A method, device and medium for detecting the running state of a screw conveyor

Abstract

The embodiment of the application discloses a spiral conveyor operation state detection method, equipment and medium, belongs to the industrial equipment operation and maintenance detection technical field, and solves the problem that the timeliness and accuracy of the existing spiral conveyor state detection are low. Including, based on the historical state parameters corresponding to the spiral shaft of the spiral conveyor in the healthy state, a dynamic running state phase space is established; the real-time value set in the spiral conveyor is mapped into a real-time state point in the dynamic running state phase space; a plurality of real-time state points in continuous time sequence are connected to generate a real-time state evolution trajectory corresponding to the spiral conveyor; the real-time state evolution trajectory is analyzed to determine a symmetry breaking index corresponding to the real-time state evolution trajectory; based on the symmetry breaking index and the double deviation degree between the healthy benchmark mode in the dynamic running state phase space, the current running state of the spiral conveyor is determined.

Description

Technical Field

[0001] This application relates to the field of intelligent operation and maintenance technology for industrial equipment, and in particular to a method, equipment and medium for detecting the operating status of a screw conveyor. Background Technology

[0002] Screw conveyors, as key equipment for the continuous conveying of powders, granules, and small lumps, are widely used in mining, metallurgy, chemical, and food processing industries. Their core moving component, the screw shaft, operates under complex conditions for extended periods, making it prone to various malfunctions such as blade wear, bearing failure, and material blockage. Unplanned downtime can result in significant economic losses. Therefore, real-time and accurate monitoring of the screw conveyor's operating status is crucial for ensuring continuous production.

[0003] Currently, the condition monitoring of screw conveyors mainly adopts the traditional method based on parameter thresholds. Specifically, sensors are installed at different parts of the screw conveyor, and the data collected by the sensors is compared with preset alarm thresholds. When the monitored parameter exceeds the preset threshold, the system issues an alarm.

[0004] However, the threshold settings in this method typically rely on experience and are mostly static. If the equipment ages or its operating conditions change, the baseline parameters under normal operating conditions will fluctuate, rendering the static thresholds inapplicable. Furthermore, traditional methods can only provide a binary judgment of abnormal or normal operation, making it difficult to distinguish specific fault modes and even more difficult to identify early signs of degradation before fault characteristics fully manifest. Consequently, the timeliness and accuracy of screw conveyor condition monitoring are relatively low. Summary of the Invention

[0005] This application provides a method, device, and medium for detecting the operating status of a screw conveyor, which addresses the following technical problem: existing methods for detecting the status of screw conveyors by setting static alarm thresholds have low timeliness and accuracy.

[0006] The embodiments of this application adopt the following technical solutions: This application provides a method for detecting the operating status of a screw conveyor. The method includes: establishing a dynamic operating status phase space based on historical state parameters of the screw shaft when the screw conveyor is in a healthy state; wherein the historical state parameters include at least one of equivalent shaft torque, vibration signal, and bearing temperature rise signal; acquiring time-series data transmitted by multiple sensors installed on the screw conveyor, determining real-time values ​​corresponding to the historical state parameters based on the time-series data, and mapping the real-time values ​​to real-time state points in the dynamic operating status phase space; connecting multiple real-time state points in a continuous time sequence to generate a real-time state evolution trajectory corresponding to the screw conveyor; analyzing the geometric shape and kinematic characteristics of the real-time state evolution trajectory within a preset time window to extract trajectory dynamics features; determining the symmetry breaking index corresponding to the real-time state evolution trajectory based on the trajectory dynamics features; and determining the current operating status of the screw conveyor based on the symmetry breaking index and the dual deviation between the symmetry breaking index and the healthy baseline pattern in the dynamic operating status phase space.

[0007] In one implementation of this application, a dynamic operating state phase space is established based on the historical state parameters of the screw shaft of the screw conveyor under healthy conditions. Specifically, this includes: constructing a joint time series of torque and vibration, and a joint time series of temperature rise and torque, based on the historical state parameters; wherein the joint time series of torque and vibration represents the response relationship between the screw shaft drive and the load, and the joint time series of temperature rise and torque represents the coupling constraint relationship between temperature and force; performing time-delay cross-correlation analysis on the joint time series of torque and vibration to determine the dynamic time delay of the vibration response relative to the torque excitation, and performing time-delay cross-correlation analysis on the joint time series of temperature rise and torque to determine the thermal inertia time delay of the accumulated temperature rise relative to the load change; reconstructing the historical state parameters by time shift based on the dynamic time delay and the thermal inertia time delay to generate a multidimensional state vector; wherein the multidimensional state vector consists at least of the equivalent shaft torque, the time-shifted vibration response, and the time-shifted temperature rise rate; mapping the multidimensional state vector to a preset constraint phase space to obtain the dynamic operating state phase space; wherein the coordinate axis direction of the preset constraint phase space is related to the direction of maximum correlation determined by the time-delay cross-correlation analysis.

[0008] In one implementation of this application, real-time values ​​are mapped to real-time state points in a dynamic operating state phase space. Specifically, this includes: outputting an equivalent flow rate based on the real-time equivalent shaft torque and the screw shaft rotation speed through a material conveying power model; wherein the equivalent flow rate is related to the material flow state; demodulating the real-time vibration signal of the screw blades at the frequency to extract the structural response components related to the interaction with the screw shaft; constructing a decoupled observation vector in the dynamic operating state phase space based on the equivalent flow rate, structural response components, and real-time temperature rise rate; and directionally projecting the decoupled observation vector onto a preset coupled manifold established by health data, with the projection point being the real-time state point; wherein the direction of the directional projection is determined by the current equivalent flow rate and is used to separate changes in material properties from state changes caused by mechanical faults.

[0009] In one implementation of this application, multiple real-time state points in a continuous time sequence are connected to generate a real-time state evolution trajectory corresponding to the screw conveyor. Specifically, this includes: using the rotation period of the screw shaft as the basic time unit, slicing the multiple real-time state points in a continuous time sequence into rotation period slices to obtain a set of periodic state points composed of state points in each rotation period; calculating the weighted centroid of each set of periodic state points to obtain the periodic feature points of the average operating state of the screw conveyor within the rotation period; generating an initial state evolution trajectory based on the temporal position of the periodic feature points in the dynamic operating state phase space; determining the state abnormal fluctuation trajectory based on the transition probability and temporal causal relationship between adjacent periodic feature points, and inserting the state abnormal fluctuation trajectory between adjacent periodic feature points; and fusing the initial state evolution trajectory with the inserted state abnormal fluctuation trajectory to generate a real-time state evolution trajectory.

