Relay contact bounce parameter identification method based on state machine and dynamic threshold

Abstract

This invention relates to the field of electrical variable measurement, specifically providing a method for identifying relay contact bounce parameters based on a state machine and dynamic thresholds. The method uses the coil voltage rising edge to lock the time reference point and simultaneously acquires contact voltage and current. It locates the initial closing point and stable point of the contact and constructs a dual-constraint dynamic threshold based on baseline statistics and stable region statistics. Within the detection interval defined by the closing point and stable point, a finite state machine combined with the minimum number of sustained points is used to segment and confirm the validity of bounce events. For each valid event, the contact voltage and current are extracted and discretely integrated to obtain the single bounce energy, and the number of bounces, total energy, maximum single energy, and energy decay characteristics are statistically output. This invention achieves robust bounce event detection through synchronous triggering and dynamic adaptive thresholds, and provides multi-dimensional quantitative parameters based on a finite state machine and energy integration, improving the automation, accuracy, and evaluation depth of relay contact bounce testing.

Description

Technical Field

[0001] This invention relates to the field of electrical variable measurement, and specifically to a method for identifying relay contact bounce parameters based on a state machine and dynamic threshold. Background Technology

[0002] When relay contacts close, due to contact elasticity and armature inertia, the contacts may experience multiple brief separations and re-contacts after the initial contact, forming a bounce process. Bounce can cause transient interruption / conduction, which may induce arcing and surges under inductive or capacitive loads, thereby affecting contact wear and electrical life assessment.

[0003] Existing bounce testing and statistical schemes can be generally categorized into the following types, and their limitations mean that there is still considerable room for improvement in the automatic identification of bounce parameters.

[0004] A) Fixed threshold or single threshold zero-crossing / over-limit counting methods. These methods typically determine and count bounces based on whether the contact voltage exceeds a fixed threshold. Their drawbacks are: significant differences in steady-state voltage drop and noise levels exist between different relay models and under different load / power supply conditions; fixed thresholds are prone to issues such as "threshold too high, missing small bounces" or "threshold too low, misjudging normal fluctuations / glitches as bounces"; and the threshold requires manual adjustment.

[0005] B) Methods that only measure bounce time or rely on manual oscilloscope readings. These methods can often provide the bounce duration, but lack quantitative support for the number of bounces, event segment consistency, and bounce severity; moreover, manual readings or single-trigger tests are not conducive to batch testing and consistency evaluation.

[0006] C) Hardware-based jump detection / counting implementation. One approach uses circuitry to detect the number of voltage jumps at the contact points to achieve jump counting. While simple to implement, the threshold depends on hardware settings and is difficult to adapt to changing operating conditions. Furthermore, it typically only outputs the number of jumps / times, lacking further characterization of the "destructive" nature of the jumps, such as energy levels. Summary of the Invention

[0007] To address the aforementioned issues, this invention provides a relay contact bounce parameter identification method based on state machines and dynamic thresholds. Robust bounce event detection is achieved through synchronous triggering and dynamic adaptive thresholds, and multi-dimensional quantization parameters are provided based on finite state machines and energy integrals, thereby improving the automation, accuracy, and evaluation depth of relay contact bounce testing.

[0008] The technical solution of this invention provides a method for identifying relay contact bounce parameters based on state machine and dynamic threshold, comprising the following steps: Using the rising edge of the relay coil voltage as the trigger source to lock a unified time reference point. ,based on Alignment is performed, and the voltage signals across the contacts and the current signals flowing through the contacts are acquired simultaneously to obtain the contact voltage sequence. With contact current sequence ; Based on contact voltage sequence Determine the first closing point of the contact. Stable point where the contact enters a stable conducting state ; based on Baseline statistics for calculating contact voltages from previous consecutive sampling points ,as well as Calculate the stable region statistics of the contact voltage at several subsequent consecutive sampling points. According to the formula Calculate the adaptive dynamic bounce threshold ,in To stabilize the additional quantity, This is the baseline scaling factor; In the and Within the defined detection range, a dynamic bounce threshold is used. As a criterion, a finite state machine is used to analyze the contact voltage sequence. Perform event detection to identify events where the number of persistent events is not less than a preset threshold. A valid bounce event; For each identified valid bounce event, extract the contact voltage sequence from its corresponding start and end time. With contact current sequence Calculate the energy value of the corresponding data segment for this bounce. ; Based on all valid bounce events, output the bounce count and energy value. Energy parameters obtained from statistics.

[0009] As can be seen from the above technical solutions, this application has the following advantages: 1) By using baseline statistics Statistics of the stable region Constructing a dual-constraint dynamic threshold This allows the rebound judgment threshold to automatically adapt to different relay models, load conditions, and steady-state voltage drops, eliminating the need for manual setting of fixed thresholds. This effectively solves the problems of "missed detection due to excessively high thresholds" or "false judgment due to excessively low thresholds," and improves the robustness and generalization ability of the detection. 2) Lock a uniform time reference point using the rising edge of the coil voltage. This achieves synchronous acquisition and time alignment of multi-channel signals, providing a unified timing reference for batch testing; through a finite state machine... and Within the defined detection interval, bounce events are segmented, and the minimum number of duration points is used as the threshold. As a valid event confirmation condition, it realizes the objective and consistent segmentation and counting of back jump events, avoids the inefficiency and subjectivity of manual reading, and supports batch automated testing and consistency evaluation. 3) After identifying each valid bounce event, extract the corresponding contact voltage and contact current data segments, and calculate the energy of a single bounce. The automatic detection method quantifies the energy dimension of the bounce event; furthermore, by outputting energy parameters such as total bounce energy, maximum single energy, and energy sequence, it can more accurately characterize the "destructive" degree of the bounce process, providing richer quantitative basis for contact wear assessment, electrical life prediction, and anomaly diagnosis, and making up for the shortcomings of traditional methods that only output the number of times / time. Attached Figure Description

[0010] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic flowchart of a relay contact bounce parameter identification method based on a state machine and dynamic threshold provided in an embodiment of the present invention.

[0012] Figure 2 A schematic diagram of the system architecture for identifying relay contact bounce parameters.

[0013] Figure 3 This is a schematic diagram of the relay activation process and bounce detection timing.

[0014] Figure 4 This is a schematic diagram of the progressive closure point positioning process.

[0015] Figure 5 Dynamic bounce threshold Calculation process diagram.

[0016] Figure 6 This is a schematic diagram of a finite state machine with a threshold hysteresis mode for backtracking detection of state transitions.

