Industrial signal stage matching method based on process map

Abstract

The invention relates to the technical field of industrial automation control, in particular to an industrial signal stage matching method based on a process map, and aims to solve the problems that in the prior art, an accurate corresponding relation between signal characteristics and process stages is difficult to capture by manually set static weights; and the time sequence offset between the actual signal and the standard process map cannot be effectively processed. The method effectively solves the problems of subjectivity, static property and deviation non-traceability of manual weight setting by constructing a standardized process map and matching industrial signal features in real time and combining dynamic weight distribution and a self-adaptive adjustment mechanism, and has the advantages of improving control precision, enhancing dynamic adaptability and improving decision reliability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial automation control, and particularly relates to an industrial signal stage matching method based on a process map. BACKGROUND

[0002] In a multi-stage industrial control process, a traditional process control method mainly relies on manual experience to set fixed weight parameters for each stage. This manual intervention method has significant defects: first, different operators have different subjective judgments on the importance of the process stage, resulting in non-uniform weight distribution standards, which directly affects the repeatability of the control result; second, when the process flow changes dynamically, the preset fixed weight cannot respond to the actual working condition changes in time, causing control deviation.

[0003] In the multi-stage control of a chemical reaction kettle, a semiconductor etching process and other precision manufacturing scenarios, this limitation is particularly prominent - the manually set static weight is difficult to capture the accurate correspondence between signal characteristics and process stages, and cannot effectively handle the timing offset problem between actual signals and standard process maps. More seriously, the traditional method lacks an adaptive adjustment mechanism for process deviations, and when abnormal situations such as equipment aging and raw material fluctuations occur, it is easy to cause control decision failure. Therefore, it is urgent to establish an intelligent control method that integrates process map standardization modeling, signal dynamic matching and weight adaptive adjustment to solve the problems of subjectivity, staticity and non-traceability of manual weight setting. SUMMARY

[0004] (I) Technical problems to be solved

[0005] To solve the above problems, the present application proposes an industrial signal stage matching method based on a process map, aiming to solve the problem that the manually set static weight in the prior art is difficult to capture the accurate correspondence between signal characteristics and process stages, and cannot effectively handle the timing offset problem between actual signals and standard process maps.

[0006] (II) Technical solutions

[0007] The industrial signal stage matching method based on a process map comprises:

[0008] Constructing a standardized process map, the process map containing multiple process stages and preset weights for each stage;

[0009] Real-time acquisition of industrial signal data, extraction of signal characteristics and matching with nodes of the process map;

[0010] Dynamic allocation of stage weights according to the matching results;

[0011] Process control decisions are made based on dynamically assigned stage weights.

[0012] (3) Beneficial effects

[0013] Compared with the prior art, the present invention has the following beneficial effects:

[0014] (1) The present invention constructs a standardized process map and matches industrial signal characteristics in real time, combined with dynamic weight allocation and adaptive adjustment mechanism, effectively solving the subjectivity, staticness and non-traceability of deviation in manually set weights, and has the advantages of improving control accuracy, enhancing dynamic adaptability and improving decision reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 The figure is a logical structure diagram of the industrial signal phase matching method based on the process map;

[0017] Figure 2 Schematic diagram of the framework structure of the overall system. DETAILED DESCRIPTION

[0018] Multi-stage process control relies on manually setting weight parameters for each stage. The limitations of this approach are particularly significant in complex industrial scenarios. Taking the semiconductor etching process as an example, the process consists of multiple interrelated stages, each with different sensitivities to parameters such as temperature and gas flow. Traditional methods require engineers to assign weights to each stage based on experience, resulting in significant differences in parameters set by different operators. When process conditions fluctuate or equipment status changes, fixed weights cannot adjust the control strategy in a timely manner, which can easily lead to quality issues such as over-etching or under-etching. This subjective and static control mode has become a key bottleneck restricting the improvement of process accuracy.

