Low-voltage bypass sequence control method and device based on phase sequence identification and fracture state

Abstract

The invention relates to the technical field of low-voltage power distribution control, and discloses a low-voltage bypass sequence control method and device based on phase sequence recognition and a fracture state. The method comprises the following steps: acquiring voltage waveform data of a low-voltage three-phase circuit through an integrated protection relay, sampling, and outputting a phase sequence identification result; generating an opening and closing control instruction based on the identification result, transmitting the opening and closing control instruction to a spring operation mechanism to execute opening and closing actions, and outputting a fracture state signal; the fracture state signal is subjected to long-distance imaging display through the large window insulation structure to obtain an imaging confirmation result of the fracture state; and generating state indication information based on an imaging confirmation result, performing comprehensive judgment by using the information in control logic, and executing low-voltage bypass switching-on operation when a judgment condition is met. According to the invention, organic combination of phase sequence identification, opening and closing control and fracture state visualization is realized, the safety and reliability of the low-voltage distribution network in the bypass operation process can be improved, and the risk of misoperation is avoided.

Description

Technical Field

[0001] This invention relates to the field of switching control technology for power distribution networks, and in particular to a low-voltage bypass sequence control method and device based on phase sequence identification and disconnection status. Background Technology

[0002] In existing technologies, bypass operations in low-voltage distribution networks typically rely on manual experience or conventional electrical testing methods to determine phase sequence and circuit breaker status. Common methods include using voltage or current detection devices to determine phase sequence and employing mechanical indicators, simple indicator lights, or secondary wiring monitoring to determine the status of switch breaks. These technologies can assist operators in controlling the opening and closing sequence to some extent, thereby ensuring basic safety and continuity of low-voltage circuits during maintenance, switching, or emergency power supply.

[0003] The aforementioned existing technologies generally suffer from insufficient reliability. First, phase sequence determination relying solely on electrical detection is prone to misjudgment under conditions of high grid noise or load fluctuations, affecting the accuracy of opening and closing. Second, most methods for detecting the status of circuit breakers are limited to local mechanical indications or electrical signal feedback, failing to provide clear and intuitive visual confirmation, making it difficult for operators to detect anomalies in a timely manner. Furthermore, the lack of a comprehensive judgment mechanism for status information during bypass operations easily leads to misoperation or delayed operation, thereby increasing the operational risks of low-voltage distribution networks.

[0004] Therefore, it is necessary to propose a new control method and device that can realize phase sequence identification and visual confirmation of break state in low-voltage bypass operation, and improve safety by combining logical judgment. Summary of the Invention

[0005] This application provides a low-voltage bypass sequence control method and device based on phase sequence identification and fault status to improve the safety and reliability of low-voltage power distribution networks during bypass operations.

[0006] This application provides a low-voltage bypass sequence control method based on phase sequence identification and fault state, including: The voltage waveform data of the low-voltage three-phase circuit is obtained by an integrated protection relay, and the phase sequence identification result is output after sampling. The phase sequence identification result is used to generate opening and closing control commands and output to the spring operating mechanism; The spring operating mechanism executes the opening and closing control command and outputs the disconnection status signal; The fracture status signal is displayed remotely by using a large-window insulation structure to obtain imaging confirmation results. Based on the imaging confirmation results, status indication information is generated, and the status indication information is used in the control logic for judgment. When the judgment conditions are met, the low-voltage bypass closing operation is performed.

[0007] This application provides a low-voltage bypass sequence control device based on phase sequence identification and fault state, comprising: The acquisition unit is used to acquire voltage waveform data of the low-voltage three-phase circuit through the integrated protection relay, and output the phase sequence identification result after sampling; The generation unit is used to generate opening and closing control commands based on the phase sequence identification results and output them to the spring operating mechanism. An execution unit is used for the spring operating mechanism to execute the opening and closing control commands and output the disconnection status signal; The display unit is used to perform long-distance imaging display of the fracture status signal through a large-window insulation structure to obtain imaging confirmation results. The determination unit is used to generate status indication information based on the imaging confirmation result, and to make a determination using the status indication information in the control logic. When the determination condition is met, the low-voltage bypass closing operation is executed.

[0008] The beneficial effects of this application mainly include: (1) By acquiring the voltage waveform of the low-voltage three-phase circuit and performing phase sequence identification through the integrated protection relay, the correct electrical connection sequence before bypass operation can be guaranteed, effectively avoiding equipment damage or power supply accidents caused by incorrect phase sequence. (2) By using the break status signal combined with the large window insulation structure for long-distance imaging display, the intuitive visualization confirmation of the break status is realized, significantly improving the operator's ability to identify and the reliability of the switch status in complex environments. (3) Based on the imaging confirmation result, the status indication information is generated and entered into the control logic judgment, which can form multiple cross-verifications at the system level, reduce the risk of misoperation, and improve the safety and intelligence level of low-voltage bypass operation. (4) By organically combining phase sequence identification, opening and closing control and break status visualization, the bypass closing operation can be performed only under strict conditions, thereby improving the continuity and stability of the low-voltage power distribution network during maintenance, switching and emergency power supply processes. Attached Figure Description

[0009] Figure 1 This is a flowchart of a low-voltage bypass sequence control method based on phase sequence identification and fault state provided in the first embodiment of this application.

[0010] Figure 2 This is a schematic diagram of a low-voltage bypass sequence control device based on phase sequence identification and break state provided in the second embodiment of this application. Detailed Implementation

[0011] Many specific details are set forth in the following description to provide a full understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this application; therefore, this application is not limited to the specific embodiments disclosed below.

[0012] The first embodiment of this application provides a low-voltage bypass sequence control method based on phase sequence identification and fault state. Please refer to... Figure 1 This figure is a flowchart of the first embodiment of this application. The following is in conjunction with... Figure 1 The first embodiment of this application provides a detailed description of a low-voltage bypass sequence control method based on phase sequence identification and fault state.

[0013] Step S101: Obtain the voltage waveform data of the low-voltage three-phase circuit through the integrated protection relay, and output the phase sequence identification result after sampling.

[0014] Step S101 relies on the integrated three-phase voltage sampling and digital signal processing capabilities built into the integrated protection relay. The relay's voltage measurement terminals are connected to phases A, B, and C of the low-voltage three-phase circuit under test, respectively. After isolation and voltage reduction via a voltage transformer or resistor divider link matched to the rated voltage, the voltage is input to the three synchronously sampled analog-to-digital converter channels. To suppress aliasing, a hardware low-pass filter is installed at the front end, with its cutoff frequency preferably twenty to thirty times the fundamental frequency. The ADC resolution should be no less than 16 bits, and the sampling rate is preferably in the range of 6.4 kS / s to 12.8 kS / s, ensuring synchronization of the three channels to avoid phase misalignment caused by multiplexing. After power-on, zero-point drift and gain calibration are first performed, recording the bias and scaling factor of each channel. The sampled values ​​are converted into equivalent phase voltage waveforms in volts, and a high-precision time stamp is added before the data enters the digital processing chain. This time stamp comes from the temperature-compensated crystal oscillator or time synchronization module inside the relay, ensuring that the phase sequence identification result can be traced back to a specific time slice.

[0015] After the sampled data enters the digital preprocessing module, a combination of notch and bandpass filtering is performed first. The notch is used to attenuate the second and higher major harmonics of the power frequency and possible DC bias. The bandpass center is set near the nominal power frequency of the system (50 Hz or 60 Hz), and the bandwidth covers ±2 Hz to ±3 Hz to accommodate fundamental frequency offset. The filter adopts a linear-phase finite impulse response structure, and its length is selected according to the sampling rate and the desired group delay to achieve a stable response within two fundamental frequency cycles. Subsequently, amplitude rationality and existence detection are performed. When the effective value of a phase is lower than a set threshold (e.g., 10% of the rated phase voltage) or a severe gap occurs intermittently for a short time, phase sequence determination is temporarily suspended and a "not sufficient for determination" status is output. At the same time, this status is written to the event log to prevent misjudgment caused by undervoltage or loose wiring.

[0016] After ensuring stable waveforms for all three phases, the fundamental phasor extraction and phase angle estimation process begins. To improve robustness under noise and distortion environments, a sliding DFT or an equivalent Goertzel algorithm is used to extract the fundamental complex phasor of each phase within a window of one to two power frequency cycles. The window function is preferably a Hanning or Blackman window to reduce spectral leakage. For each analysis window, the amplitude and phase angle estimates of phases A, B, and C are obtained, and the phase angles are normalized to the range of −180° to +180°. To handle short-term fluctuations, an exponentially weighted moving average is used to smooth the phase angle sequence. The weighting coefficients are selected based on the target response time constant, ensuring rapid convergence within two to three cycles during abrupt phase sequence changes (e.g., wiring changes) and preventing excessive jitter during normal small fluctuations. Simultaneously, the total harmonic distortion rate and signal-to-noise ratio for each window are calculated as auxiliary quantities for subsequent confidence assessment.

[0017] The core determination of phase sequence identification is based on the relative order of the three-phase phase angles and the closeness of their phase differences. Using phase A as a reference, the phase differences from A to B, B to C, and C to A are calculated and mapped to -180° to +180°. When the phase difference is close to +120° and the sum of the three is close to zero, it is determined to be in the correct sequence (ABC); when the phase difference is close to -120°, it is determined to be in the wrong sequence (ACB). To accommodate deviations in actual operating conditions, a tolerance band is set, for example, ±20°, and an adaptive tolerance tightening or loosening strategy is given in combination with the unbalance and THD: when THD is high or voltage imbalance is significant, the tolerance is moderately loosened but the number of consecutive consistent windows required for determination is increased; when the signal quality is excellent, the tolerance is tightened to improve selectivity. To prevent jumps caused by short pulse interference, the identification result must be consistent within several consecutive windows before output confirmation. The number of consecutive windows can be set to 3 to 5 cycles, equivalent to a verification time on the order of 60 ms to 100 ms in a 50 Hz system. If the three-phase phase angle difference deviates too much from the ideal 120° relationship and cannot converge within the set time, the diagnosis of "abnormal phase sequence" will be output and the existing phase sequence state will remain unchanged, waiting for the next stable window to make a judgment.