[0017] Figure 7 This is a schematic diagram for calculating the energy integral of a single bounce.

[0018] Figure 8 This is a schematic diagram of the multidimensional statistics and display interface for the bounce parameters. Detailed Implementation

[0019] To make the purpose, features, and advantages of this application more apparent and understandable, specific embodiments and accompanying drawings will be used to clearly and completely describe the technical solution protected by this application. Obviously, the embodiments described below are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this application and in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0021] Figure 1 This is a flowchart illustrating a relay contact bounce parameter identification method based on a state machine and dynamic threshold, provided by an embodiment of the present invention. The order of steps in this flowchart can be changed, and some steps can be omitted, depending on different requirements.

[0022] like Figure 1 As shown, the method includes the following steps.

[0023] SS1 uses the rising edge of the relay coil voltage as the trigger source to lock the unified time reference point. ,based on Alignment is performed, and the voltage signals across the contacts and the current signals flowing through the contacts are acquired simultaneously to obtain the contact voltage sequence. With contact current sequence .

[0024] SS2, based on contact voltage sequence Determine the first closing point of the contact. Stable point where the contact enters a stable conducting state .

[0025] SS3, based on Baseline statistics for calculating contact voltages from previous consecutive sampling points ,as well as Calculate the stable region statistics of the contact voltage at several subsequent consecutive sampling points. According to the formula Calculate the adaptive dynamic bounce threshold ,in To stabilize the additional quantity, This is the baseline scaling factor.

[0026] SS4, in the and Within the defined detection range, a dynamic bounce threshold is used. As a criterion, a finite state machine is used to analyze the contact voltage sequence. Perform event detection to identify events where the number of persistent events is not less than a preset threshold. A valid bounce event.

[0027] SS5 extracts the contact voltage sequence from the corresponding start and end times for each identified valid bounce event. With contact current sequence Calculate the energy value of the corresponding data segment for this bounce. .

[0028] SS6, based on all valid bounce events, outputs the bounce count and energy value. Energy parameters obtained from statistics.

[0029] As a refinement and extension of the specific implementation of the above embodiments, in order to fully explain the specific implementation process of this embodiment, the following will provide possible embodiments to describe the specific implementation of the above steps in a non-limiting manner.

[0030] Figure 2 This is a schematic diagram of a relay contact return parameter identification system architecture. The relay contact return parameter identification method in this embodiment is implemented based on this system architecture. Figure 2 As shown, the system can be implemented using a layered architecture: the user interface layer is used for parameter configuration, waveform display, result display and export; the business logic layer implements closed-point positioning, threshold calculation, backlash event segmentation, energy integration and statistics; the device control layer is responsible for oscilloscope communication and acquisition configuration; the hardware layer includes oscilloscope, relay drive circuit, load circuit and signal conditioning circuit.

[0031] The oscilloscope is configured for four-channel synchronous acquisition: contact voltage → channel C1, coil voltage → channel C2 (trigger source), contact current → channel C3, coil current → channel C4; rising edge triggering is used with C2, the trigger level can be 50% of the coil's rated voltage, and the pre-trigger ratio can be 10%. The sampling rate is 12.5MHz, corresponding to a sampling interval dt=80ns, a storage depth of 1.25 million points, and an acquisition time of approximately 100ms.

[0032] Figure 3 This is a schematic diagram of the relay engagement process and bounce detection timing, as shown below. Figure 3 As shown, the bounce detection can be understood according to the following timing sequence: : Pre-trigger zone, the relay is not engaged, the contacts are open, and the contact voltage remains high (close to the power supply voltage). When the coil is energized, the rising edge of the coil voltage triggers and locks the circuit. ; : Delayed engagement, contacts remain open; When the contacts close for the first time and then bounce back, the contact voltage exhibits multiple rising / falling segments. The contacts enter a stable closed state, and the contact voltage stabilizes near a low level.

[0033] This embodiment first uses the rising edge of the relay coil voltage as the trigger source to lock a unified time reference point. ,based on Alignment is performed, and the voltage signals across the contacts and the current signals flowing through the contacts are acquired simultaneously to obtain the contact voltage sequence. With contact current sequence .

[0034] Specifically, the voltage signal across the relay coil is selected as the trigger source. In the test system, this signal is typically connected to a channel of an oscilloscope or data acquisition device via a voltage probe. The trigger type is set to rising edge trigger, and the trigger level is set to a preset percentage of the coil's rated voltage. When the coil voltage rises from a low level to this trigger level, the acquisition system is triggered and locks onto that moment.

[0035] The system records the current moment at the instant it is triggered, serving as a unified time reference point for the entire data acquisition process. , defined as the trigger moment / index at which the rising edge of the coil voltage is locked. All subsequently acquired channel signals are aligned with this point as the time zero. The locking mechanism ensures that each analysis has a consistent timeline starting point, even in multiple repeated tests or batch tests, improving the reliability of subsequent comparative analysis and statistics.

[0036] In lock Simultaneously or prior to this, the system has been configured with at least two acquisition channels for synchronous acquisition. Channel C1 is used to acquire the voltage signal across the contacts, i.e., the contact voltage. This signal reflects the transient voltage change during the contact's transition from open (high level) to closed (low level) and bounce-back process. Channel C3 is used to acquire the current signal flowing through the contact, i.e., the contact current. This signal is typically converted into a voltage signal using a current probe or sampling resistor for measurement.

[0037] Optionally, coil voltage (channel C2, i.e., the trigger source itself) and coil current (channel C4) can also be acquired simultaneously for more comprehensive analysis. The system ensures that all channels use the same sampling rate and time base to achieve sampling clock synchronization.

[0038] The system uses a preset high sampling rate in Discrete sampled values ​​from each channel are recorded synchronously within a certain time window before and after the event. Ultimately, this yields... Aligned contact voltage discrete sequence and contact current discrete sequence Both sequences contain N sampling points, and each point corresponds one-to-one with the others, with synchronized timestamps. The time window can include 10% of the pre-trigger time and 90% of the post-trigger time, with a total duration of approximately 100ms.

[0039] In one specific embodiment, a four-channel high-speed oscilloscope is used. Triggering is set to the rising edge of channel C2 (coil voltage), with a trigger level of 12V. Sampling is set to a sampling rate of 12.5MHz, a record length of 1.25M points, and a time window of 100ms (including a 10ms pre-trigger time). The channel configuration includes: C1: differential probe measuring contact voltage, range ±30V; C3: current probe measuring contact current, conversion factor 100mV / A; C2: voltage probe measuring coil voltage (trigger source); C4: current probe measuring coil current. This ultimately yields four time-aligned discrete sequences. ,in and For subsequent processing.

