Elevator running state judgment method and system based on multi-source data fusion
Through state space modeling and multi-rate Kalman filters, the multi-source sensor data of elevators is combined, and the sampling frequency inconsistency and noise processing problems are solved, high-resolution elevator operation status monitoring and abnormal detection are realized, and monitoring precision and faults are improved. Identification ability.
Patent Information
- Application Number
- CN202411505874.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-28
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing elevator monitoring system relies on a single sensor to fully and accurately reflect the operating status of the elevator, and the multi-source data fusion has problems such as inconsistent sampling frequency, limited noise processing capability and inaccurate abnormal detection.
By establishing the state space modeling of the elevator, using a multi-rate Kalman filter to fuse sensor data at different sampling frequencies, extract high-resolution operating status signals, and identify abnormal states in comparison with the historical normal state.
It improves the accuracy and real-time nature of data fusion, enhances the monitoring precision and abnormal detection capabilities of elevator operating status, and can identify potential faults earlier.
Smart Images

Figure CN119018730A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of elevator monitoring and fault detection, and in particular to a method and system for determining an elevator operating state based on multi-source data fusion. Background Art
[0002] With the acceleration of modern urbanization and the popularization of high-rise buildings, the elevator system, as an important part of vertical transportation, has received widespread attention for its safety and efficiency. Traditional elevator monitoring systems rely on single sensor signals (such as acceleration, velocity or position sensors) to determine the operating status of the elevator. However, a single sensor is limited by its own sampling frequency, noise level and measurement accuracy, and it is difficult to fully and accurately reflect the operating status of the elevator. In addition, the operating environment of the elevator is complex, and the system status is easily affected by factors such as the external environment, mechanical wear and load changes, which are often difficult to obtain through the signal of a single sensor.
[0003] The existing multi-source data fusion technology still has several shortcomings in the application of elevator operation status monitoring. First, the data sampling frequencies of different sensors are often quite different. For example, the position sensor may sample once per second, while the acceleration sensor may sample at a high frequency, which makes it easy to introduce time synchronization errors or data loss during the fusion process. Secondly, the commonly used fusion algorithms are mostly simple weighted averages or Kalman filters. Although these algorithms can reduce the impact of noise to a certain extent, they have limited processing capabilities for multi-rate data and are difficult to effectively deal with data characteristics of different time scales. In addition, most of the existing monitoring systems are directly based on the raw data of the sensor for analysis, and fail to conduct in-depth modeling of the elevator's operating status. It is difficult to detect anomalies from the perspective of state estimation, resulting in slow response and inaccurate judgment of the system when dealing with complex operating conditions. Summary of the invention
[0004] In view of the above existing problems, the present invention is proposed.
[0005] Therefore, the present invention provides an elevator operation status determination method and system based on multi-source data fusion to solve the problems of difficult multi-source sensor data fusion, complex processing of different sampling frequencies and inaccurate abnormal state identification in existing elevator operation status monitoring.
[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for determining the operating state of an elevator based on multi-source data fusion, which includes collecting elevator operating state data; establishing a state space model of the elevator, defining a state vector, predicting the state of the elevator at the next moment through a state equation, and converting the state vector into a measurement vector; according to the state vector, using a multi-rate Kalman filter to fuse sensor data with different sampling frequencies to obtain a unified state estimate; extracting the operating state signal of the elevator from the state estimate, and performing interpolation processing to reconstruct a high-resolution operating state signal; identifying abnormal operating states by comparing the high-resolution operating state signal with historical normal states.
[0007] As a preferred solution of the elevator operation state determination method based on multi-source data fusion described in the present invention, wherein: establishing a state space model of the elevator, expressing the operation state of the elevator as a state vector, including the following steps, Define an elevator state vector containing elevator operation state data; The state equation is used to describe the process of evolving from the current state to the state at the next moment under the influence of control input and process noise. The expression is: ; in, is the state vector at the next moment, is the state vector at the current moment, is the state transfer matrix, is the control input matrix, For external control input, is the process noise; Using the measurement equation, the state vector is converted into a measurement vector through the observation matrix, and the expression is: ; in, is the measurement vector, is the observation matrix, To measure noise.
