Mold switching path planning and regulation method based on rule engine

Abstract

The invention discloses a mold switching path planning and regulation method based on a rule engine, and relates to the technical field of intelligent manufacturing equipment automation, and the method comprises the following steps: S100, in a cast stone continuous pouring process, collecting a continuous change value of a magma flow rate in a microsecond-level time window, extracting an instantaneous transition amplitude and a change rate of a flow rate curve, and calculating a flow rate curve; abnormal fluctuation signal features caused by equipment troubleshooting or feed recovery are identified. According to the invention, through microsecond-level flow velocity acquisition and dynamic anomaly identification, unsteady state data is effectively eliminated; the grouting state is accurately judged by combining a real-time trend map and grouting weight feedback; a path freezing mechanism prevents mistaken switching, a disturbance learning model enhances adaptivity, and a dynamic closed-loop control system is constructed. According to the overall scheme, the mold switching judgment precision and the system stability are remarkably improved, the pouring quality and the production continuity are guaranteed, and the good engineering application value and the good popularization prospect are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent manufacturing equipment automation, in particular to a mold switching path planning and regulation method based on a rule engine. BACKGROUND

[0002] The mold switching path planning and regulation based on a rule engine refers to, in the process of casting stone production, using pre-recorded structured information such as mold specifications, models and weights, combining real-time collected magma flow rate data, automatically matching and reasoning out the optimal mold switching timing and propulsion path through a rule engine, so as to accurately control the mold moving distance and replacement rhythm, and ensure that the magma pouring depth and position are always in a suitable state. When the production rhythm changes or the mold specifications are adjusted, the system can adjust the switching strategy in real time according to the rules, realize high-precision regulation and intelligent decision of the mold switching process, and avoid production deviation and resource waste caused by manual judgment errors.

[0003] The prior art has the following disadvantages:

[0004] In the process of continuous casting of casting stone, when the equipment is temporarily blocked or abnormally recovered, the magma flow rate may increase instantaneously, forming a short-period non-steady-state disturbance signal. At this time, if the rule engine does not load the dynamic stability discrimination mechanism for flow rate data, it is easy to misidentify the abnormal peak value in the process of returning to stability as an effective grouting rate, and to trigger the mold switching path planning logic, resulting in the mold exiting prematurely before reaching the pouring volume threshold. This misjudgment will cause serious under-pouring of magma in the mold, not only reducing the product density and structural strength, but also being difficult to be sensed in time in the continuous automatic switching process, thereby causing potential defects in batches of products, and finally possibly leading to large-scale rework processing and quality responsibility tracing, seriously affecting production stability and delivery reliability.

[0005] The above information disclosed in the background section is only used to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The purpose of the present application is to provide a mold switching path planning and regulation method based on a rule engine, which effectively eliminates non-steady-state data through microsecond-level flow rate collection and dynamic anomaly recognition, realizes accurate judgment of grouting state by combining real-time trend maps and grouting weight feedback, prevents mis-switching through a path freezing mechanism, enhances adaptability through a disturbance learning model, and constructs a dynamic closed-loop control system. The overall scheme significantly improves the judgment accuracy and system stability of mold switching, guarantees pouring quality and production continuity, has good engineering application value and promotion prospect, and solves the problems in the background technology.

[0007] In order to achieve the above object, the present application provides the following technical scheme: A mold switching path planning and regulation method based on a rule engine, comprising the following steps:

[0008] S100, in the continuous casting process of the stone, the continuous change value of the magma flow rate in the microsecond time window is collected, the instantaneous transition amplitude and the change rate of the flow rate curve are extracted, and the abnormal fluctuation signal characteristics caused by the device obstacle removal or the supply recovery are identified;

[0009] S200, a dynamic fluctuation screening model is constructed based on the instantaneous transition amplitude and the change rate, the non-stable flow rate value corresponding to the abnormal fluctuation signal is removed, and the main trend line of the magma flow rate filtered by the steady state is generated;

[0010] S300, the main trend line of the magma flow rate is mapped with the mold position parameters, the real-time pouring trend spectrum is generated, and the response corresponding relationship between the mold grouting state and the mold advancing path is established;

[0011] S400, based on the real-time pouring trend spectrum, the mold cumulative grouting weight increment is introduced as an auxiliary judgment basis, a time consistency judgment mechanism is established, and the matching accuracy between the mold advancing path and the actual grouting state is verified;

[0012] S500, when there is a response inconsistency between the mold advancing path and the actual grouting state, the path freezing mechanism is triggered, the mold path switching operation is stopped, and the current state parameters are fed back to re-determine whether the switching condition is met;

[0013] S600, based on the re-determination result of the switching condition, it is determined whether to continue to execute the mold path switching operation, and after the effective path switching is executed, the current magma flow rate disturbance tolerance threshold is updated, an adaptive adjustment model based on the historical disturbance characteristics is constructed, and the dynamic closed-loop regulation of the mold switching path is realized.

[0014] Preferably, the S100 process specifically comprises:

[0015] The high-frequency response type flow rate sensor is used to collect the continuous change value of the magma flow rate in the microsecond time window, and is synchronized to the mold advancing control system;

[0016] The collected continuous flow rate data is subjected to difference analysis, the instantaneous transition amplitude and the change rate are extracted, and the flow rate fluctuation characteristic spectrum is constructed;

[0017] The flow rate fluctuation characteristic spectrum is compared with the preset disturbance template library, the similarity index is calculated to judge whether it is an abnormal fluctuation;

[0018] If the similarity meets the set condition, the current segment flow rate change is identified as an abnormal characteristic section, and the related characteristic information is recorded to the flow rate behavior database.

[0019] Preferably, the S200 process specifically includes:

[0020] Based on the instantaneous transition amplitude and the change rate, a flow rate change vector sequence is constructed, fluctuation characteristic parameters are extracted, and multi-dimensional normalized expression is performed;

[0021] A dynamic screening rule based on a threshold self-adjusting mechanism is introduced to determine whether the stable range is exceeded and to mark abnormal data segments;

[0022] Weighted average, trend fitting and nonlinear interpolation methods are used for soft removal processing of the marked abnormal data segments;

[0023] A sliding window modeling method is used to generate a main trend line of the magma flow rate for the remaining stable data segments, which serves as the basic reference data for mold path determination.

[0024] Preferably, the S300 process specifically includes:

[0025] Based on a unified time coordinate system, the main trend line data of the magma flow rate and the mold position parameter data are time stamped and aligned;

[0026] A data pairing algorithm is used to one-to-one correspond the flow rate value and the current position of the mold at the same timestamp, forming a two-dimensional mapping point set;

[0027] Based on the two-dimensional mapping point set, a real-time pouring trend spectrum is generated to reflect the flow rate distribution change on the mold path;

[0028] A response corresponding relationship between the advancing distance and the grouting state is established to support the dynamic regulation logic for path determination.