[0040] This embodiment uses the electrical starting point of coil operation (voltage rising edge) as the analysis starting point, avoiding the deviation caused by using unstable events such as contact voltage jumps as the starting point; the strict synchronous acquisition of voltage and current facilitates accurate calculation of instantaneous power. Energy integration is then performed to quantify the energy dimension.

[0041] Figure 4 A schematic diagram of the progressive closure point positioning process, such as... Figure 4 As shown, this embodiment determines the first contact closure point. Specifically, this includes the following steps SS211 to SS214.

[0042] SS211, with Based on the contact voltage sequence Define the preset search range above ,in and For relative to The sampling point offset.

[0043] To efficiently and accurately locate the initial closing point of the contact. First, a reasonable search range needs to be defined in the broad contact voltage waveform to avoid meaningless global scanning and improve algorithm efficiency and anti-interference ability.

[0044] With a unified time reference point As an absolute reference, in the contact voltage sequence A predefined search interval with a closed point is defined above, and this interval is defined by the indices of the start and end points: .

[0045] in, For relative to The starting offset number is used to skip the initial stage after the coil is energized, before the contacts can close, such as mechanical response delay and electromagnetic setup time. For relative to The endpoint offset number is used to limit the latest search time and avoid invalid searches in waveform segments after the contacts have already stabilized and closed.

[0046] and The value should be based on the typical timing characteristics of the relay and the system sampling rate.

[0047] starting point This should be set to a number of points slightly smaller than the minimum expected pull-in delay. For example, if the minimum pull-in delay of the relay under test is known to be approximately... Sampling interval Then it can be calculated:

[0048] This ensures that the search does not start prematurely, avoiding misjudging pre-trigger noise as a closing event.

[0049] end This should be set to a number greater than the sum of the longest expected pull-in delay and the bounce time, to allow sufficient margin for various slow or abnormal closure cases. For example, if the total expected pull-in delay and bounce time is approximately... ,but:

[0050] The preset search interval constrains the possible locations of the closure point within a reasonable time period, significantly reducing the amount of computation and avoiding misjudgments in irrelevant waveform segments, thereby improving the accuracy and reliability of positioning.

[0051] SS212 performs the first-level positioning, including: within the search interval, calculating the voltage drop index at each sampling point location using a sliding window, and taking the location with the largest voltage drop index exceeding a preset drop threshold as the initial contact closure point. .

[0052] Specifically, within the search interval, a sliding window is used to calculate each sampling position. Voltage drop index The calculation formula is:

[0053] In the formula, The length of the sliding window. This indicates the averaging operation. In one specific embodiment, =100 points.

[0054] Front half window Represents the "current point" Previously, length was The voltage segment should be at a high level before the contacts close. (Back half window) Represents the "current point" After that, the length is The voltage segment, at the contact closure point, should begin to enter a low level. Therefore, Reflects on location At this point, the decrease in average voltage in the latter part relative to the average voltage in the former part. A significant negative jump will produce a large positive jump. value.

[0055] Identify the decline indicator Largest position If the decline indicator at this position Greater than the preset reduction criterion threshold If so, the first-level positioning is considered successful. As the first point of contact closure .

[0056] Calculate all positions within the search interval of Afterwards, find the one who The position where the maximum value is obtained :

[0057] To confirm that the location indeed represents a valid closure event and not noise disturbance, the maximum reduction value needs to be compared with a preset reduction criterion threshold. If a comparison is made, If so, the first-level positioning is considered successful, and output is generated. .

[0058] threshold Used to filter out voltage fluctuations with excessively small amplitudes. Its value should be greater than the maximum drop value caused by system noise and normal fluctuations, but less than the drop value generated by a typical closed-circuit transition. For example, it can be taken as... .

[0059] This layer's localization directly targets the essential characteristics of a closed-loop event, namely a significant voltage drop. It exhibits extremely high localization accuracy and noise immunity when waveform quality is good. Its sliding window mean calculation effectively suppresses high-frequency noise, and the strategy of finding the point of maximum voltage drop is insensitive to the steepness of the transition edge; it can be identified as long as a significant voltage drop exists.

[0060] SS213, if the first-level positioning fails, then the second-level positioning is executed, including: calculating baseline statistics based on the contact voltages of several consecutive sampling points before the starting point of the search interval. Construct a relative threshold ,in The preset proportional coefficient is used to search for the first contact voltage below a certain value within the search range. The position is used as the first closing point of the contact. .

[0061] If the first-level positioning based on voltage drop characteristics fails, this level of positioning is initiated as a supplement. This level of positioning uses the stable open-circuit voltage (baseline) before the contacts close as a reference and sets a relative threshold to identify the first significant voltage drop.

[0062] The starting point of the search interval (i.e. Before that, a continuous sampling point region where the contacts are still reliably disconnected is selected, called the baseline calculation region. This region can be selected as follows: Pre-trigger zone: For example, taking a series of consecutive sampling points from before. forward This area is in a state where the coil is not energized, the contacts are definitely open, and the voltage is stable at a high level (close to the power supply voltage). Open zone before closure: If the pre-trigger zone is insufficient in length or subject to interference, a continuous segment before the start of the search interval, but as close as possible to the expected closure point, can also be used. The condition is that the voltage fluctuation in that segment meets the open circuit characteristics, such as the voltage value continuously exceeding a certain threshold.

[0063] Calculate the baseline statistics for all contact voltage samples within the selected baseline calculation area. In one embodiment, The arithmetic mean of the voltage in this area can be taken to represent the steady-state voltage level when the contacts are open.

[0064] Based on the calculated baseline statistics Construct a relative threshold : ;in, This is a preset proportionality coefficient, and its value range is usually within... Between, for example This coefficient determines the extent to which the threshold is lowered relative to the baseline voltage. The value of must balance sensitivity and noise immunity: a value that is too small (such as 0.2) may result in a threshold that is too low and is easily triggered by noise interference; a value that is too large (such as 0.8) may result in a threshold that is too high and cannot effectively capture the small voltage drop at the beginning of the closing.

[0065] From the starting point of the search interval ( Starting from this point, scan the contact voltage sequence point by point. The position index that first meets the following conditions. Identified as a candidate for the first closure point of the contact : This condition indicates that the contact voltage has dropped below a threshold determined proportionally to the open-circuit voltage for the first time, suggesting that the contacts may have begun to make physical contact and cause a voltage drop.

