Intelligent railway fastener based on delta F judgment and burr suppression

Abstract

The invention relates to the field of railway vehicle fastener tension monitoring, in particular to a railway intelligent fastener based on delta F judgment and burr suppression, and aims to improve the monitoring precision of the railway intelligent fastener. Comprising a tension sensor and a central processing unit, the central processing unit comprises a state recognition module and a self-adaptive filtering module, and the state recognition module is connected with the tension sensor and the self-adaptive filtering module. The state recognition module collects a tension signal output by the tension sensor at a fixed period as a sampling value, calculates a difference value between the sampling value at the current moment and the sampling value at the previous moment, and sends the tension signal to the tension sensor when the difference value is larger than a first set threshold value and smaller than or equal to the first set threshold value. And respectively judging whether the railway intelligent fastener is in an installation state or an operation state. And the adaptive filtering module receives a state judgment result, and performs smoothing processing on the tension signal by adopting an exponential weighted moving average filtering algorithm and a moving average filtering algorithm respectively in an installation state and an operation state.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of rail vehicle fastener tension monitoring, in particular to a railway intelligent fastener based on delta F judgment and burr suppression. BACKGROUND

[0002] In the development process of railway transportation safety and intelligence, as a key stress monitoring device, the railway intelligent fastener has been widely used in the connection state monitoring of the running parts of freight cars, and can collect the tension change data in the installation and running process in real time, providing basis for structure loosening judgment, risk warning and maintenance scheduling, which is of great significance to improve the safety and operation efficiency of railway transportation.

[0003] The railway intelligent fastener has two working stages of installation and running, and the signal characteristics are significantly different: the tension value rises rapidly in the installation stage, showing steep and stepwise fluctuation; the tension change is gentle in the running stage, showing trend attenuation or stable fluctuation, and is easily disturbed by environmental noise, mechanical vibration and the like. At present, some commercial schemes adopt high-order algorithms such as Kalman filter, LMS adaptive filter and wavelet transform, which can theoretically consider both smoothing and response, but require high computational operations such as matrix operation and multi-level transformation, which are not suitable for small embedded platforms such as STM32 and MSP430. The existing filtering technology cannot meet the needs of the two stages, and the distortion rate is serious, resulting in a high misjudgment rate of the working state of the railway intelligent fastener. SUMMARY

[0004] The purpose of the present application is to provide a railway intelligent fastener based on delta F judgment and burr suppression, which aims to improve the monitoring accuracy of the railway intelligent fastener.

[0005] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows: The present application provides a railway intelligent fastener based on delta F judgment and burr suppression, which comprises a tension sensor and a central processing unit, the central processing unit comprising a state recognition module and an adaptive filtering module, the state recognition module being connected with the tension sensor and the adaptive filtering module respectively; the state recognition module is configured to collect the tension signal output by the tension sensor as a sampling value at a fixed period, and calculate the difference between the sampling value at the current time and the sampling value at the previous time, in the case that the difference is greater than a first set threshold, it is determined that the railway intelligent fastener is in the installation state, and in the case that the difference is less than or equal to the first set threshold, it is determined that the railway intelligent fastener is in the running state; the adaptive filtering module is configured to receive the state judgment result output by the state recognition module, and when the railway intelligent fastener is in the installation state, the exponential weighted moving average filtering algorithm is adopted to smooth the tension signal, and when the railway intelligent fastener is in the running state, the sliding average filtering algorithm is adopted to filter the tension signal.

[0006] The first set threshold is 300N-800N.

[0007] The weight coefficient of the exponential weighted moving average filtering algorithm is 0.4-0.6, and the window length of the sliding average filtering algorithm is 50-100 points.

[0008] The central processing unit further comprises a burr identification module connected with the tension sensor and the adaptive filtering module; the burr identification module is configured to: acquire the sampling value at the current moment, and calculate the difference between the sampling value at the current moment and the median of the plurality of sampling values before the current moment, determine that the sampling value at the current moment is a burr signal in the case that the difference is greater than or equal to a second set threshold, and control the adaptive filtering module to freeze the output and retain the output result before the current moment.

