Intelligent fire elevator data acquisition method and system based on intermittent classification

Abstract

The invention belongs to the technical field of data acquisition, and provides an intelligent fire elevator data acquisition method and system based on intermittent classification, and the method specifically comprises the steps: firstly, an intelligent fire elevator data acquisition assembly comprises a vibration acquisition module, the acceleration is obtained through the vibration acquisition module, a basic intermittent section is preset, and an intermittent stage number is established according to the basic intermittent section; then a loss reference effect state is obtained according to the acceleration and intermittent series decoupling analysis, then performance loss early warning factor judgment is conducted on the loss reference effect state, and finally data collection is conducted from the fire elevator in combination with a performance loss early warning factor. Based on calculation of the loss reference effect state, the influence of wind vibration or other vibration existing in a building on a motor in the intelligent fire elevator can be effectively recognized, the occurrence degree and effect of potential performance loss risks are estimated in time, and storage of unnecessary or redundant data is reduced; and negative effects of redundant data and invalid data on the constructed prediction model are further reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of data acquisition, and in particular relates to an intelligent fire elevator data acquisition method and system based on intermittent grading. Background Art

[0002] Fire elevators are specially designed for emergency situations like fires, ensuring the safe evacuation of building occupants. Fire elevators not only function as conventional elevators but also offer fire safety escort capabilities, enabling them to respond to fire, gas, or earthquake scenarios.

[0003] In everyday elevator use, elevators are often subject to significant wind vibration and vibration, particularly in high-rise buildings, where vibration is more frequent and intense. While firefighting elevators are typically designed to maintain a certain level of functionality despite long-term wind and vibration exposure, frequent and intense exposure can cause wear and tear in some of the elevator's systems (including the motor, brake system, and control panel). The cumulative effect of these wear and tear can lead to equipment degradation, increased maintenance costs, and reduced elevator response speed and stability. In firefighting scenarios, this can easily compromise the firefighting performance of firefighting elevators, including damage to structural materials, fire resistance, electrical circuits, aging components, and loose connections. Therefore, data collection on vibration-induced damage during daily firefighting elevator operation is urgently needed to develop a predictive model to mitigate performance degradation caused by firefighting elevator vibration. This predictive model can then limit the intensity of firefighting elevator operations, including suspension, restricted operating floors, and speed limits. However, due to the large number of sensors inherent in firefighting elevators and the unique characteristics of firefighting elevators across buildings, traditional data collection methods cannot accurately locate the time at which valid data occurs, leading to data redundancy and sensitivity issues in the resulting predictive models. The effective data is actually the probabilistic time point of fire performance loss. However, if only the data at the time point of vibration occurrence is collected, the timing characteristics of the vibration occurrence are easily ignored. Due to the power of the elevator motor, there is an urgent need for an intelligent fire elevator data collection method based on intermittent grading. Summary of the Invention

[0004] The purpose of the present invention is to propose an intelligent fire elevator data collection method and system based on intermittent classification to solve one or more technical problems existing in the prior art and at least provide a beneficial option or create conditions.

[0005] To achieve the above object, according to one aspect of the present invention, a method for collecting data of an intelligent fire elevator based on intermittent grading is provided, the method comprising the following steps: The intelligent fire elevator data acquisition component includes a vibration acquisition module, which obtains acceleration through the vibration acquisition module; presets a basic intermittent segment and establishes an intermittent series based on the basic intermittent segment; obtains a loss reference effective state through decoupling analysis of acceleration and intermittent series; determines the performance loss warning factor of the loss reference effective state; and collects data from the fire elevator in combination with the performance loss warning factor.

[0006] Further, the intelligent fire elevator data acquisition component includes a vibration acquisition module, the method of obtaining acceleration through the vibration acquisition module is: the intelligent fire elevator data acquisition component includes a vibration acquisition module, the vibration acquisition module is arranged in the elevator motor compartment; The reason why the vibration collection module is arranged in the motor compartment of the elevator is that the motor compartment and the elevator shaft are common setting locations for the elevator to be affected by vibration. Since the motor compartment is the location where the motor, control device and reducer are arranged during the operation of the elevator, its own operating state will bring certain vibrations. Therefore, the vibration data identified by wind vibration or building vibration is more significant and sensitive. The vibration measurement of the elevator shaft has obvious dispersion, and the walls, tracks and other components of the shaft will have a certain buffering or absorption effect on the vibration. Therefore, the monitoring of the occurrence of fire protection performance loss will inevitably have lag and adaptability problems.

[0007] The vibration acquisition module includes an acceleration sensor, which is a piezoelectric acceleration sensor, an optical fiber acceleration sensor or a MEMS acceleration sensor. The acceleration is obtained by real-time measurement using an acceleration sensor.