[0066] This layer's positioning does not rely on sharp voltage transitions, making it more adaptable to relay models with slow closing processes and gentle voltage drop edges. It achieves self-adaptation by comparing to its own baseline voltage, reducing dependence on external absolute thresholds. SS214, if the second-layer positioning also fails, then... Add a preset fixed offset number of points As the first point of contact closure .

[0067] As a robust safeguard, if both the first-layer (decline characteristic) and second-layer (relative threshold) localization fail, the fallback estimation strategy at this layer is activated to ensure that the algorithm outputs a reasonable value under any circumstances. The value allows the process to continue.

[0068] The first closing point of the contact Estimated as a unified time reference point Add a preset fixed offset number of points ,Right now .

[0069] Offset points The settings should be based on statistical or prior knowledge of typical relay pull-in delay times. Pull-in delay refers to the time from when the coil is energized (…). The typical time interval from the first mechanical contact of the contact point.

[0070] For example: if the typical pull-in delay time of a certain type of relay is known to be... And the system sampling interval Then it can be calculated:

[0071] This value should be set as a conservative estimate, typically slightly larger than the actual absorption delay of most normal samples, to ensure the estimation is accurate. It will not be too early, that is, before the actual closure point.

[0072] This step ensures that the algorithm can still run under abnormal conditions and provides usability. Estimate the timeline to avoid interrupting the entire testing process due to location failure. This may be inaccurate and could affect the precise starting position of subsequent bounce detection intervals. Therefore, setting a reasonable starting offset for the detection interval is necessary. To partially offset The impact of estimation errors may be considered, or the results of the fallback estimation may be used as a marker for the need for manual verification.

[0073] This embodiment employs a three-tiered progressive positioning approach—based on amplitude reduction characteristics, relative thresholds, and fallback estimation—to improve accurate identification while maintaining robustness to various waveform qualities and operating conditions, thus ensuring accurate initial contact closure. The positioning method has a high success rate and is practical.

[0074] This embodiment determines the stable point at which the contact enters a stable conducting state. Specifically, this includes the following steps SS221 to SS224.

[0075] SS221, in contact voltage sequence Above, at Then, a stable search interval is defined.

[0076] Determine the contact stability point The first step is to define a reasonable search range, i.e., a stable search interval. This interval should be located in time at the point of first contact closure. Then, it covers the entire bounce process and the subsequent stabilization phase.

[0077] Stable search intervals are typically based on a unified time reference point. Starting from this point, but to avoid invalid calculations during the absorbing delay phase, it is preferable to use... This will begin with a conservative starting point. In practice, the search starting point can be directly set to... itself or Then a small fixed offset, for example The search endpoint can be set to the end point of data collection, or a sufficiently late point in time that ensures a stable state has been reached, for example... point.

[0078] Limiting the search to this range excludes cases where the contact has not yet closed. ) and has long entered a period of extreme stability ( The use of irrelevant data improves computational efficiency and judgment accuracy.

[0079] SS222, within the stable search interval, uses a sliding window to advance segment by segment, and calculates the statistical characteristics of the contact voltage in each window, including the mean, variance and range.

[0080] To quantify the stability of contact voltage, this embodiment employs a sliding window method to segment and analyze the voltage sequence within the stable search interval. By calculating multiple statistical characteristics within each window, the voltage fluctuation during that period is evaluated from different dimensions.

[0081] Define a length of A sliding window (unit: number of sampling points). The window advances within the stable search interval in steps of a certain size (usually 1 point). Window length. The window should be much longer than the typical duration of a single bounce event to ensure that it covers multiple bounce cycles or a continuous period of stability, thus reflecting macroeconomic trends. At the same time, it should be much shorter than the total duration of the entire bounce process to ensure sufficient time resolution to locate the stable starting point.

[0082] For example, if the sampling interval The total bounce time is usually within a few milliseconds, which is acceptable. Point, corresponding .

[0083] For each window location, calculate the following three statistical characteristics of all contact voltage samples within the window: Window mean This reflects the average voltage level during that period. When the circuit is stably conducting, the contact voltage drop is very small. It should be close to 0V; Window variance This reflects the degree of dispersion of the voltage around the mean during that period. When stable, voltage fluctuations are small, and the variance should be close to 0. Window range This refers to the difference between the maximum and minimum values ​​within the window. When stable, the voltage is "clamped" at a low level, and the range should be very small.

[0084] SS223: When all statistical features within a sliding window meet the preset stability criteria, the time period corresponding to that window is determined to be in a stable conduction state.

[0085] A set of preset threshold values ​​is used. When a sliding window simultaneously meets all of the following conditions, it is determined that the window is in a stable conducting state: Mean condition: Window mean The absolute value is less than the preset stable voltage threshold. ,For example This condition ensures that the mean pressure drop is at a reasonable level; Variance condition: window variance Less than the preset fluctuation threshold ,For example This condition limits the range of random voltage fluctuations; Range condition: Window range Less than the preset amplitude threshold ,For example This condition directly limits the maximum instantaneous range of voltage variation.

[0086] SS224 uses the start point, midpoint, or one of the preset reference positions of the first sliding window that meets the judgment criteria as the stable point for the contact to enter a stable conducting state. .

[0087] Starting from the beginning of the stable search interval, examine each sliding window in chronological order. Mark the first sliding window that simultaneously satisfies all the above stability criteria. Determine a specific position within this first stable window as the moment when the contact enters a stable conducting state, i.e., the stable point. The specific location can be chosen from the window start point, the window midpoint, or the first sampling point within the window that meets the conditions.

[0088] Step SS3 in this embodiment is based on Baseline statistics for calculating contact voltages from previous consecutive sampling points ,as well as Calculate the stable region statistics of the contact voltage at several subsequent consecutive sampling points. Then, based on these two statistical values, an adaptive dynamic bounce threshold is calculated. This step is used to construct a bounce detection threshold that can automatically adjust according to different relay models, load conditions, and real-time signal characteristics. This is to overcome the limitations of the fixed threshold method.

[0089] The baseline region characterizes the steady-state voltage level of the relay contacts when they are fully open and unaffected by the closing process; it is typically close to the power supply voltage. When selecting this region, it should be ensured that it is located before the initial mechanical contact of the contacts. Specific selection methods include, but are not limited to, the pre-trigger region method and the pre-closing open-circuit region method. The pre-trigger region method refers to selecting a uniform time reference point. Previous consecutive At each sampling point, the region is in an unexcited coil state, and the contact is confirmed to be open. Open-circuit method before closure: Select the initial closure point. Previously, and distance A sufficiently long continuous segment Each sampling point is provided that the voltage in that segment is stable and at a high level.

