Relay protection secondary circuit virtual connection fault positioning method

Abstract

The invention discloses a relay protection secondary circuit virtual connection fault positioning method, which comprises the following steps: connecting a controllable load disturbance device in series in a relay protection secondary circuit, and arranging distributed voltage acquisition units at a plurality of key nodes of the secondary circuit; enabling the load current of the secondary circuit to jump from a reference current to a preset test current through the controllable load disturbance device; before and after the step load current is applied each time, voltage values of the corresponding monitoring nodes are synchronously acquired through the distributed voltage acquisition units, and dynamic voltage drops of the monitoring nodes are calculated; based on a response relationship between the dynamic voltage drop and the step load current, extracting a contact resistance characteristic parameter corresponding to each monitoring node; calculating a virtual connection index of each monitoring node, comparing the virtual connection index with a preset threshold value, and judging whether the corresponding monitoring node has a virtual connection fault or not; and positioning the specific position of a virtual connection fault point according to the identified virtual connection node.

Description

Technical Field

[0001] This invention relates to the field of power system relay protection technology, and in particular to a method for locating faults in the secondary circuit of relay protection. Background Technology

[0002] The secondary circuit of relay protection is a critical link in the safe operation of power systems. A loose connection fault refers to an abnormally increased contact resistance at terminals, connectors, and other connection points due to poor contact. Loose connection faults exhibit typical dynamic characteristics: good contact under light load or static conditions, but dynamic fluctuations in contact resistance occur when load current increases or environmental changes occur, leading to maloperation or failure of the protection device. Existing secondary circuit fault detection technologies mainly include: methods based on static voltage and current monitoring, which determine faults by monitoring the steady-state electrical quantities of the circuit; signal injection methods, which inject test signals into the circuit and analyze response characteristics; and methods based on smart sensor tags, which locate faults through distributed sensor nodes. These technologies are primarily designed for fault types such as open circuits, short circuits, and grounding faults, and have significant limitations in detecting loose connection faults.

[0003] The main drawback of existing technologies is that they cannot effectively identify the dynamic changes in contact resistance by using static monitoring or a single excitation method. The essence of a loose connection fault is that the contact resistance exhibits a nonlinear response with the load current. Static monitoring methods can only obtain electrical parameters under specific operating conditions and cannot capture the dynamic characteristics of the contact resistance. Although signal injection methods can apply test excitation, they mostly use constant current or constant voltage signals, lacking a step change process, making it difficult to extract the nonlinear response characteristics of the contact resistance. Summary of the Invention

[0004] The main objective of this invention is to provide a method for locating faults in the secondary circuit of a relay protection system.

[0005] Another objective of this invention is to provide a fault location system for a relay protection secondary circuit with a loose connection.

[0006] To achieve the above objectives, a first aspect of the present invention provides a method for locating a loose connection fault in a secondary circuit of a relay protection system, comprising:

[0007] S1, A controllable load disturbance device is connected in series in the secondary circuit of the relay protection, and distributed voltage acquisition units are arranged at multiple key nodes of the secondary circuit. S2, by periodically applying a step load current through the controllable load disturbance device, the load current of the secondary circuit jumps from the reference current to the preset test current. S3, before and after each step load current is applied, the voltage values ​​of the corresponding monitoring nodes are synchronously collected by each of the distributed voltage acquisition units, and the dynamic voltage drop of each monitoring node is calculated. S4. Based on the response relationship between the dynamic voltage drop and the step load current, extract the contact resistance characteristic parameters corresponding to each monitoring node. S5. Calculate the virtual connection index of each monitoring node based on the contact resistance characteristic parameters, compare the virtual connection index with a preset threshold, and determine whether there is a virtual connection fault in the corresponding monitoring node. S6. For the identified loose connection node, calculate the voltage gradient anomaly coefficient between the loose connection node and the adjacent monitoring node, and combine it with the topology information of the relay protection secondary circuit to locate the specific location of the loose connection fault point.

[0008] In one embodiment of the present invention, in step S2, the controllable load disturbance device includes a power electronic switching device, a current detection unit, and a PWM control circuit; wherein, the PWM control circuit achieves precise control of the load current by adjusting the duty cycle of the power electronic switching device, the current detection unit monitors the loop current in real time and feeds it back to the PWM control circuit to form a closed-loop control; the application of the step load current adopts a periodic triggering method, and each cycle includes a reference current stage, a current jump process, and a test current stage.

[0009] In one embodiment of the present invention, in step S3, the distributed voltage acquisition unit adopts a multi-channel synchronous sampling architecture, and the voltage sampling of each monitoring node is triggered by a unified clock signal; the method for calculating the dynamic voltage drop includes: Identify the steady-state voltage range before the application of the step load current, and calculate the average value of this steady-state voltage range as the initial voltage; Identify the steady-state voltage range after the application of a step load current, and calculate the average value of this steady-state voltage range as the final state voltage; The difference between the initial voltage and the final voltage is calculated as the dynamic voltage drop, and the complete time-domain waveform data from the start of the current jump to the voltage reaching a steady state is recorded.

