Transmission system and failure probability prediction and path planning method

Abstract

The application provides a transmission system and a fault probability prediction and path planning method, and relates to the technical field of magnetic suspension. The system comprises a conveying line body, a mover, a processing unit, an acquisition unit and a prediction model. The acquisition unit is used for acquiring operation time sequence data of a stator section. The processing unit is used for determining differential time sequence data according to the operation time sequence data acquired by the acquisition unit, determining an input data sequence according to the operation time sequence data and the differential time sequence data, determining a prediction data sequence according to the input data sequence and the prediction model, determining prediction time sequence data of the stator section according to the prediction data sequence, and determining a fault probability of the stator section according to the operation time sequence data and the prediction time sequence data. The system realizes accurate prediction of the fault probability of the stator section through the operation time sequence data and the prediction time sequence data.

Description

Technical Field

[0001] This invention relates to the field of magnetic levitation technology, and more specifically, to a transmission system and a method for fault probability prediction and path planning. Background Technology

[0002] Magnetic levitation is an advanced technology that uses the principle of magnetic force to enable objects to levitate and move. The magnetic levitation transmission system constructed by this technology uses the magnetic field generated by electromagnetic force or permanent magnets to keep the transmitted object suspended on the track, thereby realizing a contactless and low-friction transportation process. Therefore, the magnetic levitation transmission system has advantages such as high speed, cleanliness, and contactless transmission compared with traditional transmission systems. Magnetic levitation transmission systems are now widely used in automated production lines for 3C electronics, new energy, and pharmaceutical packaging. As automated production lines develop towards longer distances, multiple branches, and high concurrency, the number of stator segments in magnetic levitation transmission systems suitable for production lines is also increasing. Since the stator segment is the transmission line of the moving part, its health condition seriously affects the use of the magnetic levitation transmission system. However, existing magnetic levitation transmission systems cannot predict the failure probability of their stator segments, which can easily lead to safety issues. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a transmission system and a method for predicting the failure probability and planning the path, which can solve the technical problem that the existing magnetic levitation transmission system cannot predict the failure probability of its stator segment.

[0004] To achieve the above objectives, the technical solutions adopted in the embodiments of the present invention are as follows: In a first aspect, embodiments of the present invention provide a transmission system, the transmission system including a conveyor line, a mover, a processing unit, an acquisition unit, and a prediction model, the conveyor line being assembled from multiple stator segments, and the mover being able to move on the conveyor line under the cooperative driving of the multiple stator segments; The acquisition unit is used to acquire the running timing data of the stator segment; wherein, the running timing data includes running current timing data, running position error timing data, and running temperature timing data; The processing unit is used for: Based on the runtime timing data acquired by the acquisition unit, differential timing data is determined; wherein, the differential timing data includes current differential timing data, position error differential timing data, and temperature error differential timing data; The input data sequence is determined based on the runtime timing data and the differential timing data; Based on the input data sequence and the prediction model, determine the prediction data sequence; Based on the predicted data sequence, the predicted timing data of the stator segment is determined; the predicted timing data includes predicted current timing data, predicted position error timing data, and predicted temperature timing data. The failure probability of the stator segment is determined based on the runtime timing data and the predicted timing data.

[0005] Furthermore, this embodiment of the invention provides a first possible implementation of the first aspect, wherein determining the failure probability of the stator segment based on the runtime timing data and the predicted timing data includes: Based on the runtime timing data and the predicted timing data, the anomaly score and dynamic score threshold of the stator segment are determined; The number of consecutive anomalies in the stator segment is determined based on the anomaly score and the dynamic score threshold. The anomaly intensity of the stator segment is determined based on the number of consecutive anomalies, the anomaly score, and the dynamic score threshold. The failure probability of the stator segment is determined by normalizing the abnormal intensity using a normalization function.

[0006] Furthermore, this embodiment of the invention provides a second possible implementation of the first aspect, wherein determining the anomaly score of the stator segment based on the runtime timing data and the predicted timing data includes: Based on the runtime timing data and the predicted timing data, the runtime parameter residuals are determined; wherein, the runtime parameter residuals include current residuals, position error residuals, and temperature residuals; The anomaly score of the stator segment is determined based on the current residual, the position error residual, and the temperature residual of the stator segment.

[0007] Furthermore, this embodiment of the invention provides a third possible implementation of the first aspect, wherein determining the anomaly intensity of the stator segment based on the number of consecutive anomalies and the anomaly score includes: Calculate the ratio of the number of consecutive anomalies to the number of detections during consecutive anomaly detection to obtain the first ratio; Calculate the ratio of the abnormal score to the dynamic score threshold to obtain a second ratio; Multiplying the first ratio and the second ratio yields the abnormal intensity of the stator segment.

[0008] Furthermore, this embodiment of the invention provides a fourth possible implementation of the first aspect, wherein the processing unit is further configured to: The failure probability is compared with a first preset probability threshold, a second preset probability threshold, and a third preset probability threshold. If 0 ≤ the fault probability < the first preset probability threshold, then the stator segment is determined to be in a normal state. If the first preset probability threshold ≤ the fault probability < the second preset probability threshold, then the stator segment is determined to be at the warning level; If the second preset probability threshold ≤ the fault probability < the third preset probability threshold, then the stator segment is determined to be at the alarm level; If the third preset probability threshold ≤ the fault probability ≤ 1, then the stator segment is determined to be in a dangerous level.

[0009] Furthermore, this embodiment of the invention provides a fifth possible implementation of the first aspect, wherein the processing unit is further configured to: Based on the failure probability, the health status and health level of the stator segment are determined, and a linear extrapolation calculation is performed based on the health status to determine the remaining service life of the stator segment.