[0019] To address this issue, the inventors analyzed the inherent correlation between process parameters and stage characteristics and discovered that signal changes at key process nodes possess recognizable physical characteristics. Based on this, they proposed replacing manual empirical judgment with the construction of a standardized process model and establishing a dynamic mapping relationship between signal characteristics and process stages. Further research revealed that node matching within the process map can reflect the actual production status in real time, providing an objective basis for weight allocation. By combining the weight adjustment mechanism with real-time signal analysis, a closed-loop control system is formed, ultimately achieving adaptive optimization of process parameters.

[0020] Example 1

[0021] like Figure 1-Figure 2 As shown, the present disclosure proposes an industrial signal phase matching method based on a process map, comprising the following steps:

[0022] S100: Constructing a standardized process map, wherein the process map includes multiple process stages and preset weights for each stage;

[0023] S200, collecting industrial signal data in real time, extracting signal features and matching them with nodes of the process map;

[0024] S300, dynamically assigning stage weights based on matching results;

[0025] S400 , executing process control decision based on dynamically allocated stage weights.

[0026] Specifically, a standardized process map is a stage-by-stage model established based on engineering drawings or standard process flows. It includes the sequence of process stages and pre-set weighting parameters. This can be achieved by deconstructing the process flow chart, identifying stage boundaries by parsing standard procedures in the equipment operating manual. This map provides a benchmark framework for weight assignment, eliminating human bias.

[0027] Signal feature extraction involves identifying key points that characterize phase transitions from real-time industrial signals. Specifically, a differential algorithm can be used to detect sudden changes in signal slope, combined with an extreme value detection algorithm to locate phase boundaries. Identifying these key points ensures accurate process phase location.

[0028] Dynamically assigning stage weights involves invoking the pre-set weight parameters of the graph based on the matching process stage. This is achieved using a weight mapping table query mechanism. When a signal feature successfully matches a graph node, the pre-set weight value for the corresponding stage is automatically activated. This mechanism enables the control system to quickly respond to changes in process status.

[0029] The method first builds a graph model based on the standard process documentation provided by the equipment manufacturer, defining the duration thresholds and key parameter ranges for each stage. During operation, sensors collect process signals such as temperature and pressure in real time, and digital signal processing techniques are used to extract feature points.

[0030] When a feature point combination is detected that matches the map definition for a specific stage, the system automatically loads the preset weight parameters for that stage. The control module adjusts the operating parameters of the actuators based on the current weight values, for example, increasing the response priority of the cooling system during high-temperature stages. This entire process forms a closed-loop link from signal acquisition to control execution, ensuring that weight allocation is always synchronized with the real-time process status.

[0031] Compared to existing technologies, which rely on manual experience to set fixed weights, this solution automates the configuration of weight parameters by establishing objective standards for process maps. Existing technologies are unable to identify dynamic shifts in process stages. This solution, through a signal feature matching mechanism, automatically corrects stage boundaries and adjusts control strategies. Traditional control systems use weight parameters that are disconnected from process status. This solution creates a closed-loop system that dynamically links weight allocation to real-time production data.

[0032] Through the above technical solution, this application effectively eliminates the subjective differences in manually set weights and ensures the consistency of control parameters in different operating scenarios. Through a real-time signal feature matching mechanism, the process control system can automatically adapt to equipment state fluctuations and external interference, improving the stability of complex processes. The weight parameter system preset based on the atlas optimizes the dynamic allocation efficiency of system resources while ensuring control accuracy in key process stages.

[0033] The present application further proposes constructing a standardized process map including dividing the process stages according to engineering drawings or standard process flows, and presetting weight values ​​for each process stage, wherein the weight values ​​of key process stages are greater than 1.

[0034] Among them, engineering drawings or standard process flows refer to the design blueprints of specific equipment or production lines and verified process operation step instructions. Specifically, this can be achieved by dividing the heating, reaction, and cooling stages of a chemical reactor, or the pre-cleaning, etching, and post-processing stages of a semiconductor etching process. By directly dividing the stages based on objectively existing process design documents, subjective interference from human experience can be avoided.