[0010] In one embodiment of the present invention, in step S4, the contact resistance characteristic parameters include linearity deviation coefficient, voltage drop fluctuation coefficient, and response delay time; wherein, the extraction method for each parameter includes: When extracting the linearity deviation coefficient, the corresponding dynamic voltage drop is obtained under two different current levels: the reference current and the test current. The ratio of the two dynamic voltage drops is calculated to the ratio of the two load currents. By comparing the degree of deviation between the two ratios, the nonlinear characteristics of the contact resistance are quantified. When extracting the voltage drop fluctuation coefficient, the voltage drop time series during the test current stage is statistically analyzed, the standard deviation of the time series is calculated, and the standard deviation and the mean of the time series are normalized to obtain a dimensionless parameter reflecting the dynamic fluctuation characteristics of the contact resistance. When extracting the response delay time, the timing starts from the moment when the load current undergoes a step change, and the moment when the voltage drop curve enters a steady state is identified. The difference between the two moments is taken as the response delay time. The criterion for determining the steady state is that the voltage change rate is continuously lower than a preset threshold.

[0011] In one embodiment of the present invention, the linearity deviation coefficient The calculation formula is:

[0012] in, Indicates reference current The corresponding dynamic voltage drop, Indicates the test current The corresponding dynamic voltage drop, Indicates the reference current value. This indicates the test current value.

[0013] In one embodiment of the present invention, in step S5, the calculation of the virtual connection index H adopts a multi-dimensional feature weighted fusion algorithm, and the specific calculation formula is as follows:

[0014] in, , , These are weighting coefficients that satisfy the normalization condition; The normalized linearity deviation coefficient is obtained by using the linearity deviation coefficient. Divide by the preset maximum deviation threshold get; The normalized voltage drop fluctuation coefficient is obtained by adjusting the voltage drop fluctuation coefficient. Divide by the preset maximum fluctuation threshold get; The normalized response delay time is calculated by adjusting the response delay time. Divide by the preset maximum delay time get.

[0015] In one embodiment of the present invention, step S5, establishing a multi-level judgment mechanism based on the virtual connection index H, specifically includes: Set normal connection threshold and severe false connection threshold ; When the index is falsely connected Below the normal connection threshold When this happens, the monitoring node is determined to be in a normal connection state; When the index is falsely connected Between the normal connection threshold With severe false connection threshold If the monitoring node is determined to be in a slightly loose connection state, the system will generate an early warning record and mark the node as needing close attention. When the index is falsely connected Above the severe connection threshold When the corresponding monitoring node is determined to be in a state of serious loose connection, the system triggers the fault alarm mechanism and starts the fault location process.

[0016] In one embodiment of the present invention, in step S6, the voltage gradient anomaly coefficient The calculation method is as follows:

[0017] in, This indicates the monitoring node with the highest false connection index. express The upstream adjacent monitoring nodes, express Downstream adjacent monitoring nodes; , , These represent the dynamic voltage drops corresponding to these three nodes.

[0018] In one embodiment of the present invention, in step S1, the arrangement of key nodes follows a secondary loop topology, specifically including: a first monitoring node at the output end of the protection device, a second monitoring node at the terminal block of the protection panel, a third monitoring node at the inlet side of the operation box, a fourth monitoring node at the outlet side of the operation box, a fifth monitoring node at the intermediate junction box of the cable, a sixth monitoring node at the inlet side of the terminal box of the switchgear, and a seventh and eighth monitoring node at both ends of the coil of the operating mechanism. The distributed voltage acquisition unit uses a non-contact voltage sensor or a high-impedance voltage probe to acquire signals. The non-contact voltage sensor detects the potential distribution around the conductor based on the principle of electric field coupling, while the high-impedance voltage probe has an input impedance that is at least three orders of magnitude higher than the impedance of the circuit being measured. The distributed voltage acquisition unit is connected to the data processing center through an optical fiber communication network to achieve high-speed transmission and time synchronization of data from each monitoring node.

[0019] To achieve the above objectives, a second aspect of the present invention provides a fault location system for a relay protection secondary circuit, comprising: The controllable load disturbance module is used to periodically apply a step load current, causing the load current of the secondary circuit of the relay protection to jump between the reference current and the test current. The distributed voltage acquisition module is used to synchronously acquire the voltage of each monitoring node based on a unified clock source, and obtain the voltage values ​​of each monitoring node before and after the application of a step load current. The feature extraction module is used to calculate the dynamic voltage drop of each monitoring node and extract contact resistance feature parameters based on the response relationship between the dynamic voltage drop and the step load current. The contact resistance feature parameters include linearity deviation coefficient, voltage drop fluctuation coefficient and response delay time. The loose connection identification module is used to normalize the contact resistance characteristic parameters, calculate the loose connection index of each monitoring node using a multi-dimensional feature weighted fusion algorithm, and determine whether there is a loose connection fault based on a preset threshold. The fault location module is used to calculate the voltage gradient anomaly coefficient for the identified loose connection nodes and, in combination with the secondary circuit topology information, locate the specific location of the loose connection fault point.

