Mouse linkage wireless intelligent triggering system and method for testing mechanical characteristics of switch cabinet

Abstract

The invention relates to the technical field of power system equipment detection, and discloses a mouse linkage wireless intelligent triggering system and method for a switch cabinet mechanical characteristic test, and the system comprises a motion collection module which is used for collecting the operation parameters of a mouse, building an operation behavior feature library, and transmitting the operation behavior feature data to a signal coding module in real time; the signal coding module is used for receiving the operation behavior characteristic data of the mouse, collecting an opening and closing instruction electric signal generated by a telecontrol background, extracting behavior characteristics and electric signal characteristics, and generating a unique opening characteristic code and / or a closing characteristic code; the tester intelligent receiving module is used for receiving the opening feature code and / or the closing feature code, and starting a dual mechanism of time window verification and feature code matching verification to verify the feature code; and the cloud collaborative management platform is used for receiving data uploaded by the action acquisition module, the signal coding module and the tester intelligent receiving module in real time. The efficiency of testing the mechanical characteristics of the switch cabinet is improved.

Description

Technical Field

[0001] This invention relates to the field of power system equipment testing technology, specifically to a mouse-linked wireless intelligent triggering system and method for testing the mechanical characteristics of switchgear. Background Technology

[0002] As core equipment in power systems, switchgear's mechanical characteristics (such as opening and closing time, speed, and travel) directly affect the safe and stable operation of the power system. Therefore, regularly conducting mechanical characteristic tests on switchgear is a crucial aspect of power operation and maintenance. Traditional testing methods require manually connecting secondary test leads (such as a three-wire connection between the closing coil, opening coil, and common terminal) to each switchgear, which has many technical drawbacks.

[0003] 1. Cumbersome and inefficient operation: The wiring preparation time for a single device can be as long as 15 minutes or more. In the scenario of centralized testing during a power outage in a substation, the testing work of a large number of devices is easily constrained by the wiring efficiency, which delays the overall power outage operation progress.

[0004] 2. High risk of misoperation: During manual wiring, errors such as incorrect wiring sequence or poor terminal contact can easily lead to deviations in test data, or even cause safety accidents such as short circuits or damage to equipment.

[0005] 3. Poor compatibility: The secondary circuit interfaces of switchgear from different manufacturers and models vary greatly, requiring the use of various adapter cables, which increases maintenance costs and operational complexity.

[0006] 4. Existing wireless solutions are inadequate: Some wireless triggering solutions rely solely on the overall electrical signal acquisition from the remote control backend, which cannot accurately locate the direct trigger source of the opening and closing operation (such as mouse operation), resulting in delayed signal recognition and a high false trigger rate (usually >1%), making it difficult to meet the requirements of high-precision testing.

[0007] Based on the above problems, there is an urgent need for an intelligent triggering solution that can accurately capture operation signals, eliminate the need for secondary wiring, adapt to multiple equipment models, and support batch testing, so as to solve the pain points of traditional testing and promote the intelligent and efficient upgrading of substation equipment operation and maintenance. Summary of the Invention

[0008] The core objective of this invention is to overcome the shortcomings of traditional switchgear mechanical characteristic testing and to provide a wireless intelligent triggering system and method that combines precise mouse operation capture, multi-module collaborative linkage, and intelligent verification feedback to achieve the following objectives.

[0009] This invention is achieved through the following technical solution:

[0010] A mouse-linked wireless intelligent triggering system for testing the mechanical characteristics of switchgear includes:

[0011] The motion acquisition module is installed on the operating mouse in the substation remote control backend. It is used to collect the operation parameters of the mouse, and combine them with the electrical signal feature association code obtained from the signal encoding module to establish an operation behavior feature library based on the operation parameters and electrical signal feature association code obtained when performing opening and closing operations. The operation behavior feature data is transmitted to the signal encoding module in real time. The operation parameters include the mouse coordinate position, click pressure, click duration and operation path trajectory.

[0012] The signal encoding module, either built into or externally connected to the substation remote control backend host, is wirelessly connected to the action acquisition module. It is used to receive the operation behavior feature data of the mouse and simultaneously acquire the opening and closing command electrical signals generated by the remote control backend. Based on the operation behavior feature data and the opening and closing command electrical signals, it extracts behavioral features and electrical signal features respectively. It performs anti-interference processing and data fusion on the behavioral features and electrical signal features respectively to generate a unique opening feature code and / or closing feature code, and transmits it to the intelligent receiving module of the tester via directional wireless transmission.

[0013] The intelligent receiving module of the tester receives the opening and / or closing feature codes transmitted by the signal encoding module. It parses the feature codes through the signal decoding unit, extracts the test type identifier, and initiates a dual mechanism of time window verification and feature code matching verification to verify the feature codes. After the verification is successful, it transmits the feature codes and corresponding test type instructions to the main control unit of the switch mechanical characteristic tester, and receives the program preload ready signal fed back by the main control unit. After receiving the program preload ready signal, it controls the electromagnetic induction starter set at the start measurement button position of the switch mechanical characteristic tester to be energized, which drives the iron core to strike the start measurement button. The main control unit starts the corresponding test program according to the test type instruction corresponding to the parsed feature code.

[0014] The cloud-based collaborative management platform, built on a cloud server, is used to receive data uploaded in real time from the motion acquisition module, signal encoding module, and intelligent receiving module of the test instrument, generate visual reports and operation trajectory maps, and automatically optimize and update the operation behavior feature library after every 50 test operations.

[0015] As an optimization, the motion acquisition module includes a controller, a photoelectric sensor, a pressure sensor, a Bluetooth unit, and a power supply unit. When the mouse moves, the photoelectric sensor captures the texture changes of the contact surface with the mouse in real time, converting the light signal into an electrical signal. The controller calculates the displacement difference between adjacent frame images to calculate the coordinate changes of the mouse in the horizontal and vertical directions. Combined with the initial coordinate reference point, the real-time coordinate position of the mouse during operation is finally generated; and the operation path trajectory is calculated based on the real-time coordinate position.

[0016] When the mouse is clicked, the pressure sensor collects the click pressure. When the pressure sensor detects that the pressure value reaches the effective click threshold, it immediately sends a click start trigger signal to the controller, and the main control unit starts a timer. When the pressure sensor detects that the pressure value drops to the click end threshold, it sends a click end trigger signal to the main control unit, and the main control unit stops the timer. The difference between the start time and the end time recorded by the timer is the click duration.

[0017] As an optimization, the specific process for establishing an operation behavior feature database containing operation behavior feature data of opening and closing is as follows:

[0018] An operation behavior feature library is constructed, which includes a tripping data subset and a closing data subset. Each operation behavior feature data in the tripping data subset and the closing data subset includes several core feature dimensions. The core feature dimensions include coordinate position range, motor pressure distribution, motor market distribution, operation path trajectory, and pressure change rate. The ranges of the same type of core feature dimensions in the tripping data subset and the closing data subset are different and do not overlap.

[0019] For each switchgear, several sets of core feature dimensions corresponding to opening and closing operations are collected, filtered, standardized, and statistically analyzed. Weights are assigned to each of the core feature dimensions to obtain the core electrical signal parameters corresponding to opening and closing operations, ultimately resulting in opening data subsets and closing data subsets.

[0020] The tripping and closing data subsets are updated periodically;

[0021] After conducting mechanical characteristic tests on the switchgear, if the mean deviation rate between the core feature dimension corresponding to the actual operation during the test and the corresponding tripping data subset or closing data subset is greater than the deviation threshold, then the tripping data subset and closing data subset are updated.

[0022] When a new switchgear or action acquisition module is connected, the tripping data subset and the closing data subset are updated.

[0023] As an optimization, a two-dimensional encoding combining behavioral features and electrical signal features is adopted. After anti-interference processing of the two-dimensional encoding, the specific process for generating a unique opening feature code and / or closing feature code is as follows:

[0024] Extract behavioral features from the operation behavior feature data and establish a dynamic threshold library based on the threshold range of behavioral features and electrical signal features. The behavioral features include basic parameters and derived parameters. The basic parameters include coordinate position, click pressure and click duration. The derived parameters are calculated from the basic parameters and include coordinate trajectory curvature and pressure change rate.

[0025] The electrical signal features of the opening and closing command electrical signals are extracted, including voltage amplitude, current change rate and signal rise slope, and a correlation model between the point signal features is constructed to eliminate interference signals.

[0026] Anti-interference processing is performed on the behavioral characteristics and electrical signal characteristics respectively;

[0027] After anti-interference processing, the behavioral and electrical signal characteristics are fused to form dual-dimensional feature data. Then, a salt value is dynamically embedded into the dual-dimensional feature data to form data to be hashed. The data to be hashed is subjected to SHA-256 hash operation, and the first 24 bytes are truncated as a feature digest segment. A basic identifier segment and a check code segment are added to the feature digest to obtain the corresponding opening or closing feature code. The basic identifier segment includes the device displacement ID, operation type identifier, acquisition timestamp, and signal source identifier.

[0028] As an optimization, a dual mechanism of time window verification and feature code matching verification is implemented to verify the feature code. Specifically:

[0029] Synchronize the intelligent receiving module with the signal encoding module in time;

[0030] After the intelligent receiving module receives the feature code, it extracts the acquisition timestamp as the mouse operation trigger time, denoted as T1, and simultaneously records the precise time of local reception of the feature code, denoted as T2.

[0031] Based on the average transmission delay ΔT of the feature code from several valid historical tests, the dynamic time window range is calculated as: [T1+ΔT-10ms, T1+ΔT+10ms];

[0032] If the local time T2 falls within the range of the dynamic time window, the time window verification passes; otherwise, the time window verification fails.

[0033] Extract the operation type identifier of the basic identifier segment, retrieve the corresponding data subset according to the operation type identifier, and perform feature summary similarity verification and feature dimension consistency verification. If both pass, the feature code matching verification passes; otherwise, the feature code matching verification fails.

[0034] As an optimization, the specific process of performing feature summary similarity verification is as follows:

[0035] The operation behavior feature data of each data subset is fused and SHA-256 hashed. The first 24 bytes are extracted to obtain the feature digest segment of each operation behavior feature data in the data subset. The average value of the feature digest segments corresponding to all operation behavior feature data in the data subset is calculated to obtain the feature digest baseline value.

[0036] The similarity between the feature digest segment of the received feature code and the feature digest reference value is calculated using the Hamming distance algorithm. The formula is: similarity = (1 - Hamming distance / total number of bits) × 100%, where the Hamming distance is the number of different bits in the feature digest segment and the feature digest reference value. If the similarity is ≥ 90%, the similarity check of the feature digest segment passes.

[0037] As an optimization, the specific process of performing feature dimension consistency verification is as follows:

[0038] Obtain the behavioral features and electrical signal feature threshold ranges of the corresponding data subset from the dynamic threshold library;

[0039] Parsing the feature summary segment yields the corresponding behavioral and electrical signal features;

[0040] Determine whether the behavioral features and electrical signal features fall within the corresponding threshold ranges for behavioral features and electrical signal features. If both do, the feature dimension consistency check passes; otherwise, the feature dimension consistency check fails.

[0041] As an optimization, the intelligent receiving module of the tester also includes a three-color LED ring indicator and a miniature vibration motor. A solid blue light indicates standby mode, a flashing green light indicates successful triggering, a slow red light indicates verification in progress, and alternating red and green flashing indicates that relearning is required. The vibration motor outputs a 50ms short vibration when the trigger is successful and a 300ms long vibration when there is an operational error or verification failure. The light intensity can be automatically adjusted according to the brightness value collected by the ambient light sensor.