[0010] Furthermore, this embodiment of the invention provides a sixth possible implementation of the first aspect, wherein determining the health status and health level of the stator segment based on the failure probability, and performing linear extrapolation calculation based on the health status to determine the remaining service life of the stator segment, includes: Calculate the absolute value of the difference between the failure probability and 1 to obtain the health degree of the stator segment, and multiply the health degree by 100 to obtain the health level of the stator segment; Based on the health status and the time when the health status is detected, a straight line is obtained to show the change of the health status of the stator segment over time, and the slope of the straight line is used as the slope of the decrease in the health status of the stator segment. The health status is compared with the health failure threshold. If the health status is less than or equal to the health failure threshold, the remaining service life of the stator segment is determined to be 0. If the health status is greater than the health status failure threshold, the remaining service life of the stator segment is determined based on the health status, the health status decline slope, and the health status failure threshold; wherein, when the health status is greater than the health status failure threshold, the remaining service life is positively correlated with the health status, and the remaining service life is negatively correlated with the absolute value of the health status failure threshold and the health status decline slope, respectively.

[0011] Furthermore, this embodiment of the invention provides a seventh possible implementation of the first aspect, wherein the acquisition unit is further configured to acquire the occupancy rate of the stator segment and the position parameters of the transmission system; wherein the position parameters include the workstation position, the fork position, the mover position, and the stator segment position; The processing unit is also used for: The dynamic weight of the stator segment is determined based on the occupancy rate and the failure probability of the stator segment. Based on the dynamic weights and the position parameters, a directed graph of the transmission system is constructed, and the positions of the scheduled mover and the target workstation of the scheduled mover are determined in the directed graph. Multiple candidate paths are planned from the directed graph according to the path planning algorithm; wherein, the candidate path is the path from the position of the mover to be scheduled to the position of the target workstation; Based on the time window mechanism, candidate paths without time conflicts are selected from the candidate paths as the optimal path.

[0012] Furthermore, this embodiment of the invention provides an eighth possible implementation of the first aspect, wherein the processing unit is further configured to: When at least one stator segment in the optimal path experiences a stator segment freezing problem, the optimal path is replanned based on the current position of the mover to be scheduled. The stator segment freezing problem includes the stator segment having a failure probability greater than or equal to a second preset probability threshold, and / or the stator segment having an occupancy rate greater than a preset occupancy rate threshold, and / or the stator segment being marked as requiring manual maintenance. The replanned optimal path avoids the stator segment where the stator segment freezing problem occurs.

[0013] Secondly, embodiments of the present invention also provide a fault probability prediction method, wherein the method is applied to predict the fault probability of the stator segment in the transmission system described in any of the above claims, and the method includes: Obtain the runtime timing data of the stator segment; The differential timing data is determined based on the runtime timing data; The input data sequence is determined based on the runtime timing data and the differential timing data; The predicted data sequence is determined based on the input data sequence and the prediction model; Based on the predicted data sequence, determine the predicted time series data of the stator segment; The failure probability of the stator segment is determined based on the runtime timing data and the predicted timing data.

[0014] Thirdly, embodiments of the present invention also provide a path planning method, wherein the method is applied to planning the travel path of the moving part in the transmission system described in any of the above claims, the method comprising: Obtain the occupancy rate of the stator segment and the position parameters of the transmission system; wherein, the position parameters include the workstation position, the branch position, the mover position, and the stator segment position; The dynamic weight of the stator segment is determined based on the occupancy rate and the failure probability of the stator segment. Based on the dynamic weights and the position parameters, a directed graph of the transmission system is constructed, and the positions of the scheduled mover and the target workstation of the scheduled mover are determined in the directed graph. Multiple candidate paths are planned from the directed graph according to the path planning algorithm; wherein, the candidate path is the path from the position of the mover to be scheduled to the position of the target workstation; Based on the time window mechanism, candidate paths without time conflicts are selected from the candidate paths as the optimal path.

[0015] This invention provides a transmission system and a method for fault probability prediction and path planning. The system includes: a conveyor line, a mover, a processing unit, an acquisition unit, and a prediction model. The conveyor line is composed of multiple stator segments, and the mover can move on the conveyor line under the coordinated drive of the multiple stator segments. The acquisition unit is used to acquire the running timing data of the stator segments, including running current timing data, running position error timing data, and running temperature timing data. The processing unit is used to: determine differential timing data based on the running timing data acquired by the acquisition unit, including current differential timing data, position error differential timing data, and temperature error differential timing data; determine an input data sequence based on the running timing data and the differential timing data; determine a prediction data sequence based on the input data sequence and the prediction model; determine the predicted timing data of the stator segments based on the predicted data sequence, including predicted current timing data, predicted position error timing data, and predicted temperature timing data; and determine the fault probability of the stator segments based on the running timing data and the predicted timing data. The transmission system provided by this invention is equipped with a conveyor line composed of multiple stator segments and a mover that can move on the conveyor line. It is applicable to automated production lines in various fields. The transmission system also acquires the runtime timing data of the stator segments through an acquisition unit. The processing unit determines the input data sequence through the runtime timing data and differential timing data. The input data sequence is input into a prediction model to obtain an accurate prediction data sequence, thereby determining the prediction timing data of the stator segments. Based on the runtime timing data and prediction timing data of the stator segments, the failure probability of the stator segments is determined, realizing the prediction of the stator segment failure probability. Based on the failure probability of the stator segments, adaptive adjustments can be made to the stator segments to reduce the occurrence of safety problems.

[0016] Other features and advantages of the embodiments of the present invention will be set forth in the following description, or some features and advantages may be inferred from the description or determined without doubt, or may be learned by practicing the techniques described above in the embodiments of the present invention.

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 A schematic diagram of the structure of a transmission system provided in an embodiment of the present invention is shown; Figure 2 This diagram illustrates the main components of a transmission system provided in an embodiment of the present invention. Figure 3 This diagram illustrates the overall unit structure of a transmission system provided in an embodiment of the present invention. Figure 4 A flowchart illustrating a fault probability prediction method provided by an embodiment of the present invention is shown. Figure 5 A flowchart illustrating a path planning method provided by an embodiment of the present invention is shown; Explanation of reference numerals in the attached figures: 101-Motor; 102-Stator segment; 103-Position sensor; 104-Current sensor; 105-Temperature sensor. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be described below in conjunction with the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0021] like Figure 1As shown, this embodiment provides a transmission system, which includes a conveyor line, a mover 101, a position sensor 103, a current sensor 104, a temperature sensor 105, and a prediction model. The conveyor line is composed of multiple stator segments 102 spliced ​​together. The mover 101 can move on the conveyor line under the coordinated drive of the multiple stator segments 102. The position sensor 103 is used to detect the position of the mover 101, the current sensor 104 is used to detect the current of the stator segment 102, and the temperature sensor 105 is used to detect the temperature of the stator segment 102. Preferably, in some embodiments, the position sensor 103, the current sensor 104, and the temperature sensor 105 can be respectively set on each stator segment for independent detection, but this is not a limitation.