[0008] The vibration acquisition module transmits data via LoRa communication technology, Wi-Fi, or 5G networks. It collects real-time elevator acceleration data and sends it to a central server, which stores and processes it in real time. The server also uses this data to adjust the elevator's operating mode, including adjusting speed and limiting operating floors. The acceleration measured in real time by the accelerometer is stored in a computer as an array, with each element representing the acceleration component in different directions in three-dimensional space.

[0009] Through network functions, the environmental acquisition module and cloud servers form a distributed collaborative architecture. This enhanced data processing and real-time response capabilities improve the efficiency and reliability of the intelligent fire elevator's emergency response in changing environments. The system also supports remote control and intelligent optimization features to facilitate software updates and iterations, continuously improving system stability and intelligence.

[0010] Furthermore, the method of presetting the basic intermittent segment and establishing the intermittent series according to the basic intermittent segment is: Set the basic rest period to 1-10 minutes; The essence of the basic interval segment is the length of the time period. That is, when the basic interval segment is set to 5 minutes, the length of each basic interval segment is 5 minutes, which serves as the basic unit for modeling data collection. At any moment, there is only one corresponding basic interval segment.

[0011] An integer variable is preset as the intermittent series KL, and the value range of the intermittent series is KL∈[1,15]. For any intermittent series KL, the intermittent segment it represents is the time period of KL basic intermittent segments.

[0012] For example, when the representative interval KL is 5 and the basic interval is 3 minutes, then when the interval level is 5, the corresponding representative interval is 15 minutes.

[0013] By changing the KL value in different intervals, the data collection of more diverse time periods is enhanced so that the changing trend of elevator vibration can be analyzed within a more flexible period.

[0014] Furthermore, the method for obtaining the loss reference effective state based on the decoupling analysis of acceleration and intermittent series is: At any moment, the basic intermittent segment belonging to that moment is the current basic intermittent segment. Any intermittent series from the current basic intermittent segment is selected as the calculation level. The range of the loss characteristic data corresponding to the calculation level in the intermittent segment is calculated. The ratio of the obtained range value to the maximum value of each acceleration module length in the current basic intermittent segment is recorded as the loss reference effective state of the calculation level; the loss characteristic data is the power value of the elevator motor when it is in a uniform motion state.

[0015] The operation level corresponding to the representative intermittent segment represents a period of time from the first moment of the current intermittent segment in the reverse time direction representing the intermittent segment.

[0016] The power consumption of the motor is usually closely related to the vibration, load and operating speed of the elevator. By analyzing the changes in motor power, it is usually possible to infer whether the elevator has performance loss during operation.

[0017] Furthermore, the method for determining the performance loss warning factor of the loss reference effectiveness state is as follows: presetting a monitoring period TCr, TCr∈[2,6] hours, and recording each basic intermittent segment as a reference node within the monitoring period; Obtain the loss reference state of each benchmark node and form a sequence recorded as Wsf.Ls; Calculate the mean and standard deviation of the elements in the sequence Wsf.Ls. If the element is less than or equal to the sum of the mean and three times the standard deviation, and greater than or equal to the difference between the mean and three times the standard deviation, then record the corresponding reference node as a low-order loss point; otherwise, record it as a high-order loss point. The mathematical expression for determining whether a reference node is a low-order loss point is: Wsf.Ls.mean-3 Wsf.Ls.err≤Wsf.Ls(k1)≤Wsf.Ls.mean+3 Wsf.Ls.err; where Wsf.Ls.mean represents the mean of the elements in the sequence Wsf.Ls, Wsf.Ls.err represents the standard deviation of the elements in the sequence Wsf.Ls, Wsf.Ls(k1) represents the k1th element in the sequence Wsf.Ls, and k1 is the sequence number of the element in the sequence.

[0018] The time scale of the loss reference effective state obtained from the low-order loss point and its corresponding loss reference effective state constitute a low-loss binary group. The set constructed by the low-loss binary group within the monitoring period is interpolated with cubic splines using the plrep and splev functions to obtain a smooth curve. The corresponding fitting value of the benchmark node on the smooth curve obtained by fitting is used as the curve value of the benchmark node. The plrep and splev functions are called in Python. The absolute value of the difference between the curve value of any high-order loss point and its corresponding loss reference state is recorded as the loss difference; the loss difference of each high-order loss point is obtained and a difference evaluation sequence is constructed; the upper quartile Xes.sl, lower quartile Xes.xl and interquartile range Xes.jl in the difference evaluation sequence are obtained respectively, where the interquartile range is the difference between the upper quartile and the lower quartile; the loss stable interval is constructed according to the difference evaluation sequence as (Xes.xl-k2×Xes.jl,Xes.sl+k2×Xes.jl); where k2 is a preset amplitude coefficient, and its value range is k2∈[1,20]; The amplitude coefficient is used to compensate for the insufficient density of high-order loss points. When the time gaps between high-order loss points are more uniform, a smaller value is selected. Otherwise, a larger value is selected, thereby reducing the risk of collecting unnecessary model data due to excessive sensitivity to a certain extent.