[0009] The sliding window length adopted by the burr identification module is 15, and the second set threshold is 100N.

[0010] The railway intelligent fastener further comprises a wireless communication module connected with the adaptive filtering module; the wireless communication module is configured to: be connected with the upper computer, receive the control signal sent by the upper computer, and transmit the control signal to the adaptive filtering module; wherein the control signal comprises a first control signal and a second control signal; the adaptive filtering module is configured to: switch to the exponential weighted moving average filtering algorithm to perform smoothing processing on the tension signal when the first control signal is received, and switch to the sliding average filtering algorithm to perform filtering processing on the tension signal when the second control signal is received.

[0011] The priority of the control signal received by the adaptive filtering module is higher than the state determination result output by the state identification module.

[0012] The wireless communication module supports at least one of Bluetooth, Wi-Fi, 4G, 5G, and data radio.

[0013] The railway intelligent fastener further comprises an analog-to-digital converter, an RTC clock, a power management module, a magnetic charging head, a charging management chip, a voltage conversion chip, and a boost circuit; the tension sensor is connected with the analog-to-digital converter, the analog-to-digital converter, the RTC clock, and the wireless communication module are connected with the central processing unit; the magnetic charging head is connected with the power management module through the charging management chip and the voltage conversion chip in sequence, and the boost circuit is connected with the power management module.

[0014] The railway intelligent fastener further comprises an output module, the output module comprising an LED lamp and / or a loudspeaker; different colors of the LED lamp indicate different states of the railway intelligent fastener; the loudspeaker outputs audio information for indicating the state of the railway intelligent fastener at the current moment.

[0015] Compared with the prior art, the railway intelligent fastener has the beneficial effects that: 1. The railway intelligent fastener based on AF determination and burr suppression provided in the embodiments of the application distinguishes between installation and running states accurately through comparison of the difference AF between the sampling value at the current time and the sampling value at the previous time with the first set threshold value, and switches the exponential moving average filtering algorithm (EMA) and the moving average filtering algorithm (MA) correspondingly, taking into account the dynamic response requirement of fast change of tension during installation and the signal smoothness requirement during running, effectively reducing the distortion problem of a single filtering algorithm under different working conditions, and significantly improving the tension signal monitoring accuracy.

[0016] 2. The burr identification module identifies abnormal burr signals through the sliding window median method, freezes the filtering output and retains the historical results when it is determined to be a burr, avoids the influence of transient interference on monitoring data, ensures signal stability, and reduces the risk of misjudgment caused by noise. The wireless communication module supports interaction with the host computer, can remotely issue control signals to force switching of the filtering mode, and the control signal priority is higher than the automatic state determination result, which facilitates coping with atypical working conditions or manual intervention scenarios, and improves system adaptability and operation convenience. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 is a structural schematic diagram of a railway intelligent fastener provided in the embodiments of the application; Figure 2 is an installation state simulation schematic diagram provided in the embodiments of the application; Figure 3 is an installation state simulation waveform comparison schematic diagram provided in the embodiments of the application; Figure 4 is a running state simulation schematic diagram provided in the embodiments of the application; Figure 5 is a running state simulation waveform comparison schematic diagram provided in the embodiments of the application. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0019] The embodiments of the application provide a railway intelligent fastener based on AF determination and burr suppression, which comprises a tension sensor and a central processing unit, the central processing unit comprises a state identification module and an adaptive filtering module, and the state identification module is connected with the tension sensor and the adaptive filtering module respectively.

[0020] The state recognition module is configured to collect the tension signal F(t) output by the tension sensor as a sampling value at a fixed period T, and calculate the difference value AF between the sampling value at the current moment and the sampling value at the previous moment, that is, AF = F(t) - F(t-T), and determine that the railway intelligent fastener is in the installation state when the difference value is greater than a first set threshold Fth, that is, the tension of the railway intelligent fastener presents a rapid rising characteristic in a short time. In the case where the difference value is less than or equal to the first set threshold Fth, it is determined that the railway intelligent fastener is in the running state, that is, the tension of the railway intelligent fastener presents a slow decay or small amplitude fluctuation characteristic. Exemplarily, the first set threshold Fth is 300N-800N.

