A method and device for predicting the operating state of a reaction vessel

Abstract

This invention relates to the field of reactor operating status prediction technology, specifically a method and apparatus for predicting reactor operating status. The method includes: acquiring multiple raw signals from the reactor and preprocessing them to obtain a multi-parameter dataset; calculating the deviation between the bearing temperature change rate and the motor current change rate to obtain deviation data; fusing the deviation data with the shaft speed fluctuation rate to generate a load index; correcting the load index based on the vibration amplitude change rate to obtain fatigue accumulation parameters; calculating the acceleration rate of the fatigue accumulation parameters; performing trend quantification on the fatigue accumulation parameters and acceleration rate to obtain a comprehensive health score; and outputting the comprehensive health score and the result of a preset threshold to a terminal. This application overcomes the limitations of existing technologies that rely solely on fixed thresholds to judge the instantaneous value or simple statistical value of a single parameter, improving the early prediction capability and accuracy of potential reactor failures.

Description

Technical Field

[0001] This invention relates to the field of predicting the operating status of reactors, and more specifically, to a method and apparatus for predicting the operating status of reactors. Background Technology

[0002] In the production processes of chemical, daily chemical, adhesive coating, and new energy battery industries, reaction vessels are used as reaction equipment for material mixing, heating, cooling, and chemical reactions. However, when reaction vessels malfunction, it can affect product quality and production efficiency, and may also lead to equipment damage or safety hazards. Furthermore, under complex actual operating conditions, existing monitoring technologies are insufficient to meet the forward-looking needs of modern industry for equipment health management, which has become a key bottleneck restricting production continuity and optimizing equipment lifespan.

[0003] Existing methods for monitoring the operational status of reactors typically employ multi-sensor data acquisition and fixed threshold judgment mechanisms. These methods monitor basic physical parameters such as temperature, pressure, rotational speed, and vibration in real time, issuing alarm signals when any parameter exceeds a preset threshold range, thus prompting operators to intervene. Under relatively stable equipment operation conditions, this monitoring approach can achieve basic anomaly detection and immediate response. However, existing technologies struggle to effectively capture the dynamic, gradual changes in equipment operating status, resulting in insufficient early prediction capabilities for potential faults. Specifically, existing monitoring systems primarily rely on instantaneous parameter values ​​or simple statistical values ​​for independent judgment, making it difficult for the system to identify the changing trends of faults in their initial stages. Under complex operating conditions such as variable loads, variable temperatures, or multiphase reactions, the lack of dynamic trend acquisition capabilities often leads to delayed fault warnings, enabling only post-event alarms and failing to provide sufficient advance intervention time, thereby reducing the accuracy and practicality of predictions.

[0004] Therefore, a new method for predicting the operating status of reactors is urgently needed to solve the above problems. Summary of the Invention

[0005] The main objective of this invention is to provide a method and apparatus for predicting the operating status of a reactor, aiming to overcome the technical problem that existing technologies cannot predict the dynamic and gradual changes in the operating status of equipment.

[0006] To address the aforementioned problems, this invention proposes a method for predicting the operating status of a reactor, the method comprising: Multiple raw signals from the reactor were acquired and preprocessed to obtain a multi-parameter dataset, which included bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate. The deviation data is obtained by calculating the deviation between the bearing temperature change rate and the motor current change rate. The deviation data is then fused with the shaft speed fluctuation rate to generate load indicators. The fatigue accumulation parameter is obtained by correcting the load index based on the vibration amplitude change rate, and the acceleration rate of the fatigue accumulation parameter is calculated. The fatigue accumulation parameters and acceleration rate are subjected to trend quantification to obtain a comprehensive health score, and the comprehensive health score and the result of the preset threshold are output to the terminal.

[0007] Furthermore, the step of acquiring multiple raw signals from the reactor includes: Real-time data during the operation of the reactor is collected by multiple sensors, including temperature sensors, current sensors, speed sensors, and vibration sensors. The real-time data includes bearing temperature, motor temperature, motor current, shaft speed, and vibration amplitude. The real-time data is filtered to obtain the original signal after noise reduction.

[0008] Furthermore, the steps of preprocessing the original signal to obtain a multi-parameter dataset include: The bearing temperature change rate sequence is obtained by calculating the time derivative of the bearing temperature in the original signal, and the motor current change rate sequence is obtained by calculating the time derivative of the motor current. The shaft speed fluctuation rate sequence is obtained by calculating the absolute difference between the shaft speed and the rated speed, and the vibration amplitude change rate sequence is obtained by calculating the time derivative of the vibration amplitude. A multi-parameter dataset including bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate is obtained.

[0009] Further, the step of calculating the deviation data by comparing the bearing temperature change rate with the motor current change rate includes: The bearing temperature change rate and motor current change rate are empirically corrected, and the point-by-point difference is calculated on the corrected change rate sequence within a fixed time window to obtain the point-by-point deviation index. The point-to-point correlation deviation index sequence is summed successively within a fixed time window and statistically distributed to generate deviation data containing cumulative deviation and deviation variance.

[0010] Furthermore, the step of fusing the deviation data with the shaft speed fluctuation rate to generate the load index includes: The amplification factor is obtained by measuring the ratio of the absolute difference between the shaft speed fluctuation rate and the preset rated speed. The variance of the deviation and the shaft speed fluctuation rate are interactively calculated to obtain an interactive index; The amplification factor and interaction index are weighted and fused to obtain the load index.