[0090] Acquire the contact voltage sequence in the baseline area Subsequently, to suppress the impact of occasional noise spikes or interference pulses that may exist in the baseline region on the statistical results, one of the following robust statistical methods can be used to calculate... .

[0091] The truncated mean method involves removing a predetermined proportion of the highest and lowest values ​​from the corresponding voltage sequence and then calculating the arithmetic mean of the remaining data. Specifically, this involves removing data points from the baseline voltage sequence... All of them The sampled values ​​are arranged in ascending order of their numerical values ​​to obtain an ordered sequence. Preset a truncation ratio Calculate the number of head and tail data points that need to be truncated. ,in Indicates rounding down. From an ordered sequence In the middle, before removal The minimum value and the last The largest value. The remaining data points form the truncated sequence. , contains Points. For the truncated sequence. Calculate the arithmetic mean; this value is the baseline statistic. .

[0092] Median method: After sorting the corresponding voltage sequences by magnitude, the value at the median position is taken as the statistical value. Specifically, similar to the truncated mean method, the result is... .like If the number is odd, then the position of the median is... Baseline statistics That is (No. (values). If If the number is even, the median is usually the average of the two middle values. That is, the position... and The average value of the values.

[0093] The steady-state region characterizes the steady-state on-state voltage level of the contacts after full closure and bounce-back, i.e., the steady-state contact voltage drop. This region must be located at a predetermined steady-state point. Next, ensure that the contacts have reached a stable contact state. Typically, select... Then continuous Each sampling point is used as a stable region, for example point.

[0094] After obtaining the contact voltage sequence in the stable region, to suppress any residual minor bounces or noise interference that may remain in the stable region, one of the robust statistical methods mentioned above can also be used to calculate... The calculation process is similar to that of baseline statistics, and will not be repeated here.

[0095] Figure 5 Dynamic bounce threshold A schematic diagram of the calculation process, such as Figure 5 As shown, this embodiment comprehensively and Generate the final threshold:

[0096] in To stabilize the additional quantity, For example, the baseline scaling factor. .

[0097] The first constraint is a threshold based on the steady-state voltage drop. Statistics in the stable region Add a stable additional amount This constraint ensures the threshold. Always higher than the steady-state closure voltage drop Additional quantity This provides the necessary margin for detecting bounce events (voltage rise). Even with steady-state voltage drop... This constraint, which varies depending on the load, ensures that the threshold voltage adaptively shifts upwards, preventing the elevated steady-state voltage from being misinterpreted as a continuous rebound. If the increase is due to measurement error or anomaly... If the value is abnormally high, this constraint may cause the threshold to be too high. In this case, a second constraint will apply.

[0098] The second constraint is a threshold based on the baseline voltage. Baseline statistics Multiply by a baseline scaling factor This constraint correlates the threshold with the open-circuit voltage level, ensuring the threshold... It is always a reasonable proportion of the open-circuit voltage. Proportionality coefficient. This determines the required relative voltage drop to be considered a valid bounce, allowing the threshold to adapt to different supply voltages. If interference causes baseline measurements to be affected... If the value is abnormally low, this constraint may cause the candidate threshold to be too low; in this case, the impact can be reduced by robust statistics in the baseline area, anomaly removal, or setting a lower limit for the project. Final threshold Pick and The smaller value in the range is used to avoid the threshold being too high, which could lead to missed detections due to slight bounces. If based on steady state... Too high, while based on baseline If it's lower, then choose the lower one. As a threshold; if the baseline area is abnormally low, leading to... If the baseline is too low, the impact can be mitigated through robust baseline statistics, anomaly removal, or setting an engineering lower limit. This embodiment takes the smaller value between the two threshold generation logics "based on steady-state voltage drop" and "based on open-circuit voltage ratio" to maintain high sensitivity in bounce detection, while reducing the impact of abnormal baselines on reliability through robust statistics and engineering limits.

[0099] In this embodiment, step SS4 is performed by... and Within the defined detection range, a dynamic bounce threshold is used. As a criterion, a finite state machine is used to analyze the contact voltage sequence. Perform event detection to identify events where the number of persistent events is not less than a preset threshold. The valid bounce event specifically includes the following steps SS401 to SS405.

[0100] SS401, in contact voltage sequence Above, delineated by and Limited detection range ,in This is the preset starting offset number.

[0101] To analyze resources and avoid misjudgments, the detection interval where bounce events may occur is first defined. In this embodiment, the detection interval is... .

[0102] lower limit The first closing point of the contact As a baseline, add a preset number of starting offset points. ,For example Point. The purpose of this offset is to avoid extreme voltage transients, ringing, or measurement disturbances that may occur at the moment of closure. These transients are not true bounce events, but may trigger false detections. Offset amount It should be set based on the sampling rate and typical transient duration.

[0103] The upper limit is based on the contact stability point. This serves as the end boundary of the detection range. Once a stable state is reached, the contact voltage should remain below the threshold, and subsequent fluctuations will no longer be considered a bounce.

[0104] SS402 defines a state machine with two states: a non-rising state S0, which indicates that the current state is not in a bounce event, and a rising state S1, which indicates that the current state is in a potential bounce segment where the contact voltage exceeds a threshold.

[0105] State S0 (Non-Rising State): Indicates that the current sampling point is not in a voltage rise segment identified as a potential bounce. Typically, this corresponds to a contact voltage below or equal to a threshold. This refers to the situation where the contacts are in a stable or transitional period of "effective closure" or "open".

[0106] State S1 (Rising State): Indicates that the current sampling point is in a voltage rise segment identified as a potential bounce. This state occurs when the voltage first exceeds the threshold. The event enters at a certain time and continues until the voltage drops below the threshold. While in this state, the system is accumulating the duration of this segment to determine whether it is a valid event.

[0107] SS403 initializes the state machine to state S0 and clears the duration counter for the current segment.

[0108] Before starting the scanning detection interval, the state machine is initialized. The current state is set to Scurrent = S0, assuming the starting point is not in a bounce. The duration counter for the current segment is cleared to 0. This counter is used to accumulate the number of sampling points that have continuously exceeded the threshold in state S1.

[0109] SS404, starting from the beginning of the detection interval, measures the contact voltage sequence. The following state transition logic is executed for each sampling point.