[0021] The adaptive filtering module is configured to receive the state determination result output by the state recognition module, and to perform smoothing processing on the tension signal using an exponential weighted moving average filtering algorithm when the railway intelligent fastener is in the installation state, and to perform filtering processing on the tension signal using a sliding average filtering algorithm when the railway intelligent fastener is in the running state.

[0022] It should be understood that the EMA filter and the MA filter are included in the adaptive filtering module, and when AF is greater than Fth, the EMA filter is automatically switched to, and when AF is less than or equal to Fth, the MA filter is automatically switched to.

[0023] It should be understood that the connection mentioned in the embodiments of the present application includes direct connection and indirect connection, for example, the adaptive filtering module and the state recognition module are directly connected, and the state recognition module and the tension sensor are indirectly connected.

[0024] For the exponential weighted moving average filtering algorithm, it is mainly used to process the rapid rising process of the tension of the railway intelligent fastener in the installation stage. Since the tightening torque changes sharply in a few seconds, the tension value presents a stepwise and discontinuous jump rising trend, and the traditional average filter has a lagging response and is difficult to capture the change peak value. The EMA algorithm gives a larger weight to the new sampling point, and the weight of the old data decays according to an exponential law, so as to realize a rapid response to the signal change. Its recursive formula is: wherein a is a weight coefficient, is the current sampling value. This method can effectively improve the filtering sensitivity in the tension sudden change process, accurately describe the stress process in the installation stage, and is helpful for installation quality analysis and control optimization.

[0025] For the moving average filtering algorithm, it is mainly used to process the scene of the railway intelligent fastener in the running stage. In the running stage, the tension signal changes slowly, but there are small noise fluctuations caused by environmental disturbances, temperature fluctuations, track vibrations and other factors. The MA algorithm effectively smooths the micro-disturbance waveform and suppresses high-frequency noise by performing arithmetic averaging on historical sampling values within a time window. As a possible implementation, a fixed window length is set, for example N=70, the latest collected value is added to the window every time the sampling is performed, and the oldest value is removed, and the average of all data in the current window is taken as the output result. This way can truly reflect the stress trend of the railway intelligent fastener in the stable working state, and provide reliable data basis for daily state monitoring, maintenance strategy formulation and life prediction.

[0026] For example, the weight coefficient of the exponential weighted moving average filtering algorithm is 0.4 to 0.6; the window length of the moving average filtering algorithm is 50 to 100 points.

[0027] In the installation stage, the system selects the exponential weighted moving average filtering (EMA) algorithm. This algorithm gives a higher weight proportion to the latest collected tension sampling value (for example, the weight coefficient a takes the value 0.4), and as the time sequence moves, the weight of the historical sampling data decays by an exponential law step by step. This feature enables the filter to respond quickly to the sudden change of tension signal during installation, accurately track the instantaneous fluctuations and step changes of force value in the tightening operation, and thus provide high real-time and high-accuracy force value monitoring data for quality control of the installation process, ensuring that every subtle change in force value during the installation process can be effectively captured.

[0028] In the running stage, the system will automatically switch to the moving average filtering (MA) algorithm. This algorithm uses a relatively long sliding window (for example, the window length N is set to 50) to perform arithmetic mean operation on the continuously collected tension signal within the sliding interval. Through this processing method, it can effectively smooth the small disturbance signals caused by environmental vibration, temperature drift, mechanical noise and other factors, significantly reduce the interference of random noise on the monitoring data, and greatly improve the stability and consistency of the tension data in the running state, providing continuous and reliable basic data input for long-term state monitoring, performance degradation trend analysis and service life evaluation of the railway intelligent fastener.

[0029] The central processing unit further comprises a glitch identification module, which is connected with the tension sensor and the adaptive filtering module respectively; the glitch identification module is configured to: obtain a sampling value at a current time, and calculate a difference between the sampling value at the current time and a median value of a plurality of sampling values before the current time, and determine that the sampling value at the current time is a glitch signal in a case where the difference is greater than or equal to a second set threshold, and control the adaptive filtering module to freeze the output and retain the output result before the current time.