[0042] This invention also discloses a mouse-linked wireless intelligent triggering method for testing the mechanical characteristics of switchgear based on the aforementioned system, comprising the following steps:

[0043] System deployment and initialization: Attach the motion acquisition module to the remote control backend operating mouse, complete the connection with the signal encoding module via Bluetooth pairing wizard, attach the tester intelligent receiving module to the designated position of the switchgear mechanical characteristic tester, and input the equipment number, substation name, and switchgear model information into the cloud collaborative management platform to complete equipment registration and grouping;

[0044] During the feature learning phase, maintenance personnel perform several opening and closing operations. The action acquisition module records the coordinate position range, average click pressure, and average click duration for each operation, generating an operation behavior feature dataset. The signal encoding module simultaneously collects the opening and closing command electrical signal parameters corresponding to each operation, fuses them to generate opening and closing feature codes, and stores them in the local storage unit and the cloud feature library, providing a data foundation for subsequent feature summary benchmark values.

[0045] During the test execution phase, maintenance personnel perform opening and closing operations via mouse in the remote control backend. The action acquisition module captures the operation parameters in real time and transmits them to the signal encoding module. The signal encoding module integrates the operation parameters with the real-time electrical signal to generate a dynamic feature code, which is transmitted to the intelligent receiving module of the tester via the 915MHz frequency band. The intelligent receiving module of the tester sequentially performs time window verification and feature code matching verification on the dynamic feature code. After the dual verification is passed, the electromagnetic induction starter is powered on, and the iron core strikes the start button of the tester. At the same time, the status is fed back through LED indicator lights and vibration motors, and the test data is uploaded to the cloud platform in real time.

[0046] During the feature database update and maintenance phase, after each set number of tests, the cloud-based collaborative management platform calculates the deviation between the latest operational behavior feature data and historical features, and uses a weighted average algorithm to optimize and update the operational behavior feature database. When the deviation value of a certain device exceeds the threshold for a set number of consecutive times, a re-feature learning reminder is triggered. After the operation and maintenance personnel perform the learning operation, the opening and closing feature codes of the device stored in the feature database are updated.

[0047] As an optimization, during the test execution phase, if the intelligent receiving module of the tester does not receive the feature code transmitted by the signal encoding module within a set distance, the signal reconnection mechanism is automatically activated. After a set number of failed reconnections, a signal interruption alarm is generated and uploaded to the cloud collaborative management platform.

[0048] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0049] Based on the "initial coordinate reference point" design of the motion acquisition module (bound to the opening and closing button area, with a coordinate error of ±0.1mm), combined with the acquisition of four-dimensional parameters of "coordinate position + click pressure + duration + trajectory", the system can accurately distinguish between opening and closing operations. Even if maintenance personnel are wearing insulated gloves, the system can ensure an operation recognition accuracy of ≥99.5% by "determining the effective area by coordinate deviation ≤5px" and "filtering out false touches by pressure of 50-300g". In conjunction with the "dual-dimensional encoding (behavioral features + electrical signal features)" and SHA-256 hash operation of the signal encoding module, the system completely avoids "operation and test misalignment" (such as opening operation triggering closing test).

[0050] The tester's intelligent receiving module employs a dual mechanism of "time window verification (dynamic time window ±10ms, synchronization error ≤1ms) + feature code matching verification (similarity ≥90% is considered passing)," which reduces the false trigger rate from 1%-3% in traditional solutions to below 0.1%.

[0051] This system upgrades the traditional "manual wiring - manual triggering - paper recording" testing mode to a digital mode of "wireless linkage - automatic identification - intelligent analysis" through a full-process design of "precise acquisition of operating parameters → dual-dimensional coding for anti-interference → dual verification for precise triggering → cloud-based collaborative intelligent management". In practical application, the annual testing time for a single 110kV substation is reduced by approximately 800 hours, the equipment misjudgment rate due to false triggering is reduced to below 0.1%, and the efficiency of new equipment integration is improved by 85%. It fully meets the power industry's operation and maintenance requirements of "safety first, efficiency priority, and compatibility with existing systems", providing core technical support for the digital transformation of switchgear mechanical characteristic testing. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments. The illustrative embodiments and descriptions of this invention are only used to explain this invention and are not intended to limit this invention.

[0053] This embodiment 1 provides a mouse-linked wireless intelligent triggering system for testing the mechanical characteristics of switchgear, including:

[0054] The motion acquisition module is installed on the operating mouse in the substation remote control backend. It is used to collect the operation parameters of the mouse, and combine them with the electrical signal feature association code obtained from the signal encoding module to establish an operation behavior feature library based on the operation parameters and electrical signal feature association code obtained when performing opening and closing operations. The operation behavior feature data is transmitted to the signal encoding module in real time. The operation parameters include the mouse coordinate position, click pressure, click duration and operation path trajectory.

[0055] Built-in or external to the substation remote control backend host, wirelessly connected to the action acquisition module, used to receive the operation behavior feature data of the mouse, and simultaneously acquire the opening and closing command electrical signals generated by the remote control backend. Based on the operation behavior feature data and the opening and closing command electrical signals, behavioral features and electrical signal features are extracted respectively. Anti-interference processing and data fusion are performed on the behavioral features and electrical signal features respectively to generate a unique opening feature code and / or closing feature code, and then transmitted to the intelligent receiving module of the tester via directional wireless transmission.

[0056] The intelligent receiving module of the tester receives the opening and / or closing feature codes transmitted by the signal encoding module. It parses the feature codes through the signal decoding unit, extracts the test type identifier, and initiates a dual mechanism of time window verification and feature code matching verification to verify the feature codes. After the verification is successful, it transmits the feature codes and corresponding test type instructions to the main control unit of the switch mechanical characteristic tester, and receives the program preload ready signal fed back by the main control unit. After receiving the program preload ready signal, it controls the electromagnetic induction starter set at the start measurement button position of the switch mechanical characteristic tester to be energized, which drives the iron core to strike the start measurement button. The main control unit starts the corresponding test program according to the test type instruction corresponding to the parsed feature code.

[0057] The cloud-based collaborative management platform, built on a cloud server, is used to receive data uploaded in real time from the action acquisition module, signal encoding module, and intelligent receiving module of the tester, generate visual reports and operation trajectory maps, and automatically optimize and update the opening / closing feature library after every 50 test operations.

[0058] Next, we will introduce the functions and structure of these modules.

[0059] Motion capture module:

[0060] The motion acquisition module adopts a magnetic detachable structure and is attached to the dedicated operating mouse of the substation remote control backend. It has built-in high-precision photoelectric sensor, pressure sensor, low-power Bluetooth module and lithium battery to accurately collect mouse operation parameters, including coordinate position, click pressure, click duration and operation path trajectory, to establish a feature library of opening and closing operation behavior, and transmit the operation feature data to the remote control backend signal encoding module in real time through Bluetooth low power protocol.

[0061] The mouse-integrated motion acquisition module boasts a lithium battery life of ≥72 hours, coordinate position acquisition accuracy of ±0.1mm, an adjustable click pressure threshold range of 50-300g, a click duration acquisition range of 10-1000ms, and operation feature data transmission latency of ≤10ms. The module features a physical button on its surface for one-click learning; pressing it activates the opening / closing operation behavior feature acquisition mode. The motion acquisition module housing the one-click learning button uses a magnetically detachable structure and is directly attached to the surface of the dedicated operating mouse for the substation remote control system (e.g., on the side or top of the mouse for easy access). This button, as the core operating component of the motion acquisition module, is fixed to the mouse along with the module, forming an integrated operating unit. Maintenance personnel can directly press the button to initiate feature acquisition without needing to find or operate other devices while holding the mouse, meeting the design goal of convenient operation.

[0062] This design is not simply a function trigger, but rather addresses the core pain points of substation operation and maintenance scenarios. It optimizes the feature collection process from multiple dimensions, including operational efficiency, data collection accuracy, and system adaptability. The specific objectives are as follows:

[0063] I. Core Objective 1: Simplify feature acquisition operations and adapt to the needs of efficient operation and maintenance of substations:

[0064] (1) Solve the pain point of cumbersome traditional feature acquisition operations

[0065] Traditional wireless triggering systems often require users to access the setup mode through remote control software and navigate through multiple menu steps, such as feature acquisition, opening acquisition, and closing acquisition. Some systems also require an administrator password. The entire process takes at least 5 minutes and is prone to errors by maintenance personnel due to the complexity of the menu paths (e.g., accidentally selecting the parameter reset function). The one-click learning button replaces these multiple software steps with direct hardware triggering. Maintenance personnel do not need to be familiar with the backend software's operation logic; they only need to press the physical button on the module surface (8mm in diameter, 0.5mm raised for easy blind pressing). The feature acquisition mode is activated within 1 second, simplifying the operation from 5+ steps to 1 step. The feature acquisition preparation time is reduced from at least 5 minutes to at least 10 seconds, significantly meeting the core needs of batch testing and efficient operation during substation power outages. For example, when testing 20 devices in batches, feature acquisition preparation for all devices can be completed in just 200 seconds, saving over 90% of the time compared to traditional methods.

[0066] II. Core Objective 2: To ensure accurate and controllable feature acquisition and avoid interference from invalid data.

[0067] (1) Clearly define the data collection start point to eliminate the risk of erroneous data collection:

[0068] If no physical button is set, the system may trigger feature collection due to mouse misoperation (such as maintenance personnel accidentally clicking the backend interface), resulting in invalid operation data collected (such as the operation trajectory of accidentally clicking the parameter query button). Subsequent testing may easily lead to false triggering due to feature code mismatch.

[0069] The one-click learning button is activated by active pressing, clearly indicating the data collection start point: the button is pressed only when maintenance personnel confirm the need to collect features, in which case the collected 3 opening + 3 closing operations are conscious and standard operations, effectively filtering out invalid data generated by erroneous operations. Simultaneously, after the button is triggered, the module will provide a visual cue with a "blue flashing (1 time / second)," informing maintenance personnel that data collection mode has been entered, further enhancing operational controllability and ensuring a high degree of match between the collected feature data and the actual opening and closing operations.

[0070] In some embodiments, the motion acquisition module includes a controller, a photoelectric sensor, a pressure sensor, a Bluetooth unit, and a power supply unit. When the mouse moves, the photoelectric sensor captures the texture changes of the contact surface in real time, converting the light signal into an electrical signal. The controller calculates the displacement difference between adjacent frames to determine the coordinate changes of the mouse in the horizontal and vertical directions. Combined with an initial coordinate reference point, the controller generates the real-time coordinate position of the mouse during operation and calculates the operation path trajectory based on the real-time coordinate position.

[0071] When the mouse is clicked, the pressure sensor collects the click pressure. When the pressure sensor detects that the pressure value reaches the effective click threshold, it immediately sends a click start trigger signal to the controller, and the main control unit starts a timer. When the pressure sensor detects that the pressure value drops to the click end threshold, it sends a click end trigger signal to the main control unit, and the main control unit stops the timer. The difference between the start time and the end time recorded by the timer is the click duration.

[0072] More specifically, a detailed explanation of how to obtain the operating parameters is provided.

[0073] 1. Coordinate position: obtained through a high-precision photoelectric sensor.