[0110] Starting from the beginning of the detection interval (index i = start_idx) and ending at the end (index i = end_idx), the contact voltage sequence is analyzed. Each sampling point The following logic is executed sequentially. Let... This is the calculated dynamic bounce threshold.

[0111] a) When in state S0, if the current sampling point voltage If the condition is met, the transition to state S1 is triggered, the current index is recorded as the segment start point (start), and duration is set to 1; otherwise, state S0 is maintained.

[0112] like This indicates that a new voltage rising edge has been detected, which may be the start of a bounce. The state transitions from S0 to S1. The current index i is recorded as the start point of the potential bounce segment, start=i. The duration counter is reset and started, duration=1.

[0113] b) When in state S1, if the current sampling point voltage If so, maintain state S1 and set duration = duration + 1; if the current sampling point voltage Then determine whether the duration is greater than or equal to the preset threshold. If so, the segment from start to the current index is confirmed as a valid bounce event, the start and end information of the event is recorded, the bounce event count is incremented by one, and the process is transferred back to the S0 state. Otherwise, the segment is determined to be a glitch, and the process is transferred back to the S0 state directly.

[0114] like If the voltage continues to exceed the threshold, the current potential bounce segment continues. At this time, remain in state S1 and increase the segment duration by duration = duration + 1.

[0115] like If the voltage drops back to or below the threshold, it indicates the end of the current potential bounce segment. At this point, further checks are made to determine whether the duration of the segment meets the validity criteria, i.e., whether the duration is greater than or equal to the preset minimum duration threshold. ,For example point.

[0116] If duration≥ If the duration of the segment is long enough, it is considered a valid bounce event. At this point, we switch back from S1 to S0, record the current index i as the end point of the valid event (end=i), increment the valid bounce event counter (bounce_count=bounce_count+1), store the start and end indices [start,end] and duration of the event in the valid event list, and reset duration=0.

[0117] If duration < If the duration of the segment is too short, it is judged as a noise spike. At this time, the process is switched from S1 back to S0, the segment is discarded, no counting is performed, no event information is recorded, and duration is reset to 0.

[0118] After completing the scan of the entire detection range, the SS405 outputs the start and end time indices and corresponding duration points of all identified valid bounce events, forming a list of valid bounce events.

[0119] After scanning all sampling points in the entire detection interval, a list of valid bounce events is output. This list contains the start and end time indices and duration points of all confirmed valid bounce events, which is the total number of events in the valid event list, bounce_count.

[0120] Minimum number of persistence points Used to distinguish between valid bounce and noise glitches, its setting takes into account the system sampling rate; for example, at a sampling rate of 12.5MHz, The point corresponds to 240 ns. This is much shorter than the typical duration of a mechanical bounce (in the microsecond to millisecond range), but sufficient to filter out most nanosecond-level electrical noise spikes. In the basic state machine detection logic, a single threshold is used. This serves as the criterion for entering (S0→S1) and exiting (S1→S0) states. However, in actual signals, the contact voltage is at the threshold. There may be minor jitter or noise in the vicinity, which could cause the state machine to briefly change due to voltage fluctuations. Frequent fluctuations in state machine performance can lead to multiple invalid state transitions, potentially resulting in the missegmentation of a continuous bounce event into multiple segments or the generation of glitches. To enhance the robustness of the state machine against fluctuations near a threshold and improve the stability of event segmentation, a threshold hysteresis mode is introduced in some optional implementations, specifically including: Set hysteresis halfwidth ; Calculate the entry threshold and exit threshold ; When executing the state transition logic, the criterion for determining whether to trigger the transition to state S1 in logic a) is adjusted as follows: For logic b), the criterion for maintaining state S1 is adjusted as follows: ,when At that time, a valid bounce event is confirmed or a glitch is determined based on whether the duration is not less than a preset threshold.

[0121] Figure 6 This is a schematic diagram of the state transition for bounce detection in a finite state machine based on a threshold hysteresis pattern. Specifically, it first involves the calculation of a dynamic bounce threshold. This is achieved by superimposing a hysteresis half-width. To construct a new pair of thresholds.

[0122] Entry threshold Used to determine when a state transitions from a "non-rising state (S0)" to a "rising state (S1)", i.e., to begin identifying a new potential bounce segment whose value is higher than the base threshold. .

[0123] Exit threshold Used to determine when to exit from the "rising state (S1)," that is, to determine whether an ongoing potential bounce segment has ended when its value is below the base threshold. .

[0124] in, A preset positive value is called the hysteresis half-width. It defines... and The width of the gap between them, which is the hysteresis region or dead zone. In one specific embodiment, 0.2 V is acceptable.

[0125] When threshold hysteresis mode is enabled, the state transition logic in step SS404 is adjusted as follows to use and Replace single .

[0126] The adjusted logic a) is currently in state S0, and the judgment is based on checking... If the condition is met, it indicates that the voltage has significantly exceeded the baseline threshold, reliably indicating the start of a rebound. Trigger the S0→S1 transition, record the starting point, and set duration=1. If the condition is not met, even if... but It remains in the S0 state. This prevents the voltage from... False entries caused by fluctuations within the range.

[0127] The adjusted logic (b) is currently in state S1, and the fragment continuation judgment is based on checking... If the condition is met, it indicates that the voltage has not reliably fallen below the exit threshold. The segment continues, maintaining state S1, and executing duration = duration + 1. Wherein, when... When the current sampling point is in the hysteresis region, the state machine maintains the original S1 state to avoid the segment being truncated prematurely.

[0128] when When the voltage has reliably fallen below the exit threshold, the bounce segment ends. At this point, it is determined whether the duration is not less than a preset threshold. If the conditions are met, the segment from start to the current index is confirmed as a valid bounce event, and the start and end information of the event is recorded; if the conditions are not met, the segment is determined to be a glitch. After completing the above judgment, the process transitions back to state S0.

[0129] In some optional implementations, considering that in certain situations, due to high-frequency jitter or noise interference of the contact voltage, a voltage rise that originally belonged to the same physical bounce process may be incorrectly segmented by the state machine into two or more valid events with very short intervals, an event merging step is included after step SS405 to more accurately reflect the physical process and avoid excessive segmentation leading to an inflated bounce count. That is, after identifying the list of valid bounce events, if the interval between two adjacent valid bounce events is less than a preset merging threshold, the event merging step is performed. If so, these two valid bounce events will be merged into one valid bounce event.