[0030] For the abnormal data points identified as burrs, the system triggers a protection mechanism: the data points will be excluded from the parameter update process of the filter, while the filter output end keeps the valid output result of the last time. This design can effectively block the transmission of transient interference signals to the subsequent data processing module, avoid misjudging burrs as effective tension change signals and causing state monitoring logic disorder, thereby ensuring the continuity and reliability of the filter output, and further improving the anti-interference ability of the entire tension monitoring system under complex working conditions.

[0031] For example, the burr identification module adopts a sliding window length of 15, and the second set threshold is 100N.

[0032] The railway intelligent fastener provided by the embodiments of the application integrates a short-time sliding window median detection mechanism, which is specially used for identifying potential transient mutation interference, i.e., burrs, in a tension signal. The technical implementation logic is as follows: in each signal sampling period, the system dynamically constructs a sliding window (for example, the window length is set to 5 sampling points) containing a plurality of historical sampling values before the current sampling point, sorts the historical data in the window, extracts the median value, calculates the deviation between the current sampling value and the median value, and determines that the current sampling value is an abnormal burr signal when the absolute value of the deviation exceeds a preset threshold, for example, 100N. The current sampling value is not involved in the update of EMA or MA, and the system output keeps the normal result of the last time. This mechanism can effectively avoid the mis-triggering of high-frequency pulses or hardware jitter to subsequent logic, and ensure the signal continuity and system stability.

[0033] The railway intelligent fastener further includes a wireless communication module connected with the adaptive filtering module; the wireless communication module is configured to be connected with the upper computer, receive a control signal sent by the upper computer, and transmit the control signal to the adaptive filtering module; wherein the control signal includes a first control signal and a second control signal; the adaptive filtering module is configured to switch to an exponential weighted moving average filtering algorithm to perform smoothing processing on the tension signal when the first control signal is received, and switch to a sliding average filtering algorithm to perform filtering processing on the tension signal when the second control signal is received.

[0034] The priority of the control signal received by the adaptive filtering module is higher than the state determination result output by the state identification module. The priority of the control signal received by the adaptive filtering module is higher than the state determination result output by the state identification module, which means that the control instruction sent by the upper computer can be executed preferentially. When the upper computer sends a control signal according to the global working condition planning or specific task demand, it can not be limited by the local determination result of the state identification module, and directly switch the filtering algorithm, ensuring the rapid landing of the key filtering strategy.

[0035] With the connection of the wireless communication module and the upper computer, different control signals (first control signal or second control signal) can be sent by the upper computer according to the actual working condition requirements, and the filtering algorithm of the adaptive filtering module can be flexibly switched. For example, when the environment near the railway track is relatively frequent and unstable, the first control signal is sent to switch to the exponential weighted moving average filtering algorithm, which gives higher weight to recent data and can respond more quickly to the dynamic changes of the tension signal, effectively smoothing the interference caused by instantaneous vibration; when the environment is relatively stable and needs to be uniformly smoothed to eliminate periodic small amplitude noise, the second control signal is sent to switch to the sliding average filtering algorithm, which can calculate the average value of the data in a certain window to stably filter out noise and make the tension signal more in line with the actual situation. The processing of the tension signal can adapt to different working scenes, avoiding the problem that a single filtering algorithm has poor processing effect in complex working conditions, and improving the pertinence of signal processing.

[0036] The wireless communication module supports at least one of Bluetooth, Wi-Fi, 4G, 5G, and data radio.

[0037] The railway intelligent fastener further comprises an analog-to-digital converter, an RTC clock, a power management module, a magnetic charging head, a charging management chip, a voltage conversion chip, and a boost circuit; the tension sensor is connected with the analog-to-digital converter, the analog-to-digital converter, the RTC clock, and the wireless communication module are connected with the central processing unit; the magnetic charging head is connected with the power management module through the charging management chip and the voltage conversion chip in sequence, and the boost circuit is connected with the power management module.