[0022] In some embodiments, temperature sensor 105 can be used to detect the coil temperature and driver temperature of stator segment 102. Specifically, temperature sensor 105 may include a coil temperature sensor and a driver temperature sensor, but is not limited thereto.

[0023] In some embodiments, the position sensor 103 may be a Hall position feedback sensor for sampling the running position of the mover in real time and determining the position error of the mover, but is not limited thereto.

[0024] In some embodiments, the current sensor 104 may be a motor three-phase current sampling sensor for real-time sampling of the three-phase current of the stator section, but is not limited thereto.

[0025] In some embodiments, the predictive model may be deployed to the cloud for online operation, or deployed to a local device for offline operation, but is not limited thereto.

[0026] In some embodiments, the prediction model includes a time series prediction model, which may be the IBM Chronos-1 model, but is not limited thereto.

[0027] like Figure 2 As shown, the transmission system provided in this embodiment further includes an acquisition unit 201 and a processing unit 202; the acquisition unit 201 is used to acquire the running timing data of the stator segment; wherein, the running timing data includes running current timing data, running position error timing data, and running temperature timing data; In some embodiments, the runtime timing data may also include vibration RMS timing data, but is not limited thereto.

[0028] A programmable logic controller (PLC) program is written using Codesys software (this program can perform data acquisition, build a circular buffer, and interact with shared memory). The PLC program is downloaded to the PLC (which has the CodesysRTE runtime environment and connects to the driver via Ethernet for Control Automation Technology (EtherCAT) for input / output (IO)). The acquisition unit 201 periodically collects data from each sensor based on the PLC with the downloaded program and writes it to the circular buffer to obtain the operating current timing data (i.e., timing data composed of the three-phase current of the stator segment), the operating position error timing data (i.e., timing data of the position error between the moving part's operating position and the pre-travel position), and the operating temperature timing data (selecting the highest value between the coil temperature and the driver temperature in each sampling as the operating temperature timing data). The PLC writes the operating timing data, the number of each stator segment, the write index, and the effective data length in the circular buffer to the shared memory for subsequent use. For example, a PLC with a downloaded PLC program periodically (with a period length of 100ms) collects data from each sensor and writes it into a circular buffer (the circular buffer always retains the latest 300 points (i.e., 30 seconds) of sensor data to meet the requirements of subsequent prediction models), thus obtaining runtime sequence data.

[0029] Processing unit 202 is used for: Based on the runtime timing data acquired by the acquisition unit 201, differential timing data is determined; wherein, differential timing data includes current differential timing data, position error differential timing data, and temperature error differential timing data; The input data sequence is determined based on the runtime timing data and the differential timing data; Based on the input data sequence and the prediction model, determine the prediction data sequence; Based on the predicted data sequence, the predicted time series data of the stator segment is determined; the predicted time series data includes predicted current time series data, predicted position error time series data, and predicted temperature time series data. Based on runtime timing data and predicted timing data, determine the failure probability of the stator segment; Processing unit 202 starts the Python service. Python parses the runtime timing data from the shared memory according to the byte format strictly consistent with the PLC. Based on the timing difference algorithm, it solves the difference of the runtime timing data to obtain differential timing data. According to Python, the runtime timing data and differential timing data are subjected to Z-score standardization (solving the feature mean and standard deviation) to form a 300×6-dimensional input data sequence. Among them, the 6 dimensions include: operating current, operating temperature, operating position error, current difference, position error difference, and temperature error difference. Python loads the prediction model (including AI models in Open Neural Network Exchange (ONNX) format, including the IBM Chronos-1 model), inputs the input data sequence into the IBM Chronos-1 model, and predicts the resulting data sequence (150×6 dimensions, i.e., 6-dimensional data of the stator segment for the next 150 points (i.e., 15 seconds)). The predicted data sequence is then de-standardized to obtain 150 points of predicted time-series data for the stator segment. The failure probability of the stator segment is determined based on the operating data corresponding to the last point in the operating time-series data and the operating data corresponding to the first point in the predicted time-series data. The predicted time-series data includes predicted current time-series data, predicted position error time-series data, and predicted temperature time-series data. The prediction model also has online and offline debugging prediction capabilities.

[0030] The transmission system provided in this embodiment is equipped with a conveyor line composed of multiple stator segments and a mover that can move on the conveyor line. It can be applied to automated production lines in various fields. The transmission system also acquires the runtime timing data of the stator segments through an acquisition unit. The processing unit determines the input data sequence through the runtime timing data and differential timing data. The input data sequence is input into the prediction model to obtain an accurate prediction data sequence, thereby determining the prediction timing data of the stator segments. Based on the runtime timing data and prediction timing data of the stator segments, the failure probability of the stator segments is determined, realizing the prediction of the stator segment failure probability. Based on the failure probability of the stator segments, adaptive adjustments can be made to the stator segments to reduce the occurrence of safety problems.

[0031] In one embodiment, this embodiment provides a specific implementation method for determining the failure probability of a stator segment based on runtime timing data and predicted timing data, including: Based on runtime sequence data and predicted time sequence data, determine the anomaly score and dynamic score threshold of the stator segment; The stator segment's anomaly score is determined based on the operating data (including operating current, operating temperature, and operating position error) corresponding to the last point in the runtime timing data (i.e., this sampling) and the operating data (including predicted current, predicted temperature, and predicted position error) corresponding to the first point in the prediction timing data (i.e., the predicted next sampling); where the interval between this sampling and the next sampling is 100ms. The dynamic score threshold for the stator segment is determined based on the mean and standard deviation of the abnormal scores of the stator segment within the first preset number of times (usually every 5 seconds (50 points were sampled periodically)).