[0011] Further, the step of correcting the load index based on the vibration amplitude change rate to obtain the fatigue accumulation parameter, and calculating the acceleration rate of the fatigue accumulation parameter, includes: The vibration amplitude change rate is weighted by a sensitivity coefficient and multiplied point by point with the load index to obtain the correction term sequence; The correction term sequence and the load index sequence are added point by point within the same time window to form a cumulative value sequence; The vibration amplitude change rate is interactively calculated based on the trend of the cumulative value sequence to obtain an interaction term sequence; The fatigue accumulation parameters are obtained by calculating the second time derivative of the fatigue accumulation parameters using the interaction term sequence and the cumulative value sequence.

[0012] Furthermore, the step of performing trend quantification processing on the fatigue accumulation parameters and acceleration rate to obtain a comprehensive health score includes: The relative change calculation of the fatigue accumulation parameter is performed to obtain the trend index, wherein the relative change calculation includes calculating the ratio of the difference between the current fatigue accumulation parameter and the initial reference value. The trend index and acceleration rate are subjected to a weighted subtraction operation to obtain a weighted penalty value; The comprehensive health score is obtained by subtracting weighted penalty values ​​point by point from the preset full score to form a scoring sequence.

[0013] Furthermore, the step of outputting the comprehensive health score and the result of the preset threshold to the terminal includes: The comprehensive health score is compared with preset high, medium and low thresholds in sequence, and a corresponding state classification label sequence is generated based on the comparison results. The status classification label sequence is mapped to a preset color code and output to the terminal display interface along with the numerical sequence of the comprehensive health score.

[0014] The present invention also proposes a device for predicting the operating status of a reactor, comprising: The acquisition module is used to acquire multiple raw signals from the reactor and preprocess them to obtain a multi-parameter dataset, which includes the bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate. The calculation module is used to calculate the deviation between the bearing temperature change rate and the motor current change rate to obtain deviation data, and to fuse the deviation data with the shaft speed fluctuation rate to generate load index. The correction module is used to correct the load index according to the vibration amplitude change rate to obtain fatigue accumulation parameters, and to calculate the acceleration rate of the fatigue accumulation parameters. The output module is used to perform trend quantification processing on the fatigue accumulation parameters and acceleration rate to obtain a comprehensive health score, and output the comprehensive health score and the result of a preset threshold to the terminal.

[0015] Compared with the prior art, this application has the following beneficial effects: This application proposes a method and apparatus for predicting the operating status of a reactor, which solves the limitations of existing technologies that rely solely on fixed thresholds to judge the instantaneous value of a single parameter or simple statistical values. This approach fails to effectively capture the dynamic and gradual trends of equipment under complex operating conditions, resulting in insufficient early fault prediction, delayed warnings, and post-event alarms. The method achieves correlation and fusion among multiple parameters, enabling the identification of phenomena such as increased bearing friction, insufficient lubrication, uneven load, and accelerated fatigue accumulation caused by vibration amplification. This enhances the early prediction capability and accuracy of potential reactor faults, provides sufficient time for early intervention, extends equipment life, ensures production continuity and safety, reduces the risk of equipment damage, and improves product quality and production efficiency. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] The structures, proportions, sizes, etc., shown in the accompanying drawings are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed in the specification, and are not intended to limit the implementation conditions of this application. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size should still fall within the scope of the technical content disclosed in this application, provided that they do not affect the effects and purposes that this application can produce.

[0018] Figure 1 This is a schematic diagram of the steps of a method for predicting the operating status of a reactor in one embodiment of the present invention; Figure 2 This is a schematic block diagram of a reactor operation status prediction device according to an embodiment of the present invention; Figure 3 This is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0019] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0021] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of features, integers, steps, operations, elements, modules, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, modules, components, and / or groups thereof. It should be understood that when an element is referred to as “connected” or “coupled” to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein may include wireless connection or wireless coupling. The term “and / or” as used herein includes all or any modules and all combinations of one or more associated listed items.