[0008] As a preferred solution of the elevator operation state determination method based on multi-source data fusion described in the present invention, wherein: according to the state vector, a multi-rate Kalman filter is used to fuse sensor data with different sampling frequencies to obtain a unified state estimation, including the following steps: Define the process noise as zero-mean Gaussian noise, and use the state transfer matrix and process noise covariance matrix to predict the prediction error covariance at the next moment. The expression is: ; in, For the moment The prediction covariance matrix of For the moment The covariance matrix of is the transpose of the state transfer matrix, is the process noise covariance matrix; The measurement noise is defined as zero-mean Gaussian noise. The Kalman gain is calculated based on the prediction error covariance, the observation matrix, and the measurement noise covariance matrix. The specific formula is: ; in, For the moment The Kalman gain, is the measurement noise covariance matrix; Using the Kalman gain and the observed value, the state prediction is corrected to obtain the updated state estimate; The error covariance is updated through the Kalman gain and the observation matrix.
[0009] As a preferred solution of the elevator operation state determination method based on multi-source data fusion described in the present invention, wherein: the elevator operation state signal is extracted from the state estimation, and interpolation processing is performed to reconstruct the high-resolution operation state signal, including the following steps: Get the low frequency sensor at time and The state estimate of determines the number of time steps for interpolation; According to the linear interpolation formula, calculate the state estimation value of each interpolation point; For all time steps, the interpolation calculation is repeated until the state estimates for all interpolation points are completed.
[0010] As a preferred solution of the elevator operation state determination method based on multi-source data fusion of the present invention, wherein: the abnormal operation state is identified by comparing the high-resolution operation state signal with the historical normal state, including the following steps: Extract the state vector of the elevator in normal operation from the historical data; Use the Euclidean distance formula to quantify the difference between the current state and the normal state. The expression is:
[0011] in, represents the Euclidean distance, is the number of state vectors, It is the current moment The Kalman filter estimates the state components, is the first Normal reference value of each state component; Setting Thresholds , the calculated Euclidean distance With the set threshold Make comparisons; if , the elevator is judged to be operating abnormally; otherwise, normal operation monitoring continues.
[0012] As a preferred solution of the elevator operation status determination method based on multi-source data fusion described in the present invention, the elevator operation status data includes the speed, acceleration, vibration, load and motor current of the elevator.
[0013] In the second aspect, the present invention provides an elevator operation status determination system based on multi-source data fusion, including: a data acquisition module, responsible for collecting elevator operation status data; a state space modeling module, responsible for establishing the state space modeling of the elevator, defining the state vector, predicting the state of the elevator at the next moment through the state equation, and converting the state vector into a measurement vector; a data fusion module, responsible for fusing sensor data of different sampling frequencies according to the state vector using a multi-rate Kalman filter to obtain a unified state estimate; an interpolation and reconstruction module, responsible for extracting the elevator operation status signal from the state estimate, and interpolating and reconstructing a high-resolution operation status signal; an abnormality detection module, responsible for identifying abnormal operation status by comparing the high-resolution operation status signal with the historical normal state.
[0014] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the elevator operation status determination method based on multi-source data fusion as described in the first aspect of the present invention is implemented.
[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the elevator operation status determination method based on multi-source data fusion as described in the first aspect of the present invention.
[0016] The beneficial effects of the present invention are as follows: the dynamic evolution of an elevator under the action of control input and noise is described through a state equation, sensor data is associated with a state vector using a measurement equation, a mathematical model that can reflect the intrinsic connection between various operating states of the elevator is constructed, and a systematic description of the operating state of the elevator is realized. Sensor data of different frequencies are fused using a multi-rate Kalman filter, which solves the problem that data from multi-source sensors cannot be synchronized and fused due to inconsistent sampling frequencies, improves the accuracy and real-time performance of data fusion, reduces the impact of noise on state estimation, compensates for information loss caused by low-frequency data through time interpolation, and generates a state signal equivalent to high-frequency sampling, thereby improving the precision of elevator operating state monitoring, enabling the system to detect smaller state changes, and further enhancing the sensitivity and recognition ability to abnormal states. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.
[0018] Figure 1 This is a flow chart of the elevator operation status determination method based on multi-source data fusion in Example 1.
[0019] Figure 2 This is a system diagram of the elevator operation status determination system based on multi-source data fusion in Example 1. DETAILED DESCRIPTION
[0020] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.
[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.
[0022] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.