[0038] The railway intelligent fastener further comprises an output module, which comprises an LED lamp and / or a loudspeaker; different colors of the LED lamp indicate different states of the railway intelligent fastener; the loudspeaker outputs audio information for indicating the state of the railway intelligent fastener at the current time.

[0039] The railway intelligent fastener provided by the embodiment of the present application supports Over-The-Air Parameterization (OTAP), which allows relevant parameters of the railway intelligent fastener, such as sampling period, state switching threshold, and filter parameters, to be remotely issued and dynamically updated through a wireless communication mode, without the need for firmware recompilation or on-site manual intervention, thereby improving the maintainability and intelligent level of the system.

[0040] The embodiment of the present application also provides a comparative simulation experiment to verify the effectiveness and superiority of the adaptive combined filtering algorithm (EMA+MA), and exemplary verification is performed in the MATLAB / Simulink environment for two typical working conditions of “installation state” and “running state”: Exemplarily, as Figure 2As shown, for the comparative simulation in the installation state, the combined tension curve of step, pulse and superimposed high-frequency burr is taken as the signal input of the simulation model to simulate multiple disturbances such as torque wrench rapid tightening, assembly impact, operator pulse and the like. Comparative filter: the adaptive filtering algorithm provided in the present application automatically switches to EMA when “ΔF> threshold”; the control scheme adopts fixed window moving average filtering (MA, N=70). The simulation results obtained are as follows: The EMA curve can quickly capture the peak value of each step and pulse, and the output lag time is less than one sampling period. The MA curve is significantly smoothed at the same impact point, and the peak value is seriously attenuated, which cannot truly reflect the instantaneous tension. According to Figure 3 It can be seen that the original noise signal curve (dotted line): the curve shows a stepwise sharp rise of F(t) at the initial stage of installation, and is accompanied by high-frequency burr interference. The fixed window MA filtering curve (dotted line): the sliding average filtering with N=70 can weaken the high-frequency noise, but the smoothing effect at the signal mutation point causes the peak value to lag, which cannot accurately represent the tension value in the installation state. The adaptive filtering algorithm curve provided in the embodiments of the present application (solid line): after the installation state is determined as ΔF>Fth, the EMA mode (α=0.4) is quickly switched to, the EMA output closely follows the rising trend of the original signal, the peak value response to the step change is less than 1 sampling period, and the output is frozen when the burr is detected, avoiding the influence of abnormal data on the subsequent filtering state.

[0041] For example, as Figure 4 As shown, for the comparative simulation in the running state, the slowly decaying trend superimposed with white noise, random pulse interference and high-frequency vibration is taken as the signal input of the simulation model to simulate the tension curve in the running state of the train. Comparative filter: the adaptive filtering algorithm provided in the present application switches to MA when “ΔF≤ threshold”, and the control scheme adopts fixed weight EMA (α=0.4). The simulation results obtained are as follows: According to Figure 5 It can be seen that the original noise signal curve (dotted line): in the running stage, F(t) shows a slowly decaying trend, but is superimposed with random high-frequency noise and occasional pulse interference of about ±30N; the fixed weight EMA filtering curve (dotted line): the fixed EMA filtering with α=0.15 can better track the overall decaying trend of the signal, but the high-frequency noise suppression is insufficient. The adaptive filtering algorithm curve provided in the embodiments of the present application (solid line): when ΔF≤Fth, the MA filtering is automatically adopted, which greatly suppresses the high-frequency noise; when the abnormal point is detected due to the burr detection module, the output is frozen; compared with the fixed MA, the trend curve can be quickly restored when the occasional medium-frequency pulse occurs, and no step caused by algorithm switching is generated.

[0042] It can be seen from the simulation experiment that the adaptive filter in the EMA mode has obvious advantages in rapid and complex installation tension changes, and can ensure the high-precision response required for quality monitoring and safety evaluation in the installation process. The adaptive filtering algorithm provided in the embodiments of the application considers the rapid response of EMA to the trend and the efficient suppression of MA to the noise in the running stage, and through dynamic parameter adjustment and burr protection, a lower noise floor and a faster abnormal recovery capability are realized, which is better than the single EMA or MA filtering effect.