[0018] Considering the presence of severe phase voltage distortion and insufficient short-term stability of the phasor method in some field applications, the preprocessing module also runs in parallel with alternative criteria based on zero-crossing interpolation. Linear or quadratic interpolation is performed on the sampling pairs crossing zero points in each phase channel to obtain a more accurate zero-crossing timestamp sequence; a half-cycle time is formed between two consecutive zero crossovers, thus constructing a three-phase zero-crossing time sequence map. When the time sequence of phase A zero crossover relative to phases B and C is more consistent with the time difference corresponding to the ideal 120° electrical angle, the same forward or reverse order judgment as the phasor method is given; when the two methods are consistent, the overall confidence level is increased; when they are inconsistent, the side with lower THD is prioritized and the consistency verification window is extended to reduce the probability of false triggering. The judgment weights of the above dual-channel strategy can be automatically adjusted by online quality indicators, without the need for manual switching on-site.

[0019] To ensure that the identification results can be directly used by subsequent control logic, the relay generates a structured result object after each confirmation. This object includes the phase sequence type, current fundamental frequency estimate, three-phase RMS value, phase angle difference, THD, imbalance degree, and confidence score, along with a timestamp and sampling configuration version number for traceability. This result is published to the upper-level control task via the relay's internal bus, and the most recent records are continuously saved in non-volatile memory to support post-event analysis and anomaly verification. For special operating conditions involving incorrect wiring, the system provides a one-time reference phase binding option. For example, the physical terminals can be initially labeled as logic phases A, B, and C. If a long-term stable sequence relationship that does not conform to the labeling is detected, a "suspected wiring replacement" message is displayed, but the logic phase mapping is not arbitrarily corrected to avoid introducing unpredictable phase flips in unattended scenarios.

[0020] When the grid frequency drifts slowly, the sliding DFT center frequency adopts adaptive tracking, with the update step size of the center frequency offset limited to the order of several millihertz to tens of millihertz per cycle to balance tracking performance and stability. When the frequency changes rapidly or a power supply switch occurs, the algorithm automatically shortens the analysis window and increases the weight of the zero-crossing channel to ensure that the phase sequence results recover reliably as quickly as possible after a short disturbance. All thresholds and time constants can be configured through the relay's parameter area. The factory default parameters cover common low-voltage 400 V / 230 V, 50 Hz or 60 Hz systems, and can be fine-tuned on-site according to load characteristics and power quality levels. Thus, after reliably acquiring and processing the voltage waveform, the relay can stably output accurate phase sequence identification results within a finite time after sampling, providing a deterministic input prerequisite for subsequent opening and closing control and bypass sequence logic.

[0021] Furthermore, the step of acquiring voltage waveform data of the low-voltage three-phase circuit through an integrated protection relay and outputting phase sequence identification results after sampling includes: In the integrated protection relay, the A, B and C phases of the low-voltage three-phase circuit are sampled synchronously in three channels. Based on the internal clock and the grid phase-locked reference, time calibration is performed to generate calibration parameters and interpolate the sampling time axis to obtain the calibrated voltage waveform data. The corrected voltage waveform data is preprocessed based on DC bias elimination and main harmonic component suppression to obtain the debiased voltage waveform. At the same time, the total harmonic distortion of the voltage and the voltage imbalance are calculated as quality indicators. Sliding discrete Fourier analysis and zero-crossing interpolation calculations are performed in parallel on the debiased voltage waveform to obtain the fundamental phase angle and zero-crossing time of each phase. Weighting coefficients are adaptively generated based on the quality index, and the phase angle estimation results are weighted and fused to obtain the fused phase angle and frequency. The phase sequence is determined by integrating the phase angle input logic phase mapping and sequence determination unit. When the three-phase voltage amplitude meets the threshold condition, the phase sequence type is determined based on the phase angle difference. When any phase voltage is missing or the quality index is lower than the threshold, the symmetrical component method is used for phase angle compensation. The consistency of results within multiple consecutive analysis windows is used as the confirmation condition. Finally, the phase sequence identification result is output and structured information including phase sequence type, confidence level, frequency, phase angle difference, timestamp, and sampling configuration version number is generated.

[0022] In this embodiment, the process of acquiring voltage waveform data of a low-voltage three-phase circuit through an integrated protection relay and outputting phase sequence identification results after sampling first requires ensuring synchronous sampling of the voltage signals of the three-phase circuit. The low-voltage three-phase circuit includes phases A, B, and C, each connected to the relay sampling port via a voltage transformer or a resistor divider link. During sampling, to ensure no time offset between the three-phase signals, the relay uses a multi-channel synchronous analog-to-digital converter (ADC) to simultaneously acquire the three voltage signals. The sampling resolution is preferably 16 bits or higher, and the sampling frequency is generally set in the range of 6.4 kS / s to 12.8 kS / s to ensure the capture of the fundamental frequency and major harmonic components. Due to slight offsets between different sampling channels, the system generates time calibration parameters using an internal clock and a grid phase-locked reference, and uses an interpolation algorithm to correct the time axis, thereby ensuring strict alignment of the three sampling results in the time domain and obtaining corrected voltage waveform data.

[0023] After obtaining the corrected waveform data, preprocessing is required to eliminate DC bias and suppress major harmonic components. DC bias refers to the average value offset caused by sampling circuit or power grid waveform distortion, which interferes with the accurate calculation of phase angle. The elimination method typically involves averaging the data over one power frequency cycle and subtracting it from the waveform. Suppression of major harmonic components is achieved using digital filtering methods, commonly finite impulse response (FIR) filters or fast Fourier transform (FFT) followed by attenuation of higher-order components. The resulting debiased voltage waveform is closer to an ideal sine wave. During this process, the system calculates the total harmonic distortion (THD) and voltage imbalance as quality indicators. THD is typically calculated as the ratio of the root mean square of all harmonic components to the fundamental component, expressed as a percentage; voltage imbalance is obtained by comparing the three-phase effective values, for example, the difference between the maximum and minimum phase voltages divided by the average phase voltage as a percentage.

[0024] Based on the debiased voltage waveform, the system executes two analysis methods in parallel to improve robustness. One is sliding discrete Fourier analysis (SFT), which calculates the complex phasor of the fundamental wave using a recursive formula within a sampling window of one or more power frequency cycles to obtain the phase angle and amplitude. The other is zero-crossing interpolation calculation, which calculates the precise time of zero-crossing using linear interpolation between two adjacent sampling points when the phase voltage waveform is detected to cross zero, thereby deriving the phase. Since these two methods have their own advantages and disadvantages under different operating conditions, the system adaptively generates weighting coefficients based on previously calculated quality indicators. For example, when THD is low, the DFT method is relied upon more, while the weight of the zero-crossing method is increased when signal distortion is severe. The final result obtained through weighted fusion is the fused phase angle and frequency, thus ensuring stable and reliable output even under complex power quality conditions.

[0025] The fused phase angles are then input into the logical phase mapping and sequence determination unit. This unit maps the physical terminals of the sampling channel to logical phases A, B, and C, and determines the phase sequence type based on the phase angle difference. Under normal circumstances, the phase angle difference between the three phases should be close to 120°. When the determination result is consistent across multiple consecutive analysis windows, the phase sequence is confirmed as either positive or negative. If the three-phase voltage amplitude is insufficient to reach the set threshold, or the quality index is below the threshold, the system will automatically call the symmetrical component method for phase angle compensation. The symmetrical component method decomposes the three-phase voltage into positive, negative, and zero-sequence components, and achieves a complete judgment by reconstructing the phase angle of the missing phase. This method is particularly suitable for single-phase voltage loss or severe imbalance conditions, ensuring uninterrupted judgment. To avoid misjudgments caused by transient interference, the system stipulates that the final phase sequence identification result is only output when the results are consistent across multiple consecutive analysis windows.

[0026] Ultimately, the system outputs not just a single phase sequence conclusion, but a structured set of information. This information includes the phase sequence type, calculated confidence level, current frequency, phase angle difference between each phase, timestamp of the operation, and version number of the sampling configuration. The confidence level is a numerical value based on a comprehensive evaluation of signal quality indicators, result consistency, and the number of analysis windows, used to measure the reliability of the current phase sequence identification. The timestamp and configuration version number ensure the traceability of the results, facilitating post-analysis and recording.

[0027] Step S102: Generate opening and closing control commands using the phase sequence identification results and output them to the spring operating mechanism.

[0028] In step S102, after obtaining a stable and reliable phase sequence identification result, the system needs to convert the identification result into a closing and opening control command that can directly drive the actuator. This process is first completed by the control logic module inside the integrated protection relay, which compares the phase sequence judgment value with the preset operation strategy. For example, when the phase sequence identification result is positive and the voltage amplitude is within the allowable range, the control logic generates a closing permission signal; when it is determined to be reverse sequence or a phase voltage is missing, a closing or prohibiting closing signal is generated, thereby ensuring the correctness and safety of subsequent actions at the logic level. This logic signal is then converted into a standardized electrical control signal by the pulse drive circuit inside the relay. This signal is usually a DC voltage pulse or an isolated AC control pulse, with a voltage level generally between 24V and 220V, depending on the rated control voltage of the spring operating mechanism configured on site.

[0029] To ensure reliable transmission of control commands in noisy environments, the output port is equipped with opto-isolation or magnetic isolation modules and surge absorption circuits to prevent transient interference from the power system from damaging the control logic. During pulse generation, the pulse width is strictly controlled within the time range required for the operating mechanism to operate, typically between 50 and 200 milliseconds. This ensures reliable operation while preventing excessively long pulses from causing the operating mechanism coil to overheat. For states requiring continuous maintenance, such as prolonged closed-circuit holding, the control logic employs a combination of a pilot pulse and mechanical self-holding. After the initial pulse drive is completed, the mechanical locking mechanism of the operating mechanism maintains the switch position, thereby reducing power consumption and extending system lifespan.

