Reliable operation determination method of no-voltage release switch based on residual voltage characteristics of power grid power failure

Abstract

The application discloses a method for determining reliable action of a no-voltage release switch based on residual voltage characteristics of power grid power failure, which controls an analog-to-digital conversion module to run at a basic low-frequency sampling mode, maps and constructs three-phase alternating voltage into a three-dimensional instantaneous position vector; calls a discrete extended state observer algorithm to output smooth space speed and acceleration vectors, calculates instantaneous curvature and curvature rate; when the curvature rate is greater than a steady-state threshold, drives the underlying hardware to switch to a high-frequency sampling mode; extracts the jerk vector and calculates the space instantaneous torsion in the high-frequency mode; when the instantaneous curvature remains monotonically increasing and the space instantaneous torsion breaks through a boundary threshold, outputs a digital tripping instruction to a tripping actuator to cut off the physical connection. The application assists in distinguishing power grid fault temporary drop from residual voltage from the physical characteristic level, alleviates the determination blind area caused by the coincidence of traditional amplitude-frequency characteristics, and realizes fast feature capture and reliable action determination.

Description

Technical Field

[0001] This invention relates to the field of power system relay protection and automation control technology, specifically to a method for determining the reliable operation of a no-pressure release switch based on the residual voltage characteristics of a power grid after a power outage. Background Technology

[0002] Unloaded release switches are widely used safety protection devices in medium and low voltage power distribution networks and industrial power distribution systems. When a power grid fault causes the busbar to lose power or the voltage to drop, the unloaded release switch can drive the mechanical tripping mechanism through the monitoring and control unit to forcibly disconnect the physical connection of the main circuit. This prevents electrical equipment from being damaged due to continuous operation under low voltage conditions, or from experiencing asynchronous grid connection and unexpected startup when power is restored, thus ensuring the stable operation of the power distribution system.

[0003] In current industrial applications, the activation determination of undervoltage release switches mainly relies on traditional electrical characteristic detection technology. Conventional solutions typically use voltage transformers to continuously collect analog three-phase AC voltage data from the power grid branches, and employ digital signal processing algorithms such as Fast Fourier Transform to calculate the fundamental voltage RMS value and frequency change rate within a specific time window. When the monitoring and control unit detects that the RMS voltage value has fallen below the set undervoltage protection threshold, and the accompanying frequency change rate exceeds a preset boundary and remains there for a set delay, it determines that a power outage has occurred and then issues an isolation action command to the tripping actuator.

[0004] However, existing detection mechanisms have technical limitations in distribution branches containing inductive loads. When a real power outage occurs, the inductive load, driven by the mechanical inertia of its rotor, briefly transitions to generator operation, feeding residual voltage back to the bus. During the initial transient phase of the power outage, the amplitude and frequency of the residual voltage decay slowly, and the electrical characteristics exhibited by the residual voltage highly overlap with the voltage sag characteristics caused by a remote short-circuit fault in the power grid, even without a real power outage. Existing technology determines operation by comparing the effective voltage value and frequency with a set threshold, which cannot fundamentally distinguish between power grid fault sags and power outage residual voltage at the physical driving source. This results in a blind spot in the detection of voltage drop switches relying on effective values ​​and frequency, where characteristics overlap, potentially leading to misjudgment or failure to operate during a real power outage. Summary of the Invention

[0005] To address the problems mentioned in the background, this invention provides a method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid after a power outage. This method is executed by the underlying physical hardware system and includes the following steps:

[0006] Step S100: Perform basic low-frequency sampling and spatial vector construction, control the analog-to-digital conversion module to run at the basic low-frequency sampling rate, acquire the three-phase AC voltage discrete signal of the power grid branch, and map the three-phase AC voltage discrete signal to a mutually orthogonal three-dimensional coordinate system to construct the instantaneous position vector of the synthesized voltage spatial trajectory;

[0007] Step S200: Perform smooth reconstruction of the higher-order derivative of the spatial trajectory. Use the instantaneous position vector as the observation input to call the discrete extended state observer algorithm and output the smooth reconstructed spatial velocity vector and spatial acceleration vector.

[0008] Step S300: Perform real-time calculation of spatial differential geometric features, calculate the instantaneous curvature of the voltage space trajectory based on the spatial velocity vector and the spatial acceleration vector, and output the instantaneous curvature rate parameter based on the first-order difference derivative operation in the time domain.

[0009] Step S400: Perform the underlying hardware state switching of curvature variation drive, compare the absolute value of the instantaneous curvature variation parameter with the system preset steady-state wander threshold, and when the absolute value of the instantaneous curvature variation parameter is greater than the steady-state wander threshold, drive the analog-to-digital conversion module to switch to high-frequency sampling mode, and synchronously instruct the analog front-end processing circuit to increase the passband cutoff frequency of the anti-aliasing filter to broaden the passband bandwidth.

[0010] Step S500: Perform depth capture of high-frequency spatial topological features, update the spatial velocity vector and spatial acceleration vector in the high-frequency sampling mode, further extract the three-dimensional spatial acceleration vector, and perform a joint operation of vector dot product and cross product to calculate the instantaneous spatial torsion.

[0011] Step S600: Perform multi-dimensional feature comprehensive trip determination. Within the set high-frequency determination time window, when the instantaneous curvature remains monotonically increasing and the absolute value of the instantaneous spatial deflection exceeds the set deflection boundary threshold, output a digital trip command to the tripping actuator to cut off the physical connection of the main circuit.

[0012] Preferably, when performing basic low-frequency sampling and spatial vector construction, the instantaneous voltage values ​​of the three-phase stationary coordinate system at the same timestamp are input into a preset constant power Clark transform matrix calculation module, and the mutually orthogonal alpha axis components, beta axis components and zero sequence components are output; the above components are respectively assigned to the coordinate values ​​of three mutually perpendicular coordinate axes in the three-dimensional spatial rectangular coordinate system to synthesize the instantaneous position vector, and are synchronously written into the first-in-first-out data buffer to form a spatial position vector time series.