[0023] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and provides an elevator operation status determination method based on multi-source data fusion, comprising the following steps: S1. Collect elevator operation status data; Specifically, an acceleration sensor is installed on the top of the elevator car. It is mainly used to measure the acceleration changes during the start and stop of the elevator. The typical sampling frequency of this sensor is 100Hz, which can capture the instantaneous changes when the elevator accelerates and decelerates, and reflect the dynamic performance of the elevator; a speed sensor is installed in the elevator drive system. This sensor measures the moving speed of the elevator in real time to ensure the smooth operation of the monitored elevator. The sampling frequency of the speed sensor is generally 50Hz, which is enough to capture the speed changes of the elevator, including the state when running at a constant speed; a vibration sensor is fixed on the elevator rail to monitor the mechanical vibration during the operation of the elevator. The vibration sensor can reflect whether the elevator rail is abnormally worn or whether the mechanical parts are loose. Its sampling frequency is usually 100Hz, which can capture high-frequency vibration signals; a pressure sensor is installed at the bottom of the elevator car, which is responsible for monitoring the load of the elevator and judging whether the elevator is overloaded or unloaded. Its sampling frequency is generally 10Hz, which is suitable for detecting changes in the elevator load; a current sensor is installed on the power line of the elevator drive motor to measure the current changes of the drive motor. The sampling frequency of the current sensor is generally 10Hz, which is used to detect the running load of the motor and timely detect motor overload or failure.
[0024] It should be noted that the acquisition frequency of each sensor is set according to the characteristics of the physical quantity it monitors. For example, the sampling frequency of acceleration and vibration sensors is higher because they need to capture the transient dynamic changes of the elevator.
[0025] The data acquisition system transmits the signals of various sensors to the central processing unit (controller) through a multi-channel data acquisition card, and performs unified time stamp processing on the data of various sensors to ensure the synchronization of the data.
[0026] S2. Establish the state space model of the elevator and express the running state of the elevator as a state vector, including the following steps: S2.1. In order to construct the state space model of the elevator, we first define the elevator state vector containing the elevator operation state data. In the state space model, the elevator operation state is represented as a state vector ,in represents a discrete time step (e.g., after each sampling period), the state vector Contains the following key physical quantities:
[0027] in, The real-time speed of the elevator, reflecting the current running speed of the elevator. is the acceleration of the elevator, reflecting the acceleration or deceleration state of the elevator. The vibration amplitude of the elevator. Abnormal vibration may indicate a malfunction. The load of the elevator reflects the weight carried by the elevator. Abnormal load may cause elevator safety problems. The current of the motor. Abnormal current may indicate motor failure or overload. These physical quantities are collected in real time through sensors and reflect the operating status of the elevator.
[0028] It should be noted that the state vector It contains the core parameters of elevator operation, through which the operating status of the elevator can be fully described. It will serve as the core variable in the state space model to describe the state evolution and observation relationship of the elevator.
[0029] S2.2. The state space model consists of state equations and measurement equations, which respectively describe the dynamic changes of the elevator operating state and the sensor's observation process of the elevator operating state. This model can realize the prediction and real-time estimation of the elevator state, and is the basis of state estimation and anomaly detection.
[0030] S2.2.1. The state equation describes the evolution of the elevator operation state over time, that is, how the current state evolves to the state at the next moment under the influence of control input and process noise. The expression is,
[0031] in, is the state vector at the next moment, is the state vector at the current moment, is the state transfer matrix, which describes the changing law of the elevator operation state. is the control input matrix, which describes the influence of external input (such as motor control signal) on the elevator operation state. For external control input (such as motor control signal, brake command, etc.), is the process noise; State transition matrix Describes the natural changes in the elevator's operating state when it is not affected by control input. For example, the change in the elevator's speed over time is related to the acceleration, which may be affected by external disturbances (such as load changes). For example, for the relationship between the elevator's speed and acceleration, the state transition matrix The part is expressed as:
[0032] in, is the sampling period, which represents the relationship between velocity and acceleration; is the transfer coefficient of the vibration state, usually less than 1, indicating the natural attenuation of vibration over time. is the transfer coefficient of the load state, usually 1, indicating the stability of the load state. is the transfer matrix element of the current state, usually 1, indicating that the current state remains unchanged between time steps. If the current state is affected by dynamic factors (such as load changes or speed changes), It may be less than 1, indicating that the current will gradually adjust to a stable value. For example, when the elevator is just started, the current may be large, then gradually decrease and stabilize; Control Input Matrix Describes the external control input The effect on the elevator operation state. For example, the control signal of the motor can directly affect the acceleration and current of the elevator. Therefore, the control input matrix The form is:
[0033] in, describes the effect of control input on acceleration, The effects of control inputs on current are described; is a random noise that describes unpredictable disturbances in the elevator operation state, such as mechanical wear, environmental changes, etc. This type of noise is assumed to be zero-mean Gaussian noise, and its process noise covariance matrix is , represents the intensity and correlation of the noise, expressed as,
[0034] in, Indicates that the expected value of process noise is zero, which means that in long-term predictions, the average impact of noise can be ignored. Indicates the variance of process noise and the noise correlation between each state variable. The larger the variance, the greater the uncertainty of state prediction; The meaning of is: process noise It follows a multivariate normal distribution with zero mean.
