Intelligent protection switch control method and system for ecological restoration equipment

Abstract

The invention discloses an intelligent protection switch control method and system for ecological restoration equipment, and the method comprises the steps: building an operation parameter reference containing temperature compensation through collecting current and voltage signals of a power supply loop of the equipment, and achieving the quantitative representation of the multi-working-condition operation characteristics of the equipment; a self-adaptive monitoring strategy is generated by adopting a dynamic threshold division and periodic feature detection technology, and the sampling frequency is adjusted in real time according to the parameter deviation degree; a multi-channel synchronous detection and starting surge compensation mechanism is introduced, and real-time evaluation of a fault risk coefficient is completed through edge calculation; graded protection actions are executed according to the risk levels, reliable locking of the protection state and timely sending of fault information are ensured through electrical and mechanical dual locking and multi-link alarm reporting, and an intelligent and self-adaptive full-process protection control scheme is provided for ecological restoration equipment.

Description

Technical Field

[0001] This invention relates to the field of power protection and control technology, and in particular to an intelligent protection switch control method and system for ecological restoration equipment. Background Technology

[0002] Ecological restoration equipment is widely used in environmental protection projects such as river management, lake ecological restoration, and artificial wetland construction. This equipment typically includes various load types, such as aeration devices, circulating water pumps, mixing mechanisms, and chemical dosing systems. Operating long-term in outdoor or semi-open environments, it faces multiple challenges including power grid fluctuations, environmental temperature variations, and load shocks. Traditional circuit breakers, employing fixed threshold protection strategies, struggle to adapt to the frequent start-ups and significant load variations of ecological restoration equipment. They are prone to falsely triggering protection actions during normal startup or failing to respond promptly during the development of progressive faults.

[0003] Existing protection devices primarily rely on exceeding limits of single electrical parameters for fault identification, lacking the ability to collaboratively analyze multiple types of parameters and failing to effectively distinguish between grid-side disturbances and equipment-side anomalies. Furthermore, traditional protection schemes use a fixed monitoring frequency, which neither reduces system power consumption during normal operation nor provides sufficient data density to support fault diagnosis under abnormal conditions. In addition, some intelligent protection schemes upload fault analysis tasks to cloud servers for processing, which poses a risk of data transmission delays or even communication interruptions in scenarios with limited communication conditions in the field. This makes it difficult to meet the real-time requirement of millisecond-level response to electrical faults, and the equipment status management and remote alarm mechanisms after protection actions are also inadequate, making it difficult to transmit fault information to the operation and maintenance center in a timely manner. Summary of the Invention

[0004] This invention discloses an intelligent protection switch control method and system for ecological restoration equipment. By collecting current and voltage signals during equipment operation and establishing multi-condition parameter benchmarks, it achieves adaptive monitoring strategy configuration; it combines multi-channel synchronous detection and edge computing technology to complete fault risk quantification assessment; and it executes differentiated protection actions based on the assessment results and completes interlocking management and remote alarm, forming a complete intelligent protection control closed loop, which is suitable for the reliable protection needs of ecological restoration equipment in unattended outdoor environments.

[0005] The first aspect of this invention proposes an intelligent protection switch control method for ecological restoration equipment, comprising the following steps: Collect load current data and voltage fluctuation data of the ecological restoration equipment, and construct an operating parameter benchmark table based on the load current data and voltage fluctuation data; The monitoring cycle configuration is generated by dividing the threshold interval based on the operating parameter benchmark table, the current monitoring cycle is determined based on the monitoring cycle configuration, and a status judgment instruction is generated according to the deviation between the current monitoring cycle and the rated parameters of the equipment. Based on the state determination instruction, multi-channel synchronous detection is performed to obtain fault characteristic parameters. Start-up surge characteristics are identified for the load current data to generate a start-up current compensation amount. Based on the fault characteristic parameters and the start-up current compensation amount, a risk assessment coefficient is determined by edge computing. An abnormal triggering flag is generated based on the risk assessment coefficient. Based on the abnormal trigger flag, the protection level is determined and protection control parameters are generated. Based on the protection control parameters, threshold comparison is performed to generate the fault limit exceedance. Based on the fault limit exceedance, circuit breaker response control is performed to generate the tripping signal. The blocking action is triggered by the tripping signal to generate a blocking feedback state. An alarm reporting sequence is triggered based on the blocking feedback state to generate an alarm completion flag. A protection control command is generated based on the alarm completion flag and the tripping signal.