[0030] In actual operation, to prevent misoperation, the control logic incorporates interlocking and delay mechanisms. When it detects that a previous operation has not been completed, or the system is in a state that does not meet safety conditions, even if the phase sequence identification result is correct, a closing command will not be generated immediately. Instead, a prohibition signal will be output and the event will be recorded internally. Control command output is only allowed after all preconditions are met. Furthermore, for potential jitter signals or short-term instability in phase sequence identification, the system employs a multi-cycle confirmation mechanism. Command generation is only triggered when the identification results are consistent over several consecutive cycles, thereby significantly reducing the probability of false triggering.

[0031] During the output to the spring operating mechanism, the control signal is transmitted to the actuator through the terminal block or communication interface. If a hard-wired mode is used, the signal is directly connected to the control coil of the operating mechanism. If an intelligent operating mechanism is used, digital control commands are sent via industrial bus protocols such as Modbus or CAN bus, and then decoded and executed internally by the operating mechanism. Regardless of the mode used, the system monitors the current or voltage feedback of the output circuit in real time to confirm whether the command has been successfully delivered, and reissues the command through redundant channels if necessary, thereby ensuring reliable execution of the action.

[0032] In summary, the implementation process of step S102 starts with the logical judgment of the phase sequence identification result, goes through the conversion and isolation of electrical signals, and is finally output to the spring operating mechanism in the form of standardized and reliable pulses or communication commands. The whole process has a strict interlocking, delay and redundancy mechanism, which can ensure the accuracy and safety of opening and closing operations.

[0033] Furthermore, the step of generating opening and closing control commands using the phase sequence identification results and outputting them to the spring operating mechanism includes: Based on the phase sequence identification results, the three-phase voltage amplitude, phase angle difference and frequency stability are jointly calculated to generate a dynamic safety factor, and the dynamic safety factor is used as a prerequisite for opening and closing permits. When the dynamic safety factor meets the threshold condition, the phase sequence identification result is compared with the historical operating status to generate the opening and closing action judgment result, and the judgment data with confidence score is output. The judgment data is input into the pulse modulation unit, and the pulse amplitude, pulse width and pulse duration are adaptively adjusted according to the confidence score to form a modulated opening and closing drive pulse signal; After the opening and closing drive pulse signal is generated, the voltage level of the opening and closing drive pulse signal is matched and electromagnetic interference is suppressed by the isolation drive module, and finally output to the spring operating mechanism as a control command to drive it to perform opening and closing actions.

[0034] In this embodiment, the process of generating opening and closing control commands using the phase sequence identification results and outputting them to the spring operating mechanism is not a simple direct triggering, but rather involves multi-level calculations, comparisons, modulation, and isolation to ensure the safety, reliability, and controllability of the entire opening and closing process. First, after obtaining the phase sequence identification results, the system needs to perform joint calculations with the three-phase voltage amplitude, phase angle difference, and frequency stability. The three-phase voltage amplitude can be obtained by real-time sampling of the effective values ​​of each phase voltage, typically using a sliding window root mean square calculation method. The phase angle difference is obtained by comparing the fundamental phase angles of each phase in the phase sequence identification results, while frequency stability can be characterized by statistically analyzing the frequency fluctuation amplitude over a certain time range, for example, recording the maximum and minimum frequency values ​​within a second-level time window and calculating the difference. The joint calculation method can be a weighted summation model, where voltage amplitude deviation, phase angle difference offset, and frequency fluctuation are standardized and assigned different weights, ultimately generating a dynamic safety factor. This dynamic safety factor is used to measure whether the current power grid condition is suitable for opening and closing operations. If the factor is lower than the set threshold, the system will refuse to continue the operation, thereby avoiding malfunctions when the power quality is unstable.

[0035] After the dynamic safety factor meets the threshold condition, the system further compares the current phase sequence identification result with the historical operating status. The historical operating status is a database formed by the system's recorded fault status signals, opening and closing operation logs, and power quality data. This database reflects the typical operating characteristics of the equipment over a period of time. During the comparison process, the system evaluates whether the current phase sequence and fault characteristics are consistent with the history by retrieving historical opening and closing results under the same or similar power grid conditions. When the comparison consistency is high, the system generates an opening and closing action judgment result and adds a confidence score to this result. The confidence score is usually calculated using Bayesian inference or statistical matching methods, using the success rate under similar conditions in historical data as a basis, and adjusting the score range in combination with the current dynamic safety factor. For example, if the dynamic safety factor is close to the upper limit and the success rate of actions under similar conditions in historical data is greater than 95%, the confidence score can reach a value close to 1; if the dynamic safety factor has just reached the threshold and there are abnormal records in historical operating conditions, the confidence score will be significantly reduced.

[0036] After generating judgment data with confidence scores, the system inputs this data into the pulse modulation unit. The function of the pulse modulation unit is to adaptively adjust the parameters of the control pulse according to different confidence levels. The pulse amplitude typically determines the current in the trigger coil, the pulse width determines the duration of the trigger signal, and the pulse duration relates to whether the actuator can completely complete the opening and closing action. When the confidence score is high, the system can choose to reduce the pulse amplitude and width to reduce energy consumption and mechanical shock; while when the confidence score is low but still within the allowable range, the system will increase the pulse amplitude and lengthen the pulse width to improve the reliability of the action. This adaptive modulation method ensures that the control pulse meets the execution requirements while extending the equipment lifespan. For example, in a 220V rated voltage environment, if the confidence score is 0.95, the pulse amplitude may be set to 80% of the rated value, and the pulse width to 50 milliseconds; if the confidence score is 0.75, the pulse amplitude will be increased to 100% of the rated value, and the pulse width will be extended to 100 milliseconds to ensure reliable completion of the action.

[0037] After the modulated opening and closing drive pulse signal is generated, the signal needs to be further processed by the isolation drive module. The isolation drive module first uses opto-isolation or magnetic coupling to electrically isolate the control logic from the high-voltage execution section to prevent high-voltage interference from flowing back into the low-voltage control circuit. Secondly, the module matches the voltage level of the pulse signal, for example, boosting the 24V DC signal of the control logic to a suitable 220V AC or 110V DC signal for the spring operating mechanism. Finally, the isolation drive module also performs electromagnetic interference suppression, typically by adding an RC buffer network, surge absorption circuit, or common-mode filter at the output to ensure that the drive signal maintains waveform integrity even in environments with severe electromagnetic noise. After this series of processing steps, the final opening and closing control command is stably output to the spring operating mechanism as the direct input to drive its opening and closing actions.

[0038] Through the entire process described above, starting from the phase sequence identification results, the system generates a dynamic safety factor, compares historical states and scores confidence levels, optimizes pulse modulation, and implements isolation drive protection to ultimately form accurate and reliable opening and closing control commands. This process not only ensures the safety and reliability of low-voltage bypass operation but also significantly improves the system's adaptability to complex power grid environments through dynamic adaptive mechanisms and multiple verification logic.

[0039] Step S103: The spring operating mechanism executes the opening and closing control command and outputs the disconnection status signal.

[0040] In step S103, after receiving the opening and closing control command generated in the previous step, the spring operating mechanism needs to complete the mechanical action and output the result as a disconnection status signal. The spring operating mechanism is essentially an energy storage type actuator, typically consisting of an energy storage spring, transmission gears, connecting rods, opening and closing iron cores, a reset device, and a position detection component. After the control command is input, the drive coil is energized to release the mechanical energy stored in the energy storage spring, which then rapidly pushes the moving contact through the transmission gears and connecting rods to achieve the opening and closing operation. The entire process needs to be completed within a timescale of tens to hundreds of milliseconds to ensure the speed and reliability of the low-voltage circuit switching process.

[0041] To ensure the determinism of the action, the operating mechanism needs to confirm its energy storage state before the action begins, meaning the spring is in a compressed and locked state. Once a control signal is received, the trigger release device will quickly drive the contact to change position. After the action is completed, the final position of the contact will be captured by a mechanical limit switch or photoelectric detection device, forming a break status signal. This signal can be in the form of a switching quantity, such as "0" representing open and "1" representing closed, or it can be an analog signal representing the contact opening distance or the percentage of travel to the closing position, thus reflecting the precise position of the break status.

[0042] To avoid misjudgments caused by mechanical vibration, the break-point status signal needs to undergo anti-jitter processing, typically using time-delay confirmation or multiple consecutive sampling for consistency. For example, a valid result is only output when the detected contact maintains the same state for 50 consecutive milliseconds. This method can filter out interference signals caused by momentary bounce. If the operating mechanism fails to complete the action within the specified time, the system will output an "abnormal action" or "status unconfirmed" signal and generate an event log internally for subsequent diagnosis and maintenance.

[0043] In practical designs, the output of the break status signal is typically achieved through auxiliary contact groups. These auxiliary contacts are mechanically linked to the main contacts; when the main contacts open or close, the auxiliary contacts switch synchronously and transmit the status signal to the relay's input interface. For scenarios requiring higher precision, non-contact position detection methods can be used, such as employing Hall effect sensors, fiber optic sensors, or laser displacement sensors to directly monitor the displacement of the moving contact, thereby obtaining more accurate and real-time break status feedback.

[0044] Furthermore, to enhance system safety and reliability, the break-state signal needs to provide not only instantaneous status but also continuous monitoring capabilities. For example, in the open state, monitoring the insulation resistance or micro-current leakage between contacts further confirms reliable isolation; in the closed state, monitoring the contact clamping force or closing current waveform determines contact stability and reliability. These in-depth status monitoring signals can be acquired by sensors and superimposed on the break-state signal output, enabling subsequent logic control to make decisions based on richer data sources.