[0032] in, Let be the dynamic score threshold of the i-th stator segment at the current time t; The average of the last 50 abnormal scores corresponding to the i-th stator segment at the current time t; Let be the standard deviation of the last 50 outlier scores corresponding to the i-th stator segment at the current time t; The average of the above 50 abnormal scores is:

[0033] in, Let be the abnormal score of the i-th stator segment at the current time t; The abnormal score of the i-th stator segment at the previous time t-1; ...; The abnormal score of the i-th stator segment at time t-48; The anomaly score of the i-th stator segment at time t-49; the difference between two adjacent time points is one sampling period (i.e., 100ms); based on the mean... Determine the standard deviation .

[0034] The number of consecutive anomalies in the stator segment is determined based on the anomaly score and the dynamic score threshold. The stator segment is continuously detected for anomalies based on its anomaly score and dynamic score threshold. The number of consecutive anomalies in the stator segment is determined. Specifically, the anomaly scores of the stator segment in the last 5 times (i.e., 0.5 seconds) are compared with the dynamic score threshold of the stator segment. If there are several consecutive anomaly scores greater than the dynamic score threshold, the number of consecutive anomalies in the stator segment is determined.

[0035] The anomaly intensity of the stator segment is determined based on the number of consecutive anomalies and the anomaly score. The anomaly intensity of the stator segment is determined based on the number of consecutive anomalies, the anomaly score, and the dynamic score threshold.

[0036] The failure probability of the stator segment is determined by normalizing the anomaly intensity using a normalization function. The failure probability of the stator segment is determined by normalizing the anomaly intensity using a normalization function (including the Sigmoid function).

[0037] in, Let be the failure probability of the i-th stator segment at the current time t (i.e., this sampling); Let be the anomaly intensity of the i-th stator segment at the current time t (i.e., this sampling); *) is the Sigmoid function.

[0038] In one embodiment, this embodiment provides a specific implementation method for determining the anomaly score of a stator segment based on runtime timing data and predicted timing data, including: Based on the runtime timing data and the predicted timing data, the runtime parameter residuals are determined; among them, the runtime parameter residuals include current residuals, position error residuals, and temperature residuals. Based on the runtime timing data and the predicted timing data, the residuals of the operating parameters are determined. Specifically, the residuals of the operating parameters are determined based on the operating data (including operating current, operating temperature, and operating position error) corresponding to the last point in the runtime timing data (i.e., this sampling) and the operating data (including predicted current, predicted temperature, and predicted position error) corresponding to the first point in the predicted timing data (i.e., the predicted next sampling). ; ; ; in, This represents the current residual of the i-th stator segment at the current time t (i.e., this sampling). Let be the operating current of the i-th stator segment at the current time t (i.e., this sampling); This is the predicted current of the i-th stator segment at the next time t+1 (i.e., the next sampling). Let be the position error residual corresponding to the i-th stator segment at the current time t (i.e., this sampling); This represents the running position error of the i-th stator segment at the current time t (i.e., this sampling). The predicted position error of the i-th stator segment at the next time t+1 (i.e., the next sampling) is the predicted position error. This represents the temperature residual of the i-th stator segment at the current time t (i.e., this sampling). Let be the operating temperature of the i-th stator segment at the current time t (i.e., this sampling time); This is the predicted temperature of the i-th stator segment at the next time t+1 (i.e., the next sampling).

[0039] The anomaly score of the stator segment is determined based on the current residual, position error residual, and temperature residual of the stator segment. The anomaly score of the stator segment is determined by weighting the current residual, position error residual, and temperature residual of the stator segment: ; in, The weight corresponding to the current residual can be set to 0.5; The weight corresponding to the position error residual can be set to 0.3; The weight corresponding to the temperature residual can be set to 0.2.

[0040] In one embodiment, this embodiment provides a specific implementation method for determining the anomaly intensity of a stator segment based on the number of consecutive anomalies and anomaly scores, including: Calculate the ratio of the number of consecutive anomalies to the number of detections during consecutive anomaly detection to obtain the first ratio; Calculate the ratio of the number of consecutive anomalies to the number of detections during consecutive anomaly detection to obtain the first ratio:

[0041] in, This is the first ratio corresponding to the i-th stator segment at the current time t (i.e., this sampling); This represents the number of consecutive anomalies corresponding to the i-th stator segment at the current time t (i.e., this sampling); This refers to the number of times an anomaly is detected during continuous anomaly detection (usually 5 times).

[0042] Calculate the ratio of the abnormal score to the dynamic score threshold to obtain the second ratio; Calculate the ratio of the abnormal score to the dynamic score threshold to obtain the second ratio: ; in, The second ratio corresponding to the i-th stator segment at the current time t (i.e., this sampling).

[0043] Multiplying the first ratio and the second ratio yields the abnormal intensity of the stator segment; The anomaly intensity of the stator segment is obtained by summing the first ratio and the second ratio. .

[0044] In one embodiment, the processing unit 202 provided in this embodiment is further configured to: The failure probability is compared with the first preset probability threshold, the second preset probability threshold, and the third preset probability threshold. If 0 ≤ fault probability < first preset probability threshold, then the stator segment is determined to be in normal condition. If the first preset probability threshold ≤ the fault probability < the second preset probability threshold, then the stator segment is determined to be at the warning level. If the second preset probability threshold ≤ the fault probability < the third preset probability threshold, then the stator segment is determined to be in an alarm level. If the third preset probability threshold ≤ failure probability ≤ 1, then the stator segment is determined to be in a dangerous level; Based on the failure probability of the stator segment, the corresponding failure level (including normal level, early warning level, alarm level, and danger level) is determined. Stator segments at the normal and early warning levels can still maintain operation and will not have safety problems for the time being. However, stator segments at the alarm and danger levels are very likely to have safety problems and require manual maintenance. The failure probability of the stator segment is compared with the first preset probability threshold, the second preset probability threshold, and the third preset probability threshold. When 0 ≤ failure probability < the first preset probability threshold (which can be set to 0.5), the stator segment is determined to be at the normal level, and the stator segment at this level can maintain normal operation. When the first preset probability threshold ≤ the failure probability < the second preset probability threshold (which can be set to 0.7), the stator segment is determined to be in the warning level. The stator segment in this level can temporarily maintain normal operation, but a warning should be issued to indicate the failure probability of the stator segment. When the second preset probability threshold ≤ fault probability < the third preset probability threshold (which can be set to 0.85), the stator section is determined to be in an alarm level. The stator section in this level may fail and is prone to safety problems. The mover is controlled to bypass the stator section in the alarm level during driving and prepare to perform maintenance and repair on the stator section. When the third preset probability threshold ≤ fault probability ≤ 1, the stator section is determined to be in a dangerous level. The stator section in this level is very prone to failure, which will cause safety problems. It is necessary to disconnect the power to the stator section and arrange personnel to carry out manual maintenance on the stator section.