[0019] The principle of obtaining the loss stability interval is actually an interval defined based on the quartile analysis of the loss difference. By evaluating the loss difference of high-order loss points, the stable stage of elevator performance loss and the stage of abnormal fluctuation can be effectively identified and distinguished. The upper quartile and lower quartile represent the high and low loss fluctuation levels of the intelligent fire elevator system based on intermittent grading under normal operation, respectively, and the interquartile range reflects the volatility between the two. At the same time, the loss stability interval enhances the sensitivity of the prediction model, can accurately capture small fluctuations in elevator loss, and concentrates on analyzing the loss difference to eliminate redundant risk signals.

[0020] If the current benchmark node is a marked low-loss point, the performance loss warning factor is assigned a value of FALSE. Otherwise, if the current benchmark node is in a stable loss interval, the performance loss warning factor is assigned a value of FALSE. Otherwise, the performance loss warning factor is assigned a value of TRUE. The current benchmark node represents the benchmark node corresponding to the current moment. A FALSE performance loss warning factor indicates that no risk has occurred, while a TRUE performance loss warning factor indicates that a risk has occurred.

[0021] Since the determination of the performance loss warning factor requires processing the loss difference, it can effectively filter redundant data and quantify the performance loss risk of the intelligent fire elevator system based on intermittent classification. However, the acquisition of high-order loss points is overly dependent on the loss reference effect state, resulting in excessive data sensitivity, which causes deviations in the loss difference and unreasonable assumptions about the loss stable interval. Ultimately, it is easy to cause decision-making deviations and collect unnecessary model data. This problem is particularly prominent during periods with low density of high-order loss points. However, the existing technology cannot effectively compensate for this deviation. In order to eliminate this effect, the present invention proposes a more preferred solution as follows: Preferably, the method for determining the performance loss warning factor for the loss reference state is: The monitoring period TCr, TCr∈[2,10] hours is preset. During the monitoring period, each basic intermittent segment is recorded as a reference node. The loss reference state of each reference node during the monitoring period is obtained and a reference loss state sequence is formed. The reference nodes corresponding to each maximum value element in the reference loss state sequence are recorded as adjacent nodes; Calculate the difference between the loss reference effective state of any adjacent node and its left and right adjacent reference nodes, and select the maximum value of the difference as the loss critical value of the adjacent node; The principle of obtaining the loss critical value here is actually to obtain it by calculating the difference between the loss reference effective state of the adjacent node and its left and right adjacent reference nodes. The loss critical value directly reflects the performance loss difference between different time points in the elevator and is a reflection of the potential performance loss risk. The loss critical value quantifies the degree of elevator performance fluctuation. If the loss critical value is large, it means that the power consumption of the elevator has experienced significant abnormal fluctuations, and the probability of risk increases. If the loss critical value is small, it means that the difference in loss reference effective state between adjacent reference nodes is small, and the elevator system runs smoothly. This is the performance of the elevator under light load or continuous and stable operation. Therefore, obtaining the loss critical value is an effective evaluation method for elevator fault prediction.

[0022] The number of adjacent nodes is recorded as sdn; the time interval between any adjacent node and its previous adjacent node is recorded as the loss trend value, and the sequence of all loss trend values ​​is recorded as Gn.Ls; The DBSCAN density clustering algorithm is used to cluster the loss trend values ​​of each adjacent node, and all adjacent nodes are divided into several clusters, which are recorded as trend cluster clusters. The neighborhood radius of the DBSCAN density clustering algorithm is the upper quartile of the Gn.Ls sequence, and the minimum number of points is the intermittent level of the current benchmark node. The early warning assessment level Evyp of each trend cluster is calculated based on the edge nodes and loss critical value: ; Where j1 is the serial number of the edge node in the trend cluster, sn is the number of edge nodes in the trend cluster, Lopp j1 is the loss reference state of the j1th adjacent node in the trend cluster, MLopp is the average loss reference state of all adjacent nodes, LCocv j1 is the ordinal number of the loss critical value of the j1th adjacent node in the trend cluster corresponding to the adjacent node when it is arranged from large to small among the loss critical values ​​of each adjacent node, avg{} is the average value function, and exp() is the exponential function with the natural constant e as the base; The warning assessment level of each trend cluster is normalized using the minimum-maximum method and recorded as the loss risk status value. If the loss risk status value of a trend cluster is greater than or equal to 0.5, the trend cluster is marked as a risk type cluster, otherwise it is marked as a non-risk type cluster; the maximum loss critical value in the trend cluster is recorded as the critical peak value of the trend cluster; the Euclidean distance between the current benchmark node and the critical peak value of any trend cluster is recorded as the current critical distance, and the trend cluster with the minimum previous critical distance is selected as the target cluster. If the target cluster is a risk type cluster, the performance loss warning factor of the current benchmark node is assigned to TRUE, otherwise the performance loss warning factor is assigned to FALSE.