[0010] In one implementation of this application, the geometric shape and kinematic characteristics of the real-time state evolution trajectory within a preset time window are analyzed, and trajectory dynamics features are extracted. Specifically, this includes: dividing the real-time state evolution trajectory within the preset time window into multiple continuous trajectory segments based on the rotation period corresponding to the helical axis; performing least-squares ellipse fitting on each trajectory segment on the phase plane projection composed of equivalent shaft torque and vibration response; obtaining the coupled modal ellipse features corresponding to each trajectory segment based on the major axis tilt angle, minor-major axis ratio, and ellipse area of ​​the fitted ellipse; wherein, the major axis tilt angle is related to the dominant coupling phase of load and vibration within the period, the minor-major axis ratio is related to the nonlinearity of the vibration response, and the ellipse area is related to the energy dissipation intensity within the current rotation period; determining the trajectory morphology difference degree between adjacent sub-trajectory segments in the dynamic operating state phase space; and constructing trajectory dynamics features based on the coupled modal ellipse features and the trajectory morphology difference degree.

[0011] In one implementation of this application, a symmetry breaking index corresponding to the real-time state evolution trajectory is determined based on trajectory dynamics characteristic quantities. Specifically, this includes: selecting energy distributions corresponding to several characteristic frequencies that are multiples of the number of helical blades from the trajectory dynamics characteristic quantities; determining the offset of the centroid of the energy distribution relative to the fundamental frequency of the helical shaft rotation, and using it as the geometric symmetry breaking degree; selecting characteristic sequences reflecting the phase correlation between load changes and vibration response from the trajectory dynamics characteristic quantities, and determining the cyclic steady-state strength corresponding to the characteristic sequences; determining the benchmark value of the cyclic steady-state strength under healthy conditions, and using the attenuation rate of the real-time cyclic steady-state strength relative to the benchmark value as the functional symmetry breaking degree; and nonlinearly coupling the geometric symmetry breaking degree and the functional symmetry breaking degree to generate a symmetry breaking index.

[0012] In one implementation of this application, the current operating state of the screw conveyor is determined based on the dual deviation between the symmetry breaking index and the health benchmark pattern in the dynamic operating state phase space. Specifically, this includes: constructing a health benchmark manifold and a symmetry breaking threshold envelope surface in the dynamic operating state phase space based on the trajectory dynamics characteristics of historical health cycles; mapping the currently calculated symmetry breaking index to the dynamic operating state phase space to obtain an instantaneous state vector; determining a first deviation based on the deviation distance between the instantaneous state vector and the health benchmark manifold; determining a second deviation based on the relative positional relationship between the instantaneous state vector and the symmetry breaking threshold envelope surface; and determining the current operating state of the screw conveyor based on the local curvature change characteristics of the real-time state evolution trajectory, the first deviation, and the second deviation. The operating state includes at least one of the following: a healthy state, blade wear warning, bearing loosening alarm, and material blockage fault.

[0013] In one implementation of this application, the current operating state of the screw conveyor is determined based on the local curvature abrupt change features of the real-time state evolution trajectory, a first deviation, and a second deviation. Specifically, this includes: identifying abnormal event feature points in the dynamic operating state phase space based on the local curvature abrupt change features of the real-time state evolution trajectory; constructing a multi-dimensional anomaly representation space based on the real-time state evolution trajectory based on the first deviation and the second deviation; mapping the abnormal event feature points to the multi-dimensional anomaly representation space to form an abnormal event spatial distribution; performing multi-scale clustering analysis on the abnormal event spatial distribution to identify different fault feature clusters; and comparing the distribution of each fault feature cluster in the multi-dimensional anomaly representation space with a preset fault mode knowledge base to determine the current operating state of the screw conveyor.

[0014] This application provides a screw conveyor operation status detection device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to: establish a dynamic operation status phase space based on historical state parameters corresponding to the screw shaft of the screw conveyor in a healthy state; wherein the historical state parameters include at least one of equivalent shaft torque, vibration signal, and bearing temperature rise signal; acquire time-series data sent by multiple sensors installed on the screw conveyor, determine the real-time value corresponding to the historical state parameters based on the time-series data, and map the real-time value to a real-time state point in the dynamic operation status phase space; connect multiple real-time state points in a continuous time sequence to generate a real-time state evolution trajectory corresponding to the screw conveyor; analyze the geometric shape and kinematic characteristics of the real-time state evolution trajectory within a preset time window, and extract trajectory dynamics features; determine the symmetry breaking index corresponding to the real-time state evolution trajectory based on the trajectory dynamics features; and determine the current operation status of the screw conveyor based on the symmetry breaking index and the dual deviation between the healthy baseline mode in the dynamic operation status phase space and the healthy baseline mode.

[0015] This application provides a non-volatile computer storage medium storing computer-executable instructions. These instructions are configured to: establish a dynamic operating state phase space based on historical state parameters corresponding to the screw shaft of a screw conveyor in a healthy state; wherein the historical state parameters include at least one of equivalent shaft torque, vibration signal, and bearing temperature rise signal; acquire time-series data transmitted by multiple sensors installed on the screw conveyor, determine real-time values ​​corresponding to the historical state parameters based on the time-series data, and map the real-time values ​​to real-time state points in the dynamic operating state phase space; connect multiple real-time state points in a continuous time sequence to generate a real-time state evolution trajectory corresponding to the screw conveyor; analyze the geometric shape and kinematic characteristics of the real-time state evolution trajectory within a preset time window to extract trajectory dynamics features; determine the symmetry breaking index corresponding to the real-time state evolution trajectory based on the trajectory dynamics features; and determine the current operating state of the screw conveyor based on the symmetry breaking index and the dual deviation between the symmetry breaking index and the healthy baseline pattern in the dynamic operating state phase space.