[0074] 1. Core Components: The motion acquisition module incorporates a high-precision photoelectric sensor (resolution ≥1000dpi), which is essentially a miniature optical imaging system, including a light-emitting diode (LED), an optical lens, a photosensitive chip (CMOS), and a signal processing unit.

[0075] 2. Data Acquisition Principle:

[0076] The photoelectric sensor emits infrared light through an LED. The light is reflected by the contact surface on the bottom of the mouse (such as a mouse pad or desktop) and then focused onto the photosensitive chip by an optical lens.

[0077] When the mouse moves, the photosensitive chip captures the texture changes of the contact surface in real time, converts the light signal into an electrical signal, and the signal processing unit calculates the displacement difference between adjacent frames to calculate the coordinate changes of the mouse in the horizontal (X-axis) and vertical (Y-axis) directions.

[0078] Combining the initial coordinate reference point recorded during the initialization of the motion acquisition module, such as the origin of the coordinates of the area where the opening and closing buttons are located on the remote control backend interface, the real-time coordinate position during mouse operation is finally generated. For example, X: 250px, Y: 450px.

[0079] The initial coordinate reference point is the origin of coordinate calculation set during the initialization phase of the motion acquisition module. Its core function is to convert the relative displacement of the mouse into the absolute coordinates of the opening and closing button area on the remote control backend interface, ensuring accurate correlation between the coordinate position acquisition and the opening and closing operation area. The specific definition and implementation process are as follows:

[0080] The initial coordinate reference point is essentially a virtual coordinate origin bound to the opening and closing button area of ​​the remote control backend interface when the motion acquisition module is first installed or reset. Its core value is to establish the mapping relationship between the physical displacement of the mouse and the pixel coordinates of the remote control backend interface. Without this reference point, the photoelectric sensor can only acquire the relative displacement of the mouse, such as moving 5mm to the right or 3mm down, and cannot determine the specific position of the mouse on the interface (such as whether it falls on the opening button area). The specific characteristics are as follows:

[0081] Coordinate format: Based on the pixel coordinates of the remote control backend interface, it is expressed as a two-dimensional coordinate value of (X0,Y0) (e.g., X0=200px, Y0=400px), where px is the pixel unit of the interface and is unrelated to the resolution of the remote control backend monitor. The resolution difference will be eliminated through standardization in the future.

[0082] Binding object: The reference point is not a fixed "screen corner", but is strongly associated with the opening and closing button area. Usually, the geometric center of the opening button area is selected. For example, if the opening button is a 20×20px rectangle, the center coordinates are the reference point. Or the geometric center of the closing button area, to ensure that the coordinate acquisition directly serves to determine whether the operation is for the opening and closing buttons.

[0083] Stability: Once the reference point is set, it remains unchanged as long as the module is not reset and the remote control backend interface is not reconstructed. Even if the mouse is powered off and restarted, the module will read the stored reference point data from the local Flash without needing to be reset.

[0084] The specific implementation requires a combination of manual calibration by maintenance personnel and automatic storage by the module to ensure that the reference point perfectly matches the actual position of the opening and closing buttons. The process is as follows:

[0085] Step 1: Enter reference point calibration mode (triggering condition):

[0086] When the motion acquisition module is first attached to the mouse, or when the reference point needs to be reset due to an update to the remote control backend interface (such as an adjustment of the position of the opening and closing buttons), the calibration mode is triggered in the following ways:

[0087] Manual trigger: Press and hold the one-click learning button for 5 seconds (different from the short press of 1 second for feature acquisition), and the module indicator light will change from solid blue to alternating blue and green flashing, indicating that the coordinate reference point calibration mode has been entered;

[0088] Automatic Trigger: If the module detects that the current mouse displacement deviates from the button area corresponding to the stored reference point by more than 20px (such as when the resolution of the remote control backend interface changes from 1920×1080 to 2560×1440), it will automatically trigger a calibration reminder (slow red LED flash + short vibration of the vibration motor) to prompt the maintenance personnel to reset the reference point.

[0089] Step 2: Manually calibrate the reference point (core operation):

[0090] Maintenance personnel need to complete precise positioning and confirmation saving in the remote control backend interface, ensuring that the reference point is aligned with the opening and closing button area. The specific operation is as follows:

[0091] Locating the opening and closing button area: The maintenance personnel hold a mouse and move the mouse cursor to the opening button area on the remote control backend interface, and adjust the cursor position to the geometric center of the button (the button selection box displayed on the interface can be used to assist in positioning; some remote control backends support the button center highlighting function).

[0092] Confirm reference point: Press the left mouse button (or the confirmation sub-button on the module surface, if present). The motion acquisition module sends a reference point sampling request to the remote control backend signal encoding module via Bluetooth. The signal encoding module reads the pixel coordinates of the current cursor on the interface (e.g., the center coordinates of the trip button are (200px, 400px)) and sends the coordinates back to the motion acquisition module.

[0093] Synchronous closing button reference point: After the opening button reference point is set, the module indicator light turns "green solid", prompting to switch to closing button calibration; the maintenance personnel repeat the above operation, move the cursor to the center of the closing button, and obtain the reference point corresponding to the closing button after confirmation, such as (500px, 400px).

[0094] Step 3: Automatic storage and standardization (internal module operation):

[0095] After the module obtains the reference point corresponding to the open / close button, it automatically performs the following operations to ensure the accuracy of subsequent coordinate acquisition:

[0096] Local storage: The tripping reference point (X1,Y1), closing reference point (X2,Y2) and calibration timestamp (accurate to the second) are stored in the module's built-in 8MB Flash. The storage address is fixed (e.g., 0x00010000-0x0001001F) for easy and fast subsequent reading.

[0097] Resolution standardization: To eliminate the impact of different display resolutions (e.g., the same button has coordinates of (200, 400) in a 1920×1080 interface and (267, 533) in a 2560×1440 interface), the module converts the reference point coordinates into a proportional value relative to the interface width / height. For example, the X coordinate of the center of the circuit breaker button is 200px in a 1920×1080 interface, which is converted to a proportional value of 200 / 1920≈0.104. When collecting coordinates later, it is only necessary to reverse the calculation according to the current interface resolution to obtain a unified relative position, ensuring compatibility between different devices.

[0098] 2. Click pressure: obtained through a pressure sensor.

[0099] 1. Core Components: The module has a built-in pressure sensor (range 0~500g), which adopts the piezoresistive sensing principle. The sensor surface is equipped with an elastic sensing diaphragm, and a strain resistance network is integrated under the diaphragm.

[0100] 2. Data Acquisition Principle:

[0101] When the maintenance personnel press the mouse button, the mechanical force of the button is transmitted through the module shell to the sensing diaphragm of the pressure sensor, and the diaphragm undergoes a slight deformation (the deformation amplitude is proportional to the pressure).

[0102] The strain gauge network beneath the diaphragm changes its resistance due to deformation. This change in resistance is converted into a voltage signal (voltage range 0~3.3V) by a Wheatstone bridge circuit.

[0103] The built-in signal amplification unit of the motion acquisition module amplifies and filters the voltage signal, and then converts the analog signal into a digital signal through an A / D converter, which is finally converted into a click pressure value (e.g., 120g) to realize the function of acquiring click pressure. At the same time, the pressure threshold can be set by software (e.g., below 50g is determined to be an invalid touch), and the logic of determining that the operation is invalid if the subsequent click pressure exceeds the range can be matched.

[0104] 3. Click duration: obtained through the collaboration of pressure sensor and module main control unit.

[0105] 1. Core Components: The pressure sensor is used as the trigger source, combined with the controller built into the motion acquisition module (such as the STM32L4 series low-power MCU, which integrates a high-precision timer), with a timer accuracy of ≤1ms.

[0106] 2. Data Acquisition Principle:

[0107] When the pressure sensor detects that the pressure value has reached the effective click threshold (e.g., ≥50g), it immediately sends a click start trigger signal to the controller, and the main control unit starts the timer to begin counting.

[0108] When the pressure sensor detects that the pressure value has dropped to the click end threshold (e.g., ≤10g), it sends a click end signal to the controller, and the main control unit stops the timer.

[0109] The difference between the start and end times recorded by the timer is the click duration (e.g., 90ms). If the duration is <10ms (accidental touch) or >1000ms (long press), the subsequent operation is invalidated, thus ensuring the accuracy of the click duration collection.

[0110] IV. Operation path trajectory: Acquired through collaboration between high-precision photoelectric sensors and the controller.

[0111] 1. Core Components: Using a high-precision photoelectric sensor as the trajectory acquisition source, combined with the controller's data buffer and stitching unit, it supports ≥100 coordinate samplings per second to ensure trajectory continuity.

[0112] 2. Data Acquisition Principle:

[0113] The photoelectric sensor collects continuous coordinate points during the mouse movement in real time (e.g., 100 X / Y coordinate values ​​per second), and each coordinate point is accompanied by a collection timestamp (generated by the main control unit timer).

[0114] The controller caches continuous coordinate points and timestamp data in chronological order to local storage units (such as 8MB Flash), and uses a linear interpolation algorithm to smooth adjacent coordinate points, eliminating trajectory breakpoints caused by signal jitter.

[0115] Finally, the continuous coordinate points are spliced ​​together according to the time sequence to form a complete operation path trajectory (such as moving from coordinates (100,200) to (250,450), and then staying for 100ms before clicking). The function of collecting operation path trajectories provides trajectory dimension data support for the subsequent establishment of the opening and closing operation behavior feature library (such as distinguishing the difference between directly clicking the closing button and accidentally touching other areas before clicking).

[0116] Through the above correspondence, the motion acquisition module can accurately acquire multi-dimensional parameters of mouse operations, providing core data support for establishing a feature library of opening and closing operations, and achieving the technical goal of accurately capturing operation signals.

[0117] In some embodiments, the specific process of establishing an operation behavior feature database containing operation behavior feature data of opening and closing is as follows:

[0118] Step 1: Construct an operation behavior feature library. This library includes a subset of opening and closing data. Each operation behavior feature data set within both subsets includes several core feature dimensions. These core feature dimensions include coordinate position range, motor pressure distribution, motor market distribution, operation path trajectory, and pressure change rate. Furthermore, the ranges of the same type of core feature dimensions differ and do not overlap between the opening and closing data subsets. Define the core dimensions of the feature library, covering key parameters throughout the entire operation process. To ensure the feature library can accurately distinguish between opening and closing operations, based on the acquisition capabilities of the mouse-integrated motion acquisition module, define 5 core feature dimensions. Each dimension corresponds to two independent data subsets: opening and closing. Specific dimension and parameter requirements are as follows:

[0119] Feature Dimension Tripping operation feature sub-dimension Closing operation feature sub-dimension Data acquisition accuracy requirements Coordinate position range The coordinate range of the "shutdown button" in the remote control backend (e.g., X: 200-300px, Y: 400-500px). The coordinate range of the "close button" in the remote control backend (e.g., X: 500-600px, Y: 400-500px). Coordinate accuracy ±0.1mm, range error ≤5px Click on pressure distribution Mean and standard deviation of the trigger pressure during the tripping operation (e.g., mean 120g, standard deviation 20g). Mean and standard deviation of the closing operation's click pressure (e.g., mean 130g, standard deviation 18g). Pressure range 50-300g, sampling frequency 500Hz Click duration distribution Mean and standard deviation of the click duration for the tripping operation (e.g., mean 90ms, standard deviation 15ms). Mean and standard deviation of the closing operation click duration (e.g., mean 85ms, standard deviation 12ms). Duration range 10-1000ms, timing accuracy 1ms Operation path trajectory The mouse movement trajectory from the "initial position" to the "open button" (e.g., linear movement with no obvious offset). The mouse movement trajectory from the "initial position" to the "close button" (e.g., linear movement, offset ≤ 10px). Trajectory sampling frequency 100Hz, offset error ≤2px Pressure change rate "Pressure rise time" during circuit breaker tripping operation (the time from 10% pressure to 90% pressure, such as 20ms). "Pressure rise time" during closing operation (e.g., 18ms) Time accuracy 1ms, change rate calculation error ≤5%

[0120] Each dimension must clearly define the differential threshold between opening and closing (e.g., no overlap in coordinate intervals, pressure mean difference ≥10g) to provide a basis for differentiation in subsequent feature matching.