[0023] If the target cluster cannot be found, it is considered that the performance loss warning factor of each reference node in the monitoring period TCr is FALSE; thus, the performance loss warning factor of the current reference node is assigned FALSE.

[0024] Beneficial effect: Since the performance loss warning factor is a real-time analysis of elevator vibration data based on acceleration and intermittent series calculations, it can effectively identify the impact of wind vibration or other vibrations in the building on the motor of the intelligent fire elevator based on the calculation of the loss reference effect state, and timely estimate the extent of potential performance loss risks.

[0025] Furthermore, the method for collecting data from fire elevators in combination with the performance loss warning factor is: For any basic intermittent segment, if the performance loss warning factor corresponding to more than half of the intermittent levels is judged to be true, the risk of performance loss in this basic intermittent segment is defined; When there is a risk of performance loss, the interval level with the maximum value of the loss reference effect state is selected as the storage target interval level. The interval segment corresponding to the storage target interval level is recorded as the target interval segment, and the environmental data within the target interval segment is stored in the cloud.

[0026] The environmental data package not only includes data from the acceleration sensor, but also includes modeling data such as running speed, pressure, temperature, camera, motor power, etc. that serve the prediction model.

[0027] Preferably, all undefined variables in the present invention, if not clearly defined, can be manually set thresholds.

[0028] The present invention also provides an intelligent fire elevator data acquisition system based on intermittent grading. The intelligent fire elevator data acquisition system based on intermittent grading includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the intelligent fire elevator data acquisition method based on intermittent grading are implemented. The intelligent fire elevator data acquisition system based on intermittent grading can be run on computing devices such as desktop computers, laptop computers, PDAs, and cloud data centers. The executable system may include, but is not limited to, a processor, a memory, and a server cluster. The processor executes the computer program and runs it in the following system units: An environment arrangement unit, used for acquiring acceleration through a vibration acquisition module; Interval series unit, used to preset the basic intermittent segment and establish the intermittent series according to the basic intermittent segment; Loss analysis unit, used to obtain loss reference effective state based on acceleration and intermittent series decoupling analysis; The loss determination unit is used to determine the performance loss warning factor of the loss reference effect.

[0029] The data acquisition unit is used to collect data from the fire elevator in combination with the performance loss warning factor.

[0030] The beneficial effects of the present invention are as follows: the present invention provides an intelligent fire elevator data collection method and system based on intermittent grading, which collects and analyzes elevator vibration data in real time through performance loss warning factors, and can effectively identify the impact of wind vibration or other vibrations existing in the building on the motor in the intelligent fire elevator based on the calculation of loss reference effect state, and timely estimate the degree of potential performance loss risk, so that the data in the modeling data collection process can be dynamically and effectively screened, reducing the storage of unnecessary or redundant data, further reducing the negative effects of redundant data and invalid data on the constructed prediction model, and preventing the constructed model from having problems of insufficient adaptability or insufficient accuracy.

[0031] This method obtains elevator data when vibration occurs through an intermittent grading method, which can effectively aggregate the timing characteristics under different vibration cycles or vibration patterns, providing more timely and accurate support for subsequent fault diagnosis, performance optimization and prediction, thereby further ensuring the safety, stability and long-term operation reliability of the elevator. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The above and other features of the present invention will become more apparent through a detailed description of the embodiments shown in conjunction with the accompanying drawings. In the drawings of the present invention, the same reference numerals represent the same or similar elements. Obviously, the drawings described below are only some embodiments of the present invention. It is possible for a person skilled in the art to derive other drawings based on these drawings without inventive effort. In the drawings: Figure 1 Shown is a flow chart of an intelligent fire elevator data collection method based on intermittent classification; Figure 2 The figure shows the structure diagram of the intelligent fire elevator data acquisition system based on intermittent classification. DETAILED DESCRIPTION

[0033] The following will be combined with the embodiments and drawings to clearly and completely describe the concept, specific structure and technical effects of the present invention so as to fully understand the purpose, scheme and effect of the present invention. It should be noted that the embodiments and features in the embodiments of this application can be combined with each other unless there is a conflict.