[0016] The above-mentioned technical solutions adopted in this application embodiment can achieve the following beneficial effects: By establishing a dynamic operating state phase space, this application embodiment maps the multi-source heterogeneous state information of the screw conveyor to a high-dimensional geometric space, realizing the structured and visualized operation state, and solving the problem that traditional single-parameter analysis is difficult to obtain the overall dynamic behavior of the system. Connecting the time-series state points into an evolution trajectory realizes the dynamic evolution of the equipment state and can determine the trend of state change. Furthermore, by analyzing the geometric shape and kinematic characteristics of the trajectory, the extracted feature quantities can better reflect the inherent health status and fault mechanism of the system. In addition, by calculating the symmetry breaking index, which characterizes the degree of order disruption, the degree of equipment deviation from the healthy operating mode can be quantified, realizing sensitive early warning of early, slowly changing faults. Finally, by evaluating the dual deviation of this index from the healthy baseline mode and combining it with the local mutation characteristics of the trajectory for comprehensive decision-making, it can not only accurately determine whether the equipment is abnormal, but also effectively distinguish different fault modes, improving the accuracy of state detection, early warning capability, and the pertinence of fault diagnosis. Attached Figure Description

[0017] 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 recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 A flowchart of a method for detecting the operating status of a screw conveyor provided in this application embodiment; Figure 2 This is a schematic diagram of the structure of a screw conveyor operation status detection device provided in an embodiment of this application.

[0018] Figure label: 200: Screw conveyor operation status detection equipment; 201: Processor; 202: Memory. Detailed Implementation

[0019] This application provides a method, equipment, and medium for detecting the operating status of a screw conveyor.

[0020] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0021] Figure 1 A flowchart of a method for detecting the operating status of a screw conveyor provided in this application embodiment is shown below. Figure 1 As shown, the method for detecting the operating status of a screw conveyor includes the following steps: S101. Based on the historical state parameters of the screw shaft of the screw conveyor in a healthy state, establish a dynamic operating state phase space.

[0022] In one implementation of this application, a joint time series of torque and vibration, and a joint time series of temperature rise and torque are constructed based on historical state parameters. The joint time series of torque and vibration represents the response relationship between the helical shaft drive and the load, while the joint time series of temperature rise and torque represents the coupling constraint relationship between temperature and force. Time-delay cross-correlation analysis is performed on the joint time series of torque and vibration to determine the dynamic time delay of the vibration response relative to the torque excitation, and time-delay cross-correlation analysis is performed on the joint time series of temperature rise and torque to determine the thermal inertia time delay of the accumulated temperature rise relative to the load change. Based on the dynamic and thermal inertia time delays, the historical state parameters are reconstructed over time to generate a multidimensional state vector. This multidimensional state vector consists at least of the equivalent shaft torque, the time-shifted vibration response, and the time-shifted temperature rise rate. The multidimensional state vector is mapped to a preset constraint phase space to obtain a dynamic operating state phase space. The coordinate axis directions of the preset constraint phase space are related to the direction of maximum correlation determined by the time-delay cross-correlation analysis.

[0023] Specifically, embodiments of this application extract the equivalent shaft torque timing, vibration response timing, and bearing temperature rise timing from historical state parameters. Considering the dynamic characteristics of the screw conveyor's screw shaft, the equivalent shaft torque timing and vibration response timing are aligned with the same timestamp to form a joint timing of torque and vibration used to characterize the relationship between the drive-end excitation and the load-end mechanical response. Simultaneously, the bearing temperature rise timing is aligned with the equivalent shaft torque timing to form a joint timing of temperature rise and torque used to characterize the thermo-mechanical coupling constraint relationship between frictional heat generation and load.

[0024] A time-delay cross-correlation analysis was performed on the constructed joint time series of torque and vibration. The cross-correlation coefficients of the torque and vibration signals were calculated at different time offsets. The time offset corresponding to the first significant peak of the cross-correlation coefficient was determined as the dynamic time delay of the vibration response relative to the torque excitation. This time delay reflects the propagation time of the mechanical stress wave in the helical shaft structure. Similarly, a time-delay cross-correlation analysis was performed on the constructed joint time series of temperature rise and torque. The time offset corresponding to the peak of the cross-correlation coefficient was determined as the thermal inertia time delay of the cumulative temperature rise relative to the load change. This time delay reflects the delay from the mechanical energy doing work to its conversion into heat energy and being sensed by the sensor.

[0025] Based on the calculated dynamic and thermal inertial time delays, the original historical state parameters are time-shifted and aligned. Specifically, the vibration response time sequence is shifted forward by the duration of the mechanical time delay, and the bearing temperature rise time sequence is shifted forward by the duration of the thermal inertial time delay. After the shift, the equivalent shaft torque value at the same moment, the vibration response value after the time shift, and the temperature rise rate value after the time shift are combined into a three-dimensional vector. Traversing all historical moments generates a multi-dimensional state vector sequence composed of multiple three-dimensional vectors.

[0026] The generated multidimensional state vector sequence is used as input to calculate its covariance matrix. Eigenvalue decomposition is performed on this covariance matrix, and the two eigenvectors with the largest eigenvalues ​​are selected as the principal coordinate axes for constructing the pre-defined constraint phase space. These two principal directions correspond to the two linear combinations with the strongest correlation among torque, vibration, and temperature rise, respectively, and can best characterize the healthy operating mode of the system. Finally, all multidimensional state vectors are projected onto a two-dimensional plane formed by these two principal directions as coordinate axes. The resulting set of points and its distribution structure are defined as the dynamic operating state phase space.

[0027] S102. Obtain the timing data sent by multiple sensors set on the screw conveyor, determine the real-time value corresponding to the historical state parameters based on the timing data, and map the real-time value to the real-time state point in the dynamic operating state phase space.