[0121] Step 2: For each switchgear, collect several sets of core feature dimensions corresponding to opening and closing operations, and then filter and standardize them to obtain opening and closing data subsets. Feature data collection and filtering ensure data validity. The basic data of the feature library comes from the operation collection during the feature learning phase. The collection process needs to be triggered by one-click learning, and invalid data must be strictly filtered. The specific steps are as follows:

[0122] 1. Data Acquisition Trigger: When maintenance personnel press the one-click learning button on the integrated action acquisition module with the mouse, the module indicator light turns blue and flashes rapidly (1 time / second), entering the opening-closing feature acquisition mode. At the same time, a data acquisition start signal is sent to the remote control background signal encoding module to prepare for electrical signal feature acquisition.

[0123] 2. Circuit breaker data acquisition: The operation and maintenance personnel perform 3 standard circuit breaker operations on the remote control backend interface (each operation interval ≥ 5 seconds). The module collects 5 core dimension data of each operation at a sampling frequency of 100Hz, forming 3 sets of raw circuit breaker datasets (each set includes coordinate sequence, pressure sequence, duration value, trajectory point, and pressure change rate).

[0124] 3. Closing data acquisition: After the opening data acquisition is completed, the module indicator light will flash green, prompting the switch to closing data acquisition; the maintenance personnel will perform 3 standard closing operations to acquire 3 sets of raw data sets of the same combination of closing devices.

[0125] 4. Invalid data filtering: Based on the judgment rules, the 6 sets of original data are filtered as follows:

[0126] Operations with a pressure of <50g or >300g and a duration of <10ms or >1000ms are directly deemed invalid and need to be re-collected (e.g., if one of the three tripping operations has a pressure of 40g, it needs to be re-collected to ensure that there are ≥3 sets of valid data).

[0127] Operations whose coordinate trajectory exceeds the range of the open / close button (such as the mouse shifting to the parameter query area when clicking open) are considered invalid, and ≥3 sets of valid data are collected.

[0128] Operations with abnormal pressure change rates (such as a sudden spike to 50ms, far exceeding the normal range) are deemed invalid. If the abnormality persists after wavelet transform smoothing, re-sampling is required.

[0129] 5. Data Standardization: Standardize the filtered valid data, such as converting the coordinate position into an offset relative to the button center (eliminating the influence of different screen resolutions) and normalizing the pressure value to the 0-1 range (eliminating individual differences of sensors), to form a standardized opening / closing feature dataset.

[0130] Step 3: Feature library structure design, hierarchical storage and indexing, supporting fast matching.

[0131] The feature library adopts a dual-storage architecture of local caching + cloud backup. The local library is stored in the 8MB Flash of the mouse-integrated motion acquisition module, while the cloud library is stored in the MySQL distributed database of the cloud collaborative management platform. The two have the same structure, and the specific layered design is as follows:

[0132] (1) First layer: Device identification layer (index layer), used to quickly locate the feature data of different devices, containing 3 core fields:

[0133] Unique device ID: Consists of substation number (4 digits) + switch cabinet bay number (2 digits) + mouse module number (4 digits), such as "1101-01-0001" representing mouse module #0001 in bay #1 of substation #1 of 110kV #1.

[0134] Feature library version number: uses the format V+year+month+update number (e.g., V20250901 represents the first update in September 2025) to distinguish feature data from different periods;

[0135] Learning timestamp: Records the time of the last update of the feature library (accurate to the second), used to determine whether relearning is needed. If it has not been updated for more than 6 months, it will automatically prompt for calibration.

[0136] (2) Second layer: Feature data layer (core layer), which stores standardized feature data according to the tripping data subset and the closing data subset respectively. Each subset contains:

[0137] Statistical characteristics: mean, standard deviation, maximum and minimum values ​​of 5 core dimensions, such as the mean of the tripping click pressure of 120g, standard deviation of 20g, maximum value of 140g and minimum value of 100g;

[0138] Sample data: Retain the original standardized data of 3 sets of valid operations, such as the 3 coordinate trajectory sequences of the circuit breaker opening operation, for comparative analysis during subsequent feature updates;

[0139] Threshold range: The effective judgment threshold for each dimension, such as 200-300px for the X-axis of the gate opening coordinate. If it exceeds the threshold, it is judged as a mismatch. The threshold range is calculated by "mean ± 1.5 times standard deviation".

[0140] (3) Third layer: Verification association layer (auxiliary layer), / correspondence between closing operation and electrical signal characteristics, to ensure that behavioral characteristics and electrical signal characteristics are logically consistent:

[0141] Electrical signal feature association code (implemented in the signal encoding module): Stores the core parameters of the electrical signal corresponding to the opening / closing operation, such as the opening electrical signal voltage 220V±5%, duration 180ms±10%, and current change rate 0.5A / ms;

[0142] Matching weights: Set the matching weights for each behavioral feature dimension, such as coordinate position weight 0.3, click pressure weight 0.2, click duration weight 0.2, path trajectory weight 0.15, and pressure change rate weight 0.15, which are used for similarity calculation during subsequent feature code matching.

[0143] In other words, the data subset contains the original standardized data (core feature dimensions) as well as the mean, standard deviation, weights of each original standardized data, and electrical signal feature association codes.

[0144] Fourth step: Dynamically optimize the feature library to adapt to changes in equipment and operation.

[0145] The feature library is not static and needs to be dynamically optimized through regular updates and anomaly triggering mechanisms to ensure long-term matching accuracy. The specific mechanism is as follows:

[0146] 1. Regular updates (every 50 tests):

[0147] Triggering condition: The cloud-based collaborative management platform will automatically start updating the feature library after every 50 valid tests (one for opening and one for closing).

[0148] Update logic: Extract the operational behavior data from the most recent 50 tests, calculate the new mean and standard deviation for each dimension, and perform a weighted average with the existing feature library data (new data weight 0.6, old data weight 0.4) to update the statistical features and threshold range of the feature library;

[0149] Example: The original average pressure of the trip button was 120g, and the average of the last 50 tests was 125g. After the update, the average value is 120×0.4+125×0.6=123g. The threshold range is adjusted accordingly.

[0150] 2. Abnormal trigger update (feature deviation exceeds 15%):

[0151] Monitoring logic: After each test, the system calculates the deviation rate between the real-time operation feature and the mean value of the feature library (deviation rate = |real-time value - mean value| / mean value × 100%).

[0152] Triggering condition: When the deviation rate of a certain dimension exceeds 15% for 3 consecutive times, such as the average click duration of the gate opening is 90ms, the real-time value is 105ms, the deviation rate is 16.7%, and this happens 3 times in a row, the cloud platform will send a relearning reminder via web pop-up and APP push.

[0153] Update process: Maintenance personnel press the one-click learning button to re-collect the data from 3 opening and 3 closing operations, overwriting the original feature database data and completing the update calibration.

[0154] 3. Cross-device adaptation updates:

[0155] When new equipment is connected (such as replacing the switch cabinet or mouse module), the maintenance personnel manually upload the equipment model parameters on the cloud platform, and the platform automatically matches the preset initial feature code template (such as the trip coordinate template and pressure template of the KYN28-12 switch cabinet).

[0156] Maintenance personnel only need to perform one opening and one closing operation. The system generates a new feature library based on the initial template and actual operation data, which greatly shortens the access time of new equipment (from 2 minutes to 30 seconds).

[0157] The feature library built through the above process can achieve two core objectives:

[0158] Distinguishing features: There are significant differences in the feature dimensions of opening and closing (e.g., no overlap in coordinate intervals, and a difference in average pressure ≥10g). Similarity calculation can quickly distinguish them (similarity between opening and closing feature library <50%, and similarity with opening feature library ≥90%).

[0159] Adaptability: The dynamic update mechanism can compensate for device aging (such as pressure deviation caused by mouse sensor wear) and changes in operating habits (such as differences in click duration caused by changes in maintenance personnel), ensuring that the feature library is effective in the long term and reducing the false trigger rate to below 0.1%.

[0160] Signal encoding module:

[0161] The signal encoding module, either built-in or externally connected to the substation remote control backend host, communicates with the host via USB and establishes a Bluetooth connection with the mouse-integrated motion acquisition module. It receives mouse operation behavior data and simultaneously acquires the opening and closing command electrical signals generated by the remote control backend. Using a dual-dimensional encoding approach combining "behavioral characteristics + electrical signal characteristics," and after Manchester encoding anti-interference processing, it generates unique opening and closing feature codes. These are then transmitted wirelessly to the tester's intelligent receiving module via 915MHz UHF band directional wireless transmission. The signal encoding module's 915MHz UHF band directional wireless transmission distance is ≥50 meters, and the signal error rate is ≤0.01%. Furthermore, it has a built-in signal filtering unit that can filter 50Hz power frequency interference and 2.4GHz wireless signal interference, ensuring the stability of the opening / closing feature code transmission.

[0162] In some embodiments, a two-dimensional encoding combining behavioral features and electrical signal features is used. After anti-interference processing is applied to the two-dimensional encoding, the specific process for generating a unique opening feature code and / or closing feature code is as follows:

[0163] A1. Extract the behavioral features of the operation behavior feature data and establish a dynamic threshold library. The behavioral features include basic parameters and derived parameters. The basic parameters include coordinate position, click pressure and click duration. The derived parameters are calculated from the basic parameters and include coordinate trajectory curvature and pressure change rate.

[0164] This step involves extracting behavioral features and setting dynamic thresholds to adapt to different operating habits.

[0165] The mouse-integrated motion acquisition module collects operation behavior features in layers based on basic parameters and derived parameters. Basic parameters include coordinate position (real-time X / Y axis coordinates, sampling frequency 100Hz), click pressure (sampling frequency 500Hz), and click duration (accuracy 1ms). Derived parameters are calculated from basic parameters, such as coordinate trajectory curvature (to determine whether the operation is a straight-line click or an offset click) and pressure change rate (Δpressure / Δtime, identifying differences in operation force such as light touch and heavy press). To address the differences in operating habits among different maintenance personnel, the system establishes a dynamic threshold library through initial learning and real-time optimization. In the initial learning phase, data from 10 circuit breaker opening and closing operations by 3 maintenance personnel is collected each, and the mean and standard deviation of each parameter are calculated (e.g., mean click pressure 120g, standard deviation 20g). The initial threshold range is set to mean ± 1.5 times the standard deviation. In the real-time optimization phase, the threshold range is automatically updated every 20 tests (update increment ≤ 5%) to avoid feature extraction deviations caused by changes in operating habits.