[0006] A second aspect of this invention provides an intelligent protection switch control system for ecological restoration equipment, comprising: The data acquisition module is used to collect load current data and voltage fluctuation data of the ecological restoration equipment, and to construct an operating parameter benchmark table based on the load current data and voltage fluctuation data. The monitoring and scheduling module is used to divide the operating parameter benchmark table into threshold intervals to generate a monitoring cycle configuration, determine the current monitoring cycle based on the monitoring cycle configuration, and generate a status judgment instruction according to the deviation between the current monitoring cycle and the rated parameters of the equipment. The risk assessment module is used to perform multi-channel synchronous detection to obtain fault characteristic parameters based on the state determination instruction, perform start-up surge characteristic identification on the load current data to generate start-up current compensation amount, determine the risk assessment coefficient through edge computing based on the fault characteristic parameters and the start-up current compensation amount, and generate an abnormal triggering flag based on the risk assessment coefficient. The protection execution module is used to determine the protection level according to the abnormal trigger flag, generate protection control parameters, perform threshold comparison according to the protection control parameters to generate fault over-limit quantity, and perform circuit breaker response control according to the fault over-limit quantity to generate a tripping signal. The instruction generation module is used to trigger a blocking action via the tripping signal to generate a blocking feedback state, trigger an alarm reporting sequence based on the blocking feedback state to generate an alarm completion flag, and generate a protection control instruction based on the alarm completion flag and the tripping signal.

[0007] The beneficial effects of this invention are reflected in the following points: First, by collecting current and voltage signals during equipment operation and establishing a current-voltage mapping relationship, a temperature compensation mechanism is introduced to correct the mapping relationship in conjunction with changes in ambient temperature, forming a benchmark of operating parameters that adapts to multiple operating conditions. Based on this, periodic feature detection identifies the load operation pattern of the equipment, generating threshold boundaries that dynamically adjust with the operating phase. This enables the monitoring strategy to adaptively track the periodic operating characteristics of the equipment, automatically relaxing the judgment conditions during the startup impact period to avoid false alarms, and tightening the judgment conditions during steady-state operation to ensure fault detection sensitivity. Second, multi-channel synchronous detection technology is used to acquire time-aligned data of multi-dimensional parameters such as current, voltage, power factor, and temperature. By cross-validating the extreme points of multiple channels, synchronous anomaly characteristics are identified, effectively distinguishing between single-channel measurement noise and equipment-level fault events. Simultaneously, a surge current envelope model is established for the equipment startup process, converting the excess current during startup into a compensation amount for correcting risk assessment, avoiding misjudging normal startup processes as overcurrent faults. Real-time quantitative assessment of risk coefficients is completed locally through edge computing, reducing reliance on cloud communication. Finally, the protection level is upgraded based on the risk assessment coefficient and the accumulation of anomalies. Different tripping speeds are adapted to different fault severity levels to achieve a balance between quickly cutting off serious faults and avoiding over-response to minor anomalies. After tripping, the protection status is locked through a dual mechanism of electrical and mechanical interlocking. A multi-link redundant alarm reporting strategy is adopted to ensure that fault information can still be reliably delivered to the monitoring center when communication conditions are limited, thus providing a guarantee for timely fault handling in unattended field scenarios. Attached Figure Description

[0008] The accompanying drawings illustrate specific examples of the technical solutions described in this invention and, together with the detailed embodiments, form part of the specification, serving to explain the technical solutions, principles, and effects of this invention.