[0013] Preferably, when performing high-order derivative smoothing reconstruction of the spatial trajectory, an independent set of observer state variables is assigned to each axial voltage component. A first state variable is defined to track the actual position value, a second state variable to track the first-order derivative value, and a third state variable to track the second-order derivative and lumped disturbance value. Based on the observation position error at the current discrete sampling step and the preset feedback gain coefficient, discrete integral accumulation prediction is performed in conjunction with the working sampling period of the analog-to-digital conversion module to obtain the updated state variable value for the next discrete sampling step. The second state variable is extracted and synthesized into a spatial velocity vector, and the third state variable is extracted and synthesized into a spatial acceleration vector. The preset feedback gain coefficient is determined by configuring the bandwidth parameter of the discrete extended state observer, which is constantly configured as an integer multiple of the fundamental angular frequency of the power grid.

[0014] Preferably, when performing the low-level hardware state switching driven by curvature variation, a trigger command carrying high-frequency sampling configuration parameters is sent to the field-programmable gate array (FPGA) via a high-speed parallel data bus; the FPGA directly modifies the frequency division coefficient of the sampling rate control register inside the analog-to-digital converter module according to the trigger command, switching the sampling frequency to a high-frequency sampling frequency; at the same time, a control message is sent to the analog front-end processing circuit via the serial peripheral interface bus to switch the capacitor-resistor array of the programmable filter chip, adjusting the passband cutoff frequency of the anti-aliasing filter from narrowband to wideband mode.

[0015] Preferably, in the multi-dimensional feature comprehensive detangle determination, the instantaneous curvature value of the later time node in the sequence is strictly greater than the instantaneous curvature value of the previous time node and there is no numerical drop or zero-crossing mutation, satisfying the condition of continuous monotonous increase; the absolute value of the extracted spatial instantaneous torsion is greater than the torsion boundary threshold set according to the asymmetric component limit under the steady-state background of the system, satisfying the condition of spatial torsion disturbance exceeding the limit.

[0016] Preferably, the process of outputting a digital trip command to the trip actuator to cut off the physical connection of the main circuit is as follows: the results of the residual pressure curvature inertia criterion and the results of the spatial torsion disturbance limit exceeding criterion are converted into Boolean logic variables; the internal digital signal processor performs a Boolean logic AND operation on the two Boolean logic variables, and when the judgment conditions are met at the same time, the high-level trigger flag is written into the status configuration register of the general input / output pin to generate a digital trip command, which is output to the electrical control terminal of the trip actuator to drive the internal electromagnetic trip coil to be energized and the mechanical transmission linkage to move.

[0017] Preferably, when the set high-frequency judgment time window reaches the timing end, the logic judgment result is reviewed. If no digital tripping instruction is generated, or if the instantaneous curvature rate parameter is detected to fall back to the steady-state wandering threshold or below, the steady-state configuration word is rewritten to the analog-to-digital conversion module to restore the basic low-frequency sampling rate, and the analog front-end processing circuit is instructed to lower the passband cutoff frequency of the anti-aliasing filter to narrow the passband bandwidth, thus completing the state restoration of the underlying physical hardware system.

[0018] The present invention, by adopting the above technical solution, can bring the following beneficial effects:

[0019] 1. This invention constructs a position vector by mapping AC voltage to a three-dimensional coordinate system, calculates the spatial velocity vector and spatial acceleration vector to extract instantaneous curvature, and combines the jerk vector to calculate instantaneous deflection. The instantaneous curvature increases monotonically and the absolute value of the instantaneous deflection exceeds the deflection boundary threshold as the tripping criterion. This invention enables the auxiliary differentiation between grid fault sag and power loss residual voltage from the physical feature level, overcomes the judgment blind zone caused by the traditional reliance on effective value and frequency judgment when features overlap, and improves the reliability of the no-pressure release switch action judgment.

[0020] 2. This invention extracts the instantaneous curvature variation as an active feedback control quantity. When the absolute value of the instantaneous curvature variation is greater than the steady-state wander threshold, it drives the analog-to-digital conversion module to switch to high-frequency sampling mode and simultaneously instructs the analog front-end processing circuit to increase the passband cutoff frequency of the anti-aliasing filter. This achieves high-speed adaptive reconstruction of the sampling state and filtering bandwidth of the underlying physical hardware. Under the constraint of the high-frequency judgment time window, the system can quickly complete the high-frequency transient feature capture and action command output without relying on the traditional fixed delay criterion, thus meeting the stringent action time requirements of the circuit breaker's physical isolation.

[0021] 3. This invention uses the instantaneous position vector as the observation input to call the discrete extended state observer algorithm, and combines the discrete integral accumulation prediction with the sampling period of the analog-to-digital converter to extract the updated state variables and synthesize them into spatial velocity vector and spatial acceleration vector. This effectively suppresses the amplification effect of high-order derivative differential operation on high-frequency interference signals of the power grid while maintaining a small phase lag. It ensures that the high-order derivative input data required for spatial curvature and torsion calculation has a high signal-to-noise ratio and avoids system tripping malfunctions caused by complex electromagnetic interference. Attached Figure Description

[0022] Figure 1 This is a block diagram of the underlying physical hardware system architecture of the present invention;

[0023] Figure 2 This is a functional module and main flowchart of the pressure-free release switch reliable operation determination system of the present invention;

[0024] Figure 3This is a flowchart of the low-frequency sampling, state observation, and spatial feature extraction process of the present invention;

[0025] Figure 4 This is a flowchart of the hardware adaptive switching and multi-dimensional feature comprehensive determination process of the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Example 1

[0028] Please refer to the appendix. Figure 1 The present invention provides a reliable operation determination method for a no-pressure release switch based on the residual voltage characteristics of the power grid after power failure. The above method is executed based on a specific underlying physical hardware system, which includes a voltage sensing unit 11, an analog front-end processing circuit 12, an analog-to-digital conversion module 13, a core digital processor 14, and a tripping actuator 15.