[0166] A2. Extract the electrical signal features of the opening and closing command electrical signal, including voltage amplitude, current change rate and signal rise slope, and construct a correlation model between the point signal features to eliminate interference signals.

[0167] This step involves extracting electrical signal features and using multi-dimensional parameter correlation to eliminate environmental interference.

[0168] The signal encoding module synchronously acquires three core parameters of the opening and closing command electrical signal using a high-precision ADC (16-bit resolution, 1kHz sampling frequency): voltage amplitude (DC220V±10%), current change rate (Δcurrent / Δtime, reflecting the coil energizing speed), and signal rise slope (the time for the voltage to rise from 10% to 90% of the rated value, identifying the presence of "glitches" in the signal). A correlation model is established between the electrical signal parameters, such as a linear fit between voltage amplitude and current change rate (in a normal opening command, for every 10V increase in voltage, the current change rate increases by 0.5A / ms). If the real-time acquired parameters deviate from this model by more than 8%, it is determined to be an "interference signal," automatically triggering a secondary acquisition to avoid feature extraction errors caused by power grid fluctuations.

[0169] A3. Perform anti-interference processing on the behavioral characteristics and electrical signal characteristics respectively.

[0170] This step performs anti-interference processing on behavioral and electrical signal characteristics, and layered protection achieves a significant improvement in signal stability. The anti-interference processing process is as follows:

[0171] (a) First layer: Source filtering, based on interference suppression using adaptive notch filtering and wavelet transform.

[0172] Power frequency interference filtering: The built-in adaptive notch filter circuit dynamically adjusts the notch depth by detecting the 50Hz power frequency interference intensity in real time (sampling frequency 10kHz). When the interference intensity is >50mV, the notch depth is increased to 40dB; when the interference intensity is <10mV, the notch depth is reduced to 20dB, thus avoiding the attenuation of useful signals caused by a fixed notch depth.

[0173] High-frequency interference suppression: Wavelet transform is used to denoise the electrical signal and behavioral feature data. The electrical signal features are decomposed into three levels using the db4 wavelet basis to remove high-frequency noise components (frequency > 10kHz). The coordinate trajectory data in the behavioral features are decomposed into two levels using the sym5 wavelet basis to eliminate trajectory deviations caused by mouse mechanical jitter. The signal-to-noise ratio of the signal is improved by ≥ 20dB after denoising.

[0174] (ii) Second layer: transmission protection, multi-band fusion and frequency hopping mechanism.

[0175] Multi-band parallel transmission: Breaking through the limitations of traditional single-band transmission, it adopts a fusion transmission of "915MHz main frequency band + 2.4GHz backup frequency band". The main frequency band (915MHz) is used to transmit complete feature data (bandwidth 250kbps) and has strong penetration (can penetrate 3mm steel plate); the backup frequency band (2.4GHz) is used to transmit feature data digest (16-byte check code) and has high transmission rate (bandwidth 1Mbps).

[0176] Dynamic frequency hopping mechanism: The system detects the interference intensity of the two frequency bands in real time (judged by the Received Signal Strength Indicator (RSSI) value; RSSI < -80dBm is considered strong interference). When the interference intensity of the main frequency band exceeds the threshold, it automatically switches to the backup frequency band for transmission, with a switching time of <10ms. At the same time, it records the interference frequency band and time, generates an interference spectrum, and prioritizes avoiding high interference frequency bands during subsequent tests, reducing the transmission bit error rate from 0.01% to below 0.001%.

[0177] A4. After anti-interference processing, the behavioral features and electrical signal features are fused to form two-dimensional feature data. Then, a salt value is dynamically embedded into the two-dimensional feature data to form data to be hashed. The data to be hashed is subjected to SHA-256 hash operation, and the first 24 bytes are truncated as a feature digest segment. A basic identifier segment and a check code segment are added to the feature digest to obtain the corresponding opening feature code or closing feature code.

[0178] This step involves feature code generation, based on multi-dimensional identification and dynamic encryption for uniqueness. The specific process is as follows:

[0179] (I) Feature Code Structure Design: Multi-field hierarchical identification ensures uniqueness:

[0180] The basic identifier segment (16 bytes) includes a unique device ID (6 bytes, generated based on substation number + switchgear number), an operation type identifier (1 byte, 0x01 for closing, 0x02 for opening), a data acquisition timestamp (4 bytes, accurate to milliseconds), a signal source identifier (1 byte, 0x0A for mouse operation, 0x0B for background automatic command), and a reserved field (4 bytes, for future function expansion). After the action acquisition module captures a valid opening / closing operation, it sends the operation behavior data (coordinates, pressure, duration) to the remote control background signal encoding module, along with an operation trigger signal (a 5V high-level pulse lasting 1ms). The signal encoding module, upon receiving the operation trigger signal, immediately reads the local real-time clock (synchronized with the cloud via NTP, error ≤1ms), generates a data acquisition timestamp in a standardized format, and writes it into the basic identifier segment of the feature code. For example, when the trigger signal arrives, the local clock displays "2025-09-18 14:30:00.123", and the encoding module directly converts this time into a 4-byte hexadecimal code as the data acquisition timestamp field.

[0181] Feature Digest (24 bytes): The two-dimensional feature data is hashed using SHA-256, and the first 24 bytes are extracted as the feature digest to ensure the uniqueness of feature digests for different operations (hash collision probability < 10). -18 ); at the same time, dynamic salt values ​​are incorporated (updated once per hour, with salt values ​​distributed by the cloud platform) to prevent feature summaries from being maliciously forged.

[0182] Verification segment (8 bytes): contains dynamic threshold version number (2 bytes, identifying the current threshold library version), anti-interference processing identifier (2 bytes, recording the frequency band and filtering method used during transmission), and final check code (4 bytes, generated by XOR operation of basic identifier segment and feature digest segment).

[0183] The specific process of hashing the feature digest segment is as follows:

[0184] 1. Data Standardization and Fusion: The signal encoding module first standardizes and integrates behavioral feature data and electrical signal feature data according to a "fixed field order": Behavioral feature fields (32 bytes in total): coordinate range (X-axis minimum / maximum value, 8 bytes), click pressure mean / standard deviation (8 bytes), click duration mean / standard deviation (8 bytes), trajectory curvature (8 bytes); Electrical signal feature fields (24 bytes in total): voltage amplitude (8 bytes), current change rate (8 bytes), signal rise slope (8 bytes); After fusion, a 56-byte two-dimensional feature dataset is formed, with a fixed and unique field order, ensuring that the fused data format is completely consistent for each identical operation.

[0185] 2. Dynamic Salt Embedding: Obtain the dynamic salt value (16 bytes, such as "2025091810ABCDEF", updated once per hour) from the cloud platform for the current hour, and insert the salt value at the beginning of the fused dataset to form 16 bytes of salt value + 56 bytes of feature data = 72 bytes of data to be hashed. The salt value embedding can ensure that even if the feature data of two operations are exactly the same (such as repeated operations by the same maintenance personnel), the final hash result will be different because the salt value changes over time, thus completely avoiding malicious forgery.

[0186] 3. Overall SHA-256 hash operation: Perform a SHA-256 hash operation on the 72 bytes of data to be hashed, generating a 256-bit (32-byte) hash value. The characteristic of the SHA-256 algorithm is that even slight differences in the input data will result in completely different output hash values, ensuring the uniqueness of hash results for different operations (such as opening and closing switches, or operations by different personnel). The hash collision probability is <10. -18 .

[0187] The technical significance of truncating the first 24 bytes: balancing uniqueness, storage efficiency, and transmission performance. The SHA-256 hash operation generates a complete 32-byte (256-bit) hash value, while the feature digest only truncates the first 24 bytes (192 bits). This truncation is not arbitrary but based on a triple consideration of uniqueness assurance, storage cost, and transmission efficiency. The specific reasons are as follows:

[0188] (a) Uniqueness Guarantee: 24 bytes are sufficient to meet the low collision probability requirement:

[0189] The hash collision probability of the SHA-256 algorithm follows the birthday paradox, but even after truncating the first 24 bytes, its uniqueness is still fully guaranteed: the collision probability of a complete 32-byte hash value: when the data size is 10¹ 8 When the collision probability is <10 -18 After truncating the first 24 bytes: the collision probability formula is "P≈n² / (2×2)". 192 (n is the number of operations), even if the annual testing volume of the substation reaches 10 6 The probability of collision is approximately (10^2) 6 )² / (2×10 57 ) = 5 × 10 -45 This is far below the industry requirement of "collision probability < 10". -9 This fully satisfies the requirement that "different operation feature summaries are unique", and there is no need to use the full 32 bytes.

[0190] (ii) Storage efficiency: Reduce the amount of feature database and transmitted data:

[0191] The feature summary segment needs to be stored simultaneously in the motion acquisition module's local Flash (8MB) and the cloud-based collaborative management platform, and must be transmitted via the 915MHz frequency band. If a 32-byte summary is used: the feature library storage for 10,000 devices would be 10... 4 Unit × 32 bytes = 320KB; After truncating the first 24 bytes: storage space reduced to 10 4 Unit × 24 bytes = 240KB, storage cost reduced by 25%; in terms of transmission: the transmission bandwidth of the 915MHz band is 250kbps, and the transmission time of 24 bytes of data is 24×8bit÷250kbps≈0.77ms, which is 24% shorter than 32 bytes (1.02ms). Especially in batch testing of multiple devices (such as 16 devices transmitting simultaneously), it can significantly reduce the probability of signal congestion and improve transmission stability.

[0192] (III) Truncation Rules: The first 24 bytes are truncated to ensure consistency.

[0193] The truncation method uses a fixed first 24 bytes, rather than randomly truncating the middle or last bytes. The core is to ensure that the processing logic of the sending and receiving ends is completely consistent: the sending end (signal encoding module) truncates the first 24 bytes of the 32-byte hash value in order from the most significant byte to the least significant byte; the receiving end (intelligent receiving module of the tester) also truncates the first 24 bytes of the locally calculated 32-byte hash value during verification, and then compares it with the received 24-byte digest. If the truncation rule is not fixed (such as the sending end truncating the first 24 bytes and the receiving end truncating the last 24 bytes), it will lead to verification failure, while a fixed rule can ensure the consistency of the verification logic.

[0194] In other words, the implementation logic of the feature summary segment can be summarized into three steps: fusion, hashing, and truncation. Each step serves the core objectives of dual-dimensional encoding uniqueness and anti-interference:

[0195] Fusion computation: hashing behavioral features and electrical signal features as a whole, rather than performing separate computations, to ensure the correlation between the two-dimensional features and avoid contradictory data;

[0196] Dynamic salt value: The salt value is embedded and updated every hour to solve the problem of hash duplication of data with the same operation characteristics and prevent malicious forgery;

[0197] 24-byte truncation: While ensuring uniqueness, it balances storage efficiency and transmission performance, adapts to hardware resources, and ensures that the overall system operates efficiently and stably.

[0198] Taking the tripping operation as an example:

[0199] The opening and closing command electrical signal: The original signal collected by the signal encoding module is a complete voltage waveform of 0V→220V (rise time 5ms)→220V (last 180ms)→220V→0V (fall time 3ms), containing 200+ sampling points;

[0200] The core parameters of the electrical signal are: the key indicators extracted from the original waveform are voltage amplitude 220V±5% (average value of the stable segment), duration 180ms±10% (duration of the stable segment), and rise time 5ms±20% (time from 10% amplitude to 90% amplitude). With only 3 discrete indicator values, the difference in form and application between the two is obvious.