[0029] Preferably, the S400 process specifically includes:

[0030] Real-time acquisition of mold cumulative grouting weight data is performed, and the weight increment in a continuous period is calculated;

[0031] The mold position change and the magma flow rate trend value in the same period are extracted, the theoretical grouting volume is converted, and the mass estimate value is converted;

[0032] The actual weight increment and the theoretical estimate value are compared to determine whether they are within the consistency deviation threshold range;

[0033] According to the deviation direction and amplitude, the response consistency level of the grouting and advancing path is determined, and the switching timing strategy is adjusted accordingly.

[0034] Preferably, the S500 process specifically includes:

[0035] Based on the time consistency determination mechanism, it is determined whether there is a response inconsistency between the mold advancing path and the grouting state;

[0036] When it is determined that the responses are inconsistent, the mold path switching instruction is immediately suspended, and the mold position is maintained at the current position to enter a frozen state;

[0037] The current state parameters at the time of the freezing trigger are recorded, and the parameters are input to a switching condition re-determination process;

[0038] After the determination process is completed, it is determined whether to release the frozen state and resume the path switching operation according to the determination result.

[0039] Preferably, the S600 process specifically includes:

[0040] In the path frozen state, the re-determination of the switching condition is performed according to the collected state parameters to determine whether the switching execution requirement is met;

[0041] After the path switching is performed, the flow rate disturbance features in the time period before and after the switching are extracted and compared with historical samples to correct the disturbance tolerance threshold;

[0042] The disturbance features and the processing results thereof are written into a disturbance feature learning library to form parameter adjustment samples;

[0043] Based on an incremental modeling mechanism, the path switching strategy is dynamically optimized according to the content of the learning library in subsequent operation.

[0044] In the above technical solution, the present application provides technical effects and advantages:

[0045] The present application constructs an abnormality recognition model through microsecond-level flow rate collection and instantaneous transition analysis, ensures that the system has sufficient recognition accuracy under high-speed fluctuation conditions, further combines a dynamic screening mechanism and real-time trend spectrum construction to eliminate non-steady-state data and restore the real flow rate trend, simultaneously introduces a mold grouting weight increment and time consistency verification mechanism to realize physical order level feedback control of the mold grouting process, effectively makes up for the judgment deviation caused by the traditional rule engine relying only on logical reasoning, when the propulsion and grouting states are inconsistent, the path freezing mechanism can immediately suspend the switching operation to prevent the switching instruction from being triggered by mistake, and finally through disturbance feature learning and tolerance threshold adaptive update, a path determination model that can evolve for a long time is established to improve the adaptability of the system to complex field disturbances. The overall scheme not only enhances the judgment accuracy and operation stability of the system, but also significantly improves the quality assurance capability of the mold switching process and the reliability of continuous production, has significant engineering practical value and popularization prospect. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed in the embodiments will be briefly introduced as follows. Obviously, the accompanying drawings in the following description only represent some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained based on these drawings.

[0047] Figure 1 A method flowchart of a mold switching path planning and regulation method based on a rule engine. DETAILED DESCRIPTION

[0048] Example implementations will now be described more fully with reference to the accompanying drawings. Example implementations may, however, be implemented in many different forms and should not be construed as limited to the examples set forth herein; rather, these example implementations are provided so that this disclosure will be thorough and complete, and will fully convey the scope of example implementations to those skilled in the art.

[0049] The present application provides a mold switching path planning and regulation method based on a rule engine as shown in Figure 1 The present application provides a mold switching path planning and regulation method based on a rule engine as shown in

[0050] S100, in the continuous casting process of cast stone, the continuous change value of the magma flow rate in the microsecond time window is collected, the instantaneous transition amplitude and the change rate of the flow rate curve are extracted, and the abnormal fluctuation signal characteristics caused by the device barrier removal or the supply recovery are identified;

[0051] In order to solve the problem that the instantaneous fluctuation of magma flow rate caused by temporary barrier removal or supply recovery of the device in the continuous casting process of cast stone causes false triggering of the mold switching path, an abnormal fluctuation identification based on microsecond time accuracy is proposed, which is used to extract and determine the non-steady-state disturbance signal that may interfere with the mold path control logic in the flow rate collection. Including the following steps:

[0052] In the continuous casting process of cast stone, a high-precision fluid monitoring device is used to collect the instantaneous flow rate of the key nodes in the magma channel in real time. The monitoring device has a microsecond time resolution and can continuously record the flow rate change of the magma at a sampling frequency of not less than 1MHz. By setting a cache ring and a continuous writing mechanism, the collected flow rate data can fully reflect the dynamic behavior of the magma in a short period without affecting the real-time processing efficiency. The time stamp of the collected data is strictly synchronized with the mold advancing control system, thereby providing time consistency guarantee for subsequent behavior correlation analysis.

[0053] The high-precision fluid monitoring device is a core component for realizing microsecond-level acquisition of magma flow rate, and its composition can be optimized and configured according to actual working conditions and precision requirements, and generally includes but is not limited to a high-frequency response flow rate sensor, a signal conditioning circuit, an embedded data acquisition terminal and an industrial-grade time synchronization module. Among them, the high-frequency response flow rate sensor can be selected from an ultrasonic Doppler sensor, a laser particle velocity measuring instrument (LDV), a hot film flow rate probe or a fast response sensor based on electromagnetic induction principle, has microsecond-level time resolution and high linearity, and can work stably in a high-temperature and high-viscosity fluid environment; the signal conditioning circuit is used for filtering, amplifying and analog-to-digital converting the original electrical signal of the sensor, to ensure that the data has engineering signal characteristics before being transmitted to the acquisition terminal; the embedded data acquisition terminal has high sampling rate, high cache capacity and low time delay processing capability, supports continuous high-speed sampling and real-time data transmission or storage; and the time synchronization module is responsible for accurately timing the collected data and the control node, commonly uses IEEE 1588 PTP or GPS time service mechanism, and ensures the consistency and accuracy of subsequent multi-source data fusion analysis. The above components work cooperatively to form a high-robustness and high-real-time magma flow rate monitoring system, which meets the requirements of high-precision perception of flow state in the continuous casting process of stone.

[0054] The collected continuous flow rate data is preliminarily processed, the flow rate change between any two adjacent sampling points is calculated by using a differential analysis method, and the change rate of the continuous change curve, i.e. the first derivative of the flow rate, is further obtained. At the same time, a sliding window mechanism is introduced, the instantaneous transition amplitude and the peak value of the change rate of the curve are dynamically calculated in each window period, so as to construct a flow rate fluctuation feature spectrum in the current period. The feature spectrum takes time as the horizontal axis and flow rate value and its derivative as the vertical axis, can clearly depict the flow mutation trend of magma in the microsecond level, and identify whether there is abnormal jump behavior inconsistent with the stable running state.