[0034] like Figure 1 The following is a flow chart of the intelligent fire elevator data collection method based on intermittent classification. Figure 1 To illustrate the intelligent fire elevator data collection method based on intermittent classification according to an embodiment of the present invention, the method includes the following steps:

[0035] Example 1: The data acquisition component of an intelligent fire elevator includes a vibration acquisition module, through which acceleration is acquired; a basic intermittent segment is preset, and an intermittent series is established based on the basic intermittent segment; a loss reference effective state is obtained through decoupling analysis of acceleration and intermittent series; a performance loss warning factor is determined for the loss reference effective state; and data is collected from the fire elevator in combination with the performance loss warning factor.

[0036] Further, the intelligent fire elevator data acquisition component includes a vibration acquisition module, the method of obtaining acceleration through the vibration acquisition module is: the intelligent fire elevator data acquisition component includes a vibration acquisition module, the vibration acquisition module is arranged in the elevator motor compartment; The vibration acquisition module includes an acceleration sensor, which adopts a fiber optic acceleration sensor; The acceleration is obtained by real-time measurement using an acceleration sensor.

[0037] Furthermore, the method of presetting the basic intermittent segment and establishing the intermittent series according to the basic intermittent segment is: Set the basic rest period to 3 minutes; An integer variable is preset as the intermittent series KL, and the value of the intermittent series is 15. For any intermittent series KL, it represents an intermittent segment with a time period of KL basic intermittent segments.

[0038] Furthermore, the method for obtaining the loss reference effective state based on the decoupling analysis of acceleration and intermittent series is: At any moment, the basic intermittent segment belonging to that moment is the current basic intermittent segment. Any intermittent series from the current basic intermittent segment is selected as the calculation level. The range of the loss characteristic data corresponding to the calculation level in the intermittent segment is calculated. The ratio of the obtained range value to the maximum value of each acceleration module length in the current basic intermittent segment is recorded as the loss reference effective state of the calculation level; the loss characteristic data is the power value of the elevator motor when it is in a uniform motion state.

[0039] Furthermore, the method for determining the performance loss warning factor of the loss reference effectiveness state is as follows: a monitoring period TCr is preset, with a value of 5 hours, and each basic intermittent segment within the monitoring period is recorded as a reference node; Obtain the loss reference state of each benchmark node and form a sequence recorded as Wsf.Ls; Calculate the mean and standard deviation of the elements in the sequence Wsf.Ls. If the element is less than or equal to the sum of the mean and three times the standard deviation, and greater than or equal to the difference between the mean and three times the standard deviation, then record the corresponding reference node as a low-order loss point; otherwise, record it as a high-order loss point. The threshold of three standard deviations above and below the mean was chosen to ensure statistical significance: according to the 3σ criterion for a normal distribution, the probability of data falling within the range of μ ± 3σ is 99.7%, and values ​​outside this range are considered significant outliers. This threshold effectively distinguishes between low-order loss points with normal fluctuations and high-order loss points with extreme anomalies, ensuring that warnings are triggered only for severe vibrations that significantly affect motor performance, balancing sensitivity with false alarm rate to avoid misjudgments.

[0040] The time scale of the loss reference state obtained from the low-order loss point and its corresponding loss reference state constitute a low-loss binary group. The cubic spline interpolation of the set constructed by the low-loss binary group within the monitoring period is performed using the plrep and splev functions to obtain a smooth curve. The absolute value of the difference between the curve value of any high-order loss point and its corresponding loss reference state is recorded as the loss difference; the loss difference of each high-order loss point is obtained and a difference evaluation sequence is constructed; the upper quartile Xes.sl, lower quartile Xes.xl, and interquartile range Xes.jl in the difference evaluation sequence are obtained respectively; the loss stationary interval is constructed based on the difference evaluation sequence as (Xes.xl-k2×Xes.jl, Xes.sl+k2×Xes.jl); where k2 is the amplitude coefficient, and its value is 4; If the current benchmark node belongs to the marked low-order loss point, the performance loss warning factor is assigned to FALSE; otherwise, if the current benchmark node belongs to the loss stable interval, the performance loss warning factor is assigned to FALSE, otherwise, the performance loss warning factor is assigned to TRUE.