[0022] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0023] Reference Figure 1 This invention provides a method for predicting the operating status of a reactor, comprising the following steps: S1: Acquire multiple raw signals from the reactor and preprocess them to obtain a multi-parameter dataset, which includes bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate. In step S1, the bearing temperature is measured using a resistance temperature detector (RTD) or thermocouple temperature sensor, which is fixed to the bearing housing or near the bearing by bolts or adhesive bonding, directly contacting the heat conduction path to obtain the real-time temperature rise signal caused by friction or increased load. Motor current monitoring can be achieved using a non-contact current transformer or Hall effect current sensor, clamped to the motor power supply line or integrated into the frequency converter, enabling non-invasive continuous measurement and avoiding interference with normal motor operation. Simultaneously, shaft speed monitoring can be achieved by installing an encoder or magnetic induction speed sensor on the stirring shaft coupling or motor output end, recording rotational pulse signals to reflect the shaft's rotational stability. Vibration amplitude monitoring can be achieved using an acceleration or displacement vibration sensor, fixed to the shaft end support or vessel support, detecting minute axial or radial vibration changes. All these sensors are connected to a central data acquisition module or industrial PLC system via shielded cables to resist on-site electromagnetic interference and noise. The data acquisition module uses a relatively high fixed sampling frequency, such as 10 to 50 times per second, adjusted according to the equipment scale and the intensity of the reaction, to acquire raw signals in real time. These raw signals include bearing temperature, motor current, shaft speed, and vibration amplitude. During the acquisition process, sensor signals are sampled at the same clock, forming a unified time-series data. In actual industrial environments, reaction vessels are often exposed to high temperatures, high pressures, or corrosive media. Sensors are selected based on their temperature resistance, corrosion resistance, and explosion-proof rating to ensure long-term stable operation. The acquired raw signals are filtered, and the dynamic rate of change of each parameter is calculated based on the filtered raw time series. The bearing temperature rate of change refers to the rate of increase or decrease of bearing temperature per unit time, calculated by approximating the derivative through the difference between adjacent sampling points, expressed as: ΔT b = (T b(t) - T b (t-Δt) / Δt represents the rate of accumulation of frictional heat in the bearing. Under normal operating conditions, the rate of change is small, but it increases sharply under abnormal conditions such as insufficient lubrication. The motor current change rate ΔI is obtained by the difference between the current sequences, reflecting sudden changes in motor load or increased mechanical resistance. Because current is directly related to torque, an increase in the rate of change often indicates uneven load. The shaft speed fluctuation rate is the absolute deviation rate between the current speed and the rated speed, i.e., ΔN = |NN nominal | / N nominal , where N nominalTo determine the equipment's rated speed, it's crucial to assess whether speed instability is caused by changes in material viscosity or bearing wear. The vibration amplitude variation rate ΔV is calculated using differential methods to measure the vibration signal rate. Vibration amplitude represents the peak displacement or acceleration of the shaft; an increase in its variation rate indicates early fatigue or misalignment. These variation rates, as standardized features, eliminate the incomparability of absolute values ​​due to differences in equipment specifications. Furthermore, a sliding time window, such as data from the most recent 10 to 30 seconds, can be used in the calculation to obtain short-term dynamics rather than long-term trends, avoiding noise amplification. The multi-parameter dataset generated through the above preprocessing is a synchronized time series collection containing the original signal and the corresponding variation rate sequence. The bearing temperature variation rate and motor current variation rate exhibit thermoelectric coupling, the shaft speed fluctuation rate reflects the mechanical load, and the vibration amplitude variation rate reveals the accumulation of microscopic damage.

[0024] S2: The deviation data is obtained by calculating the deviation between the bearing temperature change rate and the motor current change rate. The deviation data is then fused with the shaft speed fluctuation rate to generate load index. In step S2, within the reactor, the stirring shaft is driven by a motor, and bearings support the shaft's rotation. The motor current represents the magnitude of the driving load. The bearing temperature is affected by frictional heat and heat conduction. Under normal and stable operation, an increase in motor current will lead to increased bearing friction through mechanical load conduction, thereby causing a corresponding rise in bearing temperature. When the rate of change of motor current increases, the rate of change of bearing temperature should increase proportionally. However, when abnormalities occur, such as poor bearing lubrication, shaft misalignment, or sudden changes in material viscosity, abnormal deviations will appear between the rate of change of bearing temperature and the rate of change of motor current. For example, the bearing temperature may rise too quickly while the current change lags behind, or vice versa. Deviation calculation involves extracting time-series data on bearing temperature change rate and motor current change rate from a multi-parameter dataset. An empirical correction coefficient is used to quantify the expected coupling ratio between the two. This correction coefficient can be pre-calibrated based on the specific specifications of the reactor and historical normal operation data, representing the expected increase in bearing temperature change rate corresponding to a unit change in motor current. For example, in a reactor, this coefficient can reflect the efficiency of heat conduction from the motor through the shaft and bearing. Deviation data is obtained by calculating the absolute difference between the bearing temperature change rate and the corrected motor current change rate point by point. This allows for the identification of positive or negative abnormal deviations. A positive deviation indicates an abnormally accelerated bearing temperature rise, possibly due to increased friction or lubrication failure. A negative deviation may indicate impeded heat conduction or poor load conduction. Statistical processing can also be performed on the deviation sequence within the time window, such as calculating cumulative deviation and deviation variance, to further quantify the persistence and volatility of the deviation, thus forming more comprehensive deviation data characteristics. In reactor operation, shaft speed fluctuations often originate from changes in material viscosity, uneven stirring resistance, or early bearing fatigue. These fluctuations amplify the deviation effects of thermoelectric parameters. For example, when the shaft speed fluctuation rate increases, even with moderate thermoelectric deviations, uneven load can lead to an increase in overall mechanical stress, potentially accelerating fatigue accumulation. This paper integrates deviation data with shaft speed fluctuation rate, using the mean or cumulative value of the deviation data as a base. Nonlinear amplification is achieved through the relative proportion of shaft speed fluctuation rate to the rated speed, while incorporating the interaction term between deviation variance and speed fluctuation to capture the dynamic amplification effect of fluctuation on deviation. In another embodiment, the short-term trend slope of the load index can be calculated to reflect the accumulation rate of load coupling effects. The fusion processing in this embodiment emphasizes the amplification and interactive modeling of speed fluctuations on thermoelectric deviations, thereby generating a comprehensive load index. This load index is sensitive to coupling anomalies in mechanical load and can more accurately characterize the reactor's operating load state under complex conditions. In practical implementation, real-time calculations can be performed using embedded processors or industrial control systems, combined with a sliding time window to ensure rapid response to dynamic changes.