[0035] If the state vector is represented as,
[0036] in, is the speed of the elevator, is the acceleration of the elevator; In this case, the process noise It is expressed as:
[0037] in, The first state variable describing the elevator operation (such as speed, acceleration) is calculated at time step The external disturbance or noise, The second state variable (such as vibration, current) described by External disturbances or noise;
[0038] Covariance matrix Then:
[0039] Among them, the diagonal elements and represents the process noise variance of velocity and acceleration, describing the uncertainty of their respective predictions; the off-diagonal elements and represents the correlation of the noise between velocity and acceleration. If there is no correlation between them, then .
[0040] For example, when an elevator is running, friction or other physical factors in the motor may cause changes in speed and acceleration. These unpredictable effects are included in process noise. If friction suddenly increases, the actual state of the elevator may deviate from the model prediction, so process noise is needed to describe this uncertainty.
[0041] S2.2.2.2, the measurement equation describes the relationship between the sensor measurement data and the elevator operation state. Since the sensor can only observe part of the elevator operation state, the measurement equation is used to convert the state vector Mapping to measurements , the expression is,
[0042] in, is the measurement vector, containing the actual measurement values obtained by the sensor, is the observation matrix, which describes the observation relationship between the sensor and the elevator operation status. To measure noise.
[0043] Observation Matrix It is a key component in the state space model and is used to transform the state vector Mapped to the output of the sensor, the measurement vector , describes which variables in the state vector can be directly observed by sensors and which are indirectly observed or unobservable; In practical applications, not all state variables can be directly measured by sensors. For example, a velocity sensor can directly measure the speed of an elevator, while an acceleration sensor can only directly measure acceleration. Vibration may not be directly measured but can be indirectly inferred through other variables (such as velocity and acceleration). Observation Matrix The structure of determines which states in the system can be directly observed. Assume that the state vector There are N state variables, the measurement vector There are M sensor outputs, so the observation matrix is an M×N matrix, where each element Indicates sensor Can state variables be observed directly? .if , then the sensor The state variables can be directly observed ;if , then the sensor State variables cannot be observed directly , but this variable may be indirectly inferred by other means.
[0044] For example, the state vector There are five corresponding sensors that can measure these state variables respectively. At this time, the observation matrix It is expressed as:
[0045] In this example, the matrix is a unit matrix, indicating that each state quantity can be directly observed by the sensor. The specific explanation is as follows: :Speed sensor can directly measure speed ; : This means that the acceleration sensor can directly measure acceleration ; :It shows that vibration sensors can directly measure vibration ; :It means that the pressure sensor can directly measure the load ; :Indicates that the motor current sensor can directly measure the motor current .
[0046] Assume vibration It cannot be measured directly by sensors, but needs to be indirectly inferred through velocity and acceleration, so the observation matrix Becomes as follows:
[0047] In this matrix, =0 means the vibration sensor cannot directly observe vibration , vibration needs to be indirectly inferred through other state variables (such as velocity and acceleration), while other sensors can still directly observe their corresponding state quantities.
[0048] represents the random noise in the measurement, such as sensor accuracy limitation, environmental interference, etc. The noise is assumed to be zero-mean Gaussian noise, and the measurement noise covariance matrix is , reflects the intensity of the measurement noise of each sensor and is expressed as,
[0049] in, The expected value of the measurement noise is zero, which means that the measurement error will not cause deviation in long-term observations. Indicates the variance of the measurement noise and the correlation between different sensor noises. The larger the variance, the greater the measurement error of the sensor. Represents the measurement noise It follows a Gaussian distribution with zero mean.