[0045] In summary, step S103 completes the mechanical action of opening and closing the circuit through the spring operating mechanism, and outputs the break status signal in real time using mechanical auxiliary contacts, optical sensors, position sensors or non-contact detection methods. After the signal is processed by the anti-jitter and confirmation mechanism, a reliable feedback is formed to ensure that the system can accurately grasp the real status of the circuit break, thereby providing solid data support for subsequent long-distance imaging display and logic judgment.

[0046] Furthermore, the spring operating mechanism executes the opening and closing control command and outputs a disconnection status signal, including: Upon receiving the opening and closing control command, the energy storage release unit is triggered, which transmits the mechanical energy in the energy storage spring to the moving contact through gear transmission and linkage mechanism, generates contact displacement trajectory data, and uses the contact displacement trajectory data as action feedback; Based on the contact displacement trajectory data, the contact velocity curve and displacement curve are calculated, and the contact action integrity parameters are generated by combining the contact pressure collected by the pressure sensor. The contact action integrity parameters are used as the criteria for determining the effectiveness of the action. When the contact action integrity parameter meets the preset range, the electrical signal output by the auxiliary contact and the fiber optic position sensor signal are cross-checked to generate break position confirmation data, and the break position confirmation data is fused with the contact action integrity parameter. The fusion result is input into the anti-jitter logic unit, and consistency verification is performed within a continuous time window to eliminate mechanical bounce signals, ultimately forming a break status signal, which is then output.

[0047] In this embodiment, after receiving the opening and closing control command, the spring operating mechanism's operation involves not only the execution of mechanical actions but also the acquisition of action trajectories, parameter calculation, state cross-verification, and signal anti-interference processing to ensure that the output break-point status signal is authentic, reliable, and traceable. First, when the opening and closing control command reaches the operating mechanism, the energy storage release unit is triggered, and the internal energy storage spring releases the pre-accumulated mechanical energy. This energy is transmitted to the moving contact through gear transmission and linkage mechanism, causing the contact to rapidly displace and complete the opening and closing action. During this process, the displacement sensor configured in the operating mechanism collects the displacement changes of the contact in real time throughout the entire action, forming contact displacement trajectory data. This trajectory data can completely reflect the dynamic process from command input to contact positioning and is returned to the control logic as an action feedback signal.

[0048] After obtaining the contact displacement trajectory data, the system further analyzes the data to calculate the contact's velocity and displacement curves. The velocity curve is typically calculated based on the time derivative of the trajectory data, while the displacement curve is a smooth curve obtained by continuously fitting the collected data points. These two types of curves reveal the contact's motion characteristics during operation, such as whether jamming or abnormal delay occurs. Simultaneously, the pressure sensor collects the contact pressure when the contact is fully closed and combines this pressure signal with the displacement and velocity curves to form contact action integrity parameters. These integrity parameters are a comprehensive measure of the degree of completion and reliability of the action, including multiple indicators such as whether the contact movement reaches the predetermined stroke, whether the speed is within the normal range, and whether the pressure meets the contact requirements. These integrity parameters serve as the primary basis for determining the effectiveness of the action. If the parameters exceed the preset range, it means that the action may have a fault or risk, and the system will determine that the action is invalid.

[0049] When the contact integrity parameters meet the preset range, the system does not immediately confirm the break state, but instead enters a multi-source signal cross-verification stage. In this stage, the auxiliary contact outputs an electrical signal indicating that the main contact has completed opening or closing; simultaneously, the fiber optic position sensor performs high-precision monitoring of the contact's spatial position. These two signals represent different dimensions of electrical state and physical position, respectively. By comparing their consistency, the accuracy of the judgment can be significantly improved. For example, if the auxiliary contact indicates closing but the fiber optic sensor indicates the contact is not fully closed, the system will mark it as abnormal to avoid erroneously outputting a closing state. After cross-verification, the system generates break position confirmation data, which includes the contact's geometric position, the auxiliary contact's electrical signal, and the consistency judgment result. This confirmation data is fused with the previously obtained contact integrity parameters to form a more complete and reliable basis for break state determination.

[0050] Finally, to address the common contact bounce phenomenon during mechanical device operation, the system inputs the above fusion results into the anti-jitter logic unit. Bounce refers to the short-term jitter caused by mechanical inertia or elasticity at the moment of contact or separation. This phenomenon manifests as high-frequency fluctuations at the millisecond level in the signal, easily leading to misjudgments. The anti-jitter logic unit sets a continuous time window, statistically analyzes the signal stability within the window, and requires the results to remain consistent across multiple consecutive sampling periods before finally confirming the state. For example, the system can set a 50-millisecond time window. If the data confirming the break location and the fusion result of the integrity parameters remain consistent within this window, the state is considered valid; otherwise, it is considered unstable and the output is delayed. After this processing, the system finally generates a break status signal, which is then used as the final output for subsequent imaging display and control logic.

[0051] Through this series of steps, the spring operating mechanism not only completes the mechanical action of opening and closing the circuit breaker, but also ensures the accuracy and robustness of the break-out status signal through displacement trajectory acquisition, speed and pressure calculation, cross signal verification, and anti-jitter processing.

[0052] Furthermore, when the contact action integrity parameter meets the preset range, the electrical signal output by the auxiliary contact and the fiber optic position sensor signal are cross-checked to generate break position confirmation data, and the break position confirmation data is fused with the contact action integrity parameter, including: The electrical signal output from the auxiliary contact is aligned with the signal from the fiber optic position sensor based on a time reference, and de-jitter filtering and noise suppression processing are performed respectively to generate an aligned signal sequence. Based on the aligned signal sequence, the event edge and displacement threshold crossing time are extracted to obtain a candidate event set, which is then used as the input for judgment. The candidate event set is checked for consistency with the contact action integrity parameters. Based on the displacement range, velocity range and contact pressure conditions in the contact action integrity parameters, fracture location confirmation data is generated. The data confirming the fracture location is weighted and fused with the parameters for the integrity of the contact action, and the fusion result is output.

[0053] In this embodiment, when the contact action integrity parameter meets the preset range, the process of cross-checking the electrical signal output by the auxiliary contact with the fiber optic position sensor signal involves signal synchronization, noise suppression, event extraction, consistency verification, and final data fusion. Its purpose is to ensure the accuracy and robustness of the break state determination. First, the electrical signal output by the auxiliary contact and the fiber optic position sensor signal need to be aligned on the same time reference. Since the two types of signals originate from different sources, their sampling frequencies and response delays also differ; direct comparison may lead to incorrect judgment. Therefore, the system uses a unified clock reference, timestamps both signals, and adjusts them to a unified time axis using interpolation algorithms or resampling techniques. For example, if the auxiliary contact signal sampling frequency is 1 kHz and the fiber optic sensor signal sampling frequency is 5 kHz, the fiber optic signal can be downsampled by an integer multiple or the auxiliary contact signal frequency can be increased through interpolation, thereby ensuring that a corresponding signal value can be found at every time point. At the same time, the auxiliary contact electrical signal is subjected to de-jitter filtering to eliminate the high-frequency jitter caused by the bouncing of the mechanical contacts during switching; the fiber optic position sensor signal is subjected to noise suppression processing, which can be done by using a low-pass filter or moving average method, thereby reducing the influence of environmental interference and random noise, and finally obtaining the aligned signal sequence.

[0054] After obtaining the aligned signal sequence, the system performs event extraction, i.e., identifies key state change points. For auxiliary contact signals, the most important are the conduction and disconnection edges, which represent changes in the electrical state of the contacts. These can be identified by monitoring voltage or current signal transitions at threshold points. For example, when a voltage signal suddenly changes from a high level to a low level, its timestamp is recorded as a disconnection event. For fiber optic position sensor signals, it is necessary to detect when the contact displacement signal crosses a certain threshold, which is usually a preset physical open or closed distance. When the displacement value detected by the fiber optic reaches or exceeds this threshold, the timestamp is recorded as a displacement crossing event. In this way, a candidate event set can be formed, where each entry contains an electrical state change and a displacement threshold crossing time. The candidate event set serves as input for subsequent judgments, providing basic data for subsequent consistency checks.

[0055] During the consistency verification phase, the candidate event set is compared with the contact action integrity parameters. These parameters are calculated based on the contact displacement trajectory, velocity curve, and contact pressure, and are used to evaluate whether the contact action meets expectations. For example, if the contact velocity remains within a reasonable range during closing, the contact pressure reaches a preset lower limit, and the displacement curve shows the contact is fully in position, the integrity parameters are considered valid. In the consistency verification, the system compares the timestamps of the candidate event set with the timing of the integrity parameters. For example, the time when the auxiliary contact shows closure should not differ from the time when the fiber optic displacement crosses the closure threshold by more than a set allowable error (e.g., 2 milliseconds); otherwise, it is considered inconsistent. Furthermore, if the contact action integrity parameters indicate that the contact is not fully in position, but the auxiliary contact is already conducting, the system will determine that a potential fault exists. In this way, the system can generate break location confirmation data, which integrates the matching of electrical and optical signals and clearly indicates the actual state of the break.

[0056] In the final step, the system needs to perform a weighted fusion of the break location confirmation data and the contact action integrity parameters. This weighted fusion process is typically based on confidence level allocation. The confidence level of the auxiliary contact electrical signal mainly comes from the determinism of its switching logic, but is limited by the mechanical bounce of the contact; the confidence level of the fiber optic position sensor signal depends on its optical detection resolution and anti-interference capability. The system assigns weights to both based on actual operating conditions. For example, in environments with severe electromagnetic interference, the weight of the fiber optic signal is increased; in environments with poor lighting or high noise from the fiber optic sensor, the weight of the auxiliary contact signal is increased. The fusion method can employ a weighted average or Kalman filter algorithm to jointly calculate the two types of signals with the integrity parameters. For example, if the confidence level of the auxiliary contact determining closure is 0.8, the confidence level of the fiber optic sensor determining closure is 0.9, and the confidence level of the integrity parameters indicating that the contact is in position is 0.95, then the final fusion result's confidence level may reach above 0.9. The fusion result not only provides the state of whether the break is closed or open, but also provides the specific value of the opening distance and the reliability of the determination.