[0029] The voltage sensing unit 11 is physically connected to the three-phase bus of the power grid branch to obtain the three-phase analog voltage on the primary side. The secondary side output terminal of the voltage sensing unit 11 is connected to the input terminal of the analog front-end processing circuit 12 through a hard wire.

[0030] The analog front-end processing circuit 12 consists of a programmable gain amplifier and a dynamically adjustable bandwidth anti-aliasing low-pass filter, which is used to perform amplitude conditioning on the three-phase analog voltage and filter out out-of-band high-frequency interference.

[0031] The analog-to-digital converter module 13 is connected to the output of the analog front-end processing circuit 12. To support the adaptive sampling technology, the analog-to-digital converter module 13 adopts an analog-to-digital converter with register dynamic configuration function. The analog-to-digital converter has an independent high-speed sampling control register. The clock division coefficient of the high-speed sampling control register is controlled by the external bus instructions issued by the core digital processor 14.

[0032] The core digital processor 14 adopts a collaborative architecture of digital signal processor and field-programmable gate array (FPGA) to execute calculation and control instructions for the no-pressure release action determination method. The FPGA is directly connected to the analog-to-digital converter (ADC) module 13 via a high-speed parallel data bus to manage the sampling timing of the underlying ADC and to allocate a first-in-first-out (FIFO) data buffer in the FPGA's internal random access memory. The digital signal processor interacts with the FPGA through an external memory interface, reads the discrete voltage sequence in the FIFO data buffer, and performs matrix operations and discretization of differential geometric equations. The general technical terms such as microprocessor and controller mentioned in the claims of this invention correspond to the specific physical hardware implementation of the aforementioned collaborative architecture of digital signal processor and FPGA in this specific embodiment.

[0033] The electrical control terminal of the trip actuator 15 is connected to the general-purpose input / output pin of the core digital processor 14, and the mechanical actuation terminal of the trip actuator 15 is connected to the main contacts of the circuit breaker. The trip actuator 15 includes an electromagnetic trip coil and a mechanical transmission link, which is used to disconnect the physical connection of the main circuit of the power grid branch after receiving a digital level signal.

[0034] Example 2

[0035] Please refer to the appendix. Figure 2 Based on Example 1, in order to support the calculation and closed-loop control of the differential geometric features of the spatial trajectory, the core digital processor 14 cooperates with the following functional modules when running firmware code, specifically including data acquisition module 101, state observation module 102, geometric parameter calculation module 103, adaptive trigger control module 104 and comprehensive judgment execution module 105.

[0036] The reliable operation determination method for no-voltage release switches based on the residual voltage characteristics of power grid failure provided by this invention is executed collaboratively by the underlying physical hardware system and its functional modules. The underlying physical hardware system monitors the spatial geometric topology evolution of transient signals of the power grid in real time, and controls the underlying hardware state based on logical closed-loop control to complete the operating condition identification and tripping isolation action. Specifically, it includes the following steps:

[0037] Step S100: Perform basic low-frequency sampling and spatial vector construction to realize spatial orthogonal mapping of transient voltage signals, transforming the one-dimensional time domain grid voltage signal into a continuous geometric trajectory in the three-dimensional spatial domain; Data acquisition module 101 issues low-level control commands to control analog-to-digital conversion module 13 to operate at the basic low-frequency sampling rate configured in steady state, and acquire the three-phase AC voltage discrete signals of the grid branch in real time; Core digital processor 14 uses orthogonal transformation matrix to map the three-phase AC voltage discrete signals to mutually orthogonal three-dimensional coordinate systems, extracts the component data of orthogonal axes, and constructs the instantaneous position vector of the synthesized voltage spatial trajectory;

[0038] Step S200: Perform high-order derivative smooth reconstruction of the spatial trajectory. The state reconstruction is achieved through the discrete extended state observer algorithm, avoiding the amplification of power grid background noise by direct differential calculation. The state observation module 102 receives the instantaneous position vector sequence. In order to overcome the amplification effect of digital differential operation on the high-frequency background noise of the power grid, the state observation module 102 calls the digital extended state observer algorithm. The digital extended state observer algorithm takes the instantaneous position vector as the observation input, performs iterative solution of the state matrix, and outputs the spatial velocity vector and spatial acceleration vector smoothly reconstructed on the time axis.

[0039] Step S300: Real-time calculation of spatial differential geometric features is performed, and transient geometric features are converted into hardware control instructions to realize dynamic reconstruction of sampling frequency and filtering bandwidth; the geometric parameter calculation module 103 retrieves the smoothed reconstructed spatial velocity vector and spatial acceleration vector, performs three-dimensional cross product operation and scalar modulus calculation of spatial vectors, obtains the instantaneous curvature of voltage space trajectory at the current discrete time node, and the core digital processor 14 performs time domain differentiation operation based on the curvature difference between adjacent time nodes, and outputs the instantaneous curvature rate parameter characterizing the three-dimensional topological distortion rate of the spatial trajectory;

[0040] Step S400: Perform the low-level hardware state switching driven by curvature variation. Extract the instantaneous spatial torsion in high-frequency sampling mode to characterize the spatial distortion of the voltage trajectory. The adaptive trigger control module 104 compares the absolute value of the instantaneous curvature variation with the system's preset steady-state wander threshold. When the absolute value of the instantaneous curvature variation is less than or equal to the steady-state wander threshold, the low-level physical hardware system maintains the basic low-frequency sampling state. When the absolute value of the instantaneous curvature variation is greater than the steady-state wander threshold, the adaptive trigger control module 104 determines that a topological change has occurred in the power grid. The core digital processor 14 sends a trigger command to the field programmable gate array through the internal data bus. The field programmable gate array rewrites the configuration word of the control register of the analog-to-digital converter module 13 according to the trigger command. This operation terminates the basic low-frequency sampling of the analog-to-digital converter module 13, drives the analog-to-digital converter module 13 to switch to high-frequency sampling mode, and synchronously instructs the analog front-end processing circuit 12 to increase the passband cutoff frequency of the anti-aliasing filter to broaden the passband bandwidth.