[0201] The opening and closing command electrical signals and their core parameters work in synergy, as follows:

[0202] Step 1: Acquisition of opening and closing command electrical signals and extraction of core parameters:

[0203] The signal encoding module performs two tasks simultaneously:

[0204] Collect the opening and closing command electrical signals (raw signals) output from the remote control backend, and perform preprocessing such as filtering and amplification;

[0205] Extract core electrical signal parameters (such as voltage and duration) from the preprocessed raw signal and synchronously transmit the parameters to the feature library for storage (as electrical signal feature association codes).

[0206] Step 2: Two-dimensional encoding generation:

[0207] The signal encoding module fuses and encodes behavioral features (such as coordinates of 200-300px and pressure of 120g for the opening and closing operation) with the preprocessed signal of the opening and closing command electrical signal. The behavioral features provide the physical behavioral basis for the operation type, and the original electrical signal provides the electrical signal basis for the operation type. After anti-interference processing such as Manchester encoding, the two generate a unique opening / closing feature code.

[0208] The relationship between the core parameters of the electrical signal and the opening / closing command electrical signal can be summarized as follows:

[0209] Homogeneity: All signals originate from the remote control backend output of the circuit breaker, ensuring the consistency and reliability of the signal source;

[0210] Different dimensions: The former is the original signal stream (complete waveform), which is the basis of the electrical signal dimension for two-dimensional encoding; the latter is the key feature index (discrete value), which is used for logical verification of the feature library and data traceability.

[0211] Synergy: Both support the dual-dimensional encoding and dual-verification logic of behavioral features and electrical signal features. Neither can be missing. If the opening and closing command electrical signal is missing, the dual-dimensional encoding will lose its electrical basis. If the core parameters of the electrical signal are missing, the feature library will not be able to complete the logical verification of behavior and electrical signal, which will eventually lead to an increase in the false trigger rate (such as misjudging interference signals as valid opening signals).

[0212] Intelligent receiving module of the tester:

[0213] The intelligent receiving module of the tester is magnetically mounted on the casing of the switch mechanical characteristic tester. An electromagnetic induction starter is built into the corresponding position of the tester's measurement start button. It receives the feature code transmitted by the remote control backend signal encoding module, and parses the opening / closing feature code through the built-in signal decoding unit. It employs a dual mechanism of time window verification and feature code matching verification. Upon successful verification, the electromagnetic induction starter is powered on, causing the iron core to strike the tester's measurement start button to initiate the test. The time window verification of the intelligent receiving module relies on a built-in precise clock synchronization module with an error of ±1ms. The set dynamic time window range is 100-500ms, and the start point of the time window is the rising edge of the signal when a mouse operation occurs, while the end point is adaptively adjusted based on the average response time of 10 historical tests. The feature code matching verification uses an Euclidean distance algorithm; a similarity of ≥90% is considered a successful match. When verification fails, an abnormal signal log is generated and uploaded to the cloud-based collaborative management platform.

[0214] It should be noted that the connection relationship and installation position of the electromagnetic induction starter, iron core, and measuring start button are set according to the actual situation. This is existing technology and will not be elaborated here.

[0215] In other words, the intelligent receiving module of the tester is installed on the switch mechanical characteristic tester and has the following functions:

[0216] 1. Receive the tripping feature code and / or closing feature code transmitted by the signal encoding module, and parse the feature code through the built-in signal decoding unit of the module to extract the test type identifier (tripping / closing).

[0217] 2. The system initiates a dual mechanism of time window verification and feature code matching verification. After the verification is successful, it transmits the feature code and the corresponding test type instruction (opening / closing) to the main control unit of the switch mechanical characteristic tester, and receives the program preload ready signal fed back by the main control unit.

[0218] 3. The module integrates an electromagnetic induction starter (or establishes a control connection with the electromagnetic induction starter located at the measurement start button of the tester). After receiving the "ready signal" from the main control unit, it controls the electromagnetic induction starter to be powered on, causing the iron core to strike the measurement start button of the tester.

[0219] 4. In coordination with the main control unit, based on the test type (opening / closing) corresponding to the parsed feature code, start the corresponding test program of the switch mechanical characteristic tester, synchronously receive the test status data (such as progress and results) fed back by the main control unit, and upload it to the cloud collaborative management platform.

[0220] In point 2, the specific process of initiating the dual mechanism of time window verification and feature code matching verification is as follows:

[0221] Verify the trigger conditions and clarify the start point of the dual mechanism. When the intelligent receiving module of the tester receives the tripping or closing feature code sent by the remote control backend signal encoding module via directional wireless transmission (such as the 915MHz band), it immediately starts the dual mechanism of "time window verification and feature code matching verification". The start point must meet the following prerequisites:

[0222] Feature code reception integrity verification: The intelligent receiving module first performs frame structure verification on the received feature code (checking whether the basic identifier segment + feature digest segment + check segment are complete, and whether the field lengths meet the preset standards, such as 16 bytes for the basic identifier segment, 24 bytes for the feature digest segment, and 8 bytes for the check segment). Only when the frame structure is complete and no fields are missing can the dual verification be triggered.

[0223] Signal strength threshold filtering: The module has a built-in signal strength detection unit (based on RSSI value). Dual verification is only initiated when the RSSI value of the received feature code is ≥-80dBm (ensuring no severe signal attenuation). If the RSSI value is <-80dBm, it is directly judged as "weak signal invalid", and the verification process is refused. At the same time, partial data retransmission is triggered (only the missing field of the feature code is retransmitted), and the number of retransmissions is ≤3.

[0224] Time window verification is a dynamic timing verification based on clock synchronization. The core of time window verification is to ensure the timing consistency between the feature code reception time and the mouse operation trigger time by synchronizing the clocks of the intelligent receiving module and the remote control backend signal encoding module. The specific process is as follows:

[0225] 1. Clock synchronization reference establishment: During system initialization, the tester's intelligent receiving module obtains the standard time (UTC time) through the cloud-based collaborative management platform and synchronizes the clock with the remote control backend signal encoding module. The synchronization error is ≤1ms (based on the NTP network time protocol, the synchronization cycle is once every 5 minutes). When generating the feature code, the signal encoding module will write the precise time triggered by the mouse operation (such as "20250918143000123", accurate to milliseconds) into the acquisition timestamp field (4 bytes) of the basic identifier segment and transmit it along with the feature code.

[0226] 2. Dynamic Time Window Calculation: After receiving the feature code, the intelligent receiving module extracts the mouse operation trigger time (denoted as T1) from the acquisition timestamp field, and simultaneously records the precise time of local reception of the feature code (denoted as T2). Based on the average transmission delay of the feature code from 10 historical valid tests (denoted as ΔT, e.g., ΔT=20ms), the module calculates the dynamic time window range: [T1+ΔT-10ms, T1+ΔT+10ms], that is, the time window width is 20ms, which can be dynamically adjusted according to the transmission environment (e.g., the time window width can be extended to 30ms under strong interference environment).

[0227] 3. Timing Consistency Determination: If the local reception time T2 falls within the dynamic time window range (T1+ΔT-10ms≤T2≤T1+ΔT+10ms), the time window verification is deemed successful; if T2 exceeds the time window range (e.g., due to signal delay, T2=T1+ΔT+15ms), the time window verification is deemed unsuccessful, and the module immediately sends a "time window out of tolerance" feedback signal to the signal encoding module, requesting retransmission of the feature code, with the number of retransmissions ≤ 2; if the error still exceeds the tolerance after 2 retransmissions, the feature code is deemed invalid, and the verification process is terminated.

[0228] Feature code matching verification is based on multi-dimensional similarity verification using a feature library. Feature code matching verification requires combining a locally cached feature library of opening / closing operation behaviors (containing feature data of opening and closing operations). The validity of the feature code is determined through multi-dimensional similarity calculation. The specific process is as follows:

[0229] 1. Feature Code Parsing and Field Extraction: The intelligent receiving module parses the feature code through the built-in signal decoding unit, extracting the operation type identifier (1 byte, 0x01 = opening, 0x02 = closing) of the basic identifier segment and the 24-byte feature digest of the feature digest segment (the first 24 bytes of the SHA-256 hash value based on the dual-dimensional feature data); according to the operation type identifier, it retrieves the corresponding feature subset in the local feature library (e.g., when the identifier is 0x01, it retrieves the opening feature subset; when the identifier is 0x02, it retrieves the closing feature subset), and extracts the feature digest baseline value (the 24-byte feature digest stored in the historical learning stage, updated once every 50 tests) and the feature dimension threshold range (e.g., coordinate position X: 200-300px, click pressure: 100-140g, which is the threshold range of behavioral features and electrical signal features).

[0230] Multi-dimensional similarity calculation: Feature digest similarity verification: The similarity between the 24-byte feature digest of the received feature code and the "feature digest baseline value" in the feature library is calculated using the Hamming distance algorithm. The formula is: similarity = (1 - Hamming distance / total number of bits) × 100%, where the Hamming distance is the number of different bits in the two-byte sequence; if the similarity is ≥ 90%, the feature digest similarity verification passes.

[0231] Feature Dimension Consistency Verification: Parse the original dual-dimensional feature data (behavioral features + electrical signal features) corresponding to the feature summary segment in the feature code, and verify whether each dimension is within the threshold range of the feature dimension in the feature library: Behavioral feature dimension: such as whether the coordinate position falls within the opening / closing button range, whether the click pressure is within the effective range of 50-300g, and whether the click duration is within 10-1000ms; Electrical signal feature dimension: such as whether the electrical signal voltage amplitude is within 220V±5%, and whether the duration is within 180ms±10%; If all dimensions are within the threshold range, the feature dimension consistency verification is deemed to have passed; If any dimension exceeds the tolerance (e.g., click pressure = 40g < 50g), the verification of that dimension is deemed to have failed.

[0232] Comprehensive matching result determination: Feature code matching verification is considered successful only when the feature summary similarity is ≥90% and all feature dimensions are within the threshold range; if the feature summary similarity is <90% (e.g., due to feature summary misalignment caused by interference), or any feature dimension exceeds the tolerance, the matching verification is considered unsuccessful. The module retrieves the 3 most recent valid feature codes from the feature library for auxiliary comparison. If the similarity with one of the feature codes is ≥85%, it is considered a suspected valid feature code, which must be manually confirmed by the maintenance personnel using the "Confirm" physical button on the module surface; if the manual confirmation is valid, the verification process continues; if it is not confirmed or the similarity is <85%, it is considered an invalid feature code, and the verification process terminates.

[0233] The intelligent receiving module sends a dual verification pass signal to the main control unit of the tester, and simultaneously transmits the operation type identifier and feature dimension data of the feature code, triggering the main control unit to preload the corresponding test program (e.g., when the identifier is 0x01, the preload test program is loaded, and the opening travel sensor and timing unit are initialized); after receiving the "program preload ready" signal from the main control unit, it controls the electromagnetic induction starter to be powered on (DC12V, current 1.2A), which drives the core impact tester measurement start button to start the corresponding test program (opening or closing test). If the time window verification or feature code matching verification fails, the module will issue a local alarm via a three-color LED indicator (red flashing rapidly, once every 0.5 seconds) and a miniature vibration motor (continuous vibration for 1 second). At the same time, the reason for the verification failure (such as time window exceeding tolerance, insufficient feature summary similarity) and the failed feature code information will be uploaded to the cloud collaborative management platform. The platform will record the failure log and generate "troubleshooting suggestions" (such as checking for wireless transmission frequency band interference if the time window exceeds tolerance). If the feature code fails to be received three times in a row, the intelligent receiving module will automatically switch the wireless transmission frequency band (such as switching from 915MHz to the 2.4GHz backup frequency band) and re-receive the feature code to ensure that the verification process can still be maintained in the interference environment.