[0045] In one embodiment, the processing unit 202 provided in this embodiment is further configured to: Based on the failure probability, the health status and health level of the stator segment are determined, and the remaining service life of the stator segment is determined by linear extrapolation based on the health status. Based on the failure probability, the health level of the stator segment is determined. The failure probability and health level are negatively correlated; the higher the failure probability, the lower the health level, and vice versa. The health level is highest when the failure probability is 0 (health level 1), and lowest when the failure probability is 1 (health level 0). Health levels include 0, 1, ..., 100, and are positively correlated with health level; the higher the health level, the higher the health level, and vice versa. A health level of 1 corresponds to a health level of 100, and a health level of 0 corresponds to a health level of 0. The health level of the stator segment at different times is determined, and linear extrapolation is performed based on the health level and each time point to determine the remaining service life of the stator segment. Based on the remaining service life of the stator segment, maintenance can be easily scheduled.

[0046] In one embodiment, this embodiment provides a specific implementation method for determining the health status and health level of a stator segment based on the failure probability, and for determining the remaining service life of the stator segment by linear extrapolation based on the health status. Calculate the absolute value of the difference between the failure probability and 1 to obtain the health status of the stator segment. Multiply the health status by 100 to obtain the health level of the stator segment. The health status of the stator segment is obtained by calculating the absolute value of the difference between the failure probability and 1:

[0047] in, The health status (0~1) of the i-th stator segment at the current time t (i.e., this sampling); Multiply the health score of the stator segment by 100 to obtain the health level of the stator segment:

[0048] in, The health level (level 0 to level 100) of the i-th stator segment at the current time t (i.e., this sampling).

[0049] Based on the health status and the time when the health status is detected, a straight line is obtained to show the change of the health status of the stator segment over time. The slope of the straight line is used as the slope of the decrease in the health status of the stator segment. Based on the stator segment health status obtained from the most recent (usually 5-20 times) and the corresponding time, a straight line is fitted to obtain a straight line showing the change of stator segment health status over time. The slope of this straight line is then determined as the slope of the health status decline (which is a negative number).

[0050] The health status is compared with the health failure threshold. If the health status is less than or equal to the health failure threshold, the remaining service life of the stator segment is determined to be 0. Comparing the health status with a pre-set health failure threshold (which can be set to 0.15) determines that the stator segment's health is too low, the failure probability is too high, and the stator segment is extremely prone to failure. In this case, the remaining service life of the stator segment can be determined to be 0, requiring manual maintenance. For example, if the health failure threshold is set to H (H=0.15), if... If so, then the remaining service life of the i-th stator segment is determined to be 0.

[0051] If the health level is greater than the health level failure threshold, the remaining service life of the stator segment is determined based on the health level, the slope of the health level decline, and the health level failure threshold. Among them, when the health level is greater than the health level failure threshold, the remaining service life is positively correlated with the health level, and the remaining service life is negatively correlated with the absolute value of the health level failure threshold and the slope of the health level decline, respectively. If the health status is greater than the health failure threshold, it is determined that the health status of the stator segment has not yet fallen below the threshold. At this point, the straight line representing the change in the stator segment's health status over time needs to be extended further using the current health status and the slope of its decline. When the stator segment's health status equals the health failure threshold, the extension of the straight line is stopped. The time it takes for the stator segment to decline from its current health status to the health failure threshold is then determined, yielding the remaining service life of the stator segment. For example, the remaining service life of the i-th stator segment can be determined as follows:

[0052] in, This represents the remaining lifespan of the i-th stator segment at the current time t (i.e., during this sampling). Let be the slope of the health status decrease of the i-th stator segment at the current time t (i.e., this sampling). From the above formula, it can be seen that when the health status of the stator segment is greater than the health failure threshold, the remaining service life is positively correlated with the health status (i.e., the higher the health status, the smaller the remaining service life). The remaining service life is negatively correlated with the health failure threshold and the absolute value of the health status decrease slope, respectively (i.e., the higher the health failure threshold, the smaller the remaining service life; the higher the absolute value of the health status decrease slope, the smaller the remaining service life). Specifically, if the health of the i-th stator segment at the current time t... The slope of the health decrease of the i-th stator segment at the current time t. (Remaining service life decreases by 0.05 per hour); Health failure threshold H = 0.15; The remaining service life of the i-th stator segment is calculated as follows: =3 hours; It is determined that the i-th stator segment can continue to be used for 3 hours. After 3 hours of use, the stator segment needs to be shut down for manual maintenance.