[0041] Furthermore, the method for collecting data from fire elevators in combination with the performance loss warning factor is: For any basic intermittent segment, if the performance loss warning factor corresponding to more than half of the intermittent levels is judged to be true, the risk of performance loss in this basic intermittent segment is defined; When there is a risk of performance loss, the interval level with the maximum value of the loss reference effect state is selected as the storage target interval level. The interval segment corresponding to the storage target interval level is recorded as the target interval segment, and the environmental data within the target interval segment is stored in the cloud.

[0042] The environmental data package includes not only the data from the acceleration sensor, but also the running speed, pressure, temperature, camera and motor power.

[0043] The purpose of using cloud-stored elevator data is to collect vibration damage data from daily fire elevator operations, build a predictive model to mitigate performance degradation caused by vibration, and use this predictive model to limit the intensity of fire elevator operations, including suspension, restricted operating floors, and speed limits. Ultimately, this will create a closed-loop monitoring and control system with adaptive microcomputer ecosystems.

[0044] Example 2: The elevator data collection method of Example 2 is the same as that of Example 1, except that the method for determining the performance loss warning factor for the loss reference state is: The monitoring period TCr is preset to 5 hours. During the monitoring period, each basic intermittent segment is recorded as a reference node. The loss reference state of each reference node during the monitoring period is obtained and a reference loss state sequence is formed. The reference nodes corresponding to each maximum value element in the reference loss state sequence are recorded as adjacent nodes; Calculate the difference between the loss reference effective state of any adjacent node and its left and right adjacent reference nodes, and select the maximum value of the difference as the loss critical value of the adjacent node; The number of adjacent nodes is recorded as sdn; the time interval between any adjacent node and its previous adjacent node is recorded as the loss trend value, and the sequence of all loss trend values ​​is recorded as Gn.Ls; The loss trend value of each adjacent node is clustered using the DBSCAN density clustering algorithm, and all adjacent nodes are divided into several clusters, which are recorded as trend cluster clusters. The early warning assessment level Evyp of each trend cluster is calculated based on the edge nodes and loss critical value: ; Where j1 is the serial number of the edge node in the trend cluster, sn is the number of edge nodes in the trend cluster, Lopp j1 is the loss reference state of the j1th adjacent node in the trend cluster, MLopp is the average loss reference state of all adjacent nodes, LCocv j1 The loss threshold of the j1th adjacent node in the trend cluster corresponding to the adjacent node is the ordinal number of the loss threshold of the adjacent node when arranged from largest to smallest among the loss thresholds of all adjacent nodes. avg{} is the average function, and exp() is the exponential function with the natural constant e as the base. All variables in this calculation process are dimensionless parameters to ensure the mathematical consistency of the formula. Statistical normalization is also used to objectively quantify the risk levels of different clusters. The loss reference effective state is a dimensionless parameter calculated from the ratio of the motor power range to the maximum acceleration modulus. It is used to quantify the correlation between vibration and motor performance fluctuations.

[0045] The warning assessment level of each trend cluster is normalized using the minimum-maximum method and recorded as the loss risk status value. If the loss risk status value of a trend cluster is greater than or equal to 0.5, the trend cluster is marked as a risk type cluster, otherwise it is marked as a non-risk type cluster; the maximum loss critical value in the trend cluster is recorded as the critical peak value of the trend cluster; the Euclidean distance between the current benchmark node and the critical peak value of any trend cluster is recorded as the current critical distance, and the trend cluster with the minimum previous critical distance is selected as the target cluster. If the target cluster is a risk type cluster, the performance loss warning factor of the current benchmark node is assigned to TRUE, otherwise the performance loss warning factor is assigned to FALSE.

[0046] The loss risk status value is processed through minimum-maximum normalization, and its value range is [0,1]. Therefore, 0.5 is selected as the threshold to achieve the purpose of evenly dividing the decision boundary between risk and non-risk clusters by using the binary midpoint split method commonly used in statistics, avoiding model deviations such as missed reports due to too high a threshold or false positives due to too low a threshold.

[0047] The embodiment of the present invention provides an intelligent fire elevator data acquisition system based on intermittent classification, such as Figure 2 FIG2 shows a structural diagram of an intermittent hierarchical intelligent fire elevator data acquisition system according to the present invention. The intermittent hierarchical intelligent fire elevator data acquisition system according to this embodiment includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the intermittent hierarchical intelligent fire elevator data acquisition method according to the embodiment are implemented.

[0048] The system includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to run in the following units of the system: An environment arrangement unit, used for acquiring acceleration through a vibration acquisition module; Interval series unit, used to preset the basic intermittent segment and establish the intermittent series according to the basic intermittent segment; Loss analysis unit, used to obtain loss reference effective state based on acceleration and intermittent series decoupling analysis; The loss determination unit is used to determine the performance loss warning factor of the loss reference effect.