[0234] By designing a dynamic time window, the limitations of the traditional fixed time window are overcome. The time window range is dynamically adjusted based on the historical average transmission delay, adapting to the signal delay changes in the complex wireless environment of substations. The time window verification pass rate is increased from 85% to 99.5%.

[0235] By combining feature summary similarity and feature dimension consistency for multi-dimensional matching and verification, the risk of misjudgment due to single-dimensional verification is avoided (e.g., relying solely on feature summary may overlook local dimension deviations). The manual confirmation mechanism is also compatible with special scenarios (e.g., temporary changes in the operating habits of maintenance personnel), reducing the false trigger rate to below 0.1%.

[0236] Cloud-based collaborative management platform:

[0237] The cloud-based collaborative management platform is built on a cloud server, supporting access via web and mobile devices. It receives data uploaded from various modules in real time, generates visual reports and operation trajectory maps, and automatically optimizes and updates the opening / closing feature library every 50 tests. It supports group management of equipment in multiple substations and adopts a master-six-slave star network structure and a time-slice rotation mechanism to achieve simultaneous testing of multiple bays. When the signal characteristic change of the equipment exceeds a 15% threshold, the cloud-based collaborative management platform sends a relearning and calibration reminder to the operation and maintenance personnel through both web pop-ups and mobile APP push notifications. It also supports manual uploading of equipment model parameters and automatically matches the preset initial feature code template, shortening the learning time for new equipment.

[0238] Example 2 discloses a mouse-linked wireless intelligent triggering method for testing the mechanical characteristics of a switchgear based on the system described in Example 1, comprising the following steps:

[0239] System deployment and initialization: Attach the motion acquisition module to the remote control backend operating mouse, complete the connection with the signal encoding module via Bluetooth pairing wizard, attach the tester intelligent receiving module to the designated position of the switchgear mechanical characteristic tester, and input the equipment number, substation name, and switchgear model information into the cloud collaborative management platform to complete equipment registration and grouping;

[0240] During the feature learning phase, maintenance personnel perform several opening and closing operations. The action acquisition module records the coordinate position range, average click pressure, and average click duration for each operation, generating an operation behavior feature dataset. The signal encoding module simultaneously collects the electrical signal parameters of the opening and closing commands corresponding to each operation, fuses them to generate opening and closing feature codes, and stores them in the local storage unit and the cloud feature library, providing a data foundation for subsequent feature summary benchmark values.

[0241] During the feature learning phase, if the click pressure or click duration of a certain opening / closing operation exceeds the preset normal range (click pressure <50g or >300g, click duration <10ms or >1000ms), the system automatically determines that the operation is invalid and prompts the maintenance personnel to re-execute it, ensuring the accuracy of the generated opening / closing feature code, and the number of valid operations completed is displayed in real time during the learning process.

[0242] During system deployment and initialization, the cloud-based collaborative management platform supports batch import of device information. By filling in the device number, IP address, and installation location fields using an Excel template, multiple devices can be automatically registered after uploading. Furthermore, when grouping devices, they can be divided into groups based on substation intervals. Each group can contain up to 16 intelligent receiving modules for testing instruments, and a relay transmission method is used within the group to enhance signal coverage.

[0243] During the feature learning phase, the interaction sequence of each module is as follows: The action acquisition module presses the 'One-Click Learning' button → the module sends a 'Acquisition Start' signal to the signal encoding module (Bluetooth, delay ≤10ms) → the signal encoding module replies 'Ready' → the action acquisition module starts acquiring 3 tripping operations (each with an interval ≥5 seconds) → after acquisition, it sends behavioral feature data to the signal encoding module → the signal encoding module synchronously acquires the tripping command electrical signal → fuses to generate feature code → and stores it in local Flash and cloud respectively (transmission delay ≤20ms).

[0244] During the test execution phase, maintenance personnel perform opening and closing operations via mouse in the remote control backend. The action acquisition module captures the operation parameters in real time and transmits them to the signal encoding module. The signal encoding module integrates the operation parameters with the real-time electrical signal to generate a dynamic feature code, which is transmitted to the intelligent receiving module of the tester via the 915MHz frequency band. The intelligent receiving module of the tester sequentially performs time window verification and feature code matching verification on the dynamic feature code. After the dual verification is passed, the electromagnetic induction starter is powered on, and the iron core strikes the start button of the tester. At the same time, the status is fed back through LED indicator lights and vibration motors, and the test data is uploaded to the cloud platform in real time.

[0245] In some embodiments, during the test execution phase, if the intelligent receiving module of the tester does not receive the feature code transmitted by the signal encoding module within a set distance, it automatically initiates a signal reconnection mechanism. After a set number of failed reconnections, a signal interruption alarm is generated and uploaded to the cloud collaborative management platform.

[0246] If the signal encoding module fails to receive behavioral feature data from the motion acquisition module for 3 consecutive times (Bluetooth disconnection), the local caching + reconnection mechanism will be activated immediately: ① The 10 most recent valid feature codes will be cached to avoid data loss; ② Bluetooth reconnection will be attempted every 5 seconds (5 times in total). If the reconnection fails, a 'module offline' alarm will be sent to the remote control backend via USB, and a red LED will be lit at the same time.

[0247] If communication between the cloud platform and the tester's receiving module is interrupted (no data is received for more than 30 seconds), the receiving module will automatically switch to local independent operation mode: ① Use the locally cached feature library for verification; ② Temporarily store test data locally (capacity ≥ 1000 records); ③ Automatically retransmit data after communication is restored to avoid test interruption.

[0248] Test execution phase sequence:

[0249] Click the trip button with the mouse → The action acquisition module acquires parameters within 100ms and sends them to the signal encoding module → The signal encoding module completes dual-dimensional encoding within 50ms → Transmits the signal to the intelligent receiving module of the tester in the 915MHz band (distance ≤ 50 meters, delay ≤ 30ms) → The receiving module completes dual verification within 100ms → Sends the trip program command to the main control unit → The main control unit preloads the program and replies 'ready' within 200ms → The receiving module controls the electromagnetic induction starter to power on (200ms) → The iron core strikes the button to start the test.

[0250] During the feature database update and maintenance phase, after each set number of tests, the cloud-based collaborative management platform calculates the deviation between the latest operational behavior feature data and historical features, and uses a weighted average algorithm to optimize and update the operational behavior feature database. When the deviation value of a certain device exceeds the threshold for a set number of consecutive times, a re-feature learning reminder is triggered. After the operation and maintenance personnel perform the learning operation, the opening and closing feature codes of the device stored in the feature database are updated.

[0251] Finally, it should be noted that this invention is based on the detection of switch opening and closing during power outages in substations. The power supply guarantee of the system and remote control backend of this invention relies on the uninterruptible power supply (UPS) system of the substation.

[0252] As a core hub of the power system, the substation's remote control backend host and operating terminals (including mice) adopt a "dual power supply + UPS uninterruptible power supply" power supply architecture. This is a mandatory design standard in the power industry (complying with GB50227-2019 "Design Code for Parallel Capacitor Devices" and DL / T5136-2012 "Technical Specification for Secondary Wiring Design of Thermal Power Plants and Substations"). The specific logic is as follows:

[0253] During normal operation: The remote control backend is powered by the substation's 10kV / 380V plant power system to ensure daily monitoring and operation functions;

[0254] During power outage testing: When it is necessary to shut down the power supply to the switchgear bay (disconnect the circuit breaker and disconnect the power supply to the primary equipment in the bay), the secondary control circuit of the remote control back-end is not interrupted. At this time, the power supply is automatically switched to the UPS system (capacity is usually ≥10kVA, and the battery life is ≥4 hours) to ensure the normal operation of the remote control back-end host, operation monitor, mouse and other equipment, and to meet the power requirements of maintenance personnel to perform opening and closing operations (trigger test) through the mouse.

[0255] Core principle: The "power outage detection" of the substation only targets the primary equipment (such as circuit breakers and busbars) of the switchgear to be tested, while the secondary control circuits (including remote control backend and relay protection devices) used for monitoring and operation need to be kept powered. This is the premise to ensure that the equipment status is controllable and the operation is executable during the testing process. The design of the system of this invention fully follows this industry standard power supply logic.

[0256] To ensure full system operation during power outage detection scenarios, except for the remote control backend which relies on UPS power supply, the core modules of this invention adopt independent power supply or low-power design, as detailed below:

[0257] Mouse integrated motion acquisition module: It uses a built-in rechargeable lithium battery (capacity 1000mAh) for independent power supply, without relying on the remote control backend or external power supply of the mouse; in the power failure detection scenario, a single charge can support continuous operation for ≥72 hours (based on testing one device every 10 minutes, it can cover testing 200+ devices), and supports real-time charging on the remote control backend host via USB Type-C interface (using the host's USB port powered by UPS), avoiding the problem of insufficient battery life;

[0258] Remote control back-end signal encoding module: If it is "external" (connected to the remote control back-end via USB Type-C), it is directly powered by the remote control back-end host (UPS powered) with 5V / 2A power; if it is "built-in" (embedded in PCIe slot), it is powered by the host power supply (UPS powered) with 12V / 1A power, and no additional wiring is required;

[0259] The intelligent receiving module of the tester adopts a dual power supply mode of "lithium battery + external power supply" - when the power is off, it can be powered independently by the built-in 2000mAh lithium battery (battery life ≥24 hours); if the test time is long, the power can also be supplemented through the USB port of the tester (the tester is powered by UPS or portable power bank) to ensure the module continues to work.

[0260] Cloud-based collaborative management platform: Deployed on Alibaba Cloud servers, with 24 / 7 uninterrupted power supply provided by the cloud service provider, unaffected by local power outages at substations, ensuring real-time data upload, feature database updates, and alarm notification functions are functioning normally.

[0261] Taking the power failure detection of the 10kV #1 bay switchgear in a 220kV substation as an example, the system power supply process is as follows:

[0262] Maintenance personnel disconnect the circuit breaker and disconnect switch of bay #1 to cut off the power supply to the primary equipment in that bay (complete the "power off").

[0263] The remote control system automatically switches to UPS power supply. The monitor, mouse, and host are operating normally. Maintenance personnel can click the "#1 bay trip" and "#1 bay close" buttons with the mouse.

[0264] The mouse-integrated motion acquisition module (powered by a lithium battery) captures mouse operation parameters and transmits them to the remote control backend signal encoding module (powered by the host USB) via Bluetooth.

[0265] The signal encoding module generates a feature code, which is transmitted to the intelligent receiving module of the tester (powered by lithium battery) via 915MHz.

[0266] After the receiving module passes verification, the tester is triggered to start (the tester is powered by a portable power bank) to complete the mechanical characteristic test.

[0267] Test data is uploaded to the cloud platform (powered by the cloud server) via 4G / 5G network to generate a test report.

[0268] Throughout the process, all modules related to operation and triggering are reliably powered, making them fully compatible with power failure detection scenarios.

[0269] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. 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.