[0041] Step S500: Perform deep capture of high-frequency spatial topology features, perform logical judgment based on multi-dimensional geometric topology features and output hardware action instructions; After the analog-to-digital conversion module 13 enters the high-frequency sampling mode, it acquires a high-resolution instantaneous position vector sequence. The comprehensive judgment execution module 105 updates the spatial velocity vector and spatial acceleration vector synchronously according to the high-frequency sequence, and further extracts the jerk vector. Based on the spatial velocity vector, spatial acceleration vector and jerk vector, the comprehensive judgment execution module 105 performs a joint operation of vector dot product and cross product to calculate the instantaneous spatial torsion. The instantaneous torsion is used to characterize the degree of spatial distortion of the voltage trajectory due to the residual voltage discharge of the inductive load or the asymmetric disturbance of the stator, which causes it to leave the two-dimensional plane.

[0042] Step S600: Perform multi-dimensional feature comprehensive trip determination and system state restoration. Within the set high-frequency determination time window, the comprehensive determination execution module 105 continuously monitors the evolution trend of instantaneous curvature and the numerical state of instantaneous torsion. When the instantaneous curvature remains monotonically increasing and the absolute value of instantaneous torsion exceeds the set torsion boundary threshold, the comprehensive determination execution module 105 outputs an action level signal to the control pin of the trip execution mechanism 15 to drive the mechanical transmission link of the trip execution mechanism 15 to cut off the physical connection of the main circuit. If the two criteria are not met simultaneously within the determination time window, or if the transient disturbance of the power grid subsides and the instantaneous curvature rate of change falls back to within the steady-state wandering threshold, the core digital processor 14 rewrites the steady-state configuration word to the analog-to-digital conversion module 13 and restores the analog-to-digital conversion module 13 to the basic low-frequency sampling state and narrowband filtering mode.

[0043] Example 3

[0044] Please refer to the appendix. Figure 1-4 Based on Embodiment 2, in step S100 above, spatial orthogonal mapping of transient voltage signals is achieved, transforming the one-dimensional time-domain grid voltage signal into a continuous geometric trajectory in the three-dimensional spatial domain. The specific steps are as follows:

[0045] In step S110, the core digital processor 14 executes the underlying hardware initialization instruction, drives the analog-to-digital conversion module 13 to run at a preset basic low-frequency sampling rate, the voltage sensing unit 11 continuously acquires the three-phase AC voltage analog quantity of the power grid branch, the analog-to-digital conversion module 13 discretizes the three-phase AC voltage analog quantity, and outputs the instantaneous discrete sequence of the three-phase voltage corresponding to the discrete time node.

[0046] In step S120, the field-programmable gate array synchronously writes the instantaneous discrete sequence of three-phase voltages into the internal first-in-first-out (FIFO) data buffer according to the timestamp order generated by the hardware sampling clock, ensuring that the three-phase voltage data at the same timestamp are strictly aligned. The digital signal processor (DSP) periodically reads the data in the FIFO data buffer through the external memory interface and calls the built-in floating-point arithmetic unit to perform spatial orthogonal transformation. Specifically, the DSP inputs the instantaneous values ​​of the three-phase stationary coordinate system voltages at the same timestamp into the preset constant power Clark transformation matrix calculation module. The DSP executes matrix multiplication instructions and outputs mutually orthogonal alpha axis components, beta axis components, and zero-sequence components. The specific coefficient configuration of the constant power Clark transformation matrix is ​​common knowledge in the field of motor control, and those skilled in the art can directly call existing algorithm libraries to implement the conversion.

[0047] In step S130, the digital signal processor constructs a spatial position vector representing the transient electromagnetic energy distribution state of the system in a three-dimensional orthogonal coordinate system based on the calculated orthogonal voltage components. The digital signal processor assigns the alpha axis component, beta axis component and zero sequence component of the aforementioned calculation output to the coordinate values ​​of three mutually perpendicular coordinate axes in a three-dimensional rectangular coordinate system. By combining the coordinate values, the three-dimensional spatial position vector corresponding to the discrete sampling time is synthesized.

[0048] In step S140, the digital signal processor continuously stores the spatial position vectors generated at each discrete sampling time into the designated address space of the random access memory, forming a time sequence of spatial position vectors arranged continuously in time. The time sequence of spatial position vectors constitutes the basic data source for subsequent differential geometric spatial trajectory calculations. During the storage process, the digital signal processor uses a cyclic overwrite mechanism to update the designated address space, maintains the time window length of the spatial position vector time sequence constant, and restricts the disorderly occupation of storage resources.

[0049] Furthermore, in step S200, the discrete spatial location vector time series is reconstructed using the discrete extended state observer algorithm to avoid amplifying the background noise of the power grid through direct differential calculation. The specific steps are as follows:

[0050] In step S210, the digital signal processor reads the spatial position vector corresponding to the current discrete sampling node from the random access memory. The spatial position vector contains three mutually orthogonal axial voltage components. The digital signal processor decomposes the three axial voltage components and uses them as independent input observation objects for the discrete extended state observer.