[0025] S3: Correct the load index according to the vibration amplitude change rate to obtain the fatigue accumulation parameter, and calculate the acceleration rate of the fatigue accumulation parameter; In step S3, in the actual operating environment of the reactor, components such as the stirring shaft, bearings, and motor are subjected to alternating loads and vibration shocks for a long time. In particular, when the viscosity of the material inside the reactor changes or the reaction exotherm is uneven, the shaft speed fluctuation will intensify and the vibration amplitude will increase. This is often an early signal of bearing wear, shaft misalignment, or insufficient lubrication. This embodiment enhances the sensitivity to these hidden damages by correcting the vibration amplitude change rate. The vibration amplitude change rate represents the dynamic change of axial or radial displacement, originating from the impact caused by increased bearing clearance or uneven material mixing. In a reactor, a high vibration amplitude change rate significantly accelerates the propagation of fatigue cracks in components. The load index is corrected by adjusting the vibration amplitude. For example, a preset adjustment factor is used to multiply the vibration amplitude change rate by the load index to form a correction term, which is then added to the load index. Simultaneously, the interaction between the trend change of the load index and vibration is considered to generate a fatigue accumulation parameter. This fatigue accumulation parameter essentially quantifies the degree of mechanical fatigue under multi-parameter interaction, simulating the nonlinear accumulation process of actual damage. Because a high vibration amplitude change rate causes the correction term to grow exponentially, it reflects the accelerating contribution of vibration to fatigue. Furthermore, when calculating the fatigue accumulation parameter, the vibration amplitude change rate is extracted from the preprocessed multi-parameter dataset and applied to the correction of the load index. For example, a vibration correction term is calculated, which equals the vibration sensitivity coefficient multiplied by the vibration amplitude change rate multiplied by the load index value. The vibration sensitivity coefficient is an empirical value preset based on equipment specifications and historical data, typically reflecting the physical efficiency of vibration amplification of the load. This correction term is added to the load index and incorporated into the interaction term between the short-term trend of the load index and the rate of change of vibration amplitude to form a complete fatigue accumulation parameter. The generated fatigue accumulation parameter includes a description of the thermoelectric speed coupling of the load index, avoiding the prediction lag problem caused by shallow fusion of multiple parameters in existing technologies. A higher value of this parameter indicates that fatigue damage is accumulating rapidly. Under long-term continuous operation or high-load conditions in the reactor, the correction mechanism can highlight the early warning role of vibration. After obtaining the fatigue accumulation parameter, its acceleration rate is calculated. The acceleration rate is obtained by taking the second derivative of the fatigue accumulation parameter, i.e., the quadratic rate of change of the fatigue accumulation parameter with respect to time, quantifying the degree of acceleration of fatigue damage. For example, when the acceleration rate increases significantly positively, it indicates that fatigue accumulation is in a rapid deterioration stage, predicting the rapid propagation of bearing fatigue cracks or the risk of shaft instability. If the acceleration rate exceeds a certain empirical threshold, it can be marked as a fatigue acceleration stage in advance. The correction and calculation process can be performed in real time in the data processing module. A sliding time window can be used to smooth and differentially calculate the parameter sequence to reduce noise interference.

[0026] S4: Perform trend quantification on the fatigue accumulation parameters and acceleration rate to obtain a comprehensive health score, and output the comprehensive health score and the result of the preset threshold to the terminal.

[0027] In step S4, the fatigue accumulation parameters are normalized to establish a relative trend index relative to an initial baseline. This baseline can be taken from the fatigue accumulation parameter values ​​collected during the normal operation phase after new installation or major overhaul of the equipment. The trend index is obtained by calculating the difference between the current fatigue accumulation parameter and the baseline value and dividing by the baseline value. A positive trend index indicates that the relative degree of fatigue accumulation is increasing, while zero or close to zero indicates that the equipment is in a healthy and stable state. Simultaneously, the acceleration rate is used for weighted fusion to generate a comprehensive health score. The score can be designed as a quantitative index with a maximum score of 100 points, where the trend index has the dominant weight because it directly reflects long-term cumulative damage, and the acceleration rate is used as an auxiliary weight. To enhance sensitivity to sudden deterioration or accelerated failure, a scoring mechanism can be implemented by multiplying the trend index by a higher deduction factor using a preset weighting coefficient, and then multiplying the acceleration rate by a lower but more warning-emphasizing deduction factor, deducting from the full score item by item to obtain the final comprehensive health score. The advantage of this scoring mechanism is that it considers not only the absolute cumulative level of fatigue but also incorporates dynamic acceleration characteristics, making the score sensitive to both slow degradation and rapid deterioration. It can provide a more refined health gradient assessment; for example, a score above 90 indicates excellent equipment health, 80-90 indicates mild fatigue requiring attention, 60-80 indicates moderate fatigue requiring planned maintenance, and below 60 indicates serious potential risk requiring immediate intervention. In another embodiment, a moving average or exponential smoothing method can be used to filter the fatigue accumulation parameter and acceleration rate sequence to improve the stability of trend quantification. In practical implementation, empirical adjustment coefficients can be introduced to optimize weight allocation. These coefficients are preset based on equipment specifications, historical operating data, or laboratory validation. For example, the trend index can be weighted higher to emphasize the cumulative effect, while the acceleration rate can be weighted slightly lower but sufficient to lower the score in the early stages of deterioration, thus achieving early warning. The generated comprehensive health score is compared with a preset threshold, and the result is output to the terminal. The preset threshold can be set according to the equipment type and process requirements. For example, the safety threshold can be set to 70 points. When the comprehensive health score is higher than the threshold, a normal status indication is output, such as a green light or health text prompt. When it is lower than the threshold, a warning message is output, such as a yellow or red alarm accompanied by an audible and visual signal. At the same time, the terminal display can include the score value, trend curve, rate of change, and the proportion analysis of specific contributing parameters. For example, it can display the deduction ratio of fatigue accumulation parameters to the score or abnormal prompts of acceleration rate, which helps operators quickly locate the root cause of the problem, realizes real-time prediction of operating status, and supports data recording and remote transmission. Through the intuitive presentation of the terminal, operators can promptly detect health degradation caused by abnormal rate of temperature rise in reactor bearings, amplified deviation of motor current load, unstable shaft speed fluctuations, or rapid changes in vibration amplitude. This avoids production interruptions or safety accidents caused by sudden failures and improves the health management level and operational reliability of reactor mechanical components.