[0055] The constructed flow rate fluctuation feature spectrum is compared with a preset disturbance template library, the template library is based on the feature extraction of the labeled device troubleshooting events, feeding fluctuation events and other actual scenes in the historical running data, and the dynamic threshold setting method is used to normalize the expression of the template features. By calculating the similarity index between the current spectrum and the disturbance template, such as waveform similarity rate, amplitude matching degree and change rate coincidence degree, it is determined whether the current collected flow rate fluctuation belongs to the known non-steady-state disturbance signal. In order to improve the accuracy of identification, a multi-feature fusion-based abnormality recognition algorithm is also introduced, which gives weighted scores to multiple dimensions such as transition intensity, duration and direction consistency, to exclude accidental and non-systematic micro-disturbance data.

[0056] When the determination result shows that there is a high similarity abnormal fluctuation feature in the current flow rate data related to device barrier removal or supply recovery, the flow rate change segment is identified as an abnormal feature segment, and its feature information, corresponding timestamp, transition amplitude, change direction, duration cycle and other key parameters are recorded in the flow rate behavior database for subsequent path planning, pouring trend analysis and disturbance mode continuous learning update. This identification mechanism not only can perceive potential non-steady-state disturbance in advance, but also can provide more stable and reliable flow rate behavior reference for subsequent mold path control, and improve the abnormal response ability and overall control accuracy under complex working conditions.

[0057] In the continuous pouring process of cast stone, the continuous change value of magma flow rate in the microsecond level time window is collected in real time, and the instantaneous transition amplitude and change rate of the flow rate curve are further extracted. Its main role is to accurately identify abnormal disturbances in the flow rate change process and prevent false judgments caused by short-period non-steady-state fluctuations. The cast stone pouring system belongs to a typical high-temperature, high-viscosity, and high-dynamic complexity fluid process. The equipment may cause abnormal peak signals of magma flow rate instantaneously increasing or sharply decreasing due to temporary barrier removal, supply fluctuation or restart during operation. Although these signals have very short duration, they are easily misidentified as "normal pouring completion" or "mold filling completion" in a rule engine system lacking identification capability, thereby triggering an incorrect mold exit instruction, causing the mold to be cut out prematurely before being filled, and seriously affecting product density and structural strength. By continuously collecting flow rate data at a microsecond level time resolution, the instantaneous fluctuation trajectory of the flow rate can be accurately restored, and the transition amplitude and change rate, two core dynamic characteristics, can be extracted. Combined with the disturbance template recognition algorithm, abnormal fluctuations caused by equipment intervention and normal speed fluctuations in the natural pouring process can be effectively distinguished, thereby achieving accurate dynamic stability judgment. This mechanism plays a key role in "pre-judgment" and "disturbance isolation" in the entire mold switching path control logic, and is a basic guarantee for intelligent and highly reliable mold path control.

[0058] S200, constructing a dynamic fluctuation screening model based on the instantaneous transition amplitude and change rate, removing the non-steady-state flow rate values corresponding to the abnormal fluctuation signals, and generating a steady-state filtered magma flow rate main trend line;

[0059] To effectively exclude the interference of magma flow rate abnormal fluctuation signals caused by equipment barrier removal, supply recovery, etc. on mold switching path judgment, on the basis of the extracted instantaneous transition amplitude and change rate, a dynamic fluctuation screening model is constructed to identify and remove non-steady-state flow rate values, and finally generate a steady-state flow rate main trend line that is real, continuous and reflects the main dynamics of the pouring process. The steps include:

[0060] With the transient jump amplitude and rate data obtained in the preceding acquisition stage, a flow rate change vector sequence is constructed in each microsecond time window, and a high-frequency change curve cluster is formed in a continuous time period. To establish a quantitative standard for fluctuation characteristics, the change standard deviation, slope change rate, and fluctuation period density of each group of data in the curve cluster are calculated in the local time domain. These statistical characteristics can intuitively reflect the stability level of the magma flow rate in a specific time range, thereby providing a basis for the subsequent parameterization of the screening model. Meanwhile, to enhance the adaptability of the model, feature learning is performed based on typical abnormal fluctuation samples in the production history data to generate multiple disturbance feature distribution templates, and a multi-dimensional feature space is constructed in a normalized manner.

[0061] After completing the feature quantification, a dynamic screening rule based on a threshold self-adjusting mechanism is introduced to compare the above-mentioned feature quantities in real-time data streams one by one. When a data segment simultaneously exceeds the stable range boundary defined by the historical disturbance templates in multiple feature dimensions, and the duration of this state exceeds the set response lag period threshold, it is determined that the segment belongs to an abnormal fluctuation section. In this screening process, the weight configuration between feature dimensions can be dynamically adjusted according to the running stage, device state, and fluid physical property conditions to ensure that the model has stable rejection accuracy and fault tolerance capability under different working conditions.

[0062] The "dynamic screening rule based on a threshold self-adjusting mechanism" refers to not using a fixed threshold in the flow rate anomaly recognition process, but dynamically generating and adjusting the judgment threshold for screening abnormal fluctuations according to the running state, historical disturbance features, and data distribution characteristics, thereby improving the recognition accuracy and adaptability of complex and nonlinear disturbance behaviors. The core role of this mechanism is that under different production stages, device operating conditions, or changes in raw material properties, it can automatically adjust the screening threshold of each dimension based on the real-time collected flow rate change characteristics (such as instantaneous jump amplitude, change rate, standard deviation, fluctuation period density, etc.), ensuring that both non-steady-state disturbances and natural fluctuations in stable states can be effectively excluded. The specific implementation steps include the following steps:

[0063] Sliding window sampling is performed on real-time flow rate characteristics, and the mean, variance, and change trend of the characteristics are calculated in each time slice and compared with historical disturbance templates or stable operation models for comparative analysis;

[0064] Based on the comparison results, the matching degree of the current flow rate state and known abnormal behaviors is evaluated, and an adaptive weight weighting mechanism is used to dynamically correct the threshold value of each feature dimension, making it more suitable for the current operating condition;

[0065] Based on the corrected threshold, real-time data segments are judged one by one, and segments that exceed the threshold and meet the multi-dimensional joint fluctuation characteristics are marked as abnormal data regions;

[0066] The data segments determined as abnormal enter the soft deletion or trend reconstruction process, and the threshold adjustment historical data is saved in the disturbance feature learning module to continuously optimize the subsequent identification accuracy. Through this mechanism, robust screening control under variable disturbance scenarios is realized, significantly improving the stability and reliability of the mold switching path decision.

[0067] For data segments determined as abnormal fluctuations, instead of being directly deleted, a "soft deletion" mechanism is used to handle them, that is, their data impact is smoothly embedded into the stable trend of the previous and subsequent time periods through weighted averaging, trend fitting, and nonlinear interpolation. In specific implementation, stable points at both ends of the abnormal segment are preferentially selected as anchor points, and a local trend compensation path is constructed through cubic Bezier curve fitting or spline smoothing algorithm to avoid trend discontinuity or physical discontinuity caused by data deletion, thereby ensuring filtering effect while preserving the physical continuity of flow rate changes.