[0130] in, The preset merging threshold is set based on the sampling rate and the physical characteristics of the relay bounce. Its value should be significantly smaller than the typical time interval between two independent physical bounces, but larger than the meaningless segmentation interval that the state machine might produce due to noise. For example... The point corresponds to approximately 400 ns.

[0131] It should be noted that the i-th and (i+1)-th events are merged into a new single event. The starting index of the new event is the start_i of the original i-th event, and the ending index is the end_i+1 of the original (i+1)-th event. The energy value of the new event is the sum of the energies of the two original events, i.e., The total number of valid bounce events is reduced by 1.

[0132] Figure 7 This is a schematic diagram of the energy integral calculation for a single bounce. Step SS5 calculates the energy value of this bounce. Specifically, this includes: for each identified valid bounce event, extracting the corresponding contact voltage and contact current data segments within the start and end time of the event; multiplying the sampled values ​​at the same time points in the voltage and current data segments point by point to obtain the instantaneous power sequence corresponding to the event; and performing discrete-time integration on the instantaneous power sequence over the duration of the event to obtain the energy value of the bounce. .

[0133] Specifically, for each valid bounce event [start, end], the contact voltage is extracted. With contact current Energy is calculated using discrete integrals: , The sampling interval is specified. In one embodiment, the contact current is obtained by dividing the sampling voltage of channel C3 by the current conversion factor, which can be set to 100; the energy can be converted into mJ units for storage. The sampling interval is specified for each... Store in the energy array (capacity can be 500), and simultaneously update: total bounce energy. Maximum single-attack capacity Statistics such as mean energy.

[0134] In this embodiment, step SS6 outputs the number of bounces and the energy value based on all valid bounce events. The energy parameters obtained through statistics, including those based on energy values The energy parameters are statistically derived, specifically including the following steps SS601 to SS604.

[0135] SS601, the single energy value corresponding to all valid bounce events. Arranged in chronological order, forming a rebound energy sequence. ,in This represents the total number of valid bounce events.

[0136] Specifically, the single-shot energy value of a valid bounce event is detected. Is it greater than the preset minimum energy threshold? If not, then remove the valid bounce event; form a bounce energy sequence from all remaining valid bounce events.

[0137] SS602 calculates energy statistics parameters, including total rebound energy, based on the rebound energy sequence. Maximum single bounce energy Average single-rebound energy .

[0138] Total rebound energy reflects the total energy consumed during the rebound process and is related to the heat accumulation from contact wear. Maximum single rebound energy is used to identify the highest-energy single event, helping to detect unusually high-energy impacts. Average single rebound energy reflects the average level of rebound energy.

[0139] SS603, if An exponential decay model is established for the bounce energy sequence, expressed as: ,in Through the and Perform linear fitting to solve for the attenuation coefficient. .

[0140] If the number of bounces This indicates that the rebound process involves multiple energy releases. Typically, the energy of a normal rebound process exhibits a decaying trend, which is quantified using an exponential decay model.

[0141] Attenuation coefficient This reflects the rate of energy decay. Taking the natural logarithm of both sides of the model, we get... .Will Treat it as a dependent variable. Treating the variable as the independent variable, a straight line is fitted using the linear least squares method. The absolute value of the slope of this line is the attenuation coefficient. . The larger the value, the faster the energy decays and the better the system damping characteristics.

[0142] SS604, the initial energy based on the bounce energy sequence With the final energy and the time span of the corresponding events. Calculate the damping time constant , is represented as:

[0143] in, and These are the time points of the first and last valid bounce events, respectively.

[0144] Damping time constant Reflects the energy from the initial value decay to the final value The corresponding equivalent time scale. The smaller the value, the faster the energy decays and the faster the system tends to stabilize.

[0145] In some optional implementations, the method also includes a step of comprehensive scoring and anomaly detection based on energy parameters, specifically including the following steps SS605 and SS606.

[0146] SS605, based on the number of bounces Total bounce energy and energy decay coefficient Calculate the overall quality score using the following formula. :

[0147] in, These are preset weighting coefficients, and the sum of the three is 1. The weights reflect the degree of influence of different dimensions on the overall quality; for example, they can be set... .

[0148] For a scoring item based on the number of bounces, it is defined as follows: , The maximum allowed number of bounces is preset based on product standards or historical qualified data.

[0149] For a scoring item based on total energy, defined as , The preset maximum allowable total energy is set based on the heat load or arc energy tolerance.

[0150] For the scoring item based on damping characteristics, it is defined as follows: , The preset attenuation coefficient reference value is set based on the typical performance of good damping characteristics.

[0151] SS606, if the overall quality score is... Below the preset qualified threshold If the maximum single bounce energy is not met, it is considered unqualified; Exceeding the preset single-use energy safety threshold If the number of bounces is too high, an energy over-limit anomaly will be triggered; Exceeding the preset frequency threshold If so, the trigger frequency will be abnormal.

[0152] Figure 8 This is a schematic diagram of the multidimensional statistics and display interface for rebound parameters. It allows maintenance of historical statistics (minimum / maximum / average, etc.) and supports exporting energy arrays for subsequent analysis. Furthermore... Lock and The stability point determination can be shared with the pull-in time test module, enabling one-time data acquisition and multiple index outputs, thus improving test efficiency.

[0153] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for identifying relay contact bounce parameters based on state machine and dynamic threshold, characterized in that, Includes the following steps: Using the rising edge of the relay coil voltage as the trigger source to lock a unified time reference point. ,based on Alignment is performed, and the voltage signals across the contacts and the current signals flowing through the contacts are acquired simultaneously to obtain the contact voltage sequence. With contact current sequence ; Based on contact voltage sequence Determine the first closing point of the contact. Stable point where the contact enters a stable conducting state ; based on Baseline statistics for calculating contact voltages from previous consecutive sampling points ,as well as Calculate the stable region statistics of the contact voltage at several subsequent consecutive sampling points. According to the formula Calculate the adaptive dynamic bounce threshold ,in To stabilize the additional quantity, This is the baseline scaling factor; In the and Within the defined detection range, a dynamic bounce threshold is used. As a criterion, a finite state machine is used to analyze the contact voltage sequence. Perform event detection to identify events where the number of persistent events is not less than a preset threshold. A valid bounce event; For each identified valid bounce event, extract the contact voltage sequence from its corresponding start and end time. With contact current sequence Calculate the energy value of the corresponding data segment for this bounce. ; Based on all valid bounce events, output the bounce count and energy value. Energy parameters obtained from statistics.