[0020] The embodiments of the present invention have the following beneficial effects: The controllable load disturbance device of this invention causes a step change in the circuit load current between the reference current and the test current. This step change process can trigger the dynamic response of the contact resistance of the virtual contact point, making the hidden contact failure characteristics visible. By synchronously collecting the voltage values ​​before and after the step change and calculating the dynamic voltage drop, the dynamic response process of the contact resistance to the load current change is fully recorded. Based on the response relationship between the dynamic voltage drop and the step load current, the characteristic parameters of the contact resistance are extracted. These parameters can quantify the nonlinear characteristics of the contact resistance. Unlike the existing technology that judges based on a single electrical quantity, this invention achieves a quantitative assessment of the degree of contact failure by calculating the virtual connection index, integrating multi-dimensional characteristic information into a unified index, and improving the accuracy of virtual connection fault identification. By calculating the voltage gradient anomaly coefficient and utilizing the abrupt distribution characteristics of the voltage drop upstream and downstream of the virtual contact point, combined with the circuit topology information, the fault point can be accurately located. Attached Figure Description

[0021] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 A flowchart of a method for locating a loose connection fault in the secondary circuit of a relay protection system, provided by an embodiment of the present invention; Figure 2 This is an architectural diagram of a method for locating a loose connection fault in a secondary circuit of a relay protection system, provided by an embodiment of the present invention. Figure 3 This is a schematic diagram illustrating the principle of applying a step load current according to an embodiment of the present invention. Figure 4 This is a schematic diagram of voltage synchronous acquisition and dynamic voltage drop calculation provided in an embodiment of the present invention; Figure 5 This is a structural diagram of a relay protection secondary circuit fault location system provided in an embodiment of the present invention. Detailed Implementation

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0024] The following describes a method for locating a loose connection fault in the secondary circuit of a relay protection system according to an embodiment of the present invention, with reference to the accompanying drawings.

[0025] Example 1 This embodiment provides a method for locating faults in the secondary circuit of a relay protection system, such as... Figure 1 and Figure 2 As shown, the method includes the following steps: S1, a controllable load disturbance device is connected in series in the secondary circuit of the relay protection, and distributed voltage acquisition units are arranged at multiple key nodes of the secondary circuit.

[0026] In this embodiment of the invention, the arrangement of key nodes follows a secondary loop topology, specifically including: a first monitoring node at the output end of the protection device, a second monitoring node at the terminal block of the protection panel, a third monitoring node at the inlet side of the control box, a fourth monitoring node at the outlet side of the control box, a fifth monitoring node at the intermediate junction box of the cable, a sixth monitoring node at the inlet side of the switchgear terminal box, and a seventh and eighth monitoring node at both ends of the coil of the operating mechanism.

[0027] Furthermore, the distributed voltage acquisition unit uses a non-contact voltage sensor or a high-impedance voltage probe to acquire signals. The non-contact voltage sensor detects the potential distribution around the conductor based on the principle of electric field coupling, while the high-impedance voltage probe has an input impedance that is at least three orders of magnitude higher than the impedance of the circuit being measured. The distributed voltage acquisition unit is connected to the data processing center through an optical fiber communication network to achieve high-speed transmission and time synchronization of data from each monitoring node. The arrangement scheme of key nodes is optimized for typical relay protection secondary circuit topologies.

[0028] Specifically, taking the tripping circuit of the main transformer protection in a 220kV substation as an example, the first monitoring node is located at the tripping output terminal of the protection device. This location is the starting point of the secondary circuit and is used to monitor the voltage signal output by the protection device. The second monitoring node is located at the terminal block of the protection panel. This location is the junction point between the internal wiring of the protection panel and the external cable. There are many terminal connections, which can easily lead to loose connections. The third and fourth monitoring nodes are located on the incoming and outgoing sides of the operating box, respectively. The operating box is an intermediate link in the secondary circuit. It contains multiple transfer terminals and is greatly affected by the ambient temperature and humidity, making it a high-incidence area for loose connection faults. The fifth monitoring node is located at the intermediate junction box of the cable. For long secondary cables, intermediate junction boxes are usually used for segmented connections. The wiring quality at this location directly affects the reliability of the circuit. The sixth monitoring node is located on the incoming side of the switchgear terminal box. This location is close to the field equipment and is subject to more severe vibration and temperature changes. The seventh and eighth monitoring nodes are located at both ends of the operating mechanism coil. These two nodes can monitor the working status of the coil itself and verify the integrity of the circuit. The non-contact voltage sensor employs a parallel-plate capacitor structure. The sensor's sensing plates are close to but not in direct contact with the conductor being measured. Voltage is measured indirectly by detecting the electric field strength around the conductor. This method has the advantage of not altering the original circuit's electrical characteristics and offering convenient installation. The high-impedance voltage probe is designed with an input impedance of over 10 megohms, ensuring that its impact on the circuit current after probe connection is less than 0.01%, which is negligible. The fiber optic communication network uses a star topology. Data from each monitoring node is converted into optical signals by a photoelectric conversion module and then transmitted via fiber optic cable to the data processing center in the central computer room. Fiber optic transmission offers advantages such as strong resistance to electromagnetic interference, long transmission distance, and large bandwidth, meeting the requirements for multi-channel high-speed data transmission and precise time synchronization.

[0029] S2, by periodically applying a step load current through the controllable load disturbance device, the load current of the secondary circuit jumps from the reference current to the preset test current.

[0030] Furthermore, such as Figure 3 As shown, the controllable load disturbance device includes a power electronic switching device, a current detection unit, and a PWM control circuit. The PWM control circuit achieves precise control of the load current by adjusting the duty cycle of the power electronic switching device. The current detection unit monitors the loop current in real time and feeds it back to the PWM control circuit to form a closed-loop control. The application of the step load current adopts a periodic triggering method, and each cycle includes a reference current stage, a current jump process, and a test current stage.