[0068] The "soft deletion" mechanism refers to, in the data processing process, for data segments determined as abnormal fluctuations, instead of directly deleting or clearing the data in this part, but through mathematical methods such as weighted averaging, trend fitting, nonlinear interpolation, etc., the abnormal segments are smoothly handled, so that their impact is naturally embedded between the stable data before and after them, thereby maintaining the continuity and physical consistency of the overall data curve. The core role of this mechanism is to effectively shield abnormal fluctuation interference while avoiding misinterpretation or oscillation of subsequent judgment models caused by data rupture, sudden drop, or "hollow". Compared with the traditional "hard deletion" method (directly deleting abnormal data), the soft deletion mechanism has stronger robustness and interpretability, and is especially suitable for process scenarios such as magma flow rate, which has viscous rheological properties and high requirements for continuity of change. In the continuous casting process, the flow rate curve not only reflects the current flow state, but also is the core basis for mold advancement path judgment. If the abnormal segment is forcibly deleted, it may disrupt the time sequence consistency of trend judgment, leading to incorrect estimation of the mold grouting state. Using the soft deletion mechanism, the trend judgment logic integrity can be preserved while significantly reducing the sudden impact of abnormal fluctuations on the logic, enabling the mold switching path planning system to maintain smooth and stable response capability in the face of complex disturbance environments, thereby achieving more accurate and reliable dynamic control.

[0069] After the above data rejection and trend compensation is completed, the remaining stable data segment is re-spliced in time sequence, and a flow rate main trend line is generated in a sliding window manner. The trend line not only truly reflects the steady-state flow rate evolution law of the magma in the continuous pouring process, but also excludes abnormal mutations that may interfere with path judgment in a short period of time, thereby providing high-credibility basic data support for subsequent mold grouting state evaluation and path switching decision-making. In addition, the dynamic fluctuation characteristic parameters of each time period of the trend line will be recorded and updated to the disturbance identification library in real time, forming a sustainable learning feature accumulation mechanism to improve the accuracy and response efficiency of model identification when similar abnormal events occur again in the future.

[0070] The generation of the flow rate main trend line by the sliding window method refers to that after the abnormal data "soft rejection" processing and the re-splicing of the remaining stable data segment, a fixed-length time window is used to slide on the stable data sequence in time sequence, local statistics or trend modeling is performed on the data in each window, thereby smoothing fluctuations, suppressing residual interference, and extracting the main line information reflecting the overall evolution trend of the magma flow rate. The specific implementation method is as follows:

[0071] First, set the window length (such as 500 microseconds, 1000 sampling points), and slide on the data sequence according to the set step (such as 10%, 50% of the window width);

[0072] In each window, the local average value, weighted median, local regression curve or polynomial fitting curve based on the least squares method of the data segment is calculated to represent the "local representative flow rate" of the time period;

[0073] Connecting the local representative flow rate points of all windows in time sequence can form a continuous, smooth and anti-interference flow rate main trend line; the most

[0074] To ensure that the trend line is not distorted at the boundary, forward extrapolation and backward mirror expansion can be used to enhance the stability of the edge window. The sliding window mechanism has the advantages of local perception and global continuity compatibility, which can accurately extract the real changes in a short period of time and suppress small fluctuations outside the trend judgment. It is a key technical step to build a high-robustness and high-credibility flow rate main trend line, and has an important supporting role for the real-time and accuracy of mold switching path judgment.

[0075] A dynamic fluctuation screening model is constructed based on the instantaneous transition amplitude and the change rate, and the non-stable flow rate values corresponding to the abnormal fluctuation signals are removed, and finally the main trend line of the magma flow rate filtered by the steady state is generated, which is used to establish a high stability and high credibility basic flow rate reference system for intelligent judgment of the mold switching path. In the continuous casting process of stone, magma is a high-temperature and high-viscosity non-Newtonian fluid, and its flow rate is not only easily affected by short-period disturbances of the feeding system, stirring mechanism and obstacle removal behavior, but also these disturbances often appear as unstable fluctuation signals such as steep rise, sharp peak and rapid drop in the sensing system. If there is no identification and removal mechanism for these abnormal fluctuations, these non-steady state data may be misidentified as effective grouting state, which directly leads to the mold being withdrawn prematurely before the pouring is completed, resulting in insufficient density, structural defects and even batch quality risks. Therefore, this step constructs a dynamic fluctuation screening model by analyzing the instantaneous transition amplitude and the change rate, which can effectively identify and remove such short-period and abnormal fluctuation behaviors at the data acquisition level, and retain the continuous flow rate section that truly reflects the stable state of the process. Further, the stable data section is spliced into the main trend line of the magma flow rate through sliding window and smoothing modeling technology, so that the subsequent mold grouting volume determination, path advancement planning and other operations have stable, continuous and predictable basis. This process not only improves the accuracy and response ability of path planning decision, but also significantly enhances the anti-interference robustness of dynamic interference, which is the key basis for ensuring the stable operation of continuous casting process and the consistency of product quality.

[0076] S300, synchronously mapping the main trend line of the magma flow rate with the mold position parameters to generate a real-time pouring trend spectrum and establish a response corresponding relationship between the mold grouting state and the mold advancement path;

[0077] To realize the precise linkage control between the mold advancement path and the actual grouting state, it is proposed to synchronously map the main trend line of the magma flow rate with the mold position parameters, to establish a response corresponding relationship between the mold grouting state and the mold advancement path by constructing a real-time pouring trend spectrum. This method fully considers the time-space coupling relationship between the dynamic evolution characteristics of magma flow and the advancement behavior of the mold on the track, thereby significantly improving the response sensitivity and judgment accuracy of path control. This method includes the following steps:

[0078] Based on the generated magma flow rate main trend line in the previous stage, a time coordinate system of flow rate data is established. In order to ensure the time consistency with the mold position data, a high-precision time synchronization mechanism is adopted to strictly align the timestamp of each sampling point in the flow rate trend line with the mold advancing control system. The mold position parameter acquisition device selects a high-resolution encoder or a laser ranging system to record the advancing distance, advancing speed and current position of the mold in the track in real time, and each position data point is also marked with a timestamp consistent with the flow rate sampling point. By constructing a unified time index matrix, the time sequence synchronization of different types of data is realized, thereby providing a basis for subsequent spatial matching.

[0079] After completing the time synchronization of the flow rate trend line and the mold position data, a data pairing algorithm is introduced to one-to-one correspond the flow rate value under the same timestamp with the current position of the mold and form a two-dimensional mapping point set. The horizontal axis of the mapping point set is the advancing distance of the mold, and the vertical axis is the corresponding magma flow rate main trend value at this position, thereby constructing the flow rate distribution atlas on the mold advancing path. Further, based on the mapping data of multiple consecutive time points, a real-time pouring trend atlas is generated by surface fitting or spatial interpolation algorithm, so that the magma flow rate on the whole advancing path continuously evolves with the spatial advancement, and the fluid injection state of the mold at any position is truly reflected.