[0051] In step S220, the state observation module 102 allocates an independent observer state variable group for each axial voltage component in its internal register. Taking the alpha axis in the three-dimensional orthogonal coordinate system as an example, the state observation module 102 defines three state variables: the first state variable is used to track the actual position value of the alpha axis voltage component, the second state variable is used to track the first derivative value of the alpha axis voltage component, and the third state variable is used to track the second derivative and lumped disturbance value of the alpha axis voltage component. The beta axis and the zero sequence axis adopt the same state variable configuration structure.

[0052] In step S230, the digital signal processor calls the arithmetic logic unit to perform the state error feedback iterative operation of the discrete extended state observer according to the set basic low-frequency sampling period. Specifically, the digital signal processor calculates the observation position error under the current discrete sampling step. The value of the observation position error is equal to the first state variable under the current discrete sampling step minus the original voltage component input value of the alpha axis. Based on the observation position error and the preset feedback gain coefficient, combined with the working sampling period of the underlying analog-to-digital converter, the digital signal processor performs discrete integral accumulation prediction on the first state variable, the second state variable and the third state variable respectively to obtain the updated value of the state variable under the next discrete sampling step. The value of the above feedback gain coefficient is determined by configuring the bandwidth parameter of the discrete extended state observer. The bandwidth parameter is set to an integer multiple of the fundamental angular frequency of the power grid to ensure that the phase lag of the voltage signal is suppressed while filtering out high-frequency noise.

[0053] In step S240, after completing the parallel iterative update of the alpha axis, beta axis, and zero-sequence axis, the digital signal processor extracts the second and third state variables of the alpha axis, beta axis, and zero-sequence axis. The digital signal processor synthesizes the second state variables of the three axes into a three-dimensional spatial velocity vector. At the same time, the digital signal processor synthesizes the third state variables of the three axes into a three-dimensional spatial acceleration vector. The spatial velocity vector corresponds to the first derivative of the spatial position vector, and the spatial acceleration vector corresponds to the second derivative of the spatial position vector.

[0054] In step S250, the digital signal processor writes the synthesized three-dimensional spatial velocity vector and three-dimensional spatial acceleration vector into a specified memory address segment. The data in the specified memory address segment is available for real-time calling by the geometric parameter calculation module 103. The three-dimensional spatial velocity vector and three-dimensional spatial acceleration vector extracted by the discrete extended state observer remain continuous in the time dimension, eliminating the interference of high-frequency differential noise.

[0055] Further, in step S300, the instantaneous curvature and curvature variation rate of the spatial trajectory are analyzed to extract the microscopic geometric features characterizing the transient topological distortion of the power grid. The specific steps are as follows:

[0056] In step S310, the geometric parameter calculation module 103 extracts the synchronously aligned three-dimensional spatial velocity vector and three-dimensional spatial acceleration vector from the specified memory address segment. The core digital processor 14 calls the floating-point vector multiplication instruction and performs the three-dimensional spatial vector cross product operation of the three-dimensional spatial velocity vector and the three-dimensional spatial acceleration vector to generate the cross product normal vector characterizing the voltage space trajectory normal disturbance. At the same time, the core digital processor 14 calculates the scalar magnitude of the three-dimensional spatial velocity vector.

[0057] In step S320, the core digital processor 14 calculates the instantaneous curvature of the voltage space trajectory at the current discrete time node based on the scalar magnitude of the cross product normal vector and the scalar magnitude of the three-dimensional spatial velocity vector. The instantaneous curvature characterizes the degree of bending of the transient voltage trajectory from the ideal straight line. Considering the overflow risk of floating-point operations in the digital chip's arithmetic logic unit, the core digital processor 14 introduces an overflow protection constant in the denominator of the division operation when calculating the instantaneous curvature. The calculation relationship is as follows:

[0058] In the formula, Represents the current discrete time node The instantaneous curvature; Represents the reconstructed three-dimensional spatial velocity vector; This represents the reconstructed three-dimensional spatial acceleration vector; Represents the vector cross product operator; Represents the scalar magnitude of a vector or a vector; Represents a very small positive real constant to prevent overflow during division by zero;

[0059] In step S330, the geometric parameter calculation module 103 stores the instantaneous curvature of the current discrete time node into the curvature time series buffer. The core digital processor 14, based on the instantaneous curvature of the current discrete time node and the instantaneous curvature of the previous discrete time node, and combined with the working sampling period of the underlying analog-to-digital conversion module 13, performs a first-order difference derivative operation in the time domain to extract the instantaneous curvature variation rate. The instantaneous curvature variation rate characterizes the distortion rate of the spatial geometric curvature as it evolves over time. The calculation relationship is as follows:

[0060] In the formula, Represents the instantaneous rate of curvature change at the current discrete time point; Represents the instantaneous curvature at the current discrete time point; Represents the instantaneous curvature at the previous discrete time node; The sampling period time interval represents the sampling period of the analog-to-digital conversion module 13 in the basic low-frequency sampling state;

[0061] In step S340, the geometric parameter calculation module 103 outputs the calculated instantaneous curvature rate of change to the adaptive trigger control module 104 as a quantitative judgment index for triggering the switching of the underlying hardware state. By mapping the electrical transient signal to pure geometric curvature features, the pressure-free release action judgment system can directly remove the judgment interference caused by the power grid amplitude fluctuation.

[0062] Furthermore, in step S400, the transient geometric features are converted into control instructions for physical hardware to achieve dynamic reconstruction of the sampling frequency and filtering bandwidth. The specific steps are as follows:

[0063] In step S410, the adaptive trigger control module 104 obtains the instantaneous curvature rate of change parameters output by the geometric parameter calculation module 103. The core digital processor 14 internally presets a steady-state wandering threshold, which is set according to the steady-state power grid background harmonic level and voltage fluctuation range.