[0053] In one embodiment, the processing unit 202 in the transmission system provided in this embodiment determines the fault probability, alarm level, and health level of the stator segment through the residual operating parameters of the stator segment. For example, ① when the current residual of the stator segment = 0.15, the position error residual = 0.01, and the temperature residual = 1.2, =0.5、 =0.3、 =0.2; then the abnormal score of the stator segment is determined as: ; The dynamic score threshold for the stator segment is set to 0.5. If no consecutive anomalies are detected after 5 consecutive anomaly detections, the anomaly intensity of the stator segment is then determined as follows: ; Determine the stator segment at this time

[0054] ; Set the first preset probability threshold to 0.5, the second preset probability threshold to 0.7, and the first preset probability threshold to 0.85; at this time, 0 ≤ fault probability (0.1) < 0.5, and determine the fault level of the stator segment as the normal level; The health level of the stator segment is determined as follows: =

[0055] The health level of the stator segment is 90. ② When the stator segment current residual = 0.5, position error residual = 0.04, and temperature residual = 3.0, =0.5、 =0.3、 =0.2; then the abnormal score of the stator segment is determined as:

[0056] The dynamic score threshold for the stator segment is set to 0.6. After 5 consecutive anomaly detections, the number of consecutive anomalies is 2. Therefore, the anomaly intensity of the stator segment is: ; Determine the stator segment at this time

[0057] ; Set the first preset probability threshold to 0.5, the second preset probability threshold to 0.7, and the first preset probability threshold to 0.85; at this time, 0.5 ≤ fault probability (0.575) < 0.7, and determine the fault level of the stator segment as the warning level; The health level of the stator segment is determined as follows: =

[0058] The health level of the stator segment is 45. ③ When the stator segment current residual = 0.9, position error residual = 0.08, and temperature residual = 5.0, =0.5、 =0.3、 =0.2; then the abnormal score of the stator segment is determined as:

[0059] The dynamic score threshold for the stator segment is set to 0.6. After 5 consecutive anomaly detections, the number of consecutive anomalies is 4. Therefore, the anomaly intensity of the stator segment is: ; Determine the stator segment at this time

[0060] ; Set the first preset probability threshold to 0.5, the second preset probability threshold to 0.7, and the first preset probability threshold to 0.85; at this time, 0.7≤fault probability (0.78)<0.85, and determine the fault level of the stator segment as the alarm level; The health level of the stator segment is determined as follows: =

[0061] The health level of the stator segment is level 22; ④ When the stator segment current residual = 1.6, position error residual = 0.15, and temperature residual = 8.0, =0.5、 =0.3、 =0.2; then the abnormal score of the stator segment is determined as:

[0062] The dynamic score threshold for the stator segment is set to 0.6. When performing 5 consecutive anomaly detections, the number of consecutive anomalies is 5. Therefore, the anomaly intensity of the stator segment is: ; Determine the stator segment at this time

[0063] ; Set the first preset probability threshold to 0.5, the second preset probability threshold to 0.7, and the first preset probability threshold to 0.85; at this time, 0.85≤fault probability (0.94)≤1, and determine the fault level of the stator segment as the dangerous level; The health level of the stator segment is determined as follows: =

[0064] The health level of the stator segment is level 6.

[0065] In one embodiment, the processing unit 202 provided in this embodiment is also used to use Python to write the branch number, fault probability, remaining service life, health level and fault level of the stator segment into the reverse shared memory for real-time display and control by the PLC; write the predicted timing data into the CSV log; and automatically generate an alarm file and provide maintenance suggestions for the stator segment when the alarm level of the stator segment is at a dangerous level (i.e., the third preset probability threshold ≤ fault probability ≤ 1). In one embodiment, the transmission system provided in this embodiment cyclically monitors all stator segments (which can be set to 50 segments) within its system at a 1-second cycle, thereby achieving 24-hour uninterrupted predictive maintenance of the stator segments.

[0066] In one embodiment, the acquisition unit 201 provided in this embodiment is further used to acquire the occupancy rate of the stator segment and the position parameters of the transmission system; wherein, the position parameters include the workstation position, the junction position, the mover position, and the stator segment position; The acquisition unit 201 is used to acquire the number of currently occupied movers (i.e., trolleys) in the stator segment (i.e., the number of movers present in the stator segment at the current moment) and the maximum number of movers that the stator segment can accommodate. Based on the number of currently occupied movers and the maximum number of movers that can be accommodated, the occupancy rate of the stator segment is determined. Specifically, the occupancy rate = the number of currently occupied movers ÷ the maximum number of movers that can be accommodated. If the occupancy rate of the stator segment = 0, there is no movement command in the stator segment, and there is no reservation time window in the stator segment, then the stator segment is determined to be in an idle state. If the occupancy rate of the stator segment ≤ 0.8, there is a mover in the stator segment that is running normally, and the time window reservation of the stator segment is normal, then the stator segment is determined to be in an occupied state and not congested. If 0.8 < occupancy rate ≤ 1, then the stator segment is determined to be in an occupied state and congested. The acquisition unit 201 is also used to transmit the position parameters of the transmission system, including the working position of the mover, the position of each stator segment, the position of each branch, and the position of each mover in the transmission system.

[0067] Processing unit 202 is also used for: The dynamic weights of the stator segments are determined based on the occupancy rate and the failure probability of the stator segments. A directed graph of the transmission system is constructed based on dynamic weights and position parameters to determine the positions of the scheduled mover and its target workstation in the directed graph. Multiple candidate paths are planned from the directed graph based on the path planning algorithm; where the candidate path is the path from the position of the mover to be scheduled to the position of the target workstation. The optimal path is selected from all candidate paths based on the time window mechanism, which selects the candidate paths that do not have time conflicts. Dynamic weights are calculated for each stator segment based on occupancy rate and stator segment failure probability to determine the dynamic weight of each stator segment. A graph model of the transmission system is constructed based on position parameters. The dynamic weights corresponding to each stator segment in the graph model are then labeled to obtain a directed graph of the transmission system. The specific positions of the mover to be scheduled (i.e., the mover that needs to be scheduled to pick up or put down materials at the target workstation) and the target workstation in the directed graph are determined. The AI ​​model (including the IBM Chronos-1 model) plans multiple candidate paths on the weighted graph by improving the A* algorithm. The candidate path is the path from the position of the mover to be scheduled to the position of the target workstation in the directed graph. Based on the time window mechanism, the usage time conflicts of the stator segments are checked, and the candidate path without time conflicts is selected as the optimal path. The mover to be scheduled is controlled to travel to the target workstation according to the optimal path to realize the scheduling of the mover.