[0050] Assume that the elevator uses two sensors, one for measuring velocity and one for measuring acceleration. The measurement noise of the sensor is expressed as:
[0051] in, and represent the measurement noise of velocity and acceleration sensors respectively; Covariance matrix of measurement noise for:
[0052] in, Represents the noise variance of the speed sensor. The larger the variance, the greater the error in the speed measured by the sensor. represents the noise variance of the acceleration sensor, and represents the correlation of the noise between the velocity sensor and the acceleration sensor. If the noise of the two sensors is independent, then .
[0053] For example, when the elevator is running, the speed sensor has a higher measurement accuracy and a noise variance of 0.01, while the acceleration sensor has a slightly lower accuracy and a noise variance of 0.05. The covariance matrix can be expressed as:
[0054] This means that the velocity sensor has less measurement noise, while the acceleration sensor has a larger measurement error.
[0055] S3. According to the state vector, a multi-rate Kalman filter is used to fuse sensor data with different sampling frequencies to obtain a unified state estimate, including the following steps: S3.1. At every moment , the Kalman filter first calculates the current state vector and control input , predict the next moment through the state equation The goal of the prediction step is to provide a priori estimation, that is, to infer the state of the next moment based on the dynamic model of the elevator operation state without measurement information.
[0056] S3.1.1, from the time Status Departure, using the state equation to predict the time Status , the expression is,
[0057] in, For the moment The predicted state vector represents the time-based The state estimate obtained for the time Prediction of status, For the moment The state vector is the optimal estimation result combined with the measurement information of the previous moment.
[0058] S3.1.2. While predicting the state, the Kalman filter also predicts the state covariance matrix at the next moment. , which reflects the uncertainty of the elevator operation state. Using the state transfer matrix and process noise covariance matrix, the prediction error covariance of the next moment is predicted, and the expression is,
[0059] in, For the moment The prediction covariance matrix of For the moment The covariance matrix of represents the state uncertainty at the current moment, is the transpose of the state transfer matrix, is the process noise covariance matrix, which describes the process noise in the elevator operation state (such as friction in elevator operation, external interference, etc.); It should be noted that the prediction step provides a priori estimates of the state for the next moment, that is, how the elevator's operating state evolves according to its own dynamic model without measurement information. This step lays the foundation for the next update step, where the predicted value will be compared with the actual measurement value to correct the state estimate.
[0060] S3.2, according to the prediction error covariance, observation matrix and measurement noise covariance matrix, calculate the Kalman gain, Kalman gain Determines the weighted ratio between the predicted value and the measured value. The specific formula is:
[0061] in, For the moment The Kalman gain determines the weight distribution between the predicted state and the measured value. is the measurement noise covariance matrix, which describes the statistical characteristics of the measurement noise (such as the measurement error of the sensor); It should be noted that the role of the Kalman gain is to balance the influence of prediction and measurement: if the prediction is more credible (that is, the uncertainty of the prediction is smaller), the Kalman gain will tend to trust the predicted value; conversely, if the measurement is more accurate (that is, the measurement noise is smaller), the Kalman gain will tend to trust the measured value.
[0062] S3.3. Using measured values To update the state vector, the expression formula is,
[0063] in, For the moment The updated state vector, which combines the prediction and measurement information, is the final optimal state estimate. For the moment The predicted state vector is For the moment The actual measured value, is the predicted measurement value, that is, through the observation matrix Map the predicted state to the observation space; It should be noted that in the state update, the error between the measured value and the predicted value It is called the measurement residual or innovation, and it reflects the difference between the actual measurement and the prediction. The Kalman gain uses the residual to modify the predicted state based on the confidence of the prediction and the accuracy of the measurement.
[0064] S3.4. While updating the state estimate, the Kalman filter also needs to update the state covariance matrix to reflect the uncertainty after the updated state. The update formula of the covariance matrix is:
[0065] in, For the moment The updated state covariance matrix reflects the uncertainty of the state vector after combining the measurement information. is the identity matrix, For the moment The Kalman gain of It should be noted that the significance of the update step is that the Kalman filter combines the actual measurement data and the system model prediction to obtain a more accurate state vector. Through the dynamic adjustment of the Kalman gain, the filter can find an optimal balance between prediction and measurement, thereby achieving the best estimate of the elevator operation state.
[0066] For sensors with a high sampling frequency, their data can be Direct use means that the data is updated in each Kalman filter cycle. For sensors with lower sampling frequency, the data is updated at the current moment. Not available. New observations are generated only at the sampling moment. When the low-frequency sensor does not generate new data, the Kalman filter only predicts the state without updating the state. When there are new observations, the Kalman filter uses the observations to update the state.