[0064] In step S420, the core digital processor 14 executes a numerical comparison instruction to compare the absolute value of the instantaneous curvature change rate with the steady-state wander threshold. When the absolute value of the instantaneous curvature change rate is less than or equal to the steady-state wander threshold, the adaptive trigger control module 104 maintains the original low-frequency sampling configuration word, and the analog-to-digital conversion module 13 continues to work at the basic low-frequency sampling rate.

[0065] In step S430, when the absolute value of the instantaneous curvature change rate is greater than the steady-state walk threshold, the core digital processor 14 determines that a transient topological change has occurred in the power grid. The core digital processor 14 sends a trigger instruction to the field programmable gate array through the internal high-speed parallel data bus. The trigger instruction carries high-frequency sampling configuration parameters.

[0066] In step S440, after receiving the trigger command, the field-programmable gate array (FPGA) performs a register rewriting operation. The FPGA directly modifies the frequency division coefficient of the sampling rate control register inside the analog-to-digital converter module 13. At the same time, the FPGA sends a control message to the analog front-end processing circuit 12 through the serial peripheral interface bus. The control message is used to switch the capacitor and resistor array of the programmable filter chip inside the analog front-end processing circuit 12, and adjust the passband cutoff frequency of the anti-aliasing filter from narrowband to wideband mode to meet the anti-aliasing requirements of analog signals under high-frequency sampling.

[0067] In step S450, the analog-to-digital conversion module 13 switches the sampling frequency to a high-frequency sampling frequency according to the updated configuration parameters. The high-frequency sampling frequency is set to an integer multiple of the basic low-frequency sampling frequency. At this time, the data filling speed of the first-in-first-out data buffer increases synchronously with the sampling frequency, ensuring that the comprehensive judgment execution module 105 obtains high-time-resolution transient sampling points of the power grid voltage.

[0068] In step S460, after the comprehensive judgment execution module 105 completes the power failure judgment or detects that the power grid has returned to a steady state, it sends a reset signal to the adaptive trigger control module 104. The core digital processor 14 operates the field programmable gate array again according to the reset signal to restore the register parameters of the analog-to-digital converter module 13 and the analog front-end processing circuit 12 to the basic low-frequency sampling configuration, thus completing the hardware closed-loop control reset.

[0069] Further, in step S500, the instantaneous spatial torsion is extracted in high-frequency sampling mode to characterize the degree of spatial distortion of the voltage trajectory. The specific steps are as follows:

[0070] In step S510, the comprehensive judgment execution module 105 obtains the instantaneous discrete voltage sequence output by the analog-to-digital conversion module 13 at the high-frequency sampling frequency, and the core digital processor 14 maintains the operation of the discrete extended state observer and outputs the three-dimensional spatial velocity vector and the three-dimensional spatial acceleration vector aligned with the time node of the high-frequency sampling frequency.

[0071] In step S520, the core digital processor 14 retrieves the three-dimensional spatial acceleration vectors of adjacent high-frequency discrete time nodes. The comprehensive judgment execution module 105, in conjunction with the high-frequency sampling period time interval, controls the core digital processor 14 to subtract the three-dimensional spatial acceleration vector of the previous time node from the three-dimensional spatial acceleration vector of the current time node, and divides the vector difference obtained by the subtraction by the high-frequency sampling period time interval to obtain the three-dimensional spatial acceleration vector.

[0072] In step S530, the comprehensive judgment execution module 105 loads the three-dimensional spatial velocity vector, the three-dimensional spatial acceleration vector, and the three-dimensional spatial jerk vector into the hardware floating-point arithmetic unit. The core digital processor 14 performs a vector cross product operation between the three-dimensional spatial velocity vector and the three-dimensional spatial acceleration vector. Then, it performs a vector dot product operation between the vector cross product operation result and the three-dimensional spatial jerk vector to obtain a spatial scalar result.

[0073] In step S540, the core digital processor 14 calculates the square of the scalar modulus of the vector cross product operation result. To prevent floating-point overflow during division, the core digital processor 14 adds an overflow protection constant as the denominator to the square of the scalar modulus to calculate the instantaneous spatial torsion. The instantaneous spatial torsion characterizes the degree of spatial distortion of the voltage trajectory due to asymmetric disturbance caused by the residual voltage discharge of the inductive load, causing it to deviate from the two-dimensional plane. The calculation formula is as follows:

[0074] In the formula, Represents the spatial instantaneous torsion at the current high-frequency discrete time node; Represents a three-dimensional velocity vector; Represents a three-dimensional acceleration vector; Represents the jerk vector in three-dimensional space; Represents the vector cross product operator; Represents the vector dot product operator; The scalar magnitude representing the computation vector; A positive real constant representing a floating-point number to prevent overflow during division by zero;

[0075] In step S550, the comprehensive judgment execution module 105 writes the calculated instantaneous spatial torsion value into the random access memory cache queue of the judgment time window, as the basis for discrete topology judgment in subsequent execution of comprehensive tripping logic.

[0076] Furthermore, in step S600, logical judgment is performed based on multi-dimensional geometric topological features, and hardware action instructions are output. The specific steps are as follows:

[0077] In step S610, the comprehensive judgment execution module 105 loads a preset high-frequency judgment time window in the internal timer. Within the high-frequency judgment time window, the comprehensive judgment execution module 105 continuously reads the instantaneous curvature sequence in the curvature time sequence buffer and synchronously retrieves the instantaneous spatial torsion value in the random access memory buffer queue.

[0078] In step S620, the comprehensive judgment execution module 105 performs residual voltage curvature inertia criterion detection. The comprehensive judgment execution module 105 traverses the instantaneous curvature sequence within the high-frequency judgment time window and determines whether the value of the instantaneous curvature satisfies the continuous monotonically increasing condition. The continuous monotonically increasing condition requires that the instantaneous curvature value of the next time node in the sequence is strictly greater than the instantaneous curvature value of the previous time node, and there is no numerical drop or zero-crossing mutation in the sequence. Under the triggering conditions described in this application, the continuous monotonically increasing condition can be used to characterize the power grid in a residual voltage nonlinear discharge condition driven by the mechanical inertia of inductive loads.