[0068] In one embodiment, the transmission system provided in this embodiment also includes an industrial control computer. The control system of the industrial control computer is divided into two parts that operate independently: one part manages real-time control (Codesys): specifically controlling the movement, stopping, and speed of the mover (i.e., the trolley) to ensure that the mover does not collide or become disorderly. When a congested stator section is detected ahead, the mover is controlled to automatically decelerate to avoid emergency braking; the other part manages intelligent AI (i.e., AI model): specifically calculating routes, predicting faults, and making decisions. The two parts of the industrial control computer's control system communicate using high-speed shared memory, resulting in extremely low latency. The industrial control computer's control system automatically scans all stator sections, movers, magnetic levitation sensors, and power supplies to confirm that they are all normal; it loads a directed graph to ensure that the main line, branches, junctions, and workstation positions of the transmission system have all been established.

[0069] In one embodiment, the processing unit 202 provided in this embodiment is further configured to: When at least one stator segment in the optimal path is frozen, the optimal path is replanned based on the current position of the mover to be scheduled. Among them, the stator segment freezing problem includes the stator segment failure probability being greater than or equal to the second preset probability threshold, and / or the stator segment occupancy rate being greater than the preset occupancy rate threshold, and / or the stator segment being marked as requiring manual maintenance. The replanned optimal path avoids stator segments where stator segment freezing occurs; When at least one (or more) stator segments in the optimal path for the scheduled trolley to reach the target workstation are frozen, the optimal path is replanned based on the current position of the scheduled mover. Specifically, if at least one of ① to ③ is satisfied, the stator segment is frozen. In this case, the optimal path of the trolley is frozen, and the optimal path is replanned based on the current position of the scheduled mover. The replanned optimal path will avoid the stator segments that are frozen.

[0070] In a real-time example, such as Figure 3 As shown, the transmission system provided in this embodiment also includes: a maintenance unit 203; The maintenance unit 203 in the transmission system is used to mark the stator section as a stator section to be repaired when the fault level of the stator section is dangerous (i.e., the third preset probability threshold ≤ fault probability ≤ 1), schedule all the moving parts (i.e. trolleys) of the stator section to be repaired to leave, and after confirming that the stator section to be repaired is not moving and there are no moving parts above it, automatically cut off the power supply to the stator section to be repaired and notify the maintenance personnel to carry out manual maintenance on the stator section to be repaired. Power to the stator section to be repaired is cut off to ensure that maintenance personnel can carry out manual maintenance (including tagging, locking, and maintenance) in a safe condition. During the maintenance period, the stator section to be repaired is completely isolated and does not participate in path planning and rotor scheduling.

[0071] In one embodiment, the transmission system provided in this embodiment further includes: a verification unit 204; The verification unit 204 in the transmission system is used to unlock the stator section to be repaired after maintenance is completed. The transmission system then powers on the stator section to be repaired again. The transmission system self-checks the current, temperature, and position (i.e., the stator section's position sensor 103 communicates normally, without loss or error, the stator section's position value is within a reasonable no-load range (no over-range, no jump), and the position synchronization deviation between adjacent stator sections is less than the allowable threshold (e.g., ≤0.1mm)) and all are normal. The AI ​​model (including the IBM Chronos-1 model) confirms that the fault probability is less than the first preset probability threshold, thus reopening the stator section and allowing it to participate in normal path planning.

[0072] In one embodiment, the industrial control computer in the transmission system provided in this embodiment has CPU core isolation when controlling the movement of the mover and performing path planning, so that the two do not interfere with each other and are not prone to lag; at the same time, the industrial control computer has sufficient memory, which is used less than half of the time, so it is not prone to crashing; the AI ​​model will not make erroneous actions, and will only automatically shut down when the stator section freezes; during the process of path planning and scheduling of the mover, the scheduling has built-in anti-conflict logic, and the travel time of the mover on the stator section is reserved through a time window reservation mechanism to avoid the mover from being blocked; during the maintenance of the stator section to be maintained, it is required that there are no movers on the stator section, the power is off, and the interlocking is in place, so as to ensure the safety of maintenance personnel during the maintenance process.

[0073] This embodiment provides a fault probability prediction method, which is applied to predict the fault probability of the stator segment in the above-mentioned transmission system, such as... Figure 4 As shown, the method includes: Step S401: Obtain the runtime sequence data of the stator segment; Step S402: Determine the differential timing data based on the runtime timing data; Step S403: Determine the input data sequence based on the runtime timing data and differential timing data; Step S404: Determine the predicted data sequence based on the input data sequence and the prediction model; Step S405: Determine the predicted time series data of the stator segment based on the predicted data sequence; Step S406: Determine the stator segment failure probability based on runtime timing data and predicted timing data.

[0074] This embodiment provides a path planning method, which is applied to planning the travel path of a moving part in the aforementioned transmission system, such as... Figure 5 As shown, the method includes: Step S501: Obtain the stator segment occupancy rate and the position parameters of the transmission system; wherein, the position parameters include the station position, the branch position, the mover position, and the stator segment position; Step S502: Determine the dynamic weight of the stator segment based on the occupancy rate and the failure probability of the stator segment; Step S503: Construct a directed graph of the transmission system based on dynamic weights and position parameters, and determine the positions of the scheduled mover and its target workstation in the directed graph. Step S504: Develop multiple candidate paths from the directed graph using a path planning algorithm; wherein, a candidate path is a path from the position of the mover to be scheduled to the position of the target workstation. Step S505: Select the candidate path without time conflict from each candidate path as the optimal path according to the time window mechanism.