[0067] S4, extracting the elevator operation status signal from the state estimation, and performing interpolation processing to reconstruct the high-resolution operation status signal, including the following steps: S4.1. Obtain the low-frequency sensor at time and The estimated value of the state: For the moment The sensor state estimate is obtained by fusing the sensor data with the Kalman filter. For the moment The sensor state estimate is the observation value provided by the low-frequency sensor at the next sampling moment, which is fused by the Kalman filter.
[0068] S4.2. Determine the number of time steps for interpolation ,in Represents the interpolation point generated between two low-frequency sampling points; Corresponding to and The first interpolation point between Corresponding to the last interpolation point, that is, close to the moment data points.
[0069] S4.3. Calculate each interpolation point according to the linear interpolation formula The state estimate of is expressed as,
[0070] in, Represents the position weight of the current interpolation point. As the interpolation point increases, , so the ratio gradually increases; The ratio Applied to time The estimated value of the state , and at the same time Applied to time The estimated value of the state , adding these two parts gives the time State estimation at interpolation points; For all time steps, the interpolation calculation is repeated until the state estimates for all interpolation points are completed.
[0071] S4.4. The interpolated state estimate It is combined with the state estimation result of the Kalman filter to form a continuous, high-resolution state vector; high resolution refers to a finer-grained time interval on the time axis, that is, more state estimation points are inserted between the originally discrete sampling points, making the state vector more continuous and smooth in time.
[0072] In a preferred technical solution, if the sampling period of the low-frequency sensor is T, the time interval of the parameters in the high-resolution state vector is 0.1T~0.5T. It should be noted that the specific time interval of the high-resolution should also be determined according to the actual usage scenario and the actual frequency of the sensor; At the moment and At each time step between This enables the state estimation of the low-frequency sensor to have a higher time resolution, thereby obtaining a smoother elevator operation status signal.
[0073] It should be noted that, for the elevator running state, the sampling frequency of low-frequency sensors (load, etc.) is generally low, and its state changes relatively slowly. The linear interpolation method can provide sufficient accuracy to smoothly transition the sampling interval of low-frequency data, for example: The data change of the load sensor usually occurs when the elevator starts or stops, and the change is relatively slow. Linear interpolation can also capture this change well.
[0074] Through the interpolation method, a higher-resolution state estimate can be generated between low-frequency sampling points, avoiding the state discontinuity problem caused by the large sampling interval of low-frequency data.
[0075] S5, by comparing the high-resolution operating status signal with the historical normal status, identifying the abnormal operating status, including the following steps, S5.1. Get the current time through Kalman filter output The state vector , which contains multiple key states of elevator operation (such as elevator speed, acceleration, load, etc.); The state vector of the elevator in normal operation is extracted from the historical data. This vector can be obtained by statistically analyzing a large amount of historical normal operation data. Usually, the mean or median is taken as the normal value of each state variable. For example, if the monitored state variables include the elevator speed , acceleration , Elevator load , the historical normal state vector is,
[0076] in, is the historical normal state vector.
[0077] S5.2. Use the Euclidean distance formula to quantify the difference between the current state and the normal state. This distance measures the degree to which the current state deviates from the normal state. The expression is:
[0078] in, represents the Euclidean distance formula, is the number of state vectors, It is the current moment Kalman filter estimation State component, is the first Normal reference value of each state component; This distance reflects the overall difference between the current state and the normal operating state. The greater the difference, the greater the degree to which the elevator operating state deviates from the normal state.
[0079] S5.3. Setting thresholds , the threshold should be determined according to the specific operating environment of the elevator and the state fluctuation range in historical data. Usually, It can be set to the mean of the Euclidean distance under historical normal conditions plus several times the standard deviation to avoid false positives.
[0080] The calculated Euclidean distance With the set threshold Make comparisons; if , the elevator is judged to be operating abnormally; otherwise, normal operation monitoring continues.
[0081] S5.4. Establish a fault warning mechanism. Exceeding the preset threshold When the elevator is running at a high speed, it indicates that the current elevator operation status deviates greatly from the normal status, and there may be a fault or abnormality. At this time, an early warning needs to be issued immediately and the maintenance personnel need to be notified to conduct an inspection.
[0082] Specifically, when When an alarm occurs, the alarm information will be recorded, including: the timestamp of the alarm, the current state vector, the calculated Euclidean distance, and the corresponding normal state vector; these data will be saved in the log to facilitate subsequent fault analysis.