[0028] In one implementation of this application, an equivalent flow rate is output through a material conveying power model based on the real-time equivalent shaft torque and the screw shaft rotation speed; wherein, the equivalent flow rate is related to the material flow state. The real-time vibration signal is demodulated by the screw blades to extract the structural response components related to the interaction with the screw shaft. In the dynamic operating state phase space, a decoupled observation vector is constructed based on the equivalent flow rate, structural response components, and real-time temperature rise rate. The decoupled observation vector is directionally projected onto a preset coupled manifold established from health data; the projection point is the real-time state point. The direction of the directional projection is determined by the current equivalent flow rate and is used to separate changes in material properties from state changes caused by mechanical faults.

[0029] Specifically, the voltage and current signals of the drive motor are acquired in real time, and the real-time output torque of the screw shaft, i.e., the equivalent shaft torque, is calculated using a lookup table method. The real-time rotational speed of the screw shaft is simultaneously acquired, and the real-time equivalent shaft torque and screw shaft speed are input into a preset material conveying power model. This model is calibrated based on the design parameters of the screw conveyor and the material characteristics. Its core formula is that the conveying power equals the torque multiplied by the angular velocity, and the equivalent material mass flow rate is inversely calculated under efficiency coefficient correction. The calculation result output by the model in real time is the equivalent flow rate characterizing the current overall material flow state. The vibration acceleration signal at the bearing housing of the screw conveyor is acquired in real time, preprocessed, and then subjected to spectrum analysis. In the spectrum diagram, a series of characteristic frequencies are identified with the screw shaft rotation frequency as the fundamental frequency and multiples of the number of screw blades, i.e., the blade passing frequency and its harmonics. Using these characteristic frequencies as the center, a reasonable narrowband filter is set to filter the original vibration signal and extract the signal components within this series of frequency bands. The component is demodulated by Hilbert transform to obtain its envelope signal. This envelope signal is the structural response component that mainly reflects the periodic impact between the propeller blade and the casing after removing low-frequency material disturbances and high-frequency noise.

[0030] In the established dynamic operating state phase space, the coordinate axes or coordinate dimensions corresponding to the equivalent flow rate, structural response component, and real-time temperature rise rate are determined. Bearing temperature signals are acquired in real-time, and their temperature rise rate is calculated. The calculated real-time equivalent flow rate value, real-time structural response component amplitude characteristics, and real-time calculated temperature rise rate value are combined according to the dimensional order defined in the phase space to form a three-dimensional observation vector, which is the decoupled observation vector. Using historical health data, a low-dimensional health state surface, i.e., a preset coupled manifold, is fitted in the dynamic operating state phase space using a manifold learning algorithm. A projection direction vector is defined, its direction determined by the current equivalent flow rate value through a predefined monotonic function, ensuring a smooth change in projection direction with the flow rate. The constructed current decoupled observation vector is projected perpendicularly onto the preset coupled manifold along the direction indicated by the projection direction vector. The intersection of this perpendicular line and the manifold surface is the mapped real-time state point, after eliminating the influence of material flow rate changes. Through this mechanism, the same mechanical state will be mapped to similar positions on the manifold under different material flow rates, while deviations caused by mechanical failures will be highlighted.

[0031] S103. Connect multiple real-time state points in a continuous time sequence to generate the real-time state evolution trajectory corresponding to the screw conveyor.

[0032] In one implementation of this application, the rotation period of the helical shaft is used as the basic time unit. Multiple real-time state points in a continuous time sequence are sliced ​​according to their rotation periods to obtain a set of periodic state points composed of state points for each rotation period. For each set of periodic state points, its weighted centroid is calculated to obtain the periodic feature points of the average operating state of the screw conveyor within the rotation period. Based on the temporal position of the periodic feature points in the dynamic operating state phase space, an initial state evolution trajectory is generated. Based on the transition probabilities and temporal causal relationships between adjacent periodic feature points, an abnormal state fluctuation trajectory is determined and inserted between adjacent periodic feature points. The initial state evolution trajectory and the inserted abnormal state fluctuation trajectory are fused to generate a real-time state evolution trajectory.

[0033] Specifically, the instantaneous phase of the screw shaft rotation is acquired in real time using an encoder or speed sensor. Based on this phase signal, all state points corresponding to each complete rotation of the screw shaft are extracted from the continuously arriving real-time state point sequence. All state points extracted within each rotation cycle are grouped into the same set, forming multiple periodic state point sets arranged in rotation cycle order. For each obtained periodic state point set, a weight is assigned to each state point. This weight is proportional to the absolute value of the sine of the screw shaft rotation phase angle corresponding to that state point, emphasizing the phase region with the strongest interaction force between the blades, material, and casing. Subsequently, the weighted average value of all state points in the periodic state point set across each coordinate dimension of the dynamic operating state phase space is calculated. The point corresponding to this weighted average value in the phase space is the periodic characteristic point of that rotation cycle. This characteristic point characterizes the average level of the screw conveyor's operating state within that cycle.

[0034] Following the chronological order of the rotation cycles, a series of periodic feature points are sequentially connected in the dynamic operating state phase space. Piecewise cubic Hermite interpolation is used for the connection to ensure the smoothness of the trajectory. The resulting continuous curve constitutes the initial state evolution trajectory, characterizing the macroscopic and stable evolution trend of the screw conveyor's operating state.

[0035] For any two adjacent periodic feature points on the obtained initial trajectory, analyze their original two corresponding periodic state point sets. Calculate the transition probability matrix from the state point of the previous period to the state point of the next period. Simultaneously, use the temporal Granger causality test to analyze whether there is a significant causal driving relationship between these two point sets. If the main diagonal elements of the transition probability matrix are significantly low, or the Granger causality relationship indicates an abnormal causal direction, then it is determined that an abnormal state fluctuation has occurred during this adjacent period. In this case, abandon simple interpolation, instead merge the two periodic state point sets, and use the optimal path generated based on the dynamic time warping algorithm that best matches the temporal evolution of all state points between these two point sets as the abnormal state fluctuation trajectory for this adjacent period. Traverse all adjacent periodic feature point pairs on the initial state evolution trajectory. For the interval determined to have abnormal fluctuations, replace the original interpolation curve segment of that interval in the initial trajectory with the calculated abnormal state fluctuation trajectory. For normal intervals not determined to have abnormal fluctuations, retain the original interpolation curve segment of the initial trajectory. By connecting the trajectory segments of all normal and abnormal fluctuation ranges in chronological order, a complete real-time state evolution trajectory that simultaneously includes macroscopic evolution trends and microscopic abnormal fluctuations is generated.