[0035] Among them, the weight value of the key process stage is greater than 1, which means that an adjustment factor higher than the baseline value is set for the key stage in the control system. Specifically, this can be achieved by setting the weight to 1.5 in the heating stage of the reactor, or setting the weight to 2.0 in the etching stage. The control intensity of the key stage is strengthened in a quantitative manner to ensure the accuracy and stability of the process parameters.

[0036] Specifically, when constructing a standardized process map, the process stages are first divided according to engineering drawings or standard process flows. For example, in the semiconductor etching process, the process stages are divided according to the three stages of pre-cleaning, etching, and post-processing defined in the equipment operation manual. A weight value is then preset for each stage, where the weight value of the key stage is greater than 1. For example, the weight of the etching stage is set to 2.0, and the weight of the remaining stages is set to 1.0. Through the objective basis of engineering drawings or standard processes, the process stage division and weight setting are based on verified production standards, eliminating subjective differences set manually. The special setting of the key stage weight value directly guides the control system to exert stronger adjustment force in the corresponding stage, such as increasing the response speed of temperature control in the etching stage, thereby ensuring that the target parameters of the core process nodes are achieved.

[0037] Compared to existing technologies, traditional methods rely on operators to manually set weights based on their experience. For example, different operators may set the weight of the etching stage to 1.8 or 2.2, resulting in inconsistent control strength. This solution divides the stages and presets weights through engineering drawings or standard processes, ensuring a unified benchmark for weight setting. For example, all production lines set the weight of the etching stage to 2.0 according to the equipment manual, eliminating human fluctuations. In addition, quantitative rules with weights greater than 1 clearly define control priorities and can be directly converted into adjustment parameters for the control system, compared to manually set fuzzy descriptions.

[0038] Through the above technical solution, the present application solves the subjectivity and inconsistency problems of manually setting the weights of process stages. The division based on engineering drawings or standard processes ensures the objectivity of stage division. For example, in chemical reactor control, different production lines are divided into heating stages according to the same drawing to avoid operators defining stage boundaries on their own. The preset weight value rules standardize the control intensity of key stages. For example, the weight is forcibly set to 2.0 in all etching processes to ensure that the temperature control intensity of this stage is consistent. As a result, the accuracy and repeatability of process control are improved. For example, in multi-batch semiconductor etching, the fluctuation range of key parameters is reduced and the process stability is significantly improved.

[0039] The present application further proposes extracting signal features including identifying slope mutation points and extreme points in the signal as stage boundary features, and aligning and matching the stage boundary features with process map nodes.

[0040] The slope mutation point refers to the critical point where the first-order derivative value in the signal waveform exceeds the preset threshold. Specifically, it can be implemented by differential operation combined with threshold comparison algorithm to capture the starting moment when the process parameters change significantly. The extreme point refers to the local maximum or minimum point in the signal waveform. Specifically, it can be implemented by sliding window extreme value detection algorithm to characterize the target state parameters of the process stage. The alignment and matching of stage boundary features and process map nodes refers to the time axis calibration of the time position of the feature points detected in real time with the predefined stage boundary nodes in the process map. Specifically, it can be implemented by dynamic time warping algorithm or minimum offset matching algorithm to eliminate random errors caused by manual judgment.

[0041] In the real-time industrial signal data stream, slope mutation points are identified by continuously calculating the first-order derivative of the signal, while a sliding window is used to scan the signal waveform to detect local extreme points. When a slope mutation point is detected, that moment is marked as the start or end node of the process stage transition; when an extreme point is detected, it is marked as the critical node at which the process parameters reach the target state of the stage. The time position information of the two types of feature points is input into the matching algorithm and aligned with the preset boundary nodes of the corresponding stage in the process map to form an objective stage division result. This process automatically completes feature extraction and matching through mathematical algorithms, avoiding the division bias caused by experience differences or visual errors when manually judging the stage boundaries.