[0057] Through the above series of processes, the system ensures that the determination of the break state does not depend on a single signal, but is confirmed by multiple verifications of electrical and optical parameters combined with integrity parameters, thereby avoiding misjudgments caused by single-point failures.

[0058] Step S104: The fracture status signal is displayed remotely through a large-window insulation structure to obtain imaging confirmation results.

[0059] In step S104, the system needs to visualize the fracture status signal output by the spring operating mechanism through long-distance imaging, allowing operators or the upper-level monitoring system to intuitively confirm the actual state of the fracture. To achieve this, an optical acquisition module is first installed near the switch fracture location. This module can use a high-definition industrial-grade camera, CCD sensor, or CMOS imaging sensor, and is fixed with a voltage-resistant insulating bracket to ensure stable operation even in high electric field environments. The lens area of ​​the acquisition module is isolated by a large-window insulating structure, which is typically made of transparent insulating material, such as polycarbonate or tempered glass, and is treated for UV resistance, arc resistance, and high temperature resistance to ensure that it will not discolor, crack, or decrease in light transmittance during long-term operation. The design of the large window needs to meet certain viewing angles and field depths, so that even when operators are several meters away from the equipment, they can clearly see the opening or closing state of the fracture.

[0060] The imaging acquisition module captures the physical state of the fracture surface in real time during operation, converting the opening gap, contact position, or indicator operation into image signals. To avoid recognition difficulties caused by insufficient ambient light, LED supplementary lights are usually installed around the viewing window. The color temperature and brightness of the light source can be automatically adjusted to adapt to different indoor and outdoor lighting conditions. The imaging signal undergoes image enhancement via a built-in processing unit, including brightness equalization, edge enhancement, and noise filtering to ensure the clarity of the fracture edges. The processed image is displayed on the monitoring terminal or relay panel, and fracture status labeling information can be superimposed, such as using color indicators to show "open" or "closed," further reducing ambiguity in manual judgment.

[0061] To improve the accuracy of imaging confirmation, the system not only relies on single-frame images but also employs continuous frame comparison and dynamic detection mechanisms. When a change in the break position from closed to open or from open to closed is detected, the imaging processing unit analyzes the continuity and speed of the contact position change to confirm whether the action has been truly completed, avoiding misjudgments caused by jitter or temporary obstruction. Simultaneously, the imaging confirmation result is cross-verified with the electrical signal from the auxiliary contact. If inconsistencies are found, the system marks the state as "abnormal" and issues a prompt at the monitoring terminal, reminding the operator to conduct further checks.

[0062] The output of long-distance imaging displays can be displayed on a local panel, transmitted via image transmission from a remote monitoring terminal, or wirelessly viewed on a mobile device. Through a network communication module, imaging confirmation results can be uploaded to the power distribution dispatch center in real time, enabling cross-regional remote monitoring and unified management. In high-reliability applications, imaging results can also be stored as historical image data for fault retrospective analysis and operational auditing, ensuring that every change in the state of a fault is traceable.

[0063] In summary, step S104 ensures the safety and visibility of fracture imaging through a large-window insulation structure, transforms the fracture state into intuitive and reliable image information using optical acquisition and imaging processing technology, and ensures the accuracy of the imaging confirmation results through real-time display, dynamic detection, and cross-verification mechanisms.

[0064] Furthermore, the step of performing long-distance imaging display of the fracture state signal through a large-window insulation structure to obtain imaging confirmation results includes: A high-resolution imaging sensor is set within the field of view of the large window insulation structure to capture images of the contact position corresponding to the break state signal and output raw imaging data. Illumination compensation and geometric distortion correction are performed on the original imaging data to generate a corrected image sequence, and the image sequence is transmitted to the fracture area detection module; The fracture edge contour is extracted in the fracture area detection module, and the fracture opening distance parameter is calculated based on the edge contour to output the fracture geometric feature data. The geometric feature data of the fracture surface is input into the dynamic recognition unit, and a fracture surface state determination result is generated by combining it with the preset fracture surface state threshold. The result is then compared with the original imaging data to output fracture surface state confirmation data. The fracture status confirmation data is fused with the timestamp and imaging quality parameters to finally generate the imaging confirmation result.

[0065] In this embodiment, the process of remotely imaging and displaying the fracture state signal through a large-viewpoint insulation structure and obtaining imaging confirmation results involves not only image acquisition but also data correction, feature extraction, state determination, and multi-parameter fusion, thereby ensuring the accuracy and reliability of the imaging confirmation results. First, a high-resolution imaging sensor is arranged within the field of view of the large-viewpoint insulation structure. This sensor can be a CCD or CMOS chip, preferably with a resolution greater than 1 million pixels to ensure clear capture of the fracture contact details. The large-viewpoint insulation structure is made of transparent, high-strength materials, such as polycarbonate or tempered glass, and is treated with UV-resistant and arc-resistant coatings to ensure light transmittance and insulation under long-term electrical operating conditions. The imaging sensor directly acquires images of the physical location of the fracture contact through this structure, obtaining raw imaging data. This data includes complete image information of the contact shape, edge contour, and surrounding lighting conditions, and serves as input for subsequent processing.

[0066] After acquiring the raw imaging data, the system performs illumination compensation and geometric distortion correction. Illumination compensation is necessary because lighting conditions at the power distribution site vary significantly; for example, insufficient ambient light or biased light sources can cause blurred contact edges. Illumination compensation typically employs histogram equalization or adaptive brightness adjustment methods to achieve a more balanced image grayscale distribution and clearer edge contrast. Geometric distortion correction primarily targets barrel or pincushion distortion caused by the lens at different angles, using calibration parameters to geometrically correct the image. For example, by inputting the imaging sensor's intrinsic parameters and distortion coefficients into the system beforehand, a polynomial model is used to inversely calculate the pixel coordinates, resulting in a distortion-free corrected image. The processed image sequence accurately reflects the geometric position and opening distance of the contacts and is transmitted to the break area detection module.

[0067] In the fracture region detection module, the system performs edge detection on the corrected image, typically using the Canny or Sobel operators to extract the contact edge contours. By identifying the boundary lines on both sides of the contact and measuring their distance, the fracture opening distance parameter can be obtained. The fracture opening distance is a geometric quantity representing the spatial gap between the moving and stationary contacts, and is an important indicator for determining whether the break is complete. The system converts pixel distance into actual physical distance through the conversion relationship between pixel pitch and lens calibration coefficient. For example, if the imaging system's calibration coefficient is 0.05 mm / pixel, when the detected distance between two edges is 200 pixels, the corresponding actual fracture opening distance is 10 mm. The fracture geometric feature data obtained in this way can accurately characterize the fracture state and serve as input for the next step of judgment.

[0068] After the geometric feature data of the fracture surface is input into the dynamic recognition unit, the system generates a fracture status determination result by combining it with a preset fracture status threshold. The preset threshold is defined according to electrical safety requirements; for example, under rated voltage conditions, the fracture surface must be greater than 8 mm to be considered effectively disconnected. If the actual opening distance is less than this threshold, it is determined to be incompletely disconnected. The dynamic recognition unit not only makes judgments based on a single frame image but also performs consistency verification on multiple consecutive frames to avoid misjudgments caused by instantaneous temporal interference. At the same time, the determination result is compared with the original imaging data for consistency, such as checking whether the edge sharpness and illumination uniformity meet the requirements. If the image quality is poor, the determination result will be marked as low confidence. Finally, the system outputs fracture status confirmation data, which includes the opening distance value, status label (open or closed), and confidence score.

[0069] In the final step, the fracture status confirmation data is fused with timestamps and imaging quality parameters to generate an imaging confirmation result. The timestamp ensures that each imaging judgment corresponds to a unique operation moment, facilitating historical review and anomaly tracking. Imaging quality parameters, including illumination level, image sharpness, and noise level, are used to assess the reliability of the image judgment. By fusing this information, the system can simultaneously provide the judgment status and confidence level when generating the imaging confirmation result. For example, the imaging confirmation result can be expressed as "fracture in place, opening distance 10.2 mm, confidence level 0.98, time 10:23:15." This result not only clearly indicates the fracture status but also provides quantifiable data support for subsequent logic control.

[0070] In summary, this process, through high-resolution imaging, illumination and distortion correction, geometric feature extraction, dynamic judgment, and multi-parameter fusion, forms a complete fracture imaging confirmation method, ensuring that the output imaging confirmation results are authentic, accurate, and traceable.

[0071] Furthermore, the step of performing illumination compensation and geometric distortion correction on the original imaging data to generate a corrected image sequence, and then transmitting the image sequence to the fracture region detection module, includes: Based on the calibration board pattern and the frame geometry of the large window insulation structure, the dark current bias, pixel response non-uniformity coefficient and illumination field estimation map are calculated to generate illumination compensation parameters. Using the illumination compensation parameters, dark current subtraction, vignetting correction and adaptive histogram equalization are sequentially performed on the original imaging data to generate an illumination-compensated image sequence. Radial and tangential distortion corrections are performed on the illumination-compensated image sequence based on the in-camera distortion coefficients, and geometric correction is completed by combining the reference markers of the large-view-window insulation structure to generate a geometrically corrected image sequence. By identifying the corner points of the frame of the large window insulation structure, the homography matrix between the fracture plane and the imaging plane is estimated, and perspective shaping and imaging straightening are performed on the geometrically corrected image sequence to generate a corrected image sequence. The corrected image sequence is passed to the fracture region detection module, and the camera intrinsic parameters, distortion coefficients, and homography matrix are attached as correction metadata.