[0036] S104. Analyze the geometric shape and kinematic characteristics of the real-time state evolution trajectory within a preset time window, and extract trajectory dynamics features.

[0037] In one implementation of this application, the real-time state evolution trajectory within a preset time window is divided into multiple continuous trajectory segments based on the rotation period corresponding to the helical shaft. Least-squares ellipse fitting is performed on each trajectory segment on the phase plane projection formed by the equivalent shaft torque and vibration response. Based on the major axis tilt angle, minor-major axis ratio, and ellipse area of ​​the fitted ellipse, the coupled modal elliptical features corresponding to each trajectory segment are obtained; wherein, the major axis tilt angle is related to the dominant coupling phase of load and vibration within that period, the minor-major axis ratio is related to the nonlinearity of the vibration response, and the ellipse area is related to the energy dissipation intensity within the current rotation period. In the dynamic operating state phase space, the trajectory morphology difference between adjacent sub-trajectory segments is determined, and trajectory dynamics features are constructed based on the coupled modal elliptical features and the trajectory morphology difference.

[0038] Specifically, the real-time rotation frequency of the helical shaft is obtained, and the number of complete rotation cycles contained within a preset time window is calculated based on the length of the window. Then, using the timestamp of the previous trajectory point as a reference, the entire trajectory within the preset time window is sequentially divided into multiple continuous trajectory segments, with the time length corresponding to each rotation of the helical shaft as a segment unit. Each trajectory segment contains all trajectory points belonging to the same rotation cycle, ensuring that the time scale of subsequent analysis is strictly synchronized with the physical motion cycle of the equipment. For each segment, the coordinate values ​​of each trajectory point in the equivalent shaft torque dimension and vibration response dimension are extracted, forming a set of scattered points on a two-dimensional plane. A least-squares ellipse fitting algorithm is used to fit this set of scattered points, obtaining an ellipse equation that best describes its distribution. This ellipse equation includes geometric parameters such as the center position of the ellipse, the lengths of the major and minor axes, and the tilt angle of the major axis relative to the torque coordinate axis.

[0039] For each trajectory segment fitted to an ellipse, three key geometric parameters are analytically derived from the ellipse equation. The first parameter is the tilt angle of the ellipse's major axis relative to the equivalent torque coordinate axis, i.e., the major axis tilt angle. This tilt angle reflects the dominant phase coupling relationship between load variation and vibration response within a rotational cycle. The second parameter is the ratio of the ellipse's minor axis length to its major axis length, i.e., the minor-major axis ratio. This ratio characterizes the nonlinearity and directional purity of the system's vibration response. The third parameter is the area of ​​the ellipse, calculated by multiplying pi by the product of the major and minor semi-axes. Its physical meaning is the area enclosed by the energy dissipation loop formed within the torque-vibration phase plane during the current rotational cycle, directly related to the intensity of mechanical energy dissipation. These three parameters for each trajectory segment collectively constitute the coupled modal elliptical characteristics of that cycle.

[0040] In the dynamic operating state phase space, for two temporally adjacent trajectory segments, the trajectory morphology difference degree between them is calculated. The set of shortest distances from any point in the preceding trajectory segment to the following trajectory segment, and the set of shortest distances from any point in the following trajectory segment to the preceding trajectory segment, are calculated separately. The maximum value of these two distance sets is taken as the distance between the two adjacent trajectory segments; this distance value is quantified as the trajectory morphology difference degree. This difference degree reflects the degree of abrupt change or drift in the equipment's operating state between adjacent cycles. The coupled modal elliptical features extracted from each trajectory segment within the current analysis window—namely, the major axis tilt angle, the ratio of the minor to the major axis, and the ellipse area—are arranged in chronological order to form three time-series sequences. Simultaneously, all trajectory morphology differences between adjacent trajectory segments are also arranged in chronological order into a sequence. Finally, these four time-series sequences are aligned on the time axis and sequentially concatenated into a multi-dimensional feature vector. This feature vector is defined as the trajectory dynamics feature quantity corresponding to the time window.

[0041] S105. Based on trajectory dynamics features, determine the symmetry breaking index corresponding to the real-time state evolution trajectory.

[0042] In one implementation of this application, energy distributions corresponding to several characteristic frequencies that are multiples of the number of helical blades are selected from the trajectory dynamics features. The offset of the centroid of the energy distribution relative to the fundamental frequency of the helical shaft rotation is determined and used as the geometric symmetry breaking degree. From the trajectory dynamics features, a feature sequence reflecting the phase correlation between load changes and vibration response is selected, and the corresponding cyclic steady-state strength is determined. A baseline value for the cyclic steady-state strength under healthy conditions is determined, and the attenuation rate of the real-time cyclic steady-state strength relative to the baseline value is used as the functional symmetry breaking degree. The geometric symmetry breaking degree and the functional symmetry breaking degree are nonlinearly coupled to generate a symmetry breaking index.

[0043] Specifically, a fast Fourier transform is performed on the vibration components of the trajectory dynamics characteristic quantities to obtain the spectrum. The fundamental rotational frequency of the helical shaft is accurately identified within the spectrum, and a series of harmonic frequencies corresponding to multiples of this fundamental frequency and the number of helical blades are located. Next, the signal energy within the frequency bands corresponding to these specific harmonic frequencies is calculated, forming an energy distribution sequence. Subsequently, the weighted average frequency of this energy distribution sequence, i.e., the energy centroid frequency, is calculated. Finally, the difference between this energy centroid frequency and the theoretical value, which is an integer multiple of the fundamental rotational frequency, is calculated and normalized by dividing it by the fundamental rotational frequency. The resulting relative offset is quantified as the geometric symmetry breaking degree, which directly reflects the degree of physical rotational imbalance caused by uneven blade wear.