[0080] The data pairing algorithm refers to the processing logic for one-to-one corresponding matching of data points with different sources but with the same timestamp under a unified time reference system in the process of multi-source time series data fusion. The role of the data pairing algorithm is to accurately match the flow rate sampling points in the magma flow rate main trend line with the position sampling points in the mold advancing process in the time dimension, thereby forming a two-dimensional mapping point set containing the "mold position-corresponding flow rate" correlation, providing a basic data structure for subsequent construction of real-time pouring trend atlas and path response mechanism. The specific implementation steps are as follows:

[0081] A unified time index system is established to sort all collected magma flow rate main trend line data points and mold position data points by timestamp, and a mapping index table is constructed using the timestamp as the primary key;

[0082] Iterate through the mold position data sequence, extract the timestamp of each position data point, and find the sampling point in the flow rate trend data that is exactly equal or closest to the timestamp;

[0083] If there is an exact matching timestamp, it is directly paired; if there is a time deviation, the nearest neighbor interpolation algorithm or linear interpolation method is used to calculate the flow rate value corresponding to the time point from the flow rate data to ensure the time synchronization accuracy;

[0084] Each pair of successful mold position and corresponding flow rate value is combined into a two-dimensional coordinate point (position, flow rate), and all points are arranged in time sequence to form a continuous two-dimensional mapping point set. The horizontal axis of the point set is the mold advancing distance, and the vertical axis is the flow rate trend value at the same time point, which can be used for real-time visualization of flow rate distribution, and thus realize the dynamic response mapping between mold grouting state and advancing path.

[0085] Based on the real-time pouring trend map, further combined with the physical parameters of the mold (such as mold volume, grouting port geometry, inclination) and the physical parameters of the magma (such as viscosity, flow resistance coefficient), a comprehensive modeling is carried out for dynamic calculation of the grouting state of the mold at different advancing positions. By linking the magma flow rate at each moment with the current position of the mold, the effective volume of the magma injected into the mold cavity at the current moment can be estimated, so as to judge whether the mold has reached the set pouring completion condition during path advancing. For example, when the pouring trend map shows a significant decrease in flow rate at a certain advancing section, and the structural parameters of the mold show that part of the effective volume has not been filled, it is judged that the mold is still in the "incomplete grouting" state, and the path switching instruction is prohibited. This modeling and judgment mechanism makes the mold advancing control no longer simply rely on linear advancing rules of time or position, but can make real-time decisions combined with actual pouring dynamics, thereby improving the accuracy of path switching and the intelligent level of regulation.

[0086] On the basis of the above calculation, by constructing a response corresponding model, a dynamic corresponding relationship between the mold advancing path and the mold grouting state is established. The model takes the mold advancing distance as the input and the cumulative grouting state as the output, sets multiple state intervals (such as initial grouting stage, stable stage, and end stage), and configures corresponding path advancing strategies for each stage. For example, acceleration is allowed in the initial grouting stage, constant speed is maintained in the stable stage, and path buffer mechanism is introduced in the end stage to prevent early exit. Through this mapping and response mechanism, not only can the mold advancing behavior accurately respond to the grouting state, but also the path strategy can be actively corrected according to the real-time trend map changes, thereby significantly improving the intelligent level and control accuracy of the mold switching process, and ensuring the consistency, stability and product forming quality of the pouring process.

[0087] The main trend line of the magma flow rate is synchronized with the mold position parameters to generate a real-time pouring trend map, and a response corresponding relationship between the mold grouting state and the mold advancing path is established, which realizes the dynamic coupling and precise collaborative control between the mold advancing behavior and the actual grouting process. In the continuous casting process of stone, only relying on the mold advancing distance or setting time threshold to determine the path is easy to cause the mold switching time deviation due to the fluctuation of magma flow rate, local filling lag and other factors, thereby causing under-grouting or over-grouting. By synchronizing the main trend line of the magma flow rate which has been stably extracted with the position parameters (such as the current position, advancing speed, cumulative travel) of the mold in the track in time and mapping the data, the corresponding real-time grouting flow rate state of the mold at each advancing position can be established, and further combined with the mold volume, pouring port characteristics and magma properties, the injection amount variation trend under unit advancing distance can be estimated to form a continuously evolving real-time pouring trend map. The map not only intuitively shows the grouting dynamics of the mold in the advancing path, but also provides quantitative basis for judging whether the pouring is completed. On this basis, a response corresponding mechanism between the mold advancing path and the grouting state can be constructed, so that each step of the advancing path can be dynamically adjusted according to the current grouting state to avoid quality risks caused by early or delayed advancing. Through the mechanism, real-time closed-loop control of the mold switching path can be realized, which significantly improves the precision, stability and automation level of the pouring process, and is an important basis for building an intelligent mold control system.

[0088] S400, based on the real-time pouring trend map, the mold cumulative grouting weight increment is introduced as an auxiliary judgment basis to establish a time consistency judgment mechanism for verifying the matching accuracy between the mold advancing path and the actual grouting state;

[0089] To further improve the matching accuracy between the mold advancing path and the actual grouting state, and prevent misjudgment caused by local data fluctuation or path instruction lag, the mold cumulative grouting weight increment is introduced as an auxiliary judgment basis based on the established real-time pouring trend map to establish a time consistency judgment mechanism. The mechanism is used to compare whether there is a response lag, advance or misplacement between the mold advancing path control logic and the actual grouting state change, thereby improving the accuracy and reliability of the mold switching time judgment. The specific embodiment includes the following steps:

[0090] The cumulative grouting weight data of the mold is collected in real time during the mold operation. The weight data is obtained by a high-precision industrial weighing unit installed below the mold bearing platform, which has a resolution of milligrams and a refresh frequency of milliseconds, ensuring stable and continuous weighing under the mold advancing state. The cumulative grouting weight refers to the total mass of magma received in the mold from the time the mold enters the grouting position to the current time point. This data can truly reflect the progress and compaction state during the actual grouting process, making up for the local errors that may be caused by flow rate and position data alone.

[0091] The cumulative grouting weight value is extracted at fixed time intervals, and the weight increment between two consecutive sampling periods is calculated, i.e. the grouting speed per unit time. At the same time, the generated real-time pouring trend map is called, and the mold position change and corresponding flow rate trend value are extracted in the same time period. The theoretical grouting volume increment is obtained by integrating the flow rate. This theoretical grouting volume can be converted into a mass estimate value in combination with the magma density parameter, and compared with the actual weighing data. Set the allowed deviation interval to determine whether the two are within the consistency threshold range. If the error range is exceeded, it is considered that there is a matching deviation between the advancing path and the grouting state.

[0092] Based on the identification of the deviation, a time consistency determination mechanism is established, i.e. judging the response consistency level according to the direction and amplitude of the deviation. For example, when the mold advancing speed is normal but the cumulative grouting speed is significantly lower than expected, it is determined as "grouting lag type misalignment", and the advancing path needs to be switched to the buffer stage for continuous pouring; on the contrary, if the grouting weight is ahead but the path is still not ready to switch, it is determined as "advancing lag type misalignment", and the switching preparation signal needs to be sent in advance. Through this mechanism, the time logic of path switching decision can be rechecked based on the dynamic actual grouting weight change, effectively avoiding misjudgment caused by abnormal disturbance or control delay.