[0083] The alarm information is notified to maintenance personnel via SMS, email or mobile application. The notification content includes the alarm time, specific description of the abnormal status (such as deviations in speed, acceleration, load, etc.), and recommended treatment measures (such as checking the elevator braking system, load sensors, etc.); maintenance personnel can take inspection and maintenance measures in time according to the alarm information to ensure the safe operation of the elevator.
[0084] The alarm status is updated to the central monitoring platform, and the alarm information is displayed on the user interface. At the same time, according to the severity of the alarm, corresponding response measures are taken, such as: if it is a minor fault, only a warning is issued and the operation continues; if it is a serious fault, the elevator operation is suspended and maintenance personnel are required to check.
[0085] According to the Euclidean distance The alarm can be divided into different levels: Low level alarm: When , it indicates a minor abnormality and maintenance personnel need to conduct a preliminary inspection.
[0086] Intermediate alarm: When When the elevator is in operation, it indicates that there may be a potential fault and maintenance personnel need to check it as soon as possible.
[0087] Advanced alarm: When When the elevator is in a critical condition, it may have a serious fault and the elevator will stop automatically, and maintenance personnel will be notified for emergency inspection.
[0088] This embodiment also provides an elevator operation state determination system based on multi-source data fusion, including: a data acquisition module, responsible for collecting elevator operation state data; a state space modeling module, responsible for establishing the state space modeling of the elevator, defining the state vector, predicting the state of the elevator at the next moment through the state equation, and converting the state vector into a measurement vector; a data fusion module, responsible for fusing sensor data with different sampling frequencies according to the state vector using a multi-rate Kalman filter to obtain a unified state estimate; an interpolation and reconstruction module, responsible for extracting the elevator operation state signal from the state estimate, and interpolating and reconstructing a high-resolution operation state signal; an abnormality detection module, responsible for identifying abnormal operation states by comparing high-resolution operation state signals with historical normal states.
[0089] This embodiment also provides a computer device, which is suitable for the elevator operation status determination method based on multi-source data fusion, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the elevator operation status determination method based on multi-source data fusion as proposed in the above embodiment.
[0090] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.
[0091] This embodiment also provides a storage medium, on which a computer program is stored. When the program is executed by a processor, the elevator operation state determination method based on multi-source data fusion proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.
[0092] In summary, the present invention covers different physical states of elevator operation through multi-dimensional data collection, ensures the comprehensiveness of data, provides the necessary data basis for subsequent state space modeling and data fusion, and can capture the complex dynamic characteristics of elevator operation. The state equation describes the dynamic evolution of the elevator under the control input and noise. The measurement equation is used to associate sensor data with the state vector, and a mathematical model that can reflect the internal connection between various operating states of the elevator is constructed to achieve a systematic description of the elevator operation state. The multi-rate Kalman filter is used to fuse sensor data of different frequencies, which solves the problem that multi-source sensors cannot synchronize and fuse data due to inconsistent sampling frequencies, improves the accuracy and real-time performance of data fusion, reduces the impact of noise on state estimation, compensates for the information loss caused by low-frequency data through time interpolation, and generates a state signal equivalent to high-frequency sampling, improves the precision of elevator operation state monitoring, enables detection of more subtle state changes, and further enhances the sensitivity and recognition ability of abnormal states. By quantifying the difference between the current state and the normal state, the fault or abnormal operation of the elevator can be effectively determined. By referring to historical data and setting thresholds, the potential fault of the elevator can be accurately identified at an early stage.
[0093] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.
Claims
1. A method for determining the operation status of an elevator based on multi-source data fusion, characterized in that: include, Collect elevator operation status data; Establish the state space model of the elevator, define the state vector, predict the state of the elevator at the next moment through the state equation, and convert the state vector into a measurement vector; According to the state vector, a multi-rate Kalman filter is used to fuse sensor data with different sampling frequencies to obtain a unified state estimate; Extract the elevator's operating status signal from the state estimation, perform interpolation processing, and reconstruct a high-resolution operating status signal; By comparing high-resolution operating status signals with historical normal states, abnormal operating conditions can be identified.