[0042] Traditional methods rely on operators to visually judge the stage transition points in the signal waveform based on their experience, which leads to the problem of inconsistent subjective judgment standards. The existing technology uses fixed time thresholds or single feature detection methods, which makes it difficult to accurately capture the dynamic change characteristics of complex process signals. This application uses a dual feature detection mechanism, combining the physical meaning of slope changes and state extremes, to construct a complementary stage boundary recognition model. Measured data in semiconductor etching process monitoring scenarios show that the stage boundary misjudgment rate is reduced from 12.3% for manual judgment to 2.1% for algorithmic judgment.

[0043] Through the above-mentioned technical solution, this application solves the technical problem of subjective bias in manually determining process stage boundaries, achieving objectivity and standardization in process stage division. In a multi-stage control scenario for chemical reactors, this technology has increased the consistency of process signal stage division results for the same batch by different operators from 68% to 98%, providing an accurate stage positioning benchmark for subsequent dynamic weight allocation and timing correction operations.

[0044] The present application further proposes a dynamic allocation of stage weights including a timing correction operation, specifically including detecting the stage time offset between the actual signal and the process map, and shifting the signal to align it to the corresponding node of the map according to the offset.

[0045] Timing correction refers to the operation of achieving timing calibration by detecting the time deviation between the actual signal and the standard process map and performing signal shift compensation. Specifically, it can be achieved by using a sliding window algorithm combined with feature point matching to eliminate stage misalignment caused by equipment response delays or process parameter fluctuations.

[0046] The stage time offset refers to the position difference between the key feature point in the actual signal and the corresponding node in the process map on the time axis. It can be quantified by calculating the time difference between the slope mutation point or extreme point to accurately capture the degree of timing deviation.

[0047] Signal shift alignment refers to moving the actual signal curve as a whole along the time axis to eliminate the detected offset. This can be achieved by using timestamp offset compensation or data resampling technology to re-establish the time synchronization relationship between the actual signal features and the graph nodes.

[0048] Specifically, in the multi-stage control scenario of a chemical reactor, the temperature signal collected in real time may have an overall time lag due to heater delay. By identifying the temperature extreme point in the signal as the stage boundary feature, the time difference between it and the end node of the heating stage preset in the process map is calculated to determine the stage delay. The timestamps of the subsequently collected signal data are shifted forward by the delay amount as a whole, so that the extreme points of the actual temperature curve are aligned with the map nodes. As a result, subsequent weight allocation and control decisions are executed based on the corrected timing benchmark to avoid premature triggering of the insulation stage control strategy due to delays in the heating stage.

[0049] Traditional methods rely on fixed timing maps or manually adjusted schedules, and are unable to adapt to timing fluctuations during production in real time. This solution automatically detects offsets and performs signal shifting, enabling dynamic timing calibration while preserving process logic. This approach eliminates the need to interrupt production flow or rely on manual experience, significantly improving the robustness of multi-stage process control.

[0050] Through the above technical solution, this application effectively eliminates the time misalignment between the actual signal and the standard spectrum, ensuring the accuracy of stage matching and the timeliness of the control strategy. In semiconductor etching process monitoring, when the etch rate fluctuates and the process stage is advanced, the signal translation alignment mechanism automatically corrects the timing reference, so that the etching endpoint detection accurately corresponds to the spectrum node, avoiding over-etching or equipment damage caused by stage misjudgment.

[0051] The present application further proposes a timing correction operation specifically as follows: when a stage delay is detected, the signal curve is shifted to the left by the offset time; when a stage advance is detected, the signal curve is shifted to the right by the offset time.