[0093] To enhance adaptability and long-term stability, the time consistency determination mechanism stores each identified grouting-path matching deviation event as a data sample in the historical behavior library, and combines magma rheological parameters, mold types, environmental temperature and other influencing factors to build a deviation prediction model, and optimize the determination parameters in the subsequent control strategy. The historical correlation between grouting weight increment and pouring trend map can also be recorded continuously to form an empirical weight adjustment rule, which realizes the self-learning and real-time adjustment of path logic under similar working conditions in the future.

[0094] Based on the real-time pouring trend map, the cumulative grouting weight increment of the mold is introduced as an auxiliary judgment basis to establish a time consistency judgment mechanism. The core function is to ensure high matching between the mold advancing path control logic and the actual grouting state from the perspective of multi-dimensional data cross verification, avoiding misjudgment caused by local sensor data fluctuations, path planning lag or abnormal disturbance. In the continuous casting process of cast stone, whether the switching condition is reached is usually judged according to the magma flow rate trend and the mold position, but there are certain limitations in relying only on these "speculative" parameters. For example, when the magma flow rate fluctuates sharply for a short time, even if the mold advances slowly, the switching logic may be triggered prematurely, causing the mold to exit prematurely without being filled. Therefore, the actual measurable "cumulative grouting weight increment of the mold", i.e. the mass change per unit time of the mold, is introduced as a physical feedback basis for the actual grouting process. By comparing the real-time weighing results with the theoretical grouting rate calculated by the pouring trend map, it is dynamically judged whether they are consistent in time and amplitude. If there is a significant deviation between the two within a certain time window, it means that the mold advancing behavior is misaligned with the grouting state, and the switching operation should not be performed. This time consistency judgment mechanism is essentially a "feedforward + feedback" fusion verification strategy, which not only enhances the ability to identify abnormal behavior, but also improves the decision-making accuracy of the mold switching instruction, preventing product quality fluctuations caused by overly single control logic. In addition, this mechanism can also learn from historical deviation data, gradually optimizing the judgment threshold and response logic, making the entire pouring control system more adaptive and intelligent. It is an indispensable key link in realizing the intelligent regulation and control loop of the mold path.

[0095] S500, when there is a response inconsistency between the mold advancing path and the actual grouting state, triggering the path freezing mechanism, stopping the mold path switching operation, and feeding back the current state parameters to re-determine whether the switching condition is met;

[0096] To prevent the automatic switching instruction of the mold advancing path from being triggered incorrectly when the grouting is not completed or the state is abnormal, causing mold underfilling, product quality defects, or even batch process out of control, a path freezing mechanism is proposed. When it is detected that there is a response inconsistency between the mold advancing path and the actual grouting state, the path switching logic is immediately stopped, and the current state parameters are fed back to the path determination unit for re-judgment and confirmation of the switching condition. This mechanism can significantly enhance the robustness and fault tolerance of the path control logic, and build a more secure and reliable intelligent grouting process control system. The path freezing mechanism includes the following steps:

[0097] A time consistency judgment mechanism is constructed based on real-time pouring trend map and accumulated injection weight increment of the mold, to continuously monitor the dynamic relationship between the mold pushing behavior and the injection state. When it is detected that a certain pushing path instruction is about to be triggered, the flow rate trend, mold position, injection increment, historical response deviation and other indicators in the current time window are jointly analyzed to determine whether they are in a normal linkage relationship. If it is found that the mold has been pushed to the critical switching position, but the actual injection weight increment is significantly lagging behind the expected level, or the trend map and the pushing path exist logical misplacement, it is determined that the response is inconsistent.

[0098] Under the premise that the response inconsistency state is confirmed, the path freezing mechanism is immediately triggered. This mechanism will forcibly stop the issuance of the current mold path switching instruction, and lock the mold driver or track execution unit at the current position, entering the frozen control state. In the frozen state, the mold pushing no longer responds to the regular control command, automatically enters the delay judgment process, and records the complete state parameters at the time of freezing trigger, including the current timestamp, mold pushing distance, accumulated injection weight, flow rate main trend value, flow rate change rate and abnormal identification label and other key indicators. These parameters will be used as context information of the freezing event, input into the subsequent re-determination process.

[0099] The re-determination logic of the switching condition is automatically activated during freezing. This determination logic will use the state parameters at the freezing time and the historical trend data of the time period before and after it for dynamic comparison, to determine whether the mold is in an acceptable injection completion interval through trend extrapolation analysis, prediction of injection curve evolution path and residual fitting. If it is confirmed through comparison that the injection state has been completed and the error is within the tolerable range, the frozen state is automatically released, allowing the path to continue switching; otherwise, if it is still in an incomplete state, the freezing is maintained, and the delay observation and repeated determination will continue until the state meets the switching requirements.

[0100] To realize the self-learning and long-term evolution of the mechanism, after each path freezing event processing is completed, the relevant state parameters, freezing duration, switching results and deviation degree are automatically written into the freezing event database. This database can be used to train the response consistency identification model, continuously optimize the determination threshold, path response tolerance and freezing trigger condition. In future similar working conditions, potential response inconsistency risks can be identified in advance, the pushing strategy can be dynamically adjusted, and the path instruction can be corrected in advance, thereby reducing the freezing trigger frequency, improving the response efficiency and the intelligent judgment ability of path switching. Through this path freezing mechanism, the stability and safety of the mold switching path control system under non-ideal working conditions are effectively improved, which is one of the core links to realize intelligent and closed-loop injection path planning and abnormal deviation correction control.

[0101] When there is a response inconsistency between the mold advancing path and the actual pouring state, the path freezing mechanism is triggered, the mold path switching operation is stopped, and the current state parameters are fed back to re-determine whether the switching conditions are met. Its role is to introduce a dynamic error correction and protection mechanism for mold switching path control, thereby improving stability, robustness and pouring consistency under complex working conditions. In the continuous casting process of cast stone, the mold advancing path is usually automatically triggered by the rule engine according to the flow rate trend and position judgment, but if it is not combined with the actual pouring state, it is easy to have the problem of "advancing too early" or "switching lag". For example, under the conditions of short-term increase in flow rate, weight measurement lag or data jitter interference, the mold may be incorrectly judged as having completed pouring, resulting in the issuance of a switching instruction, causing the mold to be pushed out in an under-filled state, and thus affecting product density, strength and overall quality. The path freezing mechanism is a control safety valve designed to address such risks. When it is detected that there is an inconsistency between the mold advancing behavior and the actual pouring state, such as the advancing distance reaching the switching position, but the pouring weight increment or flow rate trend inside the mold showing that the pouring has not been completed, the advancing operation will be immediately suspended, the frozen state will be triggered, and the incorrect path switching will be prevented. In the frozen state, the current control instruction is suspended, all state parameters are recorded, and the re-determination mechanism is activated to assess whether the switching conditions are met again. Only when it is confirmed that the pouring state is truly completed and the response relationship has returned to consistency, the freezing will be lifted and the advancing operation will continue. This mechanism not only effectively prevents incorrect switching caused by transient abnormalities or data errors, but also provides a time window for self-correction by further monitoring trend changes during freezing. At the same time, through state parameter feedback and freezing behavior database accumulation, an intelligent judgment model can be gradually formed to provide early warning or adjust the path planning strategy in similar future situations, which has important practical value and engineering significance.