[0031] Specifically, the power electronic switching devices of the controllable load disturbance device preferably use IGBT modules, with withstand voltage ratings ranging from 600V to 1200V and rated currents from 10A to 50A, to meet the typical operating voltage and current range of the relay protection secondary circuit. The core of the PWM control circuit is a digital signal processor, which generates a 20kHz pulse width modulation signal to control the gate voltage of the IGBT through the drive circuit, achieving precise switching action. The current detection unit uses a Hall current sensor with a measurement accuracy of 0.5% and a response time of less than 1 microsecond, enabling real-time capture of rapid changes in the loop current. The closed-loop control is implemented as follows: the current detection unit converts the measured current value into a digital signal and feeds it back to the PWM control circuit. The control circuit calculates the deviation between the measured current and the target current, and adjusts the duty cycle of the PWM signal through a PID control algorithm to make the actual current track the target current value. The periodic triggering time is set to 20 seconds, and the timing of each cycle is as follows: the reference current is maintained at 0.1A for the first 10 seconds, then the current jumps to the test current of 1.5A within 50 milliseconds, the test current is maintained for 5 seconds, and then it drops back to the reference current. The last 5 seconds are for data processing and system preparation. This periodic disturbance method can fully stimulate the nonlinear response characteristics of the virtual contact without affecting the normal operation of the protection device.

[0032] S3, before and after each step load current is applied, the voltage values ​​of the corresponding monitoring nodes are synchronously collected by each of the distributed voltage acquisition units, and the dynamic voltage drop of each monitoring node is calculated.

[0033] Furthermore, such as Figure 4 As shown, the distributed voltage acquisition unit adopts a multi-channel synchronous sampling architecture, and the voltage sampling of each monitoring node is triggered by a unified clock signal; the calculation method of the dynamic voltage drop includes: (1) Identify the steady-state voltage range before the step load current is applied, and calculate the average value of this steady-state voltage range as the initial voltage. ; (2) Identify the steady-state voltage segment after the step load current is applied, and calculate the average value of the steady-state voltage segment as the final voltage. ; (3) Calculate the difference between the initial voltage and the final voltage as the dynamic voltage drop, and record the complete time-domain waveform data from the start of the current jump to the voltage reaching a steady state. The calculation formula is as follows: .

[0034] In this embodiment of the invention, the multi-channel synchronous sampling architecture employs a 16-channel synchronous analog-to-digital converter chip. Each channel is configured with an independent sample-and-hold circuit to ensure that the voltage signals of each monitoring node are latched at the same time. A unified clock signal is generated by a high-precision crystal oscillator with a frequency of 50MHz. After frequency division, it provides synchronous trigger pulses for each sampling channel, achieving nanosecond-level time alignment accuracy. The identification of steady-state voltage segments adopts a sliding window determination method: the time window is set to 100 milliseconds, and when the standard deviation of the voltage data within the window is less than 2 millivolts, it is determined to be a steady-state segment. The initial voltage is calculated in the time period from 200 milliseconds to 50 milliseconds before the step occurs, and the average value is calculated by extracting the steady-state data within this segment. The final voltage is calculated after the voltage has stabilized for at least 100 milliseconds after the step occurs, and the average value is calculated by selecting a 200-millisecond data window after stabilization. The recording of complete time-domain waveform data starts 100 milliseconds before the step trigger and continues until 500 milliseconds after the step, with the sampling rate set to 50kHz to ensure that the entire process of dynamic changes in contact resistance can be captured. Data storage adopts a circular buffer mechanism, with each monitoring node allocated 32KB of storage space, which is sufficient to store waveform data for a complete cycle. The data is transferred to the data processing center at high speed via DMA.

[0035] S4. Based on the response relationship between the dynamic voltage drop and the step load current, extract the contact resistance characteristic parameters corresponding to each monitoring node.

[0036] Furthermore, the contact resistance characteristic parameters include linearity deviation coefficient, voltage drop fluctuation coefficient, and response delay time; the extraction methods for each parameter include: (1) When extracting the linearity deviation coefficient, the corresponding dynamic voltage drop is obtained under two different current levels, namely the reference current and the test current. The ratio of the two dynamic voltage drops is calculated and the ratio of the two load currents is calculated. By comparing the degree of deviation of the two ratios, the nonlinear characteristics of the contact resistance are quantified.

[0037] Specifically, in the process of extracting the linearity deviation coefficient, for the two operating points of reference current 0.1A and test current 1.5A, after calculating the dynamic voltage drop, the voltage drop ratio is first calculated as the voltage drop corresponding to the test current divided by the voltage drop corresponding to the reference current. Then, the current ratio is calculated as 15. Ideally, the two ratios should be equal. However, due to the current dependence of the contact resistance, the contact resistance under the test current will be relatively reduced at the virtual contact point, making the voltage drop ratio less than the current ratio. This degree of deviation reflects the strength of the contact nonlinearity characteristics.