[0052] Specifically, the leftward shift of the signal curve refers to adjusting the signal's time reference system toward the historical direction. This can be achieved by subtracting the offset from the timestamp as a whole, so that the signal feature points that originally appeared with a delay are moved forward to the corresponding node position on the process map. The rightward shift of the signal curve refers to adjusting the signal's time reference system toward the future. This can be achieved by adding the offset to the timestamp as a whole, so that the signal feature points that originally appeared in advance are moved backward to the corresponding node position on the process map. These two shifting operations achieve spatial alignment of signal features and process nodes by modifying the time base, solving timing misalignment problems caused by equipment response delays or abnormal material reactions.

[0053] Specifically, in the semiconductor etching process monitoring scenario, when the temperature rise is delayed due to equipment aging during the etching stage, the system detects that the actual signal curve has a positive time offset relative to the standard spectrum. At this time, the time axis of the real-time signal to be processed is shifted to the left as a whole by the offset, so that the slope mutation feature at the etching starting point is aligned with the spectrum node. During the insulation stage of the chemical reactor, if the material reaction activity is too high and the temperature peak occurs early, the system detects a negative time offset and shifts the subsequent signal time axis to the right by the offset, so that the temperature drop inflection point is aligned with the spectrum node. The translation operation achieves dynamic matching of signal characteristics and process stages by adjusting the time coordinate system, providing an accurate timing benchmark for weight allocation.

[0054] Compared to existing technologies, traditional methods typically only compensate for delays in one direction, such as by extending the phase duration or adding buffer time, but cannot address false triggering caused by premature phases. Some advanced systems use complex time scaling algorithms to adjust signal curves, but these algorithms are computationally intensive and difficult to guarantee real-time performance. This solution directly corrects the time reference through a bidirectional translation operation, without changing the signal shape or increasing computational complexity. This achieves bidirectional correction while maintaining the standard timing framework of the process map.

[0055] Through the above technical solution, this application effectively eliminates the timing deviation between industrial signals and process maps, ensuring that the stage boundary feature points are accurately aligned with the standard nodes. In delayed scenarios, this prevents the delayed execution of control instructions and prevents process parameters from deviating from the target range; in advanced scenarios, it prevents misjudgment of the end of a stage and avoids switching to the next process too early. This two-way correction mechanism provides a reliable timing synchronization foundation for multi-stage process control, improving equipment control accuracy and process stability.

[0056] This application further proposes an adaptive map adjustment method, which includes real-time comparison of the deviation between the process signal and the process map, dynamic adjustment of the map stage parameters based on the deviation, recording the adjustment results and feeding back the optimization decision rules.

[0057] Real-time comparison of the deviation between process signals and process maps refers to the detection of differences between real-time process data collected by sensors and the preset process model. It can be achieved by using a sliding window algorithm combined with a dynamic time warping algorithm to identify stage time offsets and parameter fluctuations in the actual production process.

[0058] Dynamically adjusting profile stage parameters based on deviations refers to modifying process model parameters based on detected deviations. This can be achieved using a threshold trigger mechanism coupled with a proportional-integral adjustment algorithm to eliminate the impact of inter-batch process parameter drift on control accuracy. Recording adjustment results and providing feedback for optimizing decision rules refers to storing the magnitude and effect of each parameter adjustment in a historical database. This can be achieved using a relational database coupled with an association rule mining algorithm to establish an empirical knowledge base for adaptive process parameter adjustments.

[0059] During the semiconductor etching process, when it is detected that the thickness change trend of the oxide layer deviates from the preset curve of the process map, the system automatically extends the duration parameter of the plasma excitation phase and records the thickness distribution data before and after this adjustment into the database.

[0060] In subsequent batches, the optimized decision rules prioritize the stage parameters that have the greatest impact on etch uniformity for dynamic compensation based on historical adjustment records. This forms a bidirectional iterative mechanism for process parameter adjustment and rule optimization, allowing the process map to automatically evolve in response to dynamic factors such as equipment aging and environmental changes.