2. The relay contact bounce parameter identification method based on state machine and dynamic threshold according to claim 1, characterized in that, Determine the first closing point of the contact Specifically, it includes: by Based on the contact voltage sequence Define the preset search range above ,in and For relative to The sampling point offset; Performing the first-level positioning includes: within the search interval, calculating the voltage drop index at each sampling point location using a sliding window, and taking the location with the largest voltage drop index exceeding a preset drop threshold as the initial contact closure point. ; If the first-level positioning fails, the second-level positioning is executed, including: calculating baseline statistics based on the contact voltages of several consecutive sampling points before the start of the search interval. Construct a relative threshold ,in The preset proportional coefficient is used to search for the first contact voltage below a certain value within the search range. The position is used as the first closing point of the contact. ; If the second-level positioning also fails, then Add a preset fixed offset number of points As the first point of contact closure .

3. The relay contact bounce parameter identification method based on state machine and dynamic threshold according to claim 1, characterized in that, Determine the stable point at which the contact enters a stable conducting state. Specifically, it includes: Contact voltage sequence Above, at Then define a stable search interval; Within the stable search interval, a sliding window is used to advance segment by segment, and the statistical characteristics of the contact voltage in each window are calculated. The statistical characteristics include the mean, variance and range. When all statistical features within a certain sliding window meet the preset stability criteria, the time period corresponding to that window is determined to be in a stable conduction state. The starting point, midpoint, or one of the preset reference positions of the first sliding window that meets the judgment criteria is taken as the stable point where the contact enters a stable conducting state. .

4. The relay contact bounce parameter identification method based on state machine and dynamic threshold according to claim 1, characterized in that, In the and Within the defined detection range, a dynamic bounce threshold is used. As a criterion, a finite state machine is used to analyze the contact voltage sequence. Perform event detection to identify events where the number of persistent events is not less than a preset threshold. Valid bounce events specifically include: Contact voltage sequence Above, delineated by and Limited detection range ,in The preset starting offset number; The state machine is defined to have two states: a non-rising state S0, which represents the current state not being in a bounce event, and a rising state S1, which represents the current state being in a potential bounce segment where the contact voltage exceeds a threshold. Initialize the state machine to state S0 and clear the duration counter for the current segment. Starting from the beginning of the detection interval, the contact voltage sequence is... The following state transition logic is executed for each sampling point: a) When in state S0, if the current sampling point voltage If the current index is not found, the transition to state S1 is triggered, the current index is recorded as the segment start point (start), and duration is set to 1; otherwise, state S0 is maintained. b) When in state S1, if the current sampling point voltage If so, maintain state S1 and set duration = duration + 1; if the current sampling point voltage Then determine whether the duration is greater than or equal to the preset threshold. If so, the segment from start to the current index is confirmed as a valid bounce event, the start and end information of the event is recorded, the bounce event count is incremented by one, and the process is transferred back to the S0 state; otherwise, the segment is determined to be a glitch, and the process is transferred back to the S0 state directly. After completing the scanning of the entire detection range, the start and end time indices and corresponding duration points of all identified valid bounce events are output to form a list of valid bounce events.

5. The relay contact bounce parameter identification method based on state machine and dynamic threshold according to claim 4, characterized in that, The method also includes implementing a threshold hysteresis mode, specifically including: Set hysteresis halfwidth ; Calculate the entry threshold and exit threshold ; When executing the state transition logic, the criterion for determining whether to trigger the transition to state S1 in logic a) is adjusted as follows: For logic b), the criterion for maintaining state S1 is adjusted as follows: ,when At that time, a valid bounce event is confirmed or a glitch is determined based on whether the duration is not less than a preset threshold.

6. The relay contact bounce parameter identification method based on state machine and dynamic threshold according to claim 4, characterized in that, The method also includes an event merging step, specifically including: After identifying the list of valid bounce events, if the interval between two consecutive valid bounce events is less than the preset merging threshold... If so, these two valid bounce events will be merged into one valid bounce event.

7. The relay contact bounce parameter identification method based on state machine and dynamic threshold according to claim 1, characterized in that, Based on energy value The energy parameters were statistically derived, including: The single energy value corresponding to all valid bounce events Arranged in chronological order, forming a rebound energy sequence. ,in The total number of valid bounce events; Based on the rebound energy sequence, calculate energy statistical parameters, including the total rebound energy. Maximum single bounce energy Average single-rebound energy ; like An exponential decay model is established for the bounce energy sequence, expressed as: ,in Through the and Perform linear fitting to solve for the attenuation coefficient. ; The first energy based on the bounce energy sequence With the final energy and the time span of the corresponding events. Calculate the damping time constant , is represented as: in, and These are the time points of the first and last valid bounce events, respectively.

8. The relay contact bounce parameter identification method based on state machine and dynamic threshold according to claim 7, characterized in that, The single energy value corresponding to all valid bounce events Arranged chronologically, they form a rebound energy sequence, specifically including: Single-shot energy value for detecting valid bounce events Is it greater than the preset minimum energy threshold? ; If not, then the valid bounce event will be removed; Form a bounce energy sequence from all remaining valid bounce events.

9. The relay contact bounce parameter identification method based on state machine and dynamic threshold according to claim 7, characterized in that, The method also includes steps for comprehensive scoring and anomaly detection based on energy parameters, specifically including: Based on the number of bounces Total bounce energy and energy decay coefficient Calculate the overall quality score using the following formula. : in, These are preset weighting coefficients, and the sum of the three is 1; For a scoring item based on the number of bounces, it is defined as follows: , This is the preset maximum number of bounces; For a scoring item based on total energy, defined as , This is the preset maximum allowable total energy. For the scoring item based on damping characteristics, it is defined as follows: , This is the preset reference value for the attenuation coefficient; If the overall quality score Below the preset qualified threshold If the maximum single bounce energy is not met, it is considered unqualified; Exceeding the preset single-use energy safety threshold If the number of bounces is too high, an energy over-limit anomaly will be triggered; Exceeding the preset frequency threshold If so, the trigger frequency will be abnormal.

10. The relay contact bounce parameter identification method based on state machine and dynamic threshold according to claim 1, characterized in that, Baseline statistics and / or stability zone statistics Calculated using one of the following methods: a) After removing data points of a preset proportion from the highest and lowest values ​​in the corresponding voltage sequence, calculate the arithmetic mean of the remaining data; b) After sorting the corresponding voltage sequences by size, take the value at the middle position as the statistical value.

Images