[0044] Load and vibration characteristic sequences are extracted from trajectory dynamics parameters. Second-order cyclostationary analysis is performed on these two sets of time-series data to calculate the cyclic autocorrelation function (CEC) of the vibration signal at the load characteristic frequency. The amplitude of the CEC at its cyclic frequency equal to the load characteristic frequency is defined as the cyclostationary strength, which quantifies the stability of phase lock between load changes and vibration response. A baseline value of the cyclostationary strength, statistically obtained from long-term operation of the equipment under healthy conditions, is acquired. The current cyclostationary strength is calculated in real-time, and its attenuation ratio relative to the healthy baseline value is calculated. This attenuation ratio is quantified as the functional symmetry breaking degree, which characterizes the degree of efficiency reduction in the system's energy transfer path.

[0045] Independent activation thresholds are set for geometric symmetry breaking and functional symmetry breaking. When the value of either breaking degree is below its corresponding threshold, it is considered to have a small impact on the overall system symmetry, and a linear weighted sum is used to initially fuse the two. When the value of any breaking degree exceeds its activation threshold, a nonlinear amplification mechanism is triggered. This mechanism amplifies the breaking degree component exceeding the threshold using an exponential function, and then weights and fuses the amplified component with another component. Finally, the result of the fusion calculation is mapped to the interval between zero and one using a sigmoid function, generating a comprehensive symmetry breaking index that is easy to interpret and has a unified alarm threshold set. This index can respond sensitively to both single and complex faults.

[0046] S106. Based on the symmetry breaking index and the dual deviation between the health benchmark mode in the dynamic operating state phase space, the current operating state of the screw conveyor is determined.

[0047] In one implementation of this application, in the dynamic operating state phase space, a healthy baseline manifold and a symmetry breaking threshold envelope are constructed based on the trajectory dynamics characteristics of historical health cycles. The currently calculated symmetry breaking index is mapped into the phase space to obtain an instantaneous state vector. Based on the deviation distance between the instantaneous state vector and the healthy baseline manifold, a first deviation is determined. Based on the relative positional relationship between the instantaneous state vector and the symmetry breaking threshold envelope, a second deviation is determined. Based on the local curvature abrupt change characteristics of the real-time state evolution trajectory, the first deviation, and the second deviation, the current operating state of the screw conveyor is determined; wherein the operating state includes at least one of the following: healthy state, blade wear warning, bearing loosening alarm, and material blockage fault.

[0048] Specifically, in the dynamic operating state phase space, a large number of trajectory dynamic feature sample points corresponding to the equipment being in a healthy operating phase are selected. Using manifold learning algorithms such as local linear embedding or isometric mapping, dimensionality reduction and structure learning are performed on the high-dimensional sample point cloud to fit a low-dimensional smooth surface that can characterize the core distribution of the healthy state; this surface is the healthy baseline manifold. Simultaneously, the Euclidean distance from all healthy sample points to this baseline manifold is calculated, and the distance value corresponding to a specified high quantile of its statistical distribution is taken as the threshold radius. Using this threshold radius as the normal distance, an isometric outwardly expanding surface enclosing the healthy baseline manifold is constructed; this surface is the symmetry-broken threshold envelope surface, used to define the boundary of the healthy state.

[0049] The calculated symmetry breaking index is converted into a coordinate point in phase space based on a preset mapping relationship with phase space coordinates, forming an instantaneous state vector. The shortest Euclidean distance from this instantaneous state vector to the healthy baseline manifold is calculated; this distance is defined as the first deviation, used to quantify the degree to which the current state deviates from the ideal healthy pattern. Next, it is determined whether the instantaneous state vector is inside or outside the symmetry breaking threshold envelope. If it is inside the envelope, the second deviation is zero; if it is outside the envelope, the shortest distance from the vector to the envelope is calculated and recorded as the second deviation, used to quantify whether the current state has exceeded the health tolerance range. On the current real-time state evolution trajectory, a window containing a preset number of trajectory points is extracted forward and backward, with the point corresponding to the instantaneous state vector as the analysis center. Using the coordinates of discrete points on the trajectory, the discrete curvature sequence of the trajectory line within the window is calculated. A difference operation is performed on this curvature sequence to find the abrupt change point where the absolute value of the curvature difference exceeds a preset threshold. The number of such abrupt changes within the statistical window is counted and combined with the average amplitude of curvature change at the abrupt change point to form a comprehensive local curvature abrupt change feature value, which is used to capture sharp turns or jitters in the trajectory.

[0050] Furthermore, in the dynamic operating state phase space, the local curvature of each point is calculated along the real-time state evolution trajectory. The local curvature is approximated by the reciprocal of the circumcircle radius of three adjacent points on the trajectory. A dynamic curvature change rate threshold is set. When the local curvature change rate at a point on the trajectory exceeds this threshold, it is determined that a drastic change in motion has occurred near that point, and the point is marked as a candidate abnormal event feature point. For all candidate points, the angle between the point and the line connecting it to the preceding and following trajectory points is further calculated. If the angle is less than a preset angle threshold, it indicates that the trajectory has undergone a sharp turn at that point, and the point is finally confirmed as an abnormal event feature point, and its coordinates in the phase space are recorded.

[0051] Using the tangent direction of the real-time state evolution trajectory as the first reference axis and the direction indicated by the currently calculated first deviation as the second reference axis, the outer product of the two yields the third reference axis, together forming a local orthogonal coordinate system. This coordinate system has the current trajectory point as its origin, and the directions of its three coordinate axes are defined by the trajectory direction, the direction of deviation from the healthy manifold, and the perpendicular direction of both, respectively. Its scale is dynamically calibrated by the values ​​of the first and second deviations, thus forming a local three-dimensional dynamic anomaly representation space that moves, scales, and rotates with the state point, transforming the absolute physical space into a relative anomaly feature space centered on the current state.

[0052] Each identified anomalous event feature point is transformed into a multi-dimensional anomalous representation space based on its respective state point at that time. Specifically, the coordinate difference between each feature point and the reference point at that time is calculated, and this coordinate difference is projected onto the local dynamic coordinate system at that time to obtain its relative coordinates in the anomalous representation space. The set of relative coordinates of all anomalous event feature points in this dynamic space constitutes the anomalous event spatial distribution, which expresses the spatial, temporal, and directional relationship of the anomalous event relative to the main trajectory.