[0038] In this embodiment of the invention, the linearity deviation coefficient The calculation formula is:

[0039] in, Indicates reference current The corresponding dynamic voltage drop, Indicates the test current The corresponding dynamic voltage drop, Indicates the reference current value. This indicates the test current value.

[0040] Specifically, linearity deviation coefficient This reflects the degree of deviation from the linear relationship of Ohm's law. The contact resistance of a normal connection point is a fixed value, and the dynamic voltage drop is proportional to the current. However, the contact resistance of a virtual connection point exhibits nonlinear characteristics as the current changes, resulting in an inconsistency between the voltage drop ratio and the current ratio.

[0041] In practical applications, when the reference current is 0.1A and the test current is 1.5A, the dynamic voltage drops at a normal connection point are approximately 10mV and 150mV, respectively, with a voltage drop ratio of 15, which is exactly the same as the current ratio of 15. At this point, the linearity deviation coefficient is close to zero. However, for a connection point with a loose connection, assuming its contact resistance is 0.8 ohms at 0.1A and drops to 0.6 ohms at 1.5A due to the cleaning effect of the current, the voltage drops are 80mV and 900mV, respectively, with a voltage drop ratio of 11.25, significantly less than the current ratio of 15. The calculated linearity deviation coefficient is 0.25, a value significantly higher than that of a normal connection point. The physical mechanism of this nonlinear response lies in the fact that the contact area of ​​the loose connection point changes with contact pressure and current density. The electromagnetic force and thermal effect when a large current passes through temporarily improves the contact state, but this improvement is unstable; when the current decreases, the contact resistance will rise again. By quantifying this nonlinear characteristic, the present invention can identify incomplete connection faults such as insufficient contact area or insufficient contact pressure, which are often missed under static monitoring.

[0042] (2) When extracting the voltage drop fluctuation coefficient, perform statistical analysis on the voltage drop time series during the test current stage, calculate the standard deviation of the time series, normalize the standard deviation and the mean of the time series, and obtain a dimensionless parameter that reflects the dynamic fluctuation characteristics of the contact resistance.

[0043] Specifically, the extraction of voltage drop fluctuation coefficient requires processing the voltage drop time series after the test current has stabilized. The specific method is to extract 2000 sampling points after stabilization, and calculate the standard deviation and mean of these 2000 data points. The standard deviation reflects the time fluctuation amplitude of the voltage drop, and the normalized fluctuation coefficient eliminates the influence of the absolute value of the voltage drop, making the fluctuation characteristics of different nodes comparable.

[0044] (3) When extracting the response delay time, start timing from the moment when the load current undergoes a step change, identify the moment when the voltage drop curve enters a stable state, and take the difference between the two moments as the response delay time; the criterion for determining the stable state is that the voltage change rate is continuously lower than a preset threshold.

[0045] Specifically, the response delay time is measured using a dual-threshold determination method: First, the precise moment of the current step is detected, defined as the moment when the current rises to 10% of the step amplitude; then, the change in the voltage drop curve is tracked, and when the rate of voltage change is less than 5% of the initial rate of change for 50 consecutive sampling points, it is determined to have entered a steady state. The time difference between these two moments is the response delay time. The delay time of a normal connection point is typically less than 5 milliseconds, while the delay time of a dummy connection point can reach 50 milliseconds or even longer due to the microscopic deformation of the contact surface and the arcing effect.

[0046] S5. Calculate the loose connection index of each monitoring node based on the contact resistance characteristic parameters, compare the loose connection index with a preset threshold, and determine whether the corresponding monitoring node has a loose connection fault.

[0047] Specifically, the calculation of the virtual connection index H in step S5 adopts a multi-dimensional feature weighted fusion algorithm, and the specific calculation formula is as follows:

[0048] in, , , These are weighting coefficients that satisfy the normalization condition; The normalized linearity deviation coefficient is obtained by using the linearity deviation coefficient. Divide by the preset maximum deviation threshold get; The normalized voltage drop fluctuation coefficient is obtained by adjusting the voltage drop fluctuation coefficient. Divide by the preset maximum fluctuation threshold get; The normalized response delay time is calculated by adjusting the response delay time. Divide by the preset maximum delay time get.

[0049] In this embodiment of the invention, the weighting coefficient is set based on the sensitivity of each characteristic parameter to the virtual connection fault, and the weighting coefficient of the linearity deviation coefficient is... The weighting factor of the voltage drop fluctuation coefficient is set to the highest weight. The weighting coefficient for the response delay time is set as the second highest weight. These are set as auxiliary weights. In the multi-dimensional feature weighted fusion algorithm for the loose connection index, different benchmark values ​​are used for the normalization process of each feature parameter. The maximum threshold for the linearity deviation coefficient is set to 0.5, which is determined based on statistical analysis of a large number of loose connection samples and represents the typical linearity deviation level of severe loose connections. The maximum threshold for the voltage drop fluctuation coefficient is set to 0.2, reflecting the maximum fluctuation of the loose connection point under constant current. The maximum threshold for the response delay time is set to 100 milliseconds, corresponding to the delay characteristics under severe oxidation or loosening of the contact surface. The weight coefficients are set based on sensitivity analysis and field test data. The linearity deviation coefficient is most sensitive to loose connection faults, so its weight is set to 0.4. The voltage drop fluctuation coefficient is set to 0.35, and although the response delay time can also reflect the characteristics of loose connections, it is greatly affected by measurement noise, so its weight is set to 0.25. In actual calculation, each original feature parameter is first normalized by dividing it by the corresponding maximum threshold, so that all feature values ​​fall within the range of 0 to 1. Then, the loose connection index is obtained by weighted summation according to the set weight coefficients. This normalization and weighted fusion approach preserves the independent information of each feature parameter and improves the reliability of the judgment through comprehensive evaluation, avoiding misjudgments that may occur with single feature judgment.