[0049] The data acquisition unit is used to collect data from the fire elevator in combination with the performance loss warning factor.

[0050] The intermittent hierarchical intelligent fire elevator data acquisition system can be run on computing devices such as desktop computers, laptops, PDAs, and cloud servers. Systems capable of running the intermittent hierarchical intelligent fire elevator data acquisition system may include, but are not limited to, processors and memory. Those skilled in the art will appreciate that the examples described are merely illustrative of intermittent hierarchical intelligent fire elevator data acquisition systems and do not constitute a limitation on intermittent hierarchical intelligent fire elevator data acquisition systems. Intermittent hierarchical intelligent fire elevator data acquisition systems may include more or fewer components than the examples, or combinations of certain components, or different components. For example, the intermittent hierarchical intelligent fire elevator data acquisition system may also include input and output devices, network access devices, buses, and the like.

[0051] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor. The processor serves as the control center of the intermittent hierarchical intelligent fire elevator data acquisition system operation system, and utilizes various interfaces and lines to connect various parts of the intermittent hierarchical intelligent fire elevator data acquisition system operation system.

[0052] The memory can be used to store the computer programs and / or modules. The processor implements the various functions of the intermittent hierarchical intelligent fire elevator data acquisition system by running or executing the computer programs and / or modules stored in the memory and accessing the data stored in the memory. The memory may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function (such as a sound playback function or an image playback function); the data storage area may store data generated based on the use of the mobile phone (such as audio data, a phone book, etc.). Furthermore, the memory may include high-speed random access memory and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0053] Although the present invention has been described in considerable detail and with particularity with respect to several embodiments, it is not intended to limit the present invention to any of these details or embodiments or any particular embodiment, so as to effectively encompass the intended scope of the present invention. In addition, the present invention has been described above with respect to embodiments foreseen by the inventors for the purpose of providing a useful description, and those insubstantial modifications of the present invention that are not currently foreseen may still represent equivalent modifications of the present invention.

Claims

1. An intelligent fire elevator data acquisition method based on intermittent classification is characterized by: The method comprises the following steps: a vibration acquisition module is included in the data acquisition component of the intelligent fire elevator, and acceleration is acquired through the vibration acquisition module; a basic intermittent segment is preset, and an intermittent series is established based on the basic intermittent segment; a loss reference effective state is obtained through decoupling analysis of acceleration and intermittent series; a performance loss warning factor is determined for the loss reference effective state; and data is collected from the fire elevator in combination with the performance loss warning factor. The method for determining the performance loss warning factor of the loss reference state is as follows: obtain the benchmark node through the basic intermittent segment, generate the loss reference state sequence from the benchmark node, and use its mean and standard deviation to classify low-order loss points and high-order loss points; construct a low-loss binary group from the low-order loss point and a smooth curve; obtain the loss difference from the Euclidean distance between the high-order loss point and the curve value, and generate a difference evaluation sequence from the loss difference; use the quartiles of the difference evaluation sequence to divide the loss stable interval, and obtain the warning factor by judging the low-order loss point or the loss stable interval through the benchmark node.

2. The intelligent fire elevator data acquisition method based on intermittent classification according to claim 1 is characterized in that: The method for obtaining acceleration through the vibration acquisition module is as follows: the vibration acquisition module is arranged in the motor cabin of the elevator; The vibration acquisition module includes an acceleration sensor, which is a piezoelectric acceleration sensor, an optical fiber acceleration sensor or a MEMS acceleration sensor. The acceleration is obtained by real-time measurement using an acceleration sensor.

3. The intelligent fire elevator data acquisition method based on intermittent grading according to claim 1 is characterized in that: The method of presetting the basic interval segment and establishing the interval progression based on the basic interval segment is as follows: setting the basic interval segment to 1-10 minutes; An integer variable is preset as the intermittent series KL, and the value range of the intermittent series is KL∈[1,15]. For any intermittent series KL, the intermittent segment it represents is the time period of KL basic intermittent segments.

4. The intelligent fire elevator data acquisition method based on intermittent grading according to claim 1 is characterized in that: The method for obtaining the loss reference effective state based on the decoupling analysis of acceleration and intermittent series is as follows: at any moment, the basic intermittent segment belonging to that moment is the current basic intermittent segment, and any intermittent series from the current basic intermittent segment is selected as the operation level. The range of the loss characteristic data corresponding to the operation level in the intermittent segment is calculated, and the ratio of the obtained range value to the maximum value of each acceleration modulus in the current basic intermittent segment is recorded as the loss reference effective state of the operation level. The loss characteristic data is the power value of the motor when the elevator is in a uniform motion state.