[0028] In one embodiment, the step of acquiring multiple raw signals from the reactor includes: Real-time data during the operation of the reactor is collected by multiple sensors, including temperature sensors, current sensors, speed sensors, and vibration sensors. The real-time data includes bearing temperature, motor temperature, motor current, shaft speed, and vibration amplitude. The real-time data is filtered to obtain the original signal after noise reduction.

[0029] In the above embodiments, real-time data during the operation of the reactor is collected by multiple sensors, including temperature sensors, current sensors, speed sensors, and vibration sensors. The temperature sensors are resistance temperature detectors (RTDs) or thermocouples, fixedly installed at key heat conduction points on the bearings and motor housing to capture changes in bearing and motor temperatures. The current sensors are non-contact Hall effect current sensors, clamped to the motor power supply line to monitor the instantaneous value of the motor current, thus reflecting the motor load condition. The speed sensors use encoders or Hall effect sensors, connected to the stirring shaft via a coupling to record real-time fluctuations in shaft speed. The vibration sensors are accelerometers, fixed at the shaft end support, sensitively capturing high-frequency changes in vibration amplitude. The sensors are connected to a data acquisition module via shielded cables to synchronously collect real-time data, including bearing temperature T, at a fixed sampling frequency (e.g., 10-50 times per second). b Motor temperature T m The real-time data includes motor current I, shaft speed N, and vibration amplitude V. These data are filtered to obtain the original signal after noise reduction. The filtering process can use a low-pass filter to remove high-frequency noise, and combine it with median filtering or Kalman filtering to remove outliers (such as abrupt changes that exceed physical limits) to ensure the smoothness and reliability of the signal.

[0030] In one embodiment, the step of preprocessing the original signal to obtain a multi-parameter dataset includes: The bearing temperature change rate sequence is obtained by calculating the time derivative of the bearing temperature in the original signal, and the motor current change rate sequence is obtained by calculating the time derivative of the motor current. The shaft speed fluctuation rate sequence is obtained by calculating the absolute difference between the shaft speed and the rated speed, and the vibration amplitude change rate sequence is obtained by calculating the time derivative of the vibration amplitude. A multi-parameter dataset including bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate is obtained.

[0031] In the above embodiment, the bearing temperature change rate sequence, ΔT, is obtained by calculating the time derivative of the bearing temperature in the original signal. b = dT b / dt can be calculated point-by-point within a fixed time window, reflecting the instantaneous rate of bearing temperature rise. The bearing temperature change rate is an early indicator of increased friction or insufficient lubrication. The time derivative calculation of the motor current yields the motor current change rate sequence ΔI = dI / dt, used to obtain the dynamic fluctuations of the motor load. An increase in the motor current change rate indicates an increase in material viscosity or mechanical resistance. The absolute difference between the shaft speed and the rated speed yields the shaft speed fluctuation rate sequence ΔN = |NN nominal |, where N nominal To design the rated speed of the equipment, the fluctuation rate quantifies the degree to which the speed deviates from stability. Speed ​​fluctuations often originate from uneven material distribution within the vessel or bearing fatigue, leading to unstable loads. The vibration amplitude change rate sequence ΔV = dV / dt is obtained by calculating the time derivative of the vibration amplitude. This rate of change highlights the accelerating trend of vibration, as the vibration amplitude change rate is a sensitive indicator of mechanical wear or shaft misalignment. In the above expression, d represents the mathematical differential symbol, indicating the derivative operation with respect to time. Through the above calculations, a multi-parameter dataset including the bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate is obtained.

[0032] In one embodiment, the step of calculating the deviation data by comparing the bearing temperature change rate with the motor current change rate includes: The bearing temperature change rate and motor current change rate are empirically corrected, and the point-by-point difference is calculated on the corrected change rate sequence within a fixed time window to obtain the point-by-point deviation index. The point-to-point correlation deviation index sequence is summed successively within a fixed time window and statistically distributed to generate deviation data containing cumulative deviation and deviation variance.

[0033] In the above embodiments, the bearing temperature change rate and motor current change rate are empirically corrected based on a coefficient k (e.g., 0.1-0.2℃ / A based on equipment specifications) to obtain a corrected sequence such as kΔI, representing the normal heat transfer efficiency. Within a fixed time window (e.g., the most recent 20-60 seconds), the point-by-point difference of the corrected rate of change sequence is calculated to obtain the point-by-point deviation index D. TI = |ΔT b -k*ΔI|, point-by-point deviation captures instantaneous inconsistencies, for example, if ΔT b A value significantly greater than kΔI indicates abnormal bearing friction leading to excessively rapid temperature rise; conversely, a value less than kΔI may indicate insufficient lubrication. The cumulative deviation SumD is obtained by successively adding the point-by-point deviation index sequence within a fixed time window. TI Simultaneously, statistical distribution calculations (such as calculating the mean and variance) are performed to generate deviation data containing cumulative deviation and deviation variance, where the deviation variance VarD TIThe variance indicates the volatility of the deviation, and high variance often indicates unstable coupling. The generation of deviation data represents the nonlinear interaction between the bearing temperature change rate and the motor current change rate. During normal operation, the two should be positively correlated. An increase in deviation indicates a potential fault. Furthermore, by using both cumulative and variance indicators, false positives from single threshold judgments are avoided, improving the comprehensiveness of load indicators and predictive sensitivity.