[0009] Figure 1 This is a flowchart illustrating an intelligent protection switch control method for ecological restoration equipment according to the present invention.

[0010] Figure 2 This is a schematic diagram of the intelligent protection switch structure of the present invention.

[0011] Figure 3 This is a structural block diagram of an intelligent protection switch control system for ecological restoration equipment according to the present invention.

[0012] The components are as follows: 1-Intelligent protection switch body; 2-Main circuit module; 3-Signal acquisition module; 4-Control processing module; 5-Communication module; 6-Incoming terminal; 7-Outgoing terminal; 8-Current transformer; 9-Voltage sensor; 10-Temperature sensor; 11-Power supply; 12-Ecological restoration equipment load circuit; 13-Main circuit conductor; 14-Load side busbar; 15-Distribution cabinet. Detailed Implementation

[0013] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0014] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0015] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0016] The technical solutions of the embodiments of this application will be described below.

[0017] like Figure 1 As shown, this embodiment of the invention provides an intelligent protection switch control method for ecological restoration equipment, including the following steps S110-S150: Step S110: Collect load current data and voltage fluctuation data of the ecological restoration equipment, and construct an operating parameter benchmark table based on the load current data and voltage fluctuation data.

[0018] Specifically, load current data and voltage fluctuation data of the ecological restoration equipment are collected. For example... Figure 2As shown, the intelligent protection switch body 1 includes a main circuit module 2, a signal acquisition module 3, a control processing module 4, and a communication module 5. The main circuit module 2 has an input terminal 6 and an output terminal 7. The input terminal 6 is connected to the power supply 11, and the output terminal 7 is connected to the load circuit 12 of the ecological restoration equipment. The signal acquisition module 3 includes a current transformer 8, a voltage sensor 9, and a temperature sensor 10. The current transformer 8 is mounted on the main circuit conductor 13, the voltage sensor 9 is connected in parallel to the load-side busbar 14, and the temperature sensor 10 is installed inside the distribution cabinet 15. The current transformer 8, voltage sensor 9, and temperature sensor 10 are all communicatively connected to the control processing module 4. The ecological restoration equipment typically includes load types such as aeration pumps, stirring motors, water circulation pumps, and chemical dosing devices. The secondary current signal output by the current transformer 8 is converted into a voltage signal by a sampling resistor and then sent to an analog-to-digital converter. The instantaneous current values ​​acquired within the sampling period are stored in a time sequence to form load current data. Voltage sensor 9 performs power frequency cycle analysis on the sampled voltage signal, extracts the voltage peak, valley and effective value in each cycle, and records events in which the voltage change exceeds 2% of the rated voltage as voltage fluctuation events. The occurrence time, duration, fluctuation amplitude and fluctuation direction of the voltage fluctuation events are integrated to form voltage fluctuation data.

[0019] In some embodiments, constructing an operating parameter benchmark table based on the load current data and the voltage fluctuation data includes: extracting current amplitude features from the load current data to construct a current feature set; associating the current feature set with the voltage fluctuation data to construct a current-voltage mapping relationship; analyzing temperature correlation characteristics in the current-voltage mapping relationship based on ambient temperature data to generate a temperature compensation coefficient; and correcting the current-voltage mapping relationship based on the temperature compensation coefficient to generate an operating parameter benchmark table.

[0020] A current feature set is constructed by extracting current amplitude characteristics from load current data. The load current data is analyzed periodically to extract the peak current, RMS current, and peak factor for each power frequency cycle. The peak factor is defined as the ratio of the peak current to the RMS current; the peak factor for a sinusoidal current is √2 ≈ 1.414. When the peak factor in the load current data deviates significantly from 1.414, it indicates waveform distortion. The rate of change of the RMS current value between adjacent cycles is calculated as the current volatility index. Current volatility reflects the dynamic characteristics of the load; for volumetric loads such as aeration pumps, the current volatility can reach 20% to 30% at the moment the exhaust valve opens. Fourier transform is performed on the load current data to extract harmonic components, and the total harmonic distortion (THD) is calculated. Variable frequency drive pumps and agitators in ecological restoration equipment generate characteristic harmonics such as the 5th and 7th harmonics, with THD typically ranging from 8% to 15%. The current amplitude features, such as peak current, effective current, peak factor, current fluctuation rate, and THD, are integrated and aligned according to the sampling time to form a current feature set. The current feature set uses the timestamp as the main index and each feature quantity as a column field.