[0102] S600, determining whether to continue executing the mold path switching operation based on the re-determination result of the switching condition, updating the current magma flow rate disturbance tolerance threshold after performing the effective path switching, constructing an adaptive adjustment model based on historical disturbance characteristics, and realizing dynamic closed-loop regulation and control of the mold switching path;

[0103] To realize adaptive adjustment and closed-loop control of complex disturbance conditions during the mold path switching process, a path execution decision based on the re-determination result of the switching condition is proposed, and a disturbance feature learning mechanism is introduced to dynamically update the magma flow rate disturbance tolerance threshold after each effective path switching operation, thereby constructing an adaptive adjustment model with memory and evolution capabilities, and realizing continuous optimization and closed-loop control of the mold switching path. The specific implementation includes the following steps:

[0104] During the execution of the aforementioned path freezing mechanism, a switching condition re-determination logic is activated by analyzing the mold grouting state, push position, flow rate trend changes and other state parameters collected at the freezing time. This determination process is based on a series of constraints, including whether the grouting weight reaches a preset threshold, whether the flow rate trend is in the end stage, whether the push behavior deviates from the historical grouting completion sample within the allowable range, etc. A combination judgment strategy is adopted. When multiple conditions are met at the same time, it is determined that the switching condition is established, and the path freezing state is released, preparing to perform the path switching operation; if the determination result is that the condition is not established, the frozen state is maintained, and the delay observation mode is entered, waiting for the state to be satisfied again to re-determine, ensuring that the mold is switched only after the grouting is truly completed.

[0105] When the switching condition is confirmed to be established and the mold path switching operation is performed, a disturbance tolerance updating mechanism is automatically started. This mechanism extracts the disturbance characteristics that occur in this grouting process by comparing the flow rate fluctuation amplitude, change rate, disturbance duration and other key dynamic indicators within a certain time window before and after switching. These disturbance characteristics are compared with historical disturbance samples to determine which type of disturbance (such as feed interruption disturbance, obstacle removal disturbance, equipment restart disturbance, etc.) it belongs to, and the existing disturbance tolerance threshold is locally corrected accordingly. For example, when it is found that the current disturbance has a high amplitude but a short duration, and does not cause false triggering, it means that the existing threshold setting is slightly conservative, so the corresponding determination threshold can be appropriately relaxed to improve future recognition fault tolerance.

[0106] After obtaining new disturbance characteristic parameters and correcting the threshold, the current set of disturbance characteristics and their corresponding processing results (whether it causes false judgment, whether it freezes, whether the switching is successful, etc.) are written into the disturbance characteristic learning library as training samples for subsequent model evolution. An incremental modeling strategy is adopted to gradually accumulate disturbance sample data, form a mapping relationship between disturbance intensity-response strategy-determination result, and build an adaptive parameter adjustment mechanism. In future operation, when similar disturbances occur again, they can be quickly compared and strategy recommended according to the existing samples in the learning library, so as to optimize the path judgment logic in advance, reduce the freezing rate and false judgment rate, and improve the switching efficiency.

[0107] To realize the dynamic closed-loop regulation of mold path switching control, a "judgment-execution-learning-adjustment" cycle mechanism is adopted to automatically complete the confirmation of judgment conditions, the execution of path instructions, the extraction of disturbance characteristics and the adjustment of parameters in each mold switching event. After each closed-loop iteration, the disturbance judgment threshold and the path judgment strategy will be fine-tuned and optimized based on experience, gradually adapting to long-term evolutionary factors such as equipment aging, working condition changes, and raw material differences, improving the robustness and evolution ability of the control system. Through this embodiment, the safety and timeliness of path switching judgment are guaranteed, and the long-term self-learning ability is also possessed, building a dynamically stable and highly adaptive mold path switching regulation system, effectively supporting efficient and high-quality continuous casting stone production process.

[0108] Based on the re-determination result of the switching condition, it is determined whether to continue executing the mold path switching operation, and the current magma flow rate disturbance tolerance threshold is updated after the effective path switching is executed, an adaptive adjustment model based on historical disturbance characteristics is built, and the dynamic closed-loop regulation of mold switching path is realized. Its role is to realize the self-adaptive response ability and strategy evolution ability of the mold switching control system to complex production disturbances, and to build a stable, safe and learnable path judgment mechanism. In the continuous casting stone process, due to unpredictable disturbance behaviors such as intermittent operation of equipment, fluctuation of raw material viscosity, and sudden obstacle removal operation in the production site, each mold switching needs to be judged accurately, otherwise it is easy to cause mold underfilling or mispositioning switching due to misjudgment, affecting the product density and structural performance. Therefore, before the switching path is executed, the switching condition is re-determined to ensure that each path switching is based on real-time effective, accurate and reliable data. At the same time, after confirming the switching and executing, it does not stop at the completion of the action, but continues to extract the characteristics of the flow rate fluctuation before and after the switching, the mold grouting state change, the response time delay and other behaviors, and captures the disturbance pattern in this operation. Through storage, classification and structured processing of these disturbance characteristics, the tolerance threshold in the judgment of magma flow rate can be dynamically adjusted to make the future judgment logic more consistent with the actual working condition. For example, when a certain type of high-frequency disturbance is identified multiple times and it is confirmed that it will not cause misjudgment, the trigger threshold of this type of disturbance can be appropriately relaxed to improve robustness and reduce false triggering. For disturbance patterns that have caused false switching, the threshold is tightened to strengthen protection measures. Ultimately, a stable closed-loop process is formed among path judgment, strategy execution, data learning and parameter adjustment, enabling the mold path control logic to have evolution ability and individualized adaptation ability. This not only enhances the stability and flexibility in dealing with complex working conditions, but also greatly reduces the control errors caused by parameter rigidity and strategy aging, which is a key support link for realizing intelligent mold path switching, high-quality pouring control and full-process digital evolution.