[0043] In the description of the present specification, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0044] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A railway intelligent tightener based on ΔF judgment and burr suppression, characterized in that: The railway intelligent fastener comprises a tension sensor and a central processing unit, the central processing unit comprises a state identification module and an adaptive filtering module, and the state identification module is connected with the tension sensor and the adaptive filtering module respectively; the state identification module is configured to collect a tension signal output by the tension sensor as a sampling value at a fixed period, calculate a difference value between the sampling value at a current moment and a sampling value at a previous moment, determine that the railway intelligent fastener is in an installation state when the difference value is greater than a first set threshold, and determine that the railway intelligent fastener is in a running state when the difference value is less than or equal to the first set threshold; and the adaptive filtering module is configured to receive a state determination result output by the state identification module, perform smoothing processing on the tension signal by using an exponential weighted moving average filtering algorithm when the railway intelligent fastener is in the installation state, and perform filtering processing on the tension signal by using a sliding average filtering algorithm when the railway intelligent fastener is in the running state.

2. A smart rail fastener based on delta F decision and burr suppression as claimed in claim 1, wherein: The first set threshold is 300 N-800 N.

3. A smart rail fastener based on delta F decision and burr suppression as claimed in claim 1, wherein: The weight coefficient of the exponential weighted moving average filtering algorithm is 0.4-0.6, and the window length of the sliding average filtering algorithm is 50-100 points.

4. A smart rail fastener based on delta F decision and burr suppression as claimed in claim 1, wherein: The central processing unit further comprises a burr identification module, the burr identification module is connected with the tension sensor and the adaptive filtering module respectively, the burr identification module is configured to obtain the sampling value at the current moment, calculate a difference value between the sampling value at the current moment and a median value of a plurality of sampling values before the current moment, determine that the sampling value at the current moment is a burr signal when the difference value is greater than or equal to a second set threshold, and control the adaptive filtering module to freeze output and retain an output result before the current moment.

5. A smart rail fastener based on delta F decision and burr suppression as claimed in claim 4, wherein: The sliding window length adopted by the burr identification module is 15, and the second set threshold is 100 N.

6. A smart rail fastener based on delta F decision and burr suppression as claimed in claim 1, wherein: The railway intelligent fastener further comprises a wireless communication module, the wireless communication module is connected with the adaptive filtering module, the wireless communication module is configured to be connected with an upper computer, receive a control signal sent by the upper computer, and transmit the control signal to the adaptive filtering module, wherein the control signal comprises a first control signal and a second control signal, and the adaptive filtering module is configured to switch to the exponential weighted moving average filtering algorithm to perform smoothing processing on the tension signal when the first control signal is received, and switch to the sliding average filtering algorithm to perform filtering processing on the tension signal when the second control signal is received.

7. A smart rail fastener based on delta F decision and burr suppression as claimed in claim 6, wherein: The priority of the control signal received by the adaptive filtering module is higher than that of the state determination result output by the state identification module.

8. A smart rail fastener based on delta F decision and burr suppression as claimed in claim 6, wherein: The wireless communication module supports at least one of Bluetooth, Wi-Fi, 4G, 5G and data transmission radio.

9. A smart rail fastener based on delta F decision and burr suppression as claimed in claim 1, wherein: The railway intelligent fastener further comprises an analog-digital converter, an RTC clock, a power management module, a magnetic charging head, a charging management chip, a voltage conversion chip and a boost circuit; the tension sensor is connected with the analog-digital converter, the analog-digital converter, the RTC clock and the wireless communication module are connected with the central processing unit respectively; the magnetic charging head is connected with the power management module through the charging management chip and the voltage conversion chip in sequence, and the boost circuit is connected with the power management module.

10. A smart rail fastener based on delta F decision and burr suppression as claimed in claim 1, wherein: The railway intelligent fastener further comprises an output module, the output module comprising an LED lamp and / or a horn; different colors of the LED lamp indicating different states of the railway intelligent fastener; the horn outputting audio information for indicating the state of the railway intelligent fastener at the current time.

Images