[0053] A density-based spatial clustering algorithm is employed to perform multi-scale analysis of the spatial distribution of anomalous events. First, a small neighborhood radius is set for preliminary clustering to identify dense, isolated micro-clusters that may represent transient impacts. Next, a larger neighborhood radius is set for secondary merging of the preliminary clustering results to identify events that are interconnected on a larger spatial and temporal scale, forming macro-clusters that may represent sustained anomalous processes. Finally, the geometric center, spatial distribution range, event density, and temporal span of each cluster are calculated to form a quantitative description of each fault characteristic cluster.

[0054] The feature description vector of each identified fault feature cluster is matched with a template in a pre-defined fault mode knowledge base based on similarity. The knowledge base templates are built based on historical fault data or mechanism simulations. For example, micro-clusters sparsely and uniformly distributed around the main trajectory with high curvature change points may match periodic blade scraping; macro-clusters densely distributed along a specific direction with a large distance from the main trajectory perpendicular to the main trajectory may match periodic bearing impacts; and macro-clusters with extremely high event density, large spatial distribution, long time span, and significant center deviation may match continuous material blockage. By calculating the Euclidean distance or cosine similarity between each feature cluster and different fault templates, the fault mode with the highest matching degree is taken as the judgment result for that cluster. The final operational status diagnosis of the screw conveyor is obtained by combining the judgment results of all clusters.

[0055] Figure 2 This is a schematic diagram of the structure of a screw conveyor operation status detection device provided in an embodiment of this application. Figure 2As shown, the screw conveyor operating status detection device 200 includes: at least one processor 201; and a memory 202 communicatively connected to the at least one processor 201. The memory 202 stores instructions executable by the at least one processor 201. These instructions, when executed by the at least one processor 201, enable the at least one processor 201 to: establish a dynamic operating status phase space based on historical state parameters corresponding to the screw shaft of the screw conveyor in a healthy state; wherein the historical state parameters include at least one of equivalent shaft torque, vibration signal, and bearing temperature rise signal; and acquire timing data transmitted by multiple sensors installed on the screw conveyor. The method involves determining real-time values ​​corresponding to historical state parameters based on time-series data and mapping these real-time values ​​to real-time state points in the dynamic operating state phase space. Multiple real-time state points from consecutive time series are connected to generate the real-time state evolution trajectory of the screw conveyor. The geometric shape and kinematic characteristics of the real-time state evolution trajectory within a preset time window are analyzed to extract trajectory dynamics features. Based on these trajectory dynamics features, a symmetry breaking index corresponding to the real-time state evolution trajectory is determined. Finally, based on the symmetry breaking index and the dual deviation between the index and the healthy baseline pattern in the dynamic operating state phase space, the current operating state of the screw conveyor is determined.

[0056] This application provides a non-volatile computer storage medium storing computer-executable instructions. These instructions are configured to: establish a dynamic operating state phase space based on historical state parameters corresponding to the screw shaft of a screw conveyor in a healthy state; wherein the historical state parameters include at least one of equivalent shaft torque, vibration signal, and bearing temperature rise signal; acquire time-series data transmitted by multiple sensors installed on the screw conveyor, determine real-time values ​​corresponding to the historical state parameters based on the time-series data, and map the real-time values ​​to real-time state points in the dynamic operating state phase space; connect multiple real-time state points in a continuous time sequence to generate a real-time state evolution trajectory corresponding to the screw conveyor; analyze the geometric shape and kinematic characteristics of the real-time state evolution trajectory within a preset time window to extract trajectory dynamics features; determine the symmetry breaking index corresponding to the real-time state evolution trajectory based on the trajectory dynamics features; and determine the current operating state of the screw conveyor based on the symmetry breaking index and the dual deviation between the symmetry breaking index and the healthy baseline pattern in the dynamic operating state phase space.

[0057] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0058] The above descriptions are merely embodiments of this application and are not intended to limit the scope of this application. For those skilled in the art, various modifications and variations can be made to the embodiments of this application. These modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the spirit and scope of the technical solutions in the embodiments of this application.

Claims

1. A method for detecting the operating status of a screw conveyor, characterized in that, The method includes: A dynamic operating state phase space is established based on the historical state parameters of the screw shaft of the screw conveyor under healthy conditions; wherein, the historical state parameters include at least one of the following: equivalent shaft torque, vibration signal, and bearing temperature rise signal; The system acquires time-series data sent by multiple sensors installed on the screw conveyor, determines the real-time value corresponding to the historical state parameter based on the time-series data, and maps the real-time value to a real-time state point in the dynamic operating state phase space. By connecting multiple real-time state points in a continuous time sequence, a real-time state evolution trajectory corresponding to the screw conveyor is generated. The geometric shape and kinematic characteristics of the real-time state evolution trajectory within a preset time window are analyzed to extract trajectory dynamics features. Based on the trajectory dynamics features, the symmetry breaking index corresponding to the real-time state evolution trajectory is determined. The current operating state of the screw conveyor is determined based on the dual deviation between the symmetry breaking index and the health baseline pattern in the dynamic operating state phase space.

2. The method for detecting the operating status of a screw conveyor according to claim 1, characterized in that, The establishment of a dynamic operating state phase space based on the historical state parameters of the screw shaft of the screw conveyor under healthy conditions specifically includes: Based on the historical state parameters, a joint time series of torque and vibration, and a joint time series of temperature rise and torque are constructed; wherein, the joint time series of torque and vibration is used to represent the response relationship between the helical shaft drive and the load, and the joint time series of temperature rise and torque is used to represent the coupling constraint relationship between temperature and force. A time-delay cross-correlation analysis is performed on the combined time series of the torque and vibration to determine the dynamic time delay of the vibration response relative to the torque excitation, and a time-delay cross-correlation analysis is performed on the combined time series of the temperature rise and torque to determine the thermal inertia time delay of the temperature rise accumulation relative to the load change. Based on the dynamic time delay and the thermal inertia time delay, the historical state parameters are reconstructed over time to generate a multidimensional state vector; wherein, the multidimensional state vector is composed at least of the equivalent shaft torque, the time-shifted vibration response and the time-shifted temperature rise rate; The multidimensional state vector is mapped to a preset constraint phase space to obtain the dynamic operating state phase space; wherein the coordinate axis direction of the preset constraint phase space is related to the maximum correlation direction determined by the time delay cross-correlation analysis.