2. The elevator operation status determination method based on multi-source data fusion according to claim 1, characterized in that: Establish the state space model of the elevator, predict the state of the elevator at the next moment through the state equation, and convert the state vector into the measurement vector, including the following steps: Define a state vector containing elevator operation status data; The state equation is used to describe the process of evolving from the current state to the state at the next moment under the influence of control input and process noise. The expression is: ; in, is the state vector at the next moment, is the state vector at the current moment, is the state transfer matrix, is the control input matrix, For external control input, is the process noise; Using the measurement equation, the state vector is converted into a measurement vector through the observation matrix, and the expression is: ; in, is the measurement vector, is the observation matrix, To measure noise.
3. The elevator operation status determination method based on multi-source data fusion according to claim 2, characterized in that: According to the state vector, a multi-rate Kalman filter is used to fuse sensor data with different sampling frequencies to obtain a unified state estimation, including the following steps: Define the process noise as zero-mean Gaussian noise, and use the state transfer matrix and process noise covariance matrix to predict the prediction error covariance at the next moment. The expression is: ; in, For the moment The prediction covariance matrix of For the moment The covariance matrix of is the transpose of the state transfer matrix, is the process noise covariance matrix; The measurement noise is defined as zero-mean Gaussian noise. The Kalman gain is calculated based on the prediction error covariance, the observation matrix, and the measurement noise covariance matrix. The specific formula is: ; in, For the moment The Kalman gain, is the measurement noise covariance matrix; Using the Kalman gain and the observed value, the state prediction is corrected to obtain the updated state estimate; The error covariance is updated through the Kalman gain and the observation matrix.
4. The elevator operation status determination method based on multi-source data fusion according to claim 3 is characterized in that: The method of extracting the elevator operation state signal from the state estimation, performing interpolation processing, and reconstructing the high-resolution operation state signal includes the following steps: Get the low frequency sensor at time and The state estimate of determines the number of time steps for interpolation; According to the linear interpolation formula, calculate the state estimation value of each interpolation point; For all time steps, the interpolation calculation is repeated until the state estimates for all interpolation points are completed.
5. The elevator operation status determination method based on multi-source data fusion according to claim 4, characterized in that: The method of identifying an abnormal operating state by comparing a high-resolution operating state signal with a historical normal state comprises the following steps: Extract the state vector of the elevator in normal operation from the historical data; Use the Euclidean distance formula to quantify the difference between the current state and the normal state. The expression is: ; in, represents the Euclidean distance, is the number of state vectors, It is the current moment The Kalman filter estimates the state components, is the first Normal reference value of each state component; Setting Thresholds , the calculated Euclidean distance With the set threshold Make comparisons; if , the elevator is judged to be operating abnormally; otherwise, normal operation monitoring continues.
6. The elevator operation status determination method based on multi-source data fusion according to claim 5, characterized in that: The elevator operation status data includes the speed, acceleration, vibration, load and motor current of the elevator.
7. An elevator operation status determination system based on multi-source data fusion, based on the elevator operation status determination method based on multi-source data fusion according to any one of claims 1 to 6, characterized in that: include, Data acquisition module, responsible for collecting elevator operation status data; The state space modeling module is responsible for establishing the state space modeling of the elevator, defining the state vector, predicting the state of the elevator at the next moment through the state equation, and converting the state vector into a measurement vector; The data fusion module is responsible for fusing sensor data with different sampling frequencies using a multi-rate Kalman filter according to the state vector to obtain a unified state estimate; The interpolation and reconstruction module is responsible for extracting the elevator's operating status signal from the state estimation, interpolating it, and reconstructing a high-resolution operating status signal; The anomaly detection module is responsible for identifying abnormal operating states by comparing high-resolution operating state signals with historical normal states.
Citation Information
Patent Citations
Power system dynamic state estimation method based on PMU and SCADA hybrid measurement
CN113675951A
Motion sensing running game method based on unity engine
CN116637362A
Multi-sensor data fusion method based on Kalman filtering parameter extraction and state updating
CN117313029A
Timestamp alignment method for sensor data and storage medium
CN118450483A
Elevator safety monitoring method and system based on multi-source data fusion and electronic equipment
CN118561117A
Cited By
Elevator running state evaluation method and system based on multi-source heterogeneous data
CN120430634A
An elevator operation state evaluation method and system based on multi-source heterogeneous data
CN120430634B
Elevator multi-caliper disc brake diagnostic methods, systems, apparatus, media, and products
CN120534836A
Industrial scene high-precision personnel positioning and danger early warning method and system
CN120857078A
Villa elevator fault identification method and system based on multi-modal data fusion
CN120887300A