[0021] A current-voltage mapping relationship is constructed by associating the current feature set with voltage fluctuation data. The current feature set and voltage fluctuation data are matched by timestamp to extract current feature values ​​and voltage fluctuation parameters at the same time. When a voltage drop event is recorded in the voltage fluctuation data, the corresponding effective current value and current change rate are retrieved from the current feature set. The temporal relationship between current surges in the current feature set and voltage drops in the voltage fluctuation data is analyzed. If a current surge occurs before a voltage drop and the time difference is less than 20ms, it is determined to be a voltage fluctuation caused by a load impact. If a voltage drop occurs before a current change, it is determined to be a disturbance from the grid side conducted to the load side. A quantitative correspondence between current changes and voltage changes is established. The voltage drop amplitude caused by a unit current increment is calculated as an equivalent parameter of the line impedance. This parameter reflects the electrical stiffness of the power supply line; the lower the electrical stiffness, the more severe the voltage fluctuation caused by a load impact. The correspondence between each characteristic quantity of the current feature set and each parameter of the voltage fluctuation data is classified and stored according to the operating conditions to form a current-voltage mapping relationship. The current-voltage mapping relationship includes two parts: a steady-state operating condition mapping table and a transient operating condition mapping table. The steady-state operating condition mapping table records the statistical correspondence between the current feature set and the voltage fluctuation data under normal operating conditions, while the transient operating condition mapping table records the dynamic correspondence characteristics of transient processes such as start-up, shutdown, and load sudden changes.

[0022] Temperature compensation coefficients are generated by analyzing the temperature correlation characteristics in the current-voltage mapping relationship based on ambient temperature data. Ecological restoration equipment is typically installed outdoors or in semi-open environments where ambient temperatures can range from -10℃ to 45℃. The temperature coefficient by which conductor resistance increases with temperature is approximately 0.004 / ℃. When the ambient temperature rises from 25℃ to 45℃, the line resistance increases by about 8%, and the line voltage drop under the same load current increases accordingly, causing the impedance parameter in the current-voltage mapping relationship to drift. Temperature data of the operating environment of the ecological restoration equipment is collected. Temperature sensor 10 is installed inside the distribution cabinet 15, and the sampling period is synchronized with the current and voltage sampling. The temperature data and the current-voltage mapping relationship are jointly analyzed to extract the characteristic parameter variation patterns of the current-voltage mapping relationship in different temperature ranges. The equivalent parameters of the line impedance in the current-voltage mapping relationship are temperature normalized, and the impedance deviation ratio of each measured temperature point relative to the reference temperature is calculated using 25℃ as the reference temperature. A functional relationship between impedance deviation and temperature difference is established using linear regression or polynomial fitting methods. The regression coefficient is the temperature compensation coefficient, which physically represents the offset in the current-voltage mapping relationship caused by a unit temperature change. To verify the effectiveness of the temperature compensation coefficient, the compensated current-voltage mapping relationship is compared with measured data. The compensation residual should be less than 20% of the uncompensated residual. If the operating environment temperature of the ecological restoration equipment changes drastically without introducing a temperature compensation coefficient, the prediction error of the current-voltage mapping relationship under high or low temperature conditions will significantly increase, potentially leading to malfunctions of the protection switch under normal operating conditions or failure to operate under fault conditions.