5. The intelligent fire elevator data acquisition method based on intermittent grading according to claim 1 is characterized in that: The method for determining the performance loss warning factor for the loss reference effectiveness state is as follows: presetting a monitoring period TCr, and recording each basic intermittent segment as a reference node within the monitoring period; Obtain the loss reference state of each benchmark node and form a sequence recorded as Wsf.Ls; Calculate the mean and standard deviation of the elements in the sequence Wsf.Ls. If the element is less than or equal to the sum of the mean and three times the standard deviation, and greater than or equal to the difference between the mean and three times the standard deviation, then record the corresponding reference node as a low-order loss point; otherwise, record it as a high-order loss point. The time scale of the loss reference state obtained from the low-order loss point and its corresponding loss reference state constitute a low-loss binary group. The cubic spline interpolation of the set constructed by the low-loss binary group within the monitoring period is performed using the plrep and splev functions to obtain a smooth curve. The absolute value of the difference between the curve value of any high-order loss point and its corresponding loss reference state is recorded as the loss difference; the loss difference of each high-order loss point is obtained and a difference evaluation sequence is constructed; the upper quartile Xes.sl, lower quartile Xes.xl, and interquartile range Xes.jl in the difference evaluation sequence are obtained respectively; the loss stationary interval is constructed based on the difference evaluation sequence as (Xes.xl-k2×Xes.jl, Xes.sl+k2×Xes.jl); where k2 is the preset amplitude coefficient; If the current benchmark node belongs to a marked low-loss point, the performance loss warning factor is assigned a value of FALSE; Otherwise, if the current benchmark node belongs to the loss stable interval, the performance loss warning factor is assigned a value of FALSE, otherwise the performance loss warning factor is assigned a value of TRUE.

6. The intelligent fire elevator data acquisition method based on intermittent grading according to claim 1 is characterized in that: The method for determining the performance loss warning factor for the loss reference effect state can be replaced by: presetting a monitoring period TCr, and recording each basic intermittent segment as a reference node within the monitoring period; Obtain the loss reference status of each benchmark node during the monitoring period and form a benchmark loss status sequence; The reference nodes corresponding to each maximum value element in the reference loss state sequence are recorded as adjacent nodes; Calculate the difference between the loss reference effective state of any adjacent node and its left and right adjacent reference nodes, and select the maximum value of the difference as the loss critical value of the adjacent node; The number of adjacent nodes is recorded as sdn; the time interval between any adjacent node and its previous adjacent node is recorded as the loss trend value, and the sequence of all loss trend values ​​is recorded as Gn.Ls; The loss trend value of each adjacent node is clustered using the DBSCAN density clustering algorithm to divide all adjacent nodes into several clusters, which are recorded as trend clusters; The early warning assessment level of each trend cluster is calculated based on the edge nodes and loss critical value. The early warning assessment level of each trend cluster is normalized to the minimum and maximum value and recorded as the loss risk status value. If the loss risk status value of a trend cluster is greater than or equal to 0.5, the trend cluster is marked as a risk type cluster; otherwise, it is marked as a non-risk type cluster. The maximum value of the loss critical value in the trend cluster is recorded as the critical peak value of the trend cluster. The Euclidean distance between the current benchmark node and the critical peak value of any trend cluster is recorded as the current critical distance. The trend cluster with the minimum previous critical distance is selected as the target cluster. If the target cluster is a risk type cluster, the performance loss warning factor of the current benchmark node is assigned to TRUE, otherwise the performance loss warning factor is assigned to FALSE.

7. The intelligent fire elevator data collection method based on intermittent classification according to claim 1 is characterized in that: The method of collecting data from fire elevators in combination with the performance loss warning factor is as follows: for any basic intermittent segment, if the judgment result of the performance loss warning factor corresponding to more than half of the intermittent levels is true, then the risk of performance loss occurring in the basic intermittent segment is defined; When there is a risk of performance loss, the interval level with the maximum value of the loss reference effect state is selected as the storage target interval level. The interval segment corresponding to the storage target interval level is recorded as the target interval segment, and the environmental data within the target interval segment is stored in the cloud.

8. Intelligent fire elevator data acquisition system based on intermittent classification, characterized by: The intelligent fire elevator data acquisition system based on intermittent grading includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the intelligent fire elevator data acquisition method based on intermittent grading according to any one of claims 1 to 7 are implemented. The intelligent fire elevator data acquisition system based on intermittent grading runs on a desktop computer, a laptop computer, a PDA, and a computing device in a cloud data center.

Images