[0075] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0076] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A transmission system, characterized in that, The transmission system includes a conveyor line, a mover, a processing unit, an acquisition unit, and a prediction model. The conveyor line is composed of multiple stator segments spliced ​​together, and the mover can move on the conveyor line under the coordinated drive of the multiple stator segments. The acquisition unit is used to acquire the running timing data of the stator segment; wherein, the running timing data includes running current timing data, running position error timing data, and running temperature timing data; The processing unit is used for: Based on the runtime timing data acquired by the acquisition unit, differential timing data is determined; wherein, the differential timing data includes current differential timing data, position error differential timing data, and temperature error differential timing data; The input data sequence is determined based on the runtime timing data and the differential timing data; Based on the input data sequence and the prediction model, determine the prediction data sequence; Based on the predicted data sequence, the predicted timing data of the stator segment is determined; the predicted timing data includes predicted current timing data, predicted position error timing data, and predicted temperature timing data. The failure probability of the stator segment is determined based on the runtime timing data and the predicted timing data. in: Determining the failure probability of the stator segment based on the runtime timing data and the predicted timing data includes: Based on the runtime timing data and the predicted timing data, the anomaly score and dynamic score threshold of the stator segment are determined; The number of consecutive anomalies in the stator segment is determined based on the anomaly score and the dynamic score threshold. The anomaly intensity of the stator segment is determined based on the number of consecutive anomalies, the anomaly score, and the dynamic score threshold. The failure probability of the stator segment is determined by normalizing the anomaly intensity according to the normalization function. Determining the anomaly score of the stator segment based on the runtime timing data and the predicted timing data includes: Based on the runtime timing data and the predicted timing data, the runtime parameter residuals are determined; wherein, the runtime parameter residuals include current residuals, position error residuals, and temperature residuals; The anomaly score of the stator segment is determined based on the current residual, the position error residual, and the temperature residual of the stator segment.

2. The transmission system according to claim 1, characterized in that, Determining the anomaly intensity of the stator segment based on the number of consecutive anomalies and the anomaly score includes: Calculate the ratio of the number of consecutive anomalies to the number of detections during consecutive anomaly detection to obtain the first ratio; Calculate the ratio of the abnormal score to the dynamic score threshold to obtain a second ratio; Multiplying the first ratio and the second ratio yields the abnormal intensity of the stator segment.

3. The transmission system according to claim 1, characterized in that, The processing unit is also used for: The failure probability is compared with a first preset probability threshold, a second preset probability threshold, and a third preset probability threshold. If 0 ≤ the fault probability < the first preset probability threshold, then the stator segment is determined to be in a normal state. If the first preset probability threshold ≤ the fault probability < the second preset probability threshold, then the stator segment is determined to be at the warning level; If the second preset probability threshold ≤ the fault probability < the third preset probability threshold, then the stator segment is determined to be at the alarm level; If the third preset probability threshold ≤ the fault probability ≤ 1, then the stator segment is determined to be in a dangerous level.

4. The transmission system according to claim 1, characterized in that, The processing unit is also used for: Based on the failure probability, the health status and health level of the stator segment are determined, and a linear extrapolation calculation is performed based on the health status to determine the remaining service life of the stator segment.

5. The transmission system according to claim 4, characterized in that, The step of determining the health status and health level of the stator segment based on the failure probability, and then performing a linear extrapolation calculation based on the health status to determine the remaining service life of the stator segment, includes: Calculate the absolute value of the difference between the failure probability and 1 to obtain the health degree of the stator segment, and multiply the health degree by 100 to obtain the health level of the stator segment; Based on the health status and the time when the health status is detected, a straight line is obtained to show the change of the health status of the stator segment over time, and the slope of the straight line is used as the slope of the decrease in the health status of the stator segment. The health status is compared with the health failure threshold. If the health status is less than or equal to the health failure threshold, the remaining service life of the stator segment is determined to be 0. If the health status is greater than the health status failure threshold, the remaining service life of the stator segment is determined based on the health status, the health status decline slope, and the health status failure threshold; wherein, when the health status is greater than the health status failure threshold, the remaining service life is positively correlated with the health status, and the remaining service life is negatively correlated with the absolute value of the health status failure threshold and the health status decline slope, respectively.

6. The transmission system according to claim 1, characterized in that, The acquisition unit is also used to acquire the occupancy rate of the stator segment and the position parameters of the transmission system; wherein, the position parameters include the workstation position, the junction position, the mover position, and the stator segment position; The processing unit is also used for: The dynamic weight of the stator segment is determined based on the occupancy rate and the failure probability of the stator segment. Based on the dynamic weights and the position parameters, a directed graph of the transmission system is constructed, and the positions of the scheduled mover and the target workstation of the scheduled mover are determined in the directed graph. Multiple candidate paths are planned from the directed graph according to the path planning algorithm; wherein, the candidate path is the path from the position of the mover to be scheduled to the position of the target workstation; Based on the time window mechanism, candidate paths without time conflicts are selected from the candidate paths as the optimal path.

7. The transmission system according to claim 6, characterized in that, The processing unit is also used for: When at least one stator segment in the optimal path experiences a stator segment freezing problem, the optimal path is replanned based on the current position of the mover to be scheduled. The stator segment freezing problem includes the stator segment having a failure probability greater than or equal to a second preset probability threshold, and / or the stator segment having an occupancy rate greater than a preset occupancy rate threshold, and / or the stator segment being marked as requiring manual maintenance. The replanned optimal path avoids the stator segment where the stator segment freezing problem occurs.

8. A method for predicting the probability of failure, characterized in that, The method, applied to predicting the failure probability of the stator segment in the transmission system according to any one of claims 1-7, comprises: Obtain the runtime timing data of the stator segment; The differential timing data is determined based on the runtime timing data; The input data sequence is determined based on the runtime timing data and the differential timing data; The predicted data sequence is determined based on the input data sequence and the prediction model; Based on the predicted data sequence, determine the predicted time series data of the stator segment; The failure probability of the stator segment is determined based on the runtime timing data and the predicted timing data.

9. A path planning method, characterized in that, The method, applied to planning the travel path of the moving part in the transmission system according to any one of claims 1-7, comprises: Obtain the occupancy rate of the stator segment and the position parameters of the transmission system; wherein, the position parameters include the workstation position, the branch position, the mover position, and the stator segment position; The dynamic weight of the stator segment is determined based on the occupancy rate and the failure probability of the stator segment. Based on the dynamic weights and the position parameters, a directed graph of the transmission system is constructed, and the positions of the scheduled mover and the target workstation of the scheduled mover are determined in the directed graph. Multiple candidate paths are planned from the directed graph according to the path planning algorithm; wherein, the candidate path is the path from the position of the mover to be scheduled to the position of the target workstation; Based on the time window mechanism, candidate paths without time conflicts are selected from the candidate paths as the optimal path.

Images