3. The method for detecting the operating status of a screw conveyor according to claim 1, characterized in that, The step of determining the real-time value corresponding to the historical state parameter based on the time-series data and mapping the real-time value to a real-time state point in the dynamic operating state phase space specifically includes: Based on the real-time equivalent shaft torque and screw shaft speed, the equivalent flow rate is output through the material conveying power model; wherein, the equivalent flow rate is related to the material flow state; The real-time vibration signal of the helical blade is demodulated by frequency to extract the structural response components related to the interaction with the helical shaft; In the dynamic operating state phase space, a decoupled observation vector is constructed based on the equivalent flow rate, structural response components, and real-time temperature rise rate. The decoupled observation vector is directionally projected onto a preset coupled manifold established from health data, and the projection point is the real-time state point; wherein, the direction of the directional projection is determined by the current equivalent flow rate, and is used to separate the state changes caused by material property changes and mechanical failures.

4. The method for detecting the operating status of a screw conveyor according to claim 1, characterized in that, The step of connecting multiple real-time state points in a continuous time sequence to generate the real-time state evolution trajectory corresponding to the screw conveyor specifically includes: Using the rotation period of the helical shaft as the basic time unit, the rotation period slices are performed on multiple real-time state points in a continuous time sequence to obtain a set of periodic state points composed of state points in each rotation period. For each set of periodic state points, calculate its weighted centroid to obtain the periodic characteristic points of the average operating state of the screw conveyor within the rotation cycle; Based on the temporal position of the periodic feature points in the dynamic operating state phase space, an initial state evolution trajectory is generated. Based on the transition probability and temporal causal relationship between adjacent periodic feature points, the abnormal state fluctuation trajectory is determined and inserted between the adjacent periodic feature points. The initial state evolution trajectory is fused with the inserted state anomaly fluctuation trajectory to generate the real-time state evolution trajectory.

5. The method for detecting the operating status of a screw conveyor according to claim 1, characterized in that, The step of analyzing the geometric shape and kinematic characteristics of the real-time state evolution trajectory within a preset time window and extracting trajectory dynamics features specifically includes: Based on the rotation period corresponding to the spiral axis, the real-time state evolution trajectory within the preset time window is divided into multiple continuous trajectory segments. On the phase plane projection consisting of the equivalent shaft torque and vibration response, the trajectory segments are fitted with least squares ellipse. Based on the major axis tilt angle, minor-major axis ratio, and ellipse area of ​​the fitted ellipse, the coupled mode ellipse characteristics corresponding to each trajectory segment are obtained; wherein, the major axis tilt angle is related to the dominant coupling phase of load and vibration within the cycle, the minor-major axis ratio is related to the nonlinearity of vibration response, and the ellipse area is related to the energy dissipation intensity within the current rotation cycle. In the dynamic operating state phase space, the trajectory morphology difference degree between adjacent sub-trajectory segments is determined; Based on the difference between the coupled modal elliptical features and the trajectory morphology, the trajectory dynamics feature quantity is constructed.

6. The method for detecting the operating status of a screw conveyor according to claim 1, characterized in that, The determination of the symmetry breaking index corresponding to the real-time state evolution trajectory based on the trajectory dynamics features specifically includes: Among the trajectory dynamics features, the energy distributions corresponding to several characteristic frequencies that are multiples of the number of helical blades are selected. The offset of the centroid of the energy distribution relative to the fundamental frequency of the helical axis is determined and used as the degree of geometric symmetry breaking. Among the trajectory dynamics features, feature sequences reflecting the phase correlation between load changes and vibration response are selected, and the cyclic stationary strength corresponding to the feature sequences is determined. A baseline value for the cyclic steady-state strength under healthy conditions is determined, and the decay rate of the real-time cyclic steady-state strength relative to the baseline value is used as the functional symmetry breaking degree. The geometric symmetry breaking degree and the functional symmetry breaking degree are nonlinearly coupled to generate the symmetry breaking index.

7. The method for detecting the operating status of a screw conveyor according to claim 1, characterized in that, The current operating state of the screw conveyor is determined based on the dual deviation between the symmetry breaking index and the health baseline pattern in the dynamic operating state phase space, specifically including: In the dynamic operating state phase space, a health benchmark manifold and a symmetry breaking threshold envelope are constructed based on the trajectory dynamics characteristics of historical health cycles. The symmetry breaking index obtained from the current calculation is mapped to the dynamic operating state phase space to obtain the instantaneous state vector; The first deviation is determined based on the deviation distance between the instantaneous state vector and the health baseline manifold; The second deviation is determined based on the relative positional relationship between the instantaneous state vector and the symmetry breaking threshold envelope. Based on the local curvature change characteristics of the real-time state evolution trajectory, the first deviation and the second deviation, the current operating state of the screw conveyor is determined; wherein, the operating state includes at least one of the following: health status, blade wear warning, bearing loosening alarm, and material blockage fault.

8. The method for detecting the operating status of a screw conveyor according to claim 1, characterized in that, The determination of the current operating state of the screw conveyor based on the local curvature abrupt change characteristics of the real-time state evolution trajectory, the first deviation, and the second deviation specifically includes: Based on the local curvature change characteristics of the real-time state evolution trajectory, abnormal event feature points are identified in the dynamic operating state phase space. Based on the first deviation and the second deviation, a multi-dimensional anomaly representation space is constructed with the real-time state evolution trajectory as the benchmark; The abnormal event feature points are mapped to the multi-dimensional abnormal representation space to form an abnormal event spatial distribution; Multi-scale clustering analysis was performed on the spatial distribution of the abnormal events to identify different fault feature clusters; The distribution of each fault feature cluster in the multi-dimensional anomaly representation space is compared with a preset fault mode knowledge base to determine the current operating status of the screw conveyor.

9. A device for detecting the operating status of a screw conveyor, characterized in that, The device includes a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the device is triggered to perform the method described in any one of claims 1-8.

10. A non-volatile computer storage medium storing computer-executable instructions, characterized in that, The computer-executable instructions are capable of performing the method described in any one of claims 1-8.

Images