[0072] In this embodiment, the process of performing illumination compensation and geometric distortion correction on the original imaging data to generate a corrected image sequence, and then transmitting the image sequence to the fracture region detection module, is a crucial link in the imaging verification chain. Its core lies in the systematic correction of the acquired images using physical and mathematical methods, thereby ensuring the reliability and accuracy of subsequent fracture region detection. First, before imaging begins, the system uses the calibration plate pattern and the frame geometry of the large-viewpoint insulation structure to calculate the optical error parameters of the imaging system. Dark current bias refers to the fact that the imaging sensor still outputs a non-zero signal under no-light conditions, which leads to increased imaging noise. By acquiring dark-field images under no-light conditions, the bias value of each pixel can be obtained and used as the basis for subtraction in subsequent processing. The pixel response non-uniformity coefficient refers to the difference in sensitivity of different pixels to the same illumination conditions. It can be estimated by illuminating the calibration plate pattern with a uniform light source and comparing the output values. The illumination field estimation map is calculated from the brightness distribution of the frame geometry of the large-viewpoint insulation structure under uniform illumination, reflecting the non-uniformity of illumination distribution across the entire imaging plane. These data together constitute the illumination compensation parameters.

[0073] After obtaining the illumination compensation parameters, the system uses these parameters to process the raw imaging data step by step. Dark current subtraction is the first operation, which subtracts the corresponding dark current bias from the actual output value of each pixel to eliminate background noise. Vignetting correction addresses insufficient brightness caused by image attenuation at the lens edges. It enhances the pixel brightness in edge areas using an illumination field estimation map, making the overall image brightness more uniform. Finally, adaptive histogram equalization is performed to improve the local contrast of the image, ensuring that the edges of the broken edges remain clearly visible under different lighting conditions. The images processed in this sequence are called the illumination-compensated image sequence. These images have significantly reduced errors caused by sensor defects and uneven illumination.

[0074] Next, the system performs geometric distortion correction on the illumination-compensated image sequence based on camera intrinsic parameters and distortion coefficients. Camera intrinsic parameters typically include focal length, principal point coordinates, pixel scaling, etc., while distortion coefficients are mainly divided into radial distortion and tangential distortion. Radial distortion manifests as barrel or pincushion stretching distortion, while tangential distortion originates from image tilt caused by non-parallel lens assembly. By acquiring sample images in advance under the calibration board pattern, distortion coefficients k1, k2, k3 and tangential distortion parameters p1, p2 can be calculated. In actual processing, the coordinates of image pixels are back-projected onto the ideal imaging plane through the distortion model, thereby eliminating geometric distortion. To further improve correction accuracy, the system combines the reference markers on the large-window insulation structure to align the corrected image with the actual geometric border of the structure, forming a geometrically corrected image sequence.

[0075] To eliminate the influence of different perspective angles during imaging, the system also estimates the homography matrix between the fracture plane and the imaging plane by identifying the corner points of the frame of the large-view window insulation structure. The homography matrix is ​​a geometric transformation matrix that maps one plane to another, calculated by solving for the coordinates of four or more corresponding points. In this scheme, the corner points are used as corresponding points, ensuring both the consistency of geometric features and improving mapping accuracy. After obtaining the homography matrix, the system performs perspective shaping and image straightening on the geometrically corrected image sequence, ensuring that the proportion of the fracture area in the image remains consistent with its actual physical size. The result after this processing is the corrected image sequence, which not only has accurate geometric shape but also uniform illumination and clear edges.

[0076] Finally, the corrected image sequence is transmitted to the fracture region detection module along with correction metadata. This correction metadata includes camera intrinsic parameters, distortion coefficients, and a homography matrix, serving as the basis for converting pixel distances into actual physical distances in subsequent detection. For example, when a fracture edge is detected to be d pixels apart in the image, the actual fracture distance can be directly calculated using the calibration coefficients and homography matrix, thus avoiding measurement deviations caused by image distortion or perspective effects.

[0077] Through this complete process of illumination compensation and geometric distortion correction, the system realizes the conversion from raw imaging data to high-precision corrected image sequences, providing a solid foundation for subsequent fracture area detection.

[0078] Step S105: Generate status indication information based on the imaging confirmation result, and use the status indication information to make a judgment in the control logic. When the judgment condition is met, perform the low-voltage bypass closing operation.

[0079] In step S105, the system needs to further convert the fracture confirmation result obtained from the long-distance imaging display into status indication information that can participate in logical judgment, and utilize it in the control logic to ensure that the execution conditions of the low-voltage bypass closing operation are clear and strict. After being parsed by the processing unit, the imaging confirmation result is first mapped to a clear status label, such as "disconnected in place", "closed in place", or "status abnormal". This labeling result not only reflects the physical state of the fracture, but can also include a reliability score obtained from image analysis, such as the clarity of the image, lighting conditions, and the recognition degree of the fracture edge. This information is written into the status data packet for subsequent logical comprehensive judgment.

[0080] The generated status indication information is transmitted to the control logic processing module, which is typically located in the relay's main control unit. This module uses a combination of hardware circuitry and embedded software to ensure real-time performance and safety in the decision-making process. During the logic decision-making process, the system cross-compares the status indication information with the phase sequence identification results obtained from previous steps to ensure the correct circuit connection sequence and that the break state meets the conditions for permissible operation. For example, the system will only proceed to the next step of closing permission determination if the phase sequence identification result shows a positive sequence, the break imaging confirmation result shows a complete break, and the status tag is confirmed to be consistent by the system. If any condition is not met, the logic module will prevent the issuance of the closing command and output a "closing prohibited" status, while simultaneously storing the abnormal event in the historical record for traceability.

[0081] To prevent erroneous judgments due to a single abnormal data source, the control logic incorporates a redundancy verification mechanism when processing status indication information. This involves simultaneously comparing results from auxiliary contact signals or sensor signals. If the imaging confirmation is inconsistent with the electrical signal, the logic module remains in standby mode and requires manual intervention or secondary confirmation. This redundant judgment method significantly improves the system's adaptability to complex operating conditions and avoids the risk of erroneous closing due to a single point of failure.

[0082] When the judgment conditions are met, the control logic will automatically generate a closing permission signal and transmit it to the opening and closing execution module, enabling the low-voltage bypass closing operation to proceed smoothly. This process is not only a single judgment, but also continuously tracks the status indication information during the operation to ensure that the break status remains as expected throughout the entire contact closing process. If status information is lost, imaging is blurred, or the detection signal is abnormal during execution, the system will immediately interrupt the closing operation and trigger the protection program.

[0083] As can be seen, step S105 realizes a complete closed loop from fracture imaging confirmation to status indication information generation to control logic determination, ensuring that the preconditions for low-voltage bypass closing operation are strictly verified, and the execution results are monitored in real time during the operation.

[0084] Furthermore, the step of generating status indication information based on the imaging confirmation result, and using the status indication information for judgment in the control logic, and performing a low-voltage bypass closing operation when the judgment condition is met, includes: The imaging confirmation results are subjected to feature analysis to extract the fracture opening distance parameter and the contact edge clarity index, generate fracture geometric feature data, and use the fracture geometric feature data as input for subsequent processing; The geometric feature data of the fracture surface is fused with the timestamp and imaging quality parameters in the imaging confirmation result to generate preliminary status information containing the fracture surface location accuracy and image reliability, and the preliminary status information is transmitted to the multi-source verification module. In the multi-source verification module, the preliminary state information is cross-compared with the fracture state signal and the phase sequence identification result to generate comprehensive state judgment data, and the confidence score of the comprehensive state judgment data is calculated. The comprehensive status determination data and the confidence score are input into the control logic unit, and consistency verification is performed within multiple consecutive time windows. When the consistency verification meets the preset conditions, a signal to allow closing is output; when the consistency verification fails, a signal to prohibit closing is output. After outputting the closing permission signal, the status indication information is generated and stored in the historical record database, and the low-voltage bypass closing operation is executed by the control logic unit.

[0085] In this embodiment, the process of generating status indication information based on the imaging confirmation result and using this status indication information for judgment in the control logic, and executing the low-voltage bypass closing operation when the judgment condition is met, is a multi-level verification and step-by-step confirmation chain logic process, designed to ensure that the low-voltage bypass closing operation can be safely executed under any electrical and mechanical conditions. First, the system performs feature analysis on the imaging confirmation result. The imaging confirmation result includes image data of the contact break, position parameters, and imaging quality indicators. In the feature analysis stage, the system extracts the contact opening distance and contact edge sharpness using an edge detection algorithm. The opening distance parameter is a geometric quantity that directly reflects whether the contact is completely broken; its calculation method is to convert the pixel distance and lens calibration coefficient into the actual physical distance. For example, if the imaging system calibration coefficient is 0.05 mm / pixel and the detected edge spacing is 160 pixels, then the actual opening distance is 8 mm. The contact edge sharpness indicator reflects whether the imaging is affected by blurriness or insufficient lighting, and is usually measured by gradient strength or contrast evaluation functions; a higher value indicates more reliable edge recognition. The calculated fracture geometry data is used as input for subsequent logical processing.

[0086] Next, the system fuses the fracture geometry data with the timestamp and imaging quality parameters contained in the imaging confirmation results to generate preliminary status information. The timestamp ensures that each imaging judgment result precisely corresponds to a specific operation time, enabling event traceability. Imaging quality parameters typically include brightness uniformity, noise level, and distortion correction residuals, reflecting the reliability of the image judgment. The fusion method employs a weighted fusion algorithm, using fracture opening distance and edge sharpness as primary factors, and timestamp and imaging quality parameters as correction factors, thereby generating preliminary status information containing fracture location accuracy and image reliability. For example, if the opening distance parameter is 10 mm and the sharpness is high, while the imaging quality parameters are excellent, the reliability score can approach 1; if the image is blurry and the lighting is insufficient, even if the opening distance is acceptable, the reliability score will decrease accordingly. The generated preliminary status information is then transmitted to the multi-source verification module.