[0034] In one embodiment, the step of fusing the deviation data with the shaft speed fluctuation rate to generate a load index includes: The amplification factor is obtained by measuring the ratio of the absolute difference between the shaft speed fluctuation rate and the preset rated speed. The variance of the deviation and the shaft speed fluctuation rate are interactively calculated to obtain an interactive index; The amplification factor and interaction index are weighted and fused to obtain the load index.

[0035] In the above embodiments, the ratio of the absolute difference between the shaft speed fluctuation rate and the preset rated speed is obtained to obtain the amplification factor such as (1 + ΔN / N). nominal The factor quantifies the amplification effect of rotational speed deviation on load, because increased rotational speed fluctuations often stem from changes in material viscosity or bearing fatigue, leading to uneven overall load. An interactive calculation is performed on the deviation variance and shaft speed fluctuation rate to obtain an interactive index such as α*VarD. TI *ΔN (α is a preset adjustment factor, such as 0.05-0.1) highlights the nonlinear amplification of the deviation by fluctuations; high-speed fluctuations will significantly amplify the impact of deviation variance. The amplification factor and interaction index are weighted and fused to obtain the load index L. coup , such as L coup = SumD TI *Amplification factor + interactive index, weighted fusion emphasizes the coupling amplification effect of rotational speed on thermoelectric deviation. Even if the deviation is moderate, large rotational speed fluctuations will increase the load index, indicating potential uneven load. This embodiment mainly simulates the amplification mechanism of rotational speed fluctuations on thermoelectric anomalies in actual operation, avoiding the shallow analysis of simple superposition in the prior art, thereby generating load index.

[0036] In one embodiment, the step of correcting the load index based on the vibration amplitude change rate to obtain fatigue accumulation parameters, and calculating the acceleration rate of the fatigue accumulation parameters, includes: The vibration amplitude change rate is weighted by a sensitivity coefficient and multiplied point by point with the load index to obtain the correction term sequence; The correction term sequence and the load index sequence are added point by point within the same time window to form a cumulative value sequence; The vibration amplitude change rate is interactively calculated based on the trend of the cumulative value sequence to obtain an interaction term sequence; The fatigue accumulation parameters are obtained by calculating the second time derivative of the fatigue accumulation parameters using the interaction term sequence and the cumulative value sequence.

[0037] In the above embodiments, the vibration amplitude change rate is weighted by a sensitivity coefficient. The sensitivity coefficient is a preset empirical parameter used to adjust the amplification or suppression effect of vibration on the load. For example, it is set between 0.15 and 0.3 based on the specific model of the reactor and historical operating data to ensure that the contribution ratio of vibration changes under different operating conditions is reasonable. The weighted vibration amplitude change rate is multiplied point by point with the load index sequence to generate a correction term sequence. When the vibration amplitude change rate increases, it will significantly amplify the abnormal part in the load index. The correction term sequence is added point by point to the original load index sequence within the same time window to form a cumulative value sequence. The accumulation process is a gradual accumulation of fatigue because during long-term operation of the reactor, the small stresses caused by vibration will be repeatedly superimposed, leading to material fatigue of the bearings or shafts. Based on the overall trend of the cumulative value sequence, the rate of change of vibration amplitude is interactively calculated to obtain an interaction term sequence. Specifically, this involves the product of the trend slope and the rate of change of vibration, or correlation analysis, to capture the dynamic feedback of the vibration trend on the cumulative effect. For example, if the cumulative value sequence shows an upward trend, the interaction term will further strengthen the vibration contribution, and vice versa, thus more accurately reflecting the acceleration or deceleration stage of fatigue. The interaction term sequence is combined with the cumulative value sequence to form fatigue accumulation parameters. These parameters integrate load, vibration correction, and trend interaction, providing a quantitative characterization of the mechanical fatigue state of the reactor. The second derivative of time is calculated on the fatigue accumulation parameter sequence to obtain an acceleration rate sequence. The second derivative operation represents the acceleration change of fatigue accumulation. If the acceleration rate continues to increase positively, it indicates that fatigue has entered a rapid deterioration stage, predicting the risk of bearing failure or shaft breakage.

[0038] In one embodiment, the step of performing trend quantification processing on the fatigue accumulation parameter and acceleration rate to obtain a comprehensive health score includes: The relative change calculation of the fatigue accumulation parameter is performed to obtain the trend index, wherein the relative change calculation includes calculating the ratio of the difference between the current fatigue accumulation parameter and the initial reference value. The trend index and acceleration rate are subjected to a weighted subtraction operation to obtain a weighted penalty value; The comprehensive health score is obtained by subtracting weighted penalty values ​​point by point from the preset full score to form a scoring sequence.