[0050] Furthermore, a multi-level judgment mechanism is established based on the aforementioned virtual connection index H, specifically including: Set normal connection threshold and severe false connection threshold ; When the index is falsely connected Below the normal connection threshold When this happens, the monitoring node is determined to be in a normal connection state; When the index is falsely connected Between the normal connection threshold With severe false connection threshold If the monitoring node is determined to be in a slightly loose connection state, the system will generate an early warning record and mark the node as needing close attention. When the index is falsely connected Above the severe connection threshold When the corresponding monitoring node is determined to be in a state of serious loose connection, the system triggers the fault alarm mechanism and starts the fault location process.

[0051] In this embodiment of the invention, the threshold setting of the multi-level judgment mechanism combines the reliability requirements of the relay protection secondary circuit and the operation and maintenance management needs. The normal connection threshold is set to 0.2. The principle for determining this value is to ensure that the loose connection index of a normal connection point has a probability of being lower than this threshold by more than 95%, thereby effectively distinguishing between normal connections and minor loose connections. The severe loose connection threshold is set to 0.4. Connection points higher than this threshold already have obvious poor contact problems and have a high risk of failure, requiring immediate handling. When the loose connection index is between 0.2 and 0.4, the system judges it as a minor loose connection state. Although it has not yet reached the level of immediate alarm, it has shown a trend of contact deterioration. The system will automatically generate an early warning record and mark the node in yellow on the monitoring interface, prompting operation and maintenance personnel to focus on checking it during the next maintenance. For severe loose connection states, the system not only triggers audible and visual alarms, but also automatically generates a fault work order, recording the time, location, and characteristic parameters of the fault occurrence. At the same time, it initiates the fault location process to calculate the voltage gradient anomaly coefficient to accurately locate the fault point. This hierarchical management mechanism can not only detect potential faults early and enable preventive maintenance, but also respond quickly to serious faults to avoid the escalation of accidents, significantly improving the operational reliability and maintenance efficiency of the secondary circuit.

[0052] S6. For the identified loose connection node, calculate the voltage gradient anomaly coefficient between the loose connection node and the adjacent monitoring node, and combine it with the topology information of the relay protection secondary circuit to locate the specific location of the loose connection fault point.

[0053] Furthermore, the voltage gradient anomaly coefficient The calculation method is as follows:

[0054] in, This indicates the monitoring node with the highest false connection index. express The upstream adjacent monitoring nodes, express Downstream adjacent monitoring nodes; , , These represent the dynamic voltage drops corresponding to these three nodes.

[0055] In this embodiment of the invention, the voltage gradient anomaly coefficient reflects the abrupt distribution characteristics of the voltage drop upstream and downstream of the virtual contact. When the virtual contact is located at a node... Its upstream node When the contact resistance of the connection segment increases abnormally, the voltage drop of the segment is significantly higher than that of the adjacent segment, which is manifested as the numerator being much larger than the denominator. By judging whether the voltage gradient abnormality coefficient exceeds the preset abnormal threshold, the connection fault point between adjacent monitoring nodes can be accurately located.

[0056] Furthermore, the calculation of the voltage gradient anomaly coefficient is based on the physical characteristic that a dummy connection will generate an abnormal voltage drop at its location. Under normal circumstances, the voltage drop between adjacent monitoring nodes depends on the resistance of the connecting wires. Since the resistance of the wires is uniformly distributed, the voltage drop in each segment is approximately equal, and the voltage gradient shows a gradual change. When a dummy connection exists in a certain connection segment, the total resistance of that segment increases significantly, causing the voltage drop in that segment to be much higher than that of adjacent segments, which appears as an abrupt change point on the voltage distribution curve. In the calculation, the monitoring node with the largest dummy connection index is first identified as the suspected node. Then, the dynamic voltage drop data of this node and its upstream and downstream adjacent nodes are extracted. The voltage drop difference between the suspected node and the upstream node, and the voltage drop difference between the suspected node and the downstream node are calculated. The ratio of the two differences is the voltage gradient anomaly coefficient. When the dummy connection is located between the suspected node and the upstream node, the abnormal voltage drop generated by this connection segment will cause the voltage of the suspected node to be significantly lower than that of the upstream node. However, the suspected node and the downstream node are normally connected, and the voltage drop difference is small. Therefore, the gradient anomaly coefficient will be much greater than 1. The anomaly threshold is typically set to 3. When the gradient anomaly coefficient exceeds this threshold, the specific connection segment between the suspected node and the upstream node can be identified as the virtual contact. Combined with the secondary circuit topology information, the specific terminals or posts contained in this segment can be further located, such as the outgoing terminals of the control box, cable joints, and switch cabinet terminal blocks, providing precise maintenance guidance for operation and maintenance personnel.