[0079] In step S630, the comprehensive judgment execution module 105 performs spatial torsion disturbance exceeding the limit criterion detection. The comprehensive judgment execution module 105 extracts the absolute value of the instantaneous spatial torsion and compares the absolute value of the instantaneous spatial torsion with the set torsion boundary threshold. The torsion boundary threshold is set according to the asymmetric component limit under the steady-state background of the system. The absolute value of the instantaneous spatial torsion being greater than the torsion boundary threshold can be used to characterize the voltage space trajectory being disturbed by stator discharge and leaving the steady-state two-dimensional plane.

[0080] The torsion boundary threshold is set based on the following: The amplitudes of the corresponding negative-sequence and zero-sequence voltage components are calculated according to the maximum allowable voltage imbalance during steady-state operation. After constant-power Clarke transformation, the steady-state spatial position vector in the αβγ coordinate system is obtained. Under steady-state conditions, the voltage trajectory is theoretically constrained within the αβ plane, at which point the theoretical torsion value is zero. Considering measurement errors and background harmonic disturbances, the steady-state asymmetric component limit is mapped to a per-unit value of the torsion threshold through differential geometry operations, serving as the benchmark for setting the torsion boundary threshold.

[0081] In step S640, the comprehensive judgment execution module 105 converts the results of the residual curvature inertia criterion and the spatial torsion disturbance limit violation criterion into Boolean logic variables. The digital signal processor inside the core digital processor 14 performs a Boolean logic AND operation on the two Boolean logic variables. When the instantaneous curvature satisfies the condition of continuous monotonically increasing and the absolute value of the instantaneous spatial torsion is greater than the torsion boundary threshold, the digital signal processor writes the high-level trigger flag into the status configuration register of the general-purpose input / output pin and generates a digital trip instruction.

[0082] In step S650, the core digital processor 14 outputs the digital trip command to the electrical control terminal of the trip actuator 15 through the general-purpose input / output pin. The electromagnetic trip coil inside the trip actuator 15 is energized and driven to move the mechanical transmission link, forcibly disconnecting the main contacts of the circuit breaker and cutting off the physical connection of the power grid branch.

[0083] In step S660, the comprehensive judgment execution module 105 reviews the logic judgment result when the high-frequency judgment time window reaches the timing end. If no digital tripping instruction is generated, or the adaptive trigger control module 104 detects that the instantaneous curvature change rate falls back to the steady-state wandering threshold or below, the core digital processor 14 writes the steady-state configuration word to the analog-to-digital conversion module 13. The analog-to-digital conversion module 13 restores the working state to the basic low-frequency sampling rate. At the same time, the core digital processor 14 instructs the analog front-end processing circuit 12 to lower the passband cutoff frequency of the anti-aliasing filter to narrow the passband bandwidth, thus completing the state restoration of the underlying physical hardware system.

Claims

1. A method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid after power failure, characterized in that, Includes the following steps: Step S100: Perform basic low-frequency sampling and spatial vector construction, control the analog-to-digital conversion module to run at the basic low-frequency sampling rate, acquire the three-phase AC voltage discrete signal of the power grid branch, and map the three-phase AC voltage discrete signal to a mutually orthogonal three-dimensional coordinate system to construct the instantaneous position vector of the synthesized voltage spatial trajectory; Step S200: Perform smooth reconstruction of the higher-order derivative of the spatial trajectory. Use the instantaneous position vector as the observation input to call the discrete extended state observer algorithm and output the smooth reconstructed spatial velocity vector and spatial acceleration vector. Step S300: Perform real-time calculation of spatial differential geometric features, calculate the instantaneous curvature of the voltage space trajectory based on the spatial velocity vector and the spatial acceleration vector, and output the instantaneous curvature rate parameter based on the first-order difference derivative operation in the time domain. Step S400: Perform the underlying hardware state switching of curvature variation drive, compare the absolute value of the instantaneous curvature variation parameter with the system preset steady-state wander threshold, and when the absolute value of the instantaneous curvature variation parameter is greater than the steady-state wander threshold, drive the analog-to-digital conversion module to switch to high-frequency sampling mode, and synchronously instruct the analog front-end processing circuit to increase the passband cutoff frequency of the anti-aliasing filter to broaden the passband bandwidth. Step S500: Perform depth capture of high-frequency spatial topological features, update the spatial velocity vector and spatial acceleration vector in the high-frequency sampling mode, further extract the three-dimensional spatial acceleration vector, and perform a joint operation of vector dot product and cross product to calculate the instantaneous spatial torsion. Step S600: Perform multi-dimensional feature comprehensive trip determination. Within the set high-frequency determination time window, when the instantaneous curvature remains monotonically increasing and the absolute value of the instantaneous spatial deflection exceeds the set deflection boundary threshold, output a digital trip command to the tripping actuator to cut off the physical connection of the main circuit.

2. The method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid after power failure, as described in claim 1, is characterized in that... In step S100, basic low-frequency sampling and spatial vector construction are performed, specifically including: Input the instantaneous voltage values ​​of the three-phase stationary coordinate system at the same timestamp into the preset constant power Clark transformation matrix calculation module, and output mutually orthogonal alpha axis components, beta axis components and zero sequence components; The alpha axis component, beta axis component, and zero sequence component are respectively assigned the coordinate values ​​of three mutually perpendicular coordinate axes in a three-dimensional Cartesian coordinate system to synthesize the instantaneous position vector, and the instantaneous position vector is synchronously written into the first-in-first-out data buffer to form a spatial position vector time series.