[0061] Traditional process control methods rely on manual experience to periodically adjust process parameters, which can lead to adjustment lags and the accumulation of subjective errors. However, this application utilizes a closed-loop feedback mechanism to automatically trigger pattern parameter adjustments at the initial stage of etch rate decline, eliminating the response delays associated with manual intervention. Furthermore, parameter adjustment rules are continuously optimized using historical data, overcoming the inherent lack of adaptability of fixed rules to varying equipment conditions.

[0062] Through the above technical solution, this application realizes the automated closed-loop management of process control parameter adjustment, effectively eliminating the subjective errors of manually setting weights. During the wafer manufacturing process, the impact of ion implantation dose deviation on junction depth parameters can be corrected in real time to ensure the stability of doping concentration between different batches. For changes in thin film deposition rate caused by reaction chamber pressure fluctuations, the system can automatically adjust the trigger threshold parameters of the gas flow stage to maintain film thickness uniformity within the process specification range.

[0063] The present application further proposes dynamically adjusting the atlas stage parameters, including extending or shortening the stage duration when the stage duration deviation exceeds a threshold, and adjusting the stage triggering conditions when the trend of key parameter changes deviates from the atlas.

[0064] Among them, the stage duration deviation exceeds the threshold value means that the difference between the running time of the actual process stage and the preset duration of the process map exceeds the allowable range. Specifically, the time window statistical method can be used to calculate the average value of the actual stage duration and compare it with the preset value to determine whether the stage duration parameter needs to be corrected.

[0065] Deviations from the key parameter trend map refer to deviations from the actual temperature, pressure, or concentration trajectory that align with the trigger conditions preset in the map. This can be achieved by calculating parameter derivatives within a sliding window or by using a feature point matching algorithm to identify misaligned logic at stage boundaries. Adjusting the stage trigger conditions involves modifying the parameters that determine stage switching. This can be achieved by resetting gradient thresholds or generating dynamic conditional expressions to realign the stage divisions with the actual physical process.

[0066] Specifically, if the system detects that the actual duration of a process phase exceeds a preset value, it automatically extends the duration of that phase as defined in the process map. For example, in a semiconductor etching process, if the actual duration of the main etching phase is longer than the preset value in the map, the preset duration of that phase will be adjusted from 120 seconds to 130 seconds. When the rate of change of a key parameter, such as the reactor temperature, falls below the trigger condition preset in the map, the system dynamically adjusts the phase switching conditions, for example, changing the temperature gradient threshold for entering the cooling phase from a decrease of 5 degrees Celsius per minute to a decrease of 4 degrees Celsius per minute.

[0067] These two adjustment operations act on the time dimension and state dimension of the process map respectively. By correcting the inaccuracy of the timing framework and the inaccuracy of the state logic, they ensure that subsequent process control decisions are based on the updated accurate map.

[0068] Compared to existing technologies, traditional methods rely on regular manual inspections and corrections of process parameters. These methods are unable to respond in real time to timing drift caused by equipment aging or material batch variations, and lack automated adjustments to key parameter trend deviations. This solution, through real-time detection and automated parameter correction, addresses the control inaccuracies caused by the lag and subjectivity of manual intervention, while also avoiding the one-sidedness of process models caused by single-dimensional adjustments.

[0069] Through the above technical solution, the present application effectively avoids the subsequent process chain mismatch caused by the cumulative error of stage duration, improves the consistency between stage division and actual physical process, reduces the scrap rate caused by untimely manual correction, and reduces the maintenance workload of process engineers through automated adjustment.

[0070] The present application further proposes that deviation detection adopts a sliding window algorithm, and the window width is dynamically adjusted according to the signal stability.

[0071] The sliding window algorithm captures process signal segments using a continuously moving data window on the timeline. This is achieved using a fixed-step sliding window. The process signal data covered by each window is used for real-time comparison with the process map. This algorithm solves the lag problem caused by traditional fixed inspection cycles by continuously updating the inspection data range.