[0057] Example 2 This invention also provides a relay protection secondary circuit loose connection fault location system, such as... Figure 5 As shown, the system 10 includes: The controllable load disturbance module 100 is used to periodically apply a step load current, causing the load current in the secondary circuit of the relay protection to jump between the reference current and the test current. Specifically, this module is connected in series in the secondary circuit of the relay protection and includes power electronic switching devices, a current detection module, and a PWM control circuit.

[0058] The distributed voltage acquisition module 200 is used to synchronously acquire the voltage of each monitoring node based on a unified clock source, and to obtain the voltage values ​​of each monitoring node before and after the application of a step load current. Specifically, this module is arranged at multiple key nodes in the secondary circuit and includes multiple voltage sensors and a multi-channel synchronous sampling circuit.

[0059] The feature extraction module 300 is used to calculate the dynamic voltage drop of each monitoring node and extract contact resistance characteristic parameters based on the response relationship between the dynamic voltage drop and the step load current. These contact resistance characteristic parameters include a linearity deviation coefficient, a voltage drop fluctuation coefficient, and a response delay time. Specifically, this module is connected to the distributed voltage acquisition module.

[0060] The loose connection identification module 400 is used to normalize the contact resistance characteristic parameters, calculate the loose connection index of each monitoring node using a multi-dimensional feature weighted fusion algorithm, and determine whether a loose connection fault exists based on a preset threshold. Specifically, this module is connected to the feature extraction module.

[0061] The fault location module 500 is used to calculate the voltage gradient anomaly coefficient for the identified loose connection nodes and, in conjunction with the secondary circuit topology information, locate the specific location of the loose connection fault point. Specifically, this module is connected to the loose connection identification module.

[0062] Specifically, in the overall implementation of this system, each functional module is integrated through an embedded system platform. The data processing center uses an industrial-grade embedded computer, configured with a dual-core processor with a main frequency of 1.5GHz, 4GB of memory, and a 128GB solid-state drive, running a real-time operating system to ensure the real-time performance and reliability of data processing. The feature extraction module is implemented using C language programming and employs efficient digital signal processing algorithms, capable of extracting feature parameters of all monitoring nodes within a single cycle, with a processing latency of less than 1 second. The virtual connection identification module adopts a modular design, with the feature normalization unit, weighted fusion unit, and threshold judgment unit operating independently and exchanging information via a data bus. This design facilitates functional upgrades and algorithm optimization. The topology analysis unit of the fault location module pre-stores complete topology information of the secondary circuit, including the location number of each monitoring node, adjacent relationships, connection paths, and a list of specific terminals included in that path. When the gradient calculation unit outputs the voltage gradient anomaly coefficient, the topology analysis unit immediately queries the topology database to determine the fault segment. The location output unit generates a detailed report containing the fault location, fault characteristics, and handling suggestions, which is displayed to maintenance personnel through a human-machine interface. Simultaneously, alarm information can be pushed to mobile terminals via a network interface. The system is also equipped with data storage and analysis functions. All monitoring data and fault records are automatically stored in the historical database, supporting data playback and trend analysis. Maintenance personnel can grasp the deterioration trend of each connection point by viewing the historical change curve of the loose connection index, realizing the transformation from fault diagnosis to predictive maintenance.

[0063] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0064] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0065] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for locating a loose connection fault in the secondary circuit of a relay protection system, characterized in that, Includes the following steps: S1, A controllable load disturbance device is connected in series in the secondary circuit of the relay protection, and distributed voltage acquisition units are arranged at multiple key nodes of the secondary circuit. S2, by periodically applying a step load current through the controllable load disturbance device, the load current of the secondary circuit jumps from the reference current to the preset test current. S3, before and after each step load current is applied, the voltage values ​​of the corresponding monitoring nodes are synchronously collected by each of the distributed voltage acquisition units, and the dynamic voltage drop of each monitoring node is calculated. S4. Based on the response relationship between the dynamic voltage drop and the step load current, extract the contact resistance characteristic parameters corresponding to each monitoring node. S5. Calculate the virtual connection index of each monitoring node based on the contact resistance characteristic parameters, compare the virtual connection index with a preset threshold, and determine whether there is a virtual connection fault in the corresponding monitoring node. S6. For the identified loose connection node, calculate the voltage gradient anomaly coefficient between the loose connection node and the adjacent monitoring node, and combine it with the topology information of the relay protection secondary circuit to locate the specific location of the loose connection fault point.

2. The method according to claim 1, characterized in that, In step S2, the controllable load disturbance device includes a power electronic switching device, a current detection unit, and a PWM control circuit. The PWM control circuit achieves precise control of the load current by adjusting the duty cycle of the power electronic switching device. The current detection unit monitors the loop current in real time and feeds it back to the PWM control circuit to form a closed-loop control. The application of the step load current adopts a periodic triggering method, and each cycle includes a reference current stage, a current jump process, and a test current stage.

3. The method according to claim 1, characterized in that, In step S3, the distributed voltage acquisition unit adopts a multi-channel synchronous sampling architecture, and the voltage sampling of each monitoring node is triggered by a unified clock signal; the calculation method for the dynamic voltage drop includes: Identify the steady-state voltage range before the application of the step load current, and calculate the average value of this steady-state voltage range as the initial voltage; Identify the steady-state voltage range after the application of a step load current, and calculate the average value of this steady-state voltage range as the final state voltage; The difference between the initial voltage and the final voltage is calculated as the dynamic voltage drop, and the complete time-domain waveform data from the start of the current jump to the voltage reaching a steady state is recorded.