[0087] In the multi-source verification module, the system not only relies on imaging data but also cross-compares the preliminary state information with the fracture state signal and phase sequence identification results to generate comprehensive state judgment data. The fracture state signal originates from mechanical auxiliary contacts or fiber optic sensors, representing electrical and mechanical position feedback, while the phase sequence identification result ensures the correct connection sequence of the three-phase circuit. When the data from the three sources are consistent, the comprehensive state judgment data generated by the system has the highest reliability. If there are some inconsistencies, such as the fracture state signal indicating that the circuit is closed but the imaging data showing that the fracture is not closed, the comprehensive judgment data will output an anomaly and lower the confidence score. The confidence score is usually calculated using fuzzy logic or Bayesian inference models, using the reliability of each data source as input variables for joint inference. For example, if the reliability of the imaging data is 0.9, the reliability of the fracture signal is 0.8, and the reliability of the phase sequence identification result is 1.0, the final confidence score may be 0.85.

[0088] Subsequently, the comprehensive status judgment data and confidence score are input into the control logic unit. The control logic unit does not make an immediate decision but instead requires consistency verification to be performed over multiple consecutive time windows. Consistency verification refers to whether the judgment result remains stable within a set time period (e.g., 100 milliseconds to 500 milliseconds). If the consistency verification passes, meaning the judgment result is "allow closing" in all time windows, the system outputs a "allow closing" signal. If fluctuations or inconsistencies occur within the window—for example, the judgment is "allow" at some times and "prohibit" at others—the final result is a "prohibit closing" signal, thereby preventing potential erroneous operations.

[0089] Once the system outputs a closing permission signal, it generates a formal status indication and stores it in the historical record database. The database saves the imaging confirmation results, comprehensive judgment data, confidence scores, and final execution status for each operation, providing a basis for later fault tracing and operational optimization. Finally, the control logic unit triggers the low-voltage bypass closing operation based on the confirmation signal. Through this multi-verification and consistency verification mechanism, the system ensures that closing is only permitted when the fault location is reliable, the imaging clarity is sufficient, the phase sequence is correct, and all signals are highly consistent, thereby significantly improving the safety and reliability of low-voltage bypass control.

[0090] In summary, this process, through five stages—feature analysis, data fusion, multi-source cross-comparison, consistency verification, and historical record storage—forms a complete closed loop from perception and analysis to decision-making and execution. This enables the imaging confirmation results to be transformed into high-confidence status indication information and provides a solid logical guarantee for low-voltage bypass closing.

[0091] Furthermore, in the multi-source verification module, the preliminary state information is cross-compared with the fracture state signal and the phase sequence identification result to generate comprehensive state determination data, and the confidence score of the comprehensive state determination data is calculated, including: The preliminary state information, the fracture state signal, and the phase sequence identification result are synchronized with a unified timestamp to generate an aligned multi-source signal sequence. The aligned multi-source signal sequence is input into the consistency analysis unit, and the consistency error vector is calculated based on the difference between the preliminary state information, the fracture state signal and the phase sequence identification result. The consistency error vector is then used as the input for subsequent determination. The consistency error vector is compared with a preset threshold model to generate preliminary judgment data, and the preliminary judgment data is fused with the preliminary state information to generate a judgment result. The determination result is input into the dynamic weighted calculation unit, and weights are assigned based on the stability of the fracture state signal and the continuity of the phase sequence identification result. The determination result is then weighted to generate the comprehensive state determination data. The confidence score is calculated based on the comprehensive state determination data, and the comprehensive state determination data and the confidence score are output.

[0092] In this embodiment, the process of cross-comparing the preliminary state information with the fracture state signal and the phase sequence identification result in the multi-source verification module is accomplished through unified processing of data from multiple sources under the same time reference. The aim is to form comprehensive judgment data and provide quantifiable confidence scores for subsequent control logic. First, the preliminary state information, the fracture state signal, and the phase sequence identification result need to be synchronized with unified timestamps. Since the acquisition paths, sampling frequencies, and transmission delays of the three types of signals may differ, direct comparison would cause timing misalignment. Therefore, the system design incorporates a global clock source and embeds a timestamp marking mechanism in each signal acquisition channel. For example, the preliminary state information generated by the imaging system has a millisecond-level acquisition time, while the fracture state signal and phase sequence identification result are also marked with time points using hardware timers, and finally aligned within the central processing unit. The resulting multi-source signal sequence after alignment ensures mutual correspondence under the same time reference, thereby eliminating misjudgments caused by timing drift.

[0093] After generating the aligned multi-source signal sequence, the system inputs it into the consistency analysis unit. The core task of this unit is to calculate the differences between the preliminary state information, the fracture state signal, and the phase sequence identification result. These differences are not only numerical but also include multi-dimensional parameters such as trend, amplitude, and phase. For example, if the preliminary state information indicates a fracture opening distance of 9 mm, while the fracture state signal reflects a closed state, the difference manifests as an inconsistency between the physical ranging result and the electrical signal state. To quantitatively describe these differences, the system represents them as a consistency error vector. Elements of this vector can include displacement difference, state label difference, and phase sequence angle difference. For example, the displacement difference is obtained by subtracting the fracture opening distance from the image recognition from the mechanical displacement fed back by the sensor; the phase sequence angle difference is obtained by comparing the phase sequence identification result with the historical operating phase sequence. This consistency error vector not only includes differences in single parameters but also comprehensively reflects the coupling relationship between the three types of signals and serves as an important input for subsequent judgments.

[0094] After obtaining the consistency error vector, the system compares it with a preset threshold model. The threshold model refers to the allowable error range set based on long-term operational experience and safety regulations. For example, if the fracture displacement difference is less than 1 mm and the status labels are consistent, it is considered to meet safety requirements; if the phase sequence angle difference is within 2 degrees and the voltage amplitude difference is less than 2%, it is also considered to meet the consistency standard. When an element in the consistency error vector exceeds the threshold range, an anomaly flag is output. The comparison result forms preliminary judgment data, which not only includes "normal" or "abnormal" judgment labels but also records the source and magnitude of the difference. To enhance the robustness of the judgment, the system fuses the preliminary judgment data with the preliminary status information, for example, by combining the judgment labels with image confidence weighting to form a more stable judgment result. In this way, the judgment result not only inherits the rigor of the threshold comparison but also includes the quality information of the image data, avoiding misjudgment due to a single signal deviation.

[0095] After generating the judgment result, the system inputs it into a dynamic weighted calculation unit. This unit dynamically allocates weights based on the reliability and operating characteristics of different signals. The stability of the fracture state signal can be evaluated by jitter amplitude or continuity consistency, and the continuity of the phase sequence identification result can be confirmed by whether the phase sequence judgments of multiple consecutive cycles are consistent. If a certain type of signal exhibits high stability, for example, if the fracture state signal remains unchanged over multiple consecutive sampling cycles, its weight increases accordingly; if a certain type of signal fluctuates, for example, if the phase sequence identification result is inconsistent across multiple cycles, its weight decreases accordingly. In this way, the system weights the judgment result to generate the comprehensive state judgment data. This data is equivalent to a weighted average of signals from different sources, and can more accurately reflect the true state.

[0096] Finally, the system calculates the confidence score based on the comprehensive status determination data. The confidence score is a quantitative indicator, typically between 0 and 1; a higher value indicates a more reliable determination result. The calculation method can employ a weighted confidence model, for example, using the confidence level of each type of signal as a weight to normalize the comprehensive status determination data. For instance, if the weight of the break point status signal is 0.4, the weight of the phase sequence identification result is 0.3, and the weight of the preliminary status information is 0.3, and the three types of signals are highly consistent in their determination results, then the confidence score is close to 1. If one type of signal conflicts with the other two, the confidence score will significantly decrease, possibly below 0.6. Ultimately, the comprehensive status determination data and the confidence score are output and passed to the subsequent control logic unit, providing a quantitative decision-making basis for the execution of low-voltage bypass closing.

[0097] Through the detailed steps described above, this solution ensures that under complex operating conditions, multi-source signals can not only be effectively aligned and fused, but also that a confidence level evaluation can be given in a quantitative manner, thereby significantly improving the safety and reliability of low-voltage bypass sequence control.

[0098] In the above embodiments, a low-voltage bypass sequence control method based on phase sequence identification and fault state is provided. Correspondingly, this application also provides a low-voltage bypass sequence control device based on phase sequence identification and fault state. Please refer to... Figure 2 This is a schematic diagram of an embodiment of a low-voltage bypass sequence control device based on phase sequence identification and fault state according to this application. Since this embodiment, i.e., the second embodiment, is basically similar to the method embodiment, it is described simply; relevant details can be found in the description of the method embodiment. The system embodiment described below is merely illustrative.

[0099] The second embodiment of this application provides a low-voltage bypass sequence control device based on phase sequence identification and fault state, comprising: The acquisition unit 201 is used to acquire voltage waveform data of the low-voltage three-phase circuit through the integrated protection relay, and output the phase sequence identification result after sampling; The generation unit 202 is used to generate opening and closing control commands using the phase sequence identification results and output them to the spring operating mechanism; Execution unit 203 is used for the spring operating mechanism to execute the opening and closing control command and output the disconnection status signal; Display unit 204 is used to perform long-distance imaging display of the fracture status signal through a large window insulation structure to obtain imaging confirmation results; The determination unit 205 is used to generate status indication information based on the imaging confirmation result, and to make a determination using the status indication information in the control logic. When the determination condition is met, the low-voltage bypass closing operation is performed.

[0100] Although this application discloses preferred embodiments as described above, it is not intended to limit this application. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of this application. Therefore, the scope of protection of this application should be determined by the scope defined in the claims of this application.

Claims

1. A low-voltage bypass sequential control method based on phase sequence identification and fault state, characterized in that, include: The voltage waveform data of the low-voltage three-phase circuit is obtained by an integrated protection relay, and the phase sequence identification result is output after sampling. The phase sequence identification result is used to generate opening and closing control commands and output to the spring operating mechanism; The spring operating mechanism executes the opening and closing control command and outputs the disconnection status signal; The fracture status signal is displayed remotely by using a large-window insulation structure to obtain imaging confirmation results. Based on the imaging confirmation results, status indication information is generated, and the status indication information is used in the control logic for judgment. When the judgment conditions are met, the low-voltage bypass closing operation is performed.