[0072] Dynamic window width adjustment adjusts the window coverage time based on real-time signal stability metrics. This can be assessed using variance calculation or frequency domain analysis. When increasing signal fluctuations are detected, the window width is automatically narrowed to improve the ability to capture transient deviations. When the signal stabilizes, the window width is automatically widened to suppress noise interference.

[0073] Specifically, during the process of process signal acquisition, the system continuously captures the data segments within the current sliding window and calculates the matching degree between them and the preset parameters of the corresponding stage in the process map. The moving step of the window can be set to be synchronized with the data sampling period to ensure that the window contains the latest process signal after each sliding. Signal stability is obtained by calculating the variance value of the data in the window. When the variance value exceeds the set threshold, it is determined to be an unstable signal state. At this time, the window width is shortened to half of the original length; when the variance value is lower than the stability threshold, it is determined to be a stable signal state, and the window width is expanded to twice the original length. By dynamically adjusting the window coverage, a balance is achieved between trend deviation detection under noise interference and instantaneous deviation capture under rapid fluctuations.

[0074] Through the above technical solution, this application solves the contradiction between sensitivity and noise immunity of a fixed detection window in scenarios where process signals are dynamically changing. In the injection molding process, when the mold temperature drops suddenly, the window width is dynamically reduced, allowing the system to identify temperature deviations within a short period of time and trigger cooling system adjustments. In the continuous steel rolling process, the window width automatically adjusts as the surface flatness of the strip changes, bringing the signal-to-noise ratio of the thickness detection results to the level required for stable production.

[0075] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the concept and scope of the present invention. Any modifications and improvements made to the technical solution of the present invention by a person of ordinary skill in the art without departing from the design concept of the present invention shall fall within the scope of protection of the present invention. The technical content for which protection is sought in the present invention is fully set forth in the claims.

Claims

1. An industrial signal phase matching method based on process map, characterized in that: The industrial signal phase matching method based on the process map includes: Constructing a standardized process map, which includes multiple process stages and preset weights for each stage; Collect industrial signal data in real time, extract signal features and match them with nodes of the process map; Dynamically assign stage weights based on matching results; Process control decisions are made based on dynamically assigned stage weights.

2. The industrial signal phase matching method based on process map according to claim 1 is characterized in that: The construction of the standardized process map includes: Divide the process stages according to engineering drawings or standard process flow; A weight value is preset for each process stage, where the weight value of the key process stage is greater than 1.

3. The industrial signal phase matching method based on process map according to claim 2 is characterized in that: The extracting signal features comprises: Identify slope mutation points and extreme points in the signal as phase boundary features; The stage boundary features are aligned and matched with the process map nodes.

4. The industrial signal phase matching method based on process map according to claim 1 or 3, characterized in that: The dynamic allocation stage weight includes timing correction operation: Detect the phase time offset between the actual signal and the process map; The signal is shifted and aligned to the corresponding node in the spectrum according to the offset.

5. The industrial signal phase matching method based on process map according to claim 4 is characterized in that: The timing correction operation is specifically as follows: When a phase delay is detected, the signal curve is shifted to the left by the offset time; When a phase advance is detected, the signal curve is shifted to the right by the offset duration.

6. The industrial signal phase matching method based on process map according to claim 1 is characterized in that: Also includes adaptive atlas adjustments: Compare the deviation between process signal and process map in real time; Dynamically adjust the atlas stage parameters based on the deviation; Record adjustment results and provide feedback to optimize decision rules.

7. The industrial signal phase matching method based on process map according to claim 6 is characterized in that: The parameters of the dynamic adjustment graph stage include: When the phase duration deviation exceeds the threshold, the phase duration is extended or shortened; When the trend of key parameter changes deviates from the graph, the adjustment phase triggers the conditions.

8. The industrial signal phase matching method based on process map according to claim 6 is characterized in that: The deviation detection adopts a sliding window algorithm, and the window width is dynamically adjusted according to the signal stability.

Images