4. The method according to claim 1, characterized in that, In step S4, the contact resistance characteristic parameters include linearity deviation coefficient, voltage drop fluctuation coefficient, and response delay time; wherein, the extraction methods for each parameter include: When extracting the linearity deviation coefficient, the corresponding dynamic voltage drop is obtained under two different current levels: the reference current and the test current. The ratio of the two dynamic voltage drops is calculated to the ratio of the two load currents. By comparing the degree of deviation between the two ratios, the nonlinear characteristics of the contact resistance are quantified. When extracting the voltage drop fluctuation coefficient, the voltage drop time series during the test current stage is statistically analyzed, the standard deviation of the time series is calculated, and the standard deviation and the mean of the time series are normalized to obtain a dimensionless parameter reflecting the dynamic fluctuation characteristics of the contact resistance. When extracting the response delay time, the timing starts from the moment when the load current undergoes a step change, and the moment when the voltage drop curve enters a steady state is identified. The difference between the two moments is taken as the response delay time. The criterion for determining the steady state is that the voltage change rate is continuously lower than a preset threshold.

5. The method according to claim 4, characterized in that, The linearity deviation coefficient The calculation formula is: in, Indicates reference current The corresponding dynamic voltage drop, Indicates the test current The corresponding dynamic voltage drop, Indicates the reference current value. This indicates the test current value.

6. The method according to claim 4, characterized in that, In step S5, the virtual connection index H is calculated using a multi-dimensional feature weighted fusion algorithm, and the specific calculation formula is as follows: in, , , These are weighting coefficients that satisfy the normalization condition; The normalized linearity deviation coefficient is obtained by using the linearity deviation coefficient. Divide by the preset maximum deviation threshold get; The normalized voltage drop fluctuation coefficient is obtained by adjusting the voltage drop fluctuation coefficient. Divide by the preset maximum fluctuation threshold get; The normalized response delay time is calculated by adjusting the response delay time. Divide by the preset maximum delay time get.

7. The method according to claim 6, characterized in that, In step S5, a multi-level judgment mechanism is established based on the virtual connection index H, specifically including: Set normal connection threshold and severe false connection threshold ; When the index is falsely connected Below the normal connection threshold When this happens, the monitoring node is determined to be in a normal connection state; When the index is falsely connected Between the normal connection threshold With severe false connection threshold If the monitoring node is determined to be in a slightly loose connection state, the system will generate an early warning record and mark the node as needing close attention. When the index is falsely connected Above the severe connection threshold When the corresponding monitoring node is determined to be in a state of serious loose connection, the system triggers the fault alarm mechanism and starts the fault location process.

8. The method according to claim 1, characterized in that, In step S6, the voltage gradient anomaly coefficient The calculation method is as follows: in, This indicates the monitoring node with the highest false connection index. express The upstream adjacent monitoring nodes, express Downstream adjacent monitoring nodes; , , These represent the dynamic voltage drops corresponding to these three nodes.

9. The method according to claim 1, characterized in that, In step S1, the arrangement of key nodes follows the secondary loop topology, specifically including: the first monitoring node at the output end of the protection device, the second monitoring node at the terminal block of the protection panel, the third monitoring node at the inlet side of the control box, the fourth monitoring node at the outlet side of the control box, the fifth monitoring node at the intermediate junction box of the cable, the sixth monitoring node at the inlet side of the switchgear terminal box, and the seventh and eighth monitoring nodes at both ends of the coil of the operating mechanism. The distributed voltage acquisition unit uses a non-contact voltage sensor or a high-impedance voltage probe to acquire signals. The non-contact voltage sensor detects the potential distribution around the conductor based on the principle of electric field coupling, while the high-impedance voltage probe has an input impedance that is at least three orders of magnitude higher than the impedance of the circuit being measured. The distributed voltage acquisition unit is connected to the data processing center through an optical fiber communication network to achieve high-speed transmission and time synchronization of data from each monitoring node.

10. A fault location system for a relay protection secondary circuit, used to implement the method described in claims 1-9, characterized in that, include: The controllable load disturbance module is used to periodically apply a step load current, causing the load current of the secondary circuit of the relay protection to jump between the reference current and the test current. The distributed voltage acquisition module is used to synchronously acquire the voltage of each monitoring node based on a unified clock source, and obtain the voltage values ​​of each monitoring node before and after the application of a step load current. The feature extraction module is used to calculate the dynamic voltage drop of each monitoring node and extract contact resistance feature parameters based on the response relationship between the dynamic voltage drop and the step load current. The contact resistance feature parameters include linearity deviation coefficient, voltage drop fluctuation coefficient and response delay time. The loose connection identification module is used to normalize the contact resistance characteristic parameters, calculate the loose connection index of each monitoring node using a multi-dimensional feature weighted fusion algorithm, and determine whether there is a loose connection fault based on a preset threshold. The fault location module is used to calculate the voltage gradient anomaly coefficient for the identified loose connection nodes and, in combination with the secondary circuit topology information, locate the specific location of the loose connection fault point.

Images