and deviation amplitude, and generates a state monitoring data table; The demand comparison submodule extracts unmet action corresponding time points based on the state monitoring data table, identifies remaining capacity and instantaneous gaps, matches target gaps and capacity sections, and generates a demand comparison result table; The dynamic adjustment submodule fills in abnormal gaps by adjusting action allocation order based on the demand comparison result table, identifies capacity gap nodes and response lag sections, updates action output timing and liquid level curve, and generates a safety interlocking global optimization execution scheme.

10. A method for real-time control of molten metal pouring level based on a robot arm, characterized by, The mechanical arm-based molten metal casting liquid level real-time control system according to any one of claims 1-9 performs, including the following steps: S1: Extract liquid level key parameters and target values based on the liquid level target value in the pouring pool and the initial position distribution curve, normalize liquid level fluctuation influence factors, match liquid level and mechanical arm motion relationships, and generate a liquid level adaptation table; S2: Based on the liquid level adaptation table, extract the target and real-time liquid level deviation value, analyze the offset amplitude and response duration, filter the liquid level change trend and response time data, and generate a liquid level correction data set; S3: Based on the liquid level correction data set, extract the mechanical arm action frequency and peak node per unit time, associate the liquid level response value to identify abnormal transition points and stable recovery points, extract the echo time and jump boundary in the fluctuation interval, and generate a mechanical arm action level matching table; S4: Based on the mechanical arm action level matching table, analyze the high-frequency liquid level priority section and fluctuation peak position, identify the step change node and reconstruct the main channel and compensation path, and build a liquid level driving anti-disturbance control scheme; S5: Based on the liquid level driving anti-disturbance control scheme, filter the action demand parameters and key node abnormal state values that do not meet the requirements, extract the offset frequency peak value and correct the action adjustment logic, and generate a safety interlocking global optimization execution scheme.

Images