[0039] In the above embodiments, the fatigue accumulation parameters are processed by relative change calculation to obtain a trend index. The relative change calculation specifically includes calculating the difference between the current fatigue accumulation parameter value and the initial baseline value, and then dividing it by the initial baseline value to obtain the ratio, that is, trend index = (current fatigue accumulation parameter - initial baseline value) / initial baseline value, where the initial baseline value is taken from the normal operation stage data after the equipment is newly installed or overhauled, the purpose of which is to normalize the fatigue accumulation so that it is not affected by the absolute scale. When the trend index is close to 0, it indicates that the equipment is healthy, and when it gradually increases, it indicates that fatigue is accumulating. The trend index and acceleration rate are weighted and subtracted to obtain a weighted penalty value. The weighted subtraction form is weighted penalty value = α * trend index - β * (1 - normalized acceleration rate). However, in practice, it is more often reflected as a positive accumulation of penalty values. That is, a penalty deduction mechanism that combines trend dominance and acceleration warning is used. For example, the trend index is the main weight (about 70%) and the acceleration rate is the auxiliary weight (about 30%). The subtraction highlights that if the acceleration rate is negative (fatigue slows down), the penalty value is reduced, and if it is positive, the penalty value is increased. This generates a comprehensive penalty deduction index. The penalty value represents the dual impact of static accumulation and dynamic deterioration of fatigue, avoiding the risk of ignoring sudden acceleration when only looking at the trend. A comprehensive health score is obtained by progressively subtracting weighted penalty values ​​from a preset maximum score to form a scoring sequence. The preset maximum score is typically 100 points, representing an ideal health state. The point-by-point subtraction operation ensures that the scoring sequence is dynamically updated over time. For example, in the initial stage, the penalty value is small, and the score is close to 100. As the running time increases, the trend index and acceleration rate increase, and the accumulated penalty value causes the score to decrease. A score below 80 can be considered mild fatigue, below 60 moderate fatigue, and below 40 severe fatigue. The core of this step lies in the combination of trend quantification and weighted penalty deduction, achieving a smooth mapping from fatigue parameters to health scores. Through relative changes and weighting mechanisms, the adaptability to individual differences in different reactors is improved, and it can be directly used for terminal display and status classification.

[0040] In one embodiment, the step of outputting the comprehensive health score and the result of a preset threshold to a terminal includes: The comprehensive health score is compared with preset high, medium and low thresholds in sequence, and a corresponding state classification label sequence is generated based on the comparison results. The status classification label sequence is mapped to a preset color code and output to the terminal display interface along with the numerical sequence of the comprehensive health score.

[0041] In the above embodiments, the preset thresholds include high, medium, and low thresholds. For example, the high threshold is set to 80 points, the medium threshold to 60 points, and the low threshold to 40 points. These thresholds are preset based on equipment specifications, historical fault data, and safety regulations to classify different risk levels. The comprehensive health score sequence is compared point by point with these preset thresholds. For example, for each time point, if the score is greater than the high threshold, it is classified as normal; if it is between the medium and high thresholds, it is classified as slightly abnormal; if it is between the low and medium thresholds, it is classified as moderate; and if it is below the low threshold, it is classified as severe. Based on this comparison result, a corresponding status classification label sequence is generated. The label sequence can be health, attention, warning, fault, or a numerical code for easy subsequent mapping and display. The status classification label sequence is mapped to a preset color code. For example, health is mapped to green, attention to yellow, warning to orange, and fault to red. The color code is based on the human-computer interaction principle, which facilitates operators to quickly identify risks. At the same time, it is output to the terminal display interface along with the comprehensive health score numerical sequence. The terminal can be an industrial touch screen, a computer monitoring system, or a mobile APP. The displayed content includes real-time numerical curves, color bars, trend arrows, and historical records, thus forming a complete visualization interface. In another embodiment, the output step can be extended to triggering an alarm. If the score falls below a low threshold or the label indicates a fault, an audible and visual alarm or push notification is triggered to ensure timely intervention. This embodiment transforms health scores into intuitive classifications and visual feedback, avoiding the burden on operators of directly interpreting complex parameters. Simultaneously, through multi-threshold comparisons and color mapping, it improves the accuracy and response speed of status recognition, achieving closed-loop prediction and display of the reactor's operating status, and enhancing the equipment's preventative maintenance capabilities. Reference Figure 2 A device for predicting the operating status of a reactor, comprising: The acquisition module 100 is used to acquire multiple raw signals from the reactor and preprocess them to obtain a multi-parameter dataset, which includes bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate. The calculation module 200 is used to calculate the deviation between the bearing temperature change rate and the motor current change rate to obtain deviation data, and to fuse the deviation data with the shaft speed fluctuation rate to generate load index. The correction module 300 is used to correct the load index according to the vibration amplitude change rate to obtain fatigue accumulation parameters and calculate the acceleration rate of the fatigue accumulation parameters. The output module 400 is used to perform trend quantification processing on the fatigue accumulation parameters and acceleration rate to obtain a comprehensive health score, and output the comprehensive health score and the result of a preset threshold to the terminal.

[0042] One embodiment of this application also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements a caster fatigue life testing method, including the following steps: acquiring multiple raw signals from a reaction vessel and preprocessing them to obtain a multi-parameter dataset, the multi-parameter dataset including bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate; calculating the deviation between the bearing temperature change rate and the motor current change rate to obtain deviation data; fusing the deviation data with the shaft speed fluctuation rate to generate a load index; correcting the load index according to the vibration amplitude change rate to obtain a fatigue accumulation parameter, and calculating the acceleration rate of the fatigue accumulation parameter; performing trend quantification processing on the fatigue accumulation parameter and the acceleration rate to obtain a comprehensive health score; and outputting the comprehensive health score and the result of a preset threshold to a terminal.