Claims

1. A mouse linkage wireless intelligent triggering system for mechanical property testing of a switch cabinet, characterized in that, The application relates to a switch mechanical characteristic testing system, which comprises the following parts: an action collecting module, which is installed on a mouse of a substation remote background operation and is used for collecting operation parameters of the mouse, combining with electric signal characteristic association codes obtained from a signal coding module, establishing an operation behavior characteristic library based on behavior characteristic data formed by the operation parameters and the electric signal characteristic association codes obtained during opening and closing operation, and transmitting the operation behavior characteristic data to the signal coding module in real time, wherein the operation parameters include coordinate positions of the mouse, click pressure, click time length and operation path trajectories; a signal coding module, which is built-in or externally connected to a substation remote background host and is wirelessly connected to the action collecting module, is used for receiving operation behavior characteristic data of the mouse, simultaneously collecting opening and closing instruction electric signals generated by the remote background, extracting behavior characteristics and electric signal characteristics based on the operation behavior characteristic data and the opening and closing instruction electric signals respectively, carrying out anti-interference processing and data fusion on the behavior characteristics and the electric signal characteristics respectively, generating unique opening characteristic codes and / or closing characteristic codes, and transmitting the opening characteristic codes and / or the closing characteristic codes to a tester intelligent receiving module through directional wireless transmission; a tester intelligent receiving module, which receives the opening characteristic codes and / or the closing characteristic codes transmitted by the signal coding module, analyzes the characteristic codes through a signal decoding unit, extracts a test type identifier, starts a time window verification and a characteristic code matching verification double mechanism to verify the characteristic codes, after the verification passes, transmits the characteristic codes and corresponding test type instructions to a main control unit of a switch mechanical characteristic tester, receives a program preloading ready signal fed back by the main control unit, controls an electromagnetic induction starter arranged at a position of a start measurement button of the switch mechanical characteristic tester to be electrified after the program preloading ready signal is received, drives an iron core to hit the start measurement button, and the main control unit starts corresponding test programs according to the test type instructions corresponding to the analyzed characteristic codes; a cloud collaborative management platform, which is built based on a cloud server and is used for receiving data uploaded by the action collecting module, the signal coding module and the tester intelligent receiving module in real time, generating visual reports and operation trajectory maps, and automatically optimizing and updating an operation behavior characteristic library after 50 test operations are completed.

2. The mouse linkage wireless intelligent trigger system for mechanical property test of switch cabinet according to claim 1, characterized in that, The action collecting module comprises a controller, a photoelectric sensor, a pressure sensor, a Bluetooth unit and a power supply unit, the photoelectric sensor is used for capturing texture changes of a contact surface in contact with the mouse in real time when the mouse moves, converting optical signals into electric signals, the controller calculates displacement differences of adjacent image frames, converts coordinate changes of the mouse in horizontal and vertical directions, combines with initial coordinate reference points, and finally generates real-time coordinate positions of the mouse during mouse operation, and calculates operation path trajectories based on the real-time coordinate positions. ​ When the mouse is clicked, the click pressure is collected by the pressure sensor, and when the pressure sensor detects that the pressure value reaches the effective click threshold, a trigger signal of click start is sent to the controller immediately, the main control unit starts the timer to start timing; when the pressure sensor detects that the pressure value drops to the click end threshold, a trigger signal of click end is sent to the main control unit, the main control unit stops the timer, and the difference between the start time and the end time recorded by the timer is the click duration.

3. The mouse linkage wireless intelligent trigger system for mechanical property test of switch cabinet according to claim 1, characterized in that, The specific process of establishing the operation behavior characteristic library containing the operation behavior characteristic data of opening and closing is as follows: An operation behavior characteristic library is constructed, the operation behavior characteristic library includes an opening data subset and a closing data subset, and each operation behavior characteristic data in the opening data subset and the closing data subset includes a plurality of core feature dimensions, the core feature dimensions include a coordinate position range, a motor pressure distribution, a motor market distribution, an operation path trajectory and a pressure change rate, and the core feature dimensions of the same type in the opening data subset and the closing data subset are different in range and do not overlap; For each switch cabinet, a plurality of groups of core feature dimensions corresponding to opening and closing operations are collected and filtered, standardized and statistical features are collected, and each core feature dimension is assigned a weight to obtain electrical signal core parameters corresponding to opening and closing operations, and finally an opening data subset and a closing data subset are obtained; The opening data subset and the closing data subset are updated regularly; After the mechanical property test of the switch cabinet is performed, if the deviation rate of the core feature dimension corresponding to the actual operation during the test from the mean value of the corresponding opening data subset or closing data subset is greater than a deviation threshold, the opening data subset and the closing data subset are updated; When a new switch cabinet or action collection module is accessed, the opening data subset and the closing data subset are updated.

4. The mouse linkage wireless intelligent trigger system for mechanical property test of switch cabinet according to claim 1, characterized in that, The specific process of generating a unique opening feature code and / or closing feature code by adopting a dual-dimension encoding of behavior characteristics combined with electrical signal characteristics and performing anti-interference processing on the dual-dimension encoding is as follows: The behavior characteristics of the operation behavior characteristic data are extracted, and a dynamic threshold library based on behavior characteristics and electrical signal characteristic threshold ranges is established, the behavior characteristics include basic parameters and derived parameters, the basic parameters include coordinate positions, click pressures and click durations, the derived parameters are calculated from the basic parameters and include coordinate trajectory curvatures and pressure change rates; The electrical signal characteristics of the opening and closing instruction electrical signals are extracted, the electrical signal characteristics include voltage amplitudes, current change rates and signal rising edge slopes, and a correlation model between the electrical signal characteristics is constructed to eliminate interference signals; The behavior characteristics and the electrical signal characteristics are respectively subjected to anti-interference processing. The behavior features and electrical signal features after interference processing are data fused to form two-dimensional feature data, salt values are dynamically embedded into the two-dimensional feature data to form to-be-hashed data, SHA-256 hash operation is performed on the to-be-hashed data, the first 24 bytes are intercepted as a feature digest segment, a basic identification segment and a check code segment are added in the feature digest, and thus corresponding opening feature codes or closing feature codes are obtained, the basic identification segment includes a device displacement ID, an operation type identification, a collection timestamp and a signal source identification.

5. The mouse linkage wireless intelligent trigger system for mechanical property test of switch cabinet according to claim 4, characterized in that, The starting time window verification and feature code matching verification dual mechanisms for verifying the feature codes are as follows: The intelligent receiving module and the signal coding module are time synchronized; After the intelligent receiving module receives the feature code, the collection timestamp is extracted as a mouse operation trigger time, denoted as T1, and the accurate time of receiving the feature code locally is recorded, denoted as T2; According to the feature code transmission delay mean ΔT of historical several times of effective tests, the dynamic time window range is calculated: [T1+ΔT-10ms, T1+ΔT+10ms]; If the local time T2 falls within the dynamic time window range, the time window verification passes, otherwise, the time window verification fails; The operation type identification of the basic identification segment is extracted, the corresponding data subset is recalled according to the operation type identification, and feature digest similarity verification and feature dimension consistency verification are performed, if both pass, the feature code matching verification passes, otherwise, the feature code matching verification fails.

6. The mouse linkage wireless intelligent trigger system for mechanical property test of switch cabinet according to claim 5, characterized in that, The specific process of the feature digest similarity verification is as follows: The operation behavior feature data in the data subset is fused and subjected to SHA-256 hash operation, the first 24 bytes are intercepted to obtain the feature digest segment of each operation behavior feature data in the data subset, the average value of the feature digest segments corresponding to all the operation behavior feature data in the data subset is calculated to obtain a feature digest reference value; The similarity between the feature digest segment of the received feature code and the feature digest reference value is calculated by using the Hamming distance algorithm, the formula is: similarity = (1-Hamming distance / total number of bits) x 100%, wherein the Hamming distance is the number of different bits in the feature digest segment and the feature digest reference value; if the similarity is greater than or equal to 90%, the similarity verification of the feature digest segment passes.

7. The mouse linkage wireless intelligent trigger system for mechanical property test of switch cabinet according to claim 5, characterized in that, The specific process of the feature dimension consistency verification is as follows: The behavior feature and electrical signal feature threshold range of the corresponding data subset is obtained from the dynamic threshold library; The corresponding behavior feature and electrical signal feature are obtained by analyzing the feature digest segment; It is judged whether the behavior feature and electrical signal feature fall within the corresponding behavior feature and electrical signal feature threshold range, if both are, the feature dimension consistency verification passes, otherwise, the feature dimension consistency verification fails.

8. The mouse linkage wireless intelligent trigger system for mechanical property test of switch cabinet according to claim 1, characterized in that, The tester intelligent receiving module further comprises a three-color LED ring indicator light and a micro vibration motor, a blue constant light indicating a standby state, a green flicker indicating a successful trigger, a red slow flash indicating a verification, and a red-green alternating flash indicating a need for re-learning; the vibration motor outputs a 50 ms short vibration when the trigger is successful, and outputs a 300 ms long vibration when the operation is wrong or the verification fails, and the light intensity can be automatically adjusted according to the brightness value collected by the ambient light sensor.

9. A mouse linkage wireless intelligent triggering method for switch cabinet mechanical property testing based on the system of any one of claims 1-8, characterized in that, The method comprises the following steps: System deployment and initialization, the action acquisition module is adsorbed on the remote background operation mouse, the connection with the signal coding module is completed through the Bluetooth pairing wizard, the tester intelligent receiving module is adsorbed on the specified position of the switch cabinet mechanical property tester, the device number, substation name and switch cabinet model information are input into the cloud collaborative management platform, the device registration and grouping are completed; In the feature learning stage, the operation and maintenance personnel perform a plurality of times of opening and closing operations, the action acquisition module records the coordinate position range, the click pressure average value and the click duration average value of each operation, generates an operation behavior feature data set, the signal coding module synchronously collects the opening and closing instruction electric signal parameters corresponding to each operation, fuses to generate opening and closing feature codes, and stores them into the local storage unit and the cloud feature library, thereby providing a data basis for the subsequent feature abstract reference value; In the test execution stage, the operation and maintenance personnel perform the opening and closing operation through the mouse in the remote background, the action acquisition module captures the operation parameters in real time and transmits them to the signal coding module, the signal coding module fuses the operation parameters and the real-time electric signal to generate a dynamic feature code, transmits the dynamic feature code to the tester intelligent receiving module through the 915 MHz frequency band, and the tester intelligent receiving module sequentially performs time window verification and feature code matching verification on the dynamic feature code, controls the electromagnetic induction starter to be powered on after the double verification is passed, the iron core hits the tester start button, the state is fed back through the LED indicator light and the vibration motor, and the test data is uploaded to the cloud platform in real time; In the feature library updating and maintaining stage, after a set number of tests are completed, the cloud collaborative management platform calculates the deviation value of the latest operation behavior feature data and the historical features, optimizes and updates the operation behavior feature database by using a weighted average algorithm, triggers a re-feature learning reminder when the deviation value of a certain device is continuously above the threshold value for a set number of times, and updates the opening and closing feature codes of the device stored in the feature library after the learning operation of the operation and maintenance personnel.

10. The mouse linkage wireless intelligent triggering method for mechanical property testing of a switch cabinet according to claim 9, characterized in that, In the test execution stage, if the tester intelligent receiving module does not receive the feature code transmitted by the signal coding module within a set distance, a signal reconnection mechanism is automatically started, and after a set number of reconnection failures, a signal interruption alarm is generated and uploaded to the cloud collaborative management platform.