[0109] The mold switching path planning and regulation method based on the rule engine can effectively solve the problem of mold switching in advance caused by misidentification of abnormal disturbance signals of magma flow rate as effective grouting state, and realize dynamic perception, intelligent judgment and closed-loop control of the mold switching path. The abnormal identification model is constructed through microsecond-level flow rate collection and instantaneous transition analysis, ensuring that the system has sufficient identification accuracy under high-speed fluctuation conditions. Further, the non-steady-state data is eliminated and the real flow rate trend is restored by combining a dynamic screening mechanism with real-time trend spectrum construction. Meanwhile, the mold grouting weight increment and time consistency verification mechanism are introduced to realize the physical quantity level feedback control of the mold grouting process, effectively making up for the judgment deviation caused by the traditional rule engine relying only on logical reasoning. When the advancing and grouting states are inconsistent, the path freezing mechanism can promptly stop the switching operation to prevent the mis-triggering of the switching instruction. Finally, through disturbance feature learning and tolerance threshold adaptive updating, a long-term evolving path judgment model is established to improve the adaptability of the system to complex field disturbances. The overall scheme not only enhances the judgment accuracy and operation stability of the system, but also significantly improves the quality assurance capability of the mold switching process and the reliability of continuous production, and has significant engineering practical value and popularization prospect.

[0110] The above formulas are dimensionless values calculated, and the formulas are obtained by software simulation of a large amount of data to obtain a formula of the most recent real situation. The preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0111] The above only describes some exemplary embodiments of the present application by way of illustration, and it is self-evident that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present application. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present application.

[0112] It should be noted that in this paper, if there are relationship terms such as first and second, they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. Without more limitations, the element defined by the statement "includes a" does not exclude the presence of another identical element in the process, method, article or equipment including the element.

[0113] It should be understood that the size of the sequence number of the above processes does not mean the order of execution in various embodiments of the present application, and the execution order of the processes should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0114] Those skilled in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0115] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0116] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiment of the present application according to actual needs.

[0117] In addition, the functional units in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present alone, or two or more units can be integrated in one unit.

[0118] The above is only a specific implementation 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 in 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.

[0119] The above only describes some exemplary embodiments of the present application by way of illustration, and it is needless to say that those skilled in the art can modify the described embodiments in various ways without deviating from the spirit and scope of the present application. Therefore, the above figures and descriptions are illustrative in nature and should not be understood as limiting the scope of protection of the claims of the present application.

Claims

1. A rule engine based mold switching path planning and regulation method, characterized in that, The method comprises the following steps: S100, in the continuous casting process of cast stone, the continuous change value of the magma flow rate in the microsecond time window is collected, the instantaneous transition amplitude and the change rate of the flow rate curve are extracted, and the abnormal fluctuation signal characteristics caused by the device barrier removal or the supply recovery are identified; S200, a dynamic fluctuation screening model is constructed based on the instantaneous transition amplitude and the change rate, the non-stable flow rate value corresponding to the abnormal fluctuation signal is removed, and a main trend line of the magma flow rate after steady-state filtering is generated; S300, the main trend line of the magma flow rate is synchronously mapped with the mold position parameters, a real-time pouring trend spectrum is generated, and a response corresponding relationship between the mold grouting state and the mold advancing path is established; S400, based on the real-time pouring trend spectrum, the mold cumulative grouting weight increment is introduced as an auxiliary judgment basis, a time consistency judgment mechanism is established, and the matching accuracy between the mold advancing path and the actual grouting state is verified; S500, when there is a response inconsistency between the mold advancing path and the actual grouting state, a path freezing mechanism is triggered, the mold path switching operation is stopped, and the current state parameters are fed back to re-determine whether the switching condition is met; S600, whether to continue to execute the mold path switching operation is determined based on the re-determination result of the switching condition, the current magma flow rate disturbance tolerance threshold is updated after the effective path switching is executed, an adaptive adjustment model based on historical disturbance characteristics is constructed, and dynamic closed-loop regulation and control of the mold switching path is realized.

2. The method of claim 1, wherein, The S100 process specifically comprises: A high-frequency response type flow rate sensor is used to collect the continuous change value of the magma flow rate in the microsecond time window, and is synchronously connected to the mold advancing control system; The collected continuous flow rate data is subjected to difference analysis, the instantaneous transition amplitude and the change rate are extracted, and a flow rate fluctuation characteristic spectrum is constructed; The flow rate fluctuation characteristic spectrum is compared with a preset disturbance template library, a similarity index is calculated to determine whether it is an abnormal fluctuation, and if the similarity meets the set condition, the current segment flow rate change is identified as an abnormal characteristic section, and the related characteristic information is recorded to a flow rate behavior database. The S200 process specifically comprises:

3. The method of claim 1, wherein, A flow rate change vector sequence is constructed based on the instantaneous transition amplitude and the change rate, fluctuation characteristic parameters are extracted and subjected to multi-dimensional normalization expression; A dynamic screening rule based on a threshold self-adjusting mechanism is introduced to determine whether it is out of the stable range and to mark the abnormal data segment; The marked abnormal data segment is subjected to soft removal processing by using a weighted average, a trend fitting and a nonlinear interpolation method; The remaining stable data segment is subjected to sliding window modeling to generate a main trend line of the magma flow rate, which is used as the basic reference data for mold path judgment. The S300 process specifically comprises:

4. The method of claim 1, wherein, Based on a unified time coordinate system, the timestamp of the magma flow rate main trend line data and the mold position parameter data is aligned; A data pairing algorithm is used to correspond the flow rate value with the current mold position at the same timestamp to form a two-dimensional mapping point set; A real-time pouring trend spectrum is generated based on the two-dimensional mapping point set to reflect the flow rate distribution change on the mold path; A response corresponding relationship between the advancing distance and the grouting state is established to support the dynamic regulation and control logic of the path judgment. ​ 5. The method of claim 1, wherein, The S400 process specifically includes: Real-time acquisition of cumulative grouting weight data of the mold and calculation of weight increment in a continuous period; Extracting the mold position change and magma flow rate trend value in the same period, converting the theoretical grouting volume and converting it into a mass estimate value; Comparing the actual weight increment with the theoretical estimate value to determine whether it is within the consistency deviation threshold; According to the deviation direction and amplitude, determine the response consistency level of the grouting and advancing path, and adjust the switching timing strategy accordingly.

6. The method of claim 1, wherein, The S500 process specifically includes: Based on the time consistency determination mechanism, determine whether there is a response inconsistency between the mold advancing path and the grouting state; When it is determined that the response is inconsistent, immediately suspend the mold path switching instruction, and keep the mold position at the current position and enter the frozen state; Record the current state parameters at the time of freezing trigger, and input the parameters into the switching condition re-determination process; After the determination process is completed, decide whether to release the frozen state and restore the path switching operation according to the judgment result.

7. The method of claim 1, wherein, The S600 process specifically includes: In the path frozen state, re-determine the switching condition according to the collected state parameters to determine whether the switching execution requirements are met; After executing the path switching, extract the flow rate disturbance characteristics in the time period before and after the switching and compare them with the historical samples to correct the disturbance tolerance threshold; Write the disturbance characteristics and their processing results into the disturbance characteristic learning library to form parameter adjustment samples; Based on the incremental modeling mechanism, dynamically optimize the path switching strategy in subsequent operation according to the content of the learning library.

Images