[0043] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for predicting the operating status of a reactor, characterized in that, include: Multiple raw signals from the reactor were acquired and preprocessed to obtain a multi-parameter dataset, which included bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate. The deviation data is obtained by calculating the deviation between the bearing temperature change rate and the motor current change rate. The deviation data is then fused with the shaft speed fluctuation rate to generate load indicators. The fatigue accumulation parameter is obtained by correcting the load index based on the vibration amplitude change rate, and the acceleration rate of the fatigue accumulation parameter is calculated. The fatigue accumulation parameters and acceleration rate are subjected to trend quantification to obtain a comprehensive health score, and the comprehensive health score and the result of the preset threshold are output to the terminal.

2. The method for predicting the operating status of a reactor according to claim 1, characterized in that, The step of acquiring multiple raw signals from the reactor includes: Real-time data during the operation of the reactor is collected by multiple sensors, including temperature sensors, current sensors, speed sensors, and vibration sensors. The real-time data includes bearing temperature, motor temperature, motor current, shaft speed, and vibration amplitude. The real-time data is filtered to obtain the original signal after noise reduction.

3. The method for predicting the operating status of a reactor according to claim 1, characterized in that, The steps for preprocessing the original signal to obtain a multi-parameter dataset include: The bearing temperature change rate sequence is obtained by calculating the time derivative of the bearing temperature in the original signal, and the motor current change rate sequence is obtained by calculating the time derivative of the motor current. The shaft speed fluctuation rate sequence is obtained by calculating the absolute difference between the shaft speed and the rated speed, and the vibration amplitude change rate sequence is obtained by calculating the time derivative of the vibration amplitude. A multi-parameter dataset including bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate is obtained.

4. The method for predicting the operating status of a reactor according to claim 1, characterized in that, The step of calculating the deviation data by comparing the bearing temperature change rate with the motor current change rate includes: The bearing temperature change rate and motor current change rate are empirically corrected, and the point-by-point difference is calculated on the corrected change rate sequence within a fixed time window to obtain the point-by-point deviation index. The point-to-point correlation deviation index sequence is summed successively within a fixed time window and statistically distributed to generate deviation data containing cumulative deviation and deviation variance.

5. The method for predicting the operating status of a reactor according to claim 4, characterized in that, The step of fusing the deviation data with the shaft speed fluctuation rate to generate load indicators includes: The amplification factor is obtained by measuring the ratio of the absolute difference between the shaft speed fluctuation rate and the preset rated speed. The variance of the deviation and the shaft speed fluctuation rate are interactively calculated to obtain an interactive index; The amplification factor and interaction index are weighted and fused to obtain the load index.

6. The method for predicting the operating status of a reactor according to claim 1, characterized in that, The step of correcting the load index based on the vibration amplitude change rate to obtain the fatigue accumulation parameter, and calculating the acceleration rate of the fatigue accumulation parameter, includes: The vibration amplitude change rate is weighted by a sensitivity coefficient and multiplied point by point with the load index to obtain the correction term sequence; The correction term sequence and the load index sequence are added point by point within the same time window to form a cumulative value sequence; The vibration amplitude change rate is interactively calculated based on the trend of the cumulative value sequence to obtain an interaction term sequence; The fatigue accumulation parameters are obtained by calculating the second time derivative of the fatigue accumulation parameters using the interaction term sequence and the cumulative value sequence.

7. The method for predicting the operating status of a reactor according to claim 1, characterized in that, The step of performing trend quantification processing on the fatigue accumulation parameters and acceleration rate to obtain a comprehensive health score includes: The relative change calculation of the fatigue accumulation parameter is performed to obtain the trend index, wherein the relative change calculation includes calculating the ratio of the difference between the current fatigue accumulation parameter and the initial reference value. The trend index and acceleration rate are subjected to a weighted subtraction operation to obtain a weighted penalty value; The comprehensive health score is obtained by subtracting weighted penalty values ​​point by point from the preset full score to form a scoring sequence.

8. The method for predicting the operating status of a reactor according to claim 1, characterized in that, The step of outputting the comprehensive health score and the result of the preset threshold to the terminal includes: The comprehensive health score is compared with preset high, medium and low thresholds in sequence, and a corresponding state classification label sequence is generated based on the comparison results. The status classification label sequence is mapped to a preset color code and output to the terminal display interface along with the numerical sequence of the comprehensive health score.

9. A device for predicting the operating status of a reactor, characterized in that, include: The acquisition module is used to acquire multiple raw signals from the reactor and preprocess them to obtain a multi-parameter dataset, which includes the bearing temperature change rate, motor current change rate, shaft speed fluctuation rate, and vibration amplitude change rate. The calculation module is used to calculate the deviation between the bearing temperature change rate and the motor current change rate to obtain deviation data, and to fuse the deviation data with the shaft speed fluctuation rate to generate load index. The correction module is used to correct the load index according to the vibration amplitude change rate to obtain fatigue accumulation parameters, and to calculate the acceleration rate of the fatigue accumulation parameters. The output module is used to perform trend quantification processing on the fatigue accumulation parameters and acceleration rate to obtain a comprehensive health score, and output the comprehensive health score and the result of a preset threshold to the terminal.

10. A computer device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in any one of claims 1 to 8.

Images