2. The low-voltage bypass sequence control method based on phase sequence identification and fault state according to claim 1, characterized in that, The process of acquiring voltage waveform data of a low-voltage three-phase circuit through an integrated protection relay and outputting phase sequence identification results after sampling includes: In the integrated protection relay, the A, B and C phases of the low-voltage three-phase circuit are sampled synchronously in three channels. Based on the internal clock and the grid phase-locked reference, time calibration is performed to generate calibration parameters and interpolate the sampling time axis to obtain the calibrated voltage waveform data. The corrected voltage waveform data is preprocessed based on DC bias elimination and main harmonic component suppression to obtain the debiased voltage waveform. At the same time, the total harmonic distortion of the voltage and the voltage imbalance are calculated as quality indicators. Sliding discrete Fourier analysis and zero-crossing interpolation calculations are performed in parallel on the debiased voltage waveform to obtain the fundamental phase angle and zero-crossing time of each phase. Weighting coefficients are adaptively generated based on the quality index, and the phase angle estimation results are weighted and fused to obtain the fused phase angle and frequency. The phase sequence is determined by integrating the phase angle input logic phase mapping and sequence determination unit. When the three-phase voltage amplitude meets the threshold condition, the phase sequence type is determined based on the phase angle difference. When any phase voltage is missing or the quality index is lower than the threshold, the symmetrical component method is used for phase angle compensation. The consistency of results within multiple consecutive analysis windows is used as the confirmation condition. Finally, the phase sequence identification result is output and structured information including phase sequence type, confidence level, frequency, phase angle difference, timestamp, and sampling configuration version number is generated.

3. The low-voltage bypass sequence control method based on phase sequence identification and fault state according to claim 1, characterized in that, The step of generating opening and closing control commands using the phase sequence identification results and outputting them to the spring operating mechanism includes: Based on the phase sequence identification results, the three-phase voltage amplitude, phase angle difference and frequency stability are jointly calculated to generate a dynamic safety factor, and the dynamic safety factor is used as a prerequisite for opening and closing permits. When the dynamic safety factor meets the threshold condition, the phase sequence identification result is compared with the historical operating status to generate the opening and closing action judgment result, and the judgment data with confidence score is output. The judgment data is input into the pulse modulation unit, and the pulse amplitude, pulse width and pulse duration are adaptively adjusted according to the confidence score to form a modulated opening and closing drive pulse signal; After the opening and closing drive pulse signal is generated, the voltage level of the opening and closing drive pulse signal is matched and electromagnetic interference is suppressed by the isolation drive module, and finally output to the spring operating mechanism as a control command to drive it to perform opening and closing actions.

4. The low-voltage bypass sequence control method based on phase sequence identification and fault state according to claim 1, characterized in that, The spring operating mechanism executes the opening and closing control command and outputs the disconnection status signal, including: Upon receiving the opening and closing control command, the energy storage release unit is triggered, which transmits the mechanical energy in the energy storage spring to the moving contact through gear transmission and linkage mechanism, generates contact displacement trajectory data, and uses the contact displacement trajectory data as action feedback; Based on the contact displacement trajectory data, the contact velocity curve and displacement curve are calculated, and the contact action integrity parameters are generated by combining the contact pressure collected by the pressure sensor. The contact action integrity parameters are used as the criteria for determining the effectiveness of the action. When the contact action integrity parameter meets the preset range, the electrical signal output by the auxiliary contact and the fiber optic position sensor signal are cross-checked to generate break position confirmation data, and the break position confirmation data is fused with the contact action integrity parameter. The fusion result is input into the anti-jitter logic unit, and consistency verification is performed within a continuous time window to eliminate mechanical bounce signals, ultimately forming a break status signal, which is then output.

5. The low-voltage bypass sequence control method based on phase sequence identification and fault state according to claim 1, characterized in that, The step of performing long-distance imaging display of the fracture state signal through a large-viewing-window insulation structure to obtain imaging confirmation results includes: A high-resolution imaging sensor is set within the field of view of the large window insulation structure to capture images of the contact position corresponding to the break state signal and output raw imaging data. Illumination compensation and geometric distortion correction are performed on the original imaging data to generate a corrected image sequence, and the image sequence is transmitted to the fracture area detection module; The fracture edge contour is extracted in the fracture area detection module, and the fracture opening distance parameter is calculated based on the edge contour to output the fracture geometric feature data. The geometric feature data of the fracture surface is input into the dynamic recognition unit, and a fracture surface state determination result is generated by combining it with the preset fracture surface state threshold. The result is then compared with the original imaging data to output fracture surface state confirmation data. The fracture status confirmation data is fused with the timestamp and imaging quality parameters to finally generate the imaging confirmation result.

6. The low-voltage bypass sequence control method based on phase sequence identification and fault state according to claim 1, characterized in that, The process of generating status indication information based on the imaging confirmation result and using the status indication information for judgment in the control logic, and performing a low-voltage bypass closing operation when the judgment condition is met, includes: The imaging confirmation results are subjected to feature analysis to extract the fracture opening distance parameter and the contact edge clarity index, generate fracture geometric feature data, and use the fracture geometric feature data as input for subsequent processing; The geometric feature data of the fracture surface is fused with the timestamp and imaging quality parameters in the imaging confirmation result to generate preliminary status information containing the fracture surface location accuracy and image reliability, and the preliminary status information is transmitted to the multi-source verification module. In the multi-source verification module, the preliminary state information is cross-compared with the fracture state signal and the phase sequence identification result to generate comprehensive state judgment data, and the confidence score of the comprehensive state judgment data is calculated. The comprehensive status determination data and the confidence score are input into the control logic unit, and consistency verification is performed within multiple consecutive time windows. When the consistency verification meets the preset conditions, a signal to allow closing is output; when the consistency verification fails, a signal to prohibit closing is output. After outputting the closing permission signal, the status indication information is generated and stored in the historical record database, and the low-voltage bypass closing operation is executed by the control logic unit.

7. The low-voltage bypass sequence control method based on phase sequence identification and fault state according to claim 4, characterized in that, When the contact action integrity parameter meets the preset range, the electrical signal output by the auxiliary contact and the fiber optic position sensor signal are cross-checked to generate break position confirmation data. The break position confirmation data is then fused with the contact action integrity parameter, including: The electrical signal output from the auxiliary contact is aligned with the signal from the fiber optic position sensor based on a time reference, and de-jitter filtering and noise suppression processing are performed respectively to generate an aligned signal sequence. Based on the aligned signal sequence, the event edge and displacement threshold crossing time are extracted to obtain a candidate event set, which is then used as the input for judgment. The candidate event set is checked for consistency with the contact action integrity parameters. Based on the displacement range, velocity range and contact pressure conditions in the contact action integrity parameters, fracture location confirmation data is generated. The data confirming the fracture location is weighted and fused with the parameters for the integrity of the contact action, and the fusion result is output.

8. The low-voltage bypass sequence control method based on phase sequence identification and fault state according to claim 5, characterized in that, The step of performing illumination compensation and geometric distortion correction on the original imaging data to generate a corrected image sequence, and then transmitting the image sequence to the fracture region detection module, includes: Based on the calibration board pattern and the frame geometry of the large window insulation structure, the dark current bias, pixel response non-uniformity coefficient and illumination field estimation map are calculated to generate illumination compensation parameters. Using the illumination compensation parameters, dark current subtraction, vignetting correction and adaptive histogram equalization are sequentially performed on the original imaging data to generate an illumination-compensated image sequence. Radial and tangential distortion corrections are performed on the illumination-compensated image sequence based on the in-camera distortion coefficients, and geometric correction is completed by combining the reference markers of the large-view-window insulation structure to generate a geometrically corrected image sequence. By identifying the corner points of the frame of the large window insulation structure, the homography matrix between the fracture plane and the imaging plane is estimated, and perspective shaping and imaging straightening are performed on the geometrically corrected image sequence to generate a corrected image sequence. The corrected image sequence is passed to the fracture region detection module, and the camera intrinsic parameters, distortion coefficients, and homography matrix are attached as correction metadata.

9. The low-voltage bypass sequence control method based on phase sequence identification and fault state according to claim 6, characterized in that, In the multi-source verification module, the preliminary state information is cross-compared with the fracture state signal and the phase sequence identification result to generate comprehensive state determination data, and the confidence score of the comprehensive state determination data is calculated, including: The preliminary state information, the fracture state signal, and the phase sequence identification result are synchronized with a unified timestamp to generate an aligned multi-source signal sequence. The aligned multi-source signal sequence is input into the consistency analysis unit, and the consistency error vector is calculated based on the difference between the preliminary state information, the fracture state signal and the phase sequence identification result. The consistency error vector is then used as the input for subsequent determination. The consistency error vector is compared with a preset threshold model to generate preliminary judgment data, and the preliminary judgment data is fused with the preliminary state information to generate a judgment result. The determination result is input into the dynamic weighted calculation unit, and weights are assigned based on the stability of the fracture state signal and the continuity of the phase sequence identification result. The determination result is then weighted to generate the comprehensive state determination data. The confidence score is calculated based on the comprehensive state determination data, and the comprehensive state determination data and the confidence score are output.

10. A low-voltage bypass sequence control device based on phase sequence identification and fault state, characterized in that, include: The acquisition unit is used to acquire voltage waveform data of the low-voltage three-phase circuit through the integrated protection relay, and output the phase sequence identification result after sampling; The generation unit is used to generate opening and closing control commands based on the phase sequence identification results and output them to the spring operating mechanism. An execution unit is used for the spring operating mechanism to execute the opening and closing control commands and output the disconnection status signal; The display unit is used to perform long-distance imaging display of the fracture status signal through a large-window insulation structure to obtain imaging confirmation results. The determination unit is used to generate status indication information based on the imaging confirmation result, and to make a determination using the status indication information in the control logic. When the determination condition is met, the low-voltage bypass closing operation is executed.

Images