3. The method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid after power failure, as described in claim 1, is characterized in that... In step S200, the higher-order derivative smoothing reconstruction of the spatial trajectory is performed, specifically including: For each axial voltage component, an independent group of observer state variables is assigned. A first state variable is defined to track the actual position value, a second state variable is defined to track the first derivative value, and a third state variable is defined to track the second derivative and lumped disturbance value. Based on the observation position error under the current discrete sampling step and the preset feedback gain coefficient, and combined with the working sampling period of the analog-to-digital conversion module, discrete integral accumulation prediction is performed to obtain the updated value of the state variable under the next discrete sampling step. The second state variable is extracted and synthesized into the spatial velocity vector, and the third state variable is extracted and synthesized into the spatial acceleration vector.

4. The method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid failure, as described in claim 1, is characterized in that... In step S300, real-time calculation of spatial differential geometric features is performed, specifically including: Perform a three-dimensional spatial vector cross product operation between the spatial velocity vector and the spatial acceleration vector to generate a cross product normal vector. Then, calculate the instantaneous curvature based on the scalar magnitude of the cross product normal vector and the scalar magnitude of the spatial velocity vector. The calculation relationship is as follows: In the formula, Represents the instantaneous curvature at the current discrete time point; Represents the reconstructed three-dimensional spatial velocity vector; This represents the reconstructed three-dimensional spatial acceleration vector; Represents the vector cross product operator; The scalar magnitude representing the computation vector; It represents a very small positive real constant that prevents overflow when dividing by zero during floating-point operations.

5. The method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid after power failure, as described in claim 1, is characterized in that... In step S400, the underlying hardware state switch driven by curvature variation is performed, specifically including: Trigger commands carrying high-frequency sampling configuration parameters are sent to the field-programmable gate array via a high-speed parallel data bus. The field-programmable gate array directly modifies the frequency division coefficient of the sampling rate control register inside the analog-to-digital converter module according to the trigger instruction, and switches the sampling frequency to a high-frequency sampling frequency; Simultaneously, a control message is sent to the analog front-end processing circuit via the serial peripheral interface bus to switch the capacitor and resistor array of the programmable filter chip, thereby adjusting the passband cutoff frequency of the anti-aliasing filter from narrowband to wideband mode.

6. The method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid failure, as described in claim 1, is characterized in that... In step S500, extracting the three-dimensional spatial acceleration vector and calculating the instantaneous spatial torsion specifically includes: The three-dimensional spatial acceleration vector is obtained by subtracting the three-dimensional spatial acceleration vector of the previous time node from the three-dimensional spatial acceleration vector of the current time node, and dividing the vector difference obtained by the subtraction by the high-frequency sampling period time interval. The cross product of the spatial velocity vector and the spatial acceleration vector is then multiplied by the dot product of the three-dimensional spatial acceleration vector to calculate the instantaneous spatial torsion. The calculation formula is as follows: In the formula, Represents the spatial instantaneous torsion at the current high-frequency discrete time node; Represents the jerk vector in three-dimensional space; Represents the vector dot product operator; This represents a positive real constant to prevent overflow during division by zero in floating-point operations.

7. The method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid after power failure, as described in claim 1, is characterized in that... In step S500, the logical conditions for executing the multi-dimensional feature comprehensive deactivation determination are specifically limited as follows: Within the high-frequency determination time window, the instantaneous curvature value of the next time node in the sequence is strictly greater than the instantaneous curvature value of the previous time node and there is no numerical drop or zero-crossing abrupt change, so as to satisfy the continuous monotonically increasing condition. Under this condition, the residual pressure nonlinear discharge condition can be characterized. The absolute value of the instantaneous spatial torsion is extracted. The absolute value of the instantaneous spatial torsion is greater than the torsion boundary threshold set according to the limit of the asymmetric component under the steady-state background of the system. Under this condition, the spatial torsion disturbance exceeding the limit of the steady-state two-dimensional plane can be characterized.

8. The method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid after power failure, as described in claim 1, is characterized in that... In step S600, outputting a digital trip command to the tripping actuator to disconnect the physical connection of the main circuit specifically includes: The results of the residual curvature inertia criterion and the spatial torsion disturbance limit violation criterion are converted into Boolean logic variables; The internal digital signal processor performs a Boolean AND operation on two Boolean logic variables. When the judgment conditions are met simultaneously, a high-level trigger flag is written into the status configuration register of the general-purpose input / output pin to generate the digital trip instruction. The digital trip command is output to the electrical control terminal of the trip actuator via a general-purpose input / output pin, driving the internal electromagnetic trip coil to be energized and the mechanical transmission linkage to move.

9. The method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid after power failure, as described in claim 1, is characterized in that... The method also includes the step of performing system state restoration: When the high-frequency determination time window reaches the timing end, the logic determination result is reviewed. If the digital tripping command is not generated, or if the instantaneous curvature rate parameter is detected to fall back to or below the steady-state wandering threshold; The steady-state configuration word is rewritten to the analog-to-digital conversion module to restore the basic low-frequency sampling rate, and the analog front-end processing circuit is instructed to lower the passband cutoff frequency of the anti-aliasing filter to narrow the passband bandwidth, thus completing the state restoration of the underlying physical hardware system.

10. The method for determining the reliable operation of a no-voltage release switch based on the residual voltage characteristics of a power grid failure, as described in claim 3, is characterized in that... The bandwidth limitation in performing discrete integral accumulation prediction is specifically as follows: The value of the preset feedback gain coefficient is determined by configuring the bandwidth parameter of the discrete extended state observer. The bandwidth parameter is constantly configured as an integer multiple of the fundamental angular frequency of the power grid, so as to suppress the phase lag of the voltage signal on the time axis while filtering out high-frequency noise.

Images