[0023] A reference table of operating parameters is generated by correcting the current-voltage mapping relationship based on the temperature compensation coefficient. The current ambient temperature is read, and the difference ΔT between the current temperature and the reference temperature of 25℃ is calculated. The temperature compensation coefficient K_T is multiplied by the temperature difference ΔT to obtain the compensation correction amount ΔR, calculated as ΔR = K_T × ΔT, where K_T is the temperature compensation coefficient (unit: 1 / ℃), ΔT is the temperature deviation (unit: ℃), and ΔR is the dimensionless relative correction amount. Temperature correction is applied to the equivalent line impedance parameter in the current-voltage mapping relationship. The correction formula is R_corrected = R_measured × (1 + ΔR), where R_corrected is the temperature-corrected equivalent impedance, R_measured is the equivalent impedance at the measured temperature, and 1 + ΔR is the temperature correction factor. Based on the corrected impedance, the correspondence between current and voltage changes under each operating condition is recalculated. The temperature-corrected current-voltage mapping relationship is converted into a protection action criterion format. Current thresholds, voltage thresholds, and action delay parameters for each operating condition are extracted. The current threshold is set to 1.2 to 1.5 times the steady-state average current as the overload protection trigger condition, and the voltage threshold is set to 85% to 90% of the rated voltage as the undervoltage protection trigger condition. The protection criterion parameters for each operating condition are organized into an operating parameter reference table according to equipment type and temperature range. The operating parameter reference table includes three sets of parameters for normal temperature, high temperature, and low temperature ranges. The protection switch automatically selects the corresponding parameter set from the operating parameter reference table based on the real-time temperature for protection judgment.

[0024] Step S120: Divide the operating parameter reference table into threshold intervals to generate a monitoring cycle configuration, determine the current monitoring cycle based on the monitoring cycle configuration, and generate a status judgment instruction according to the deviation between the current monitoring cycle and the rated parameters of the equipment.

[0025] In some embodiments, the step of dividing the operating parameter benchmark table into threshold intervals to generate a monitoring cycle configuration includes: dividing the operating parameter benchmark table into normal parameter segments and early warning parameter segments; performing periodic feature detection on the operating parameter benchmark table to generate a load operating cycle; performing adaptive boundary adjustment on the normal parameter segments and the early warning parameter segments according to the load operating cycle to generate a dynamic threshold boundary; and locating the optimal division position in the dynamic threshold boundary to generate a monitoring cycle configuration.

[0026] The operating parameter benchmark table is divided into normal parameter segments and warning parameter segments. Statistical analysis is performed on the parameters of each device under each operating condition in the operating parameter benchmark table, calculating the parameter mean, standard deviation, and extreme value range. The parameter mean reflects the typical operating state of the equipment, and the standard deviation reflects the natural fluctuation of the parameter. The boundary of the normal parameter segment is defined with the parameter mean as the center and a radius of 2 times the standard deviation. Parameter values ​​falling within the normal parameter segment correspond to the normal operating state of the equipment; statistically, approximately 95.4% of normal operating data points should fall within this range. The range outside the normal parameter segment boundary to 3 times the standard deviation is defined as the warning parameter segment. Parameter values ​​falling into the warning parameter segment indicate that the equipment's operating state has deviated from the normal range but has not yet reached the level of a fault. The current and voltage parameters in the operating parameter benchmark table are also divided into normal and warning parameter segments respectively. The standard deviation of current parameters is typically 3% to 8% of the mean, and the standard deviation of voltage parameters is typically 1% to 3% of the mean, with differences in the range width between the two types of parameters. The boundary values ​​of the normal and warning parameter segments are organized and stored according to equipment type and parameter type, forming a preliminary threshold range division result. The dosing device in the ecological restoration equipment adopts an intermittent operation mode. During operation, the current is concentrated near the rated value, and during shutdown, the current is close to zero. If the two-standard-deviation rule is simply applied to such equipment, the normal parameter range will be too wide and lose its monitoring significance. Therefore, it is necessary to first identify the operating mode of the equipment and then divide it into segments.

[0027] Periodic feature detection is performed on the operating parameter baseline table to generate load operation cycles. The operation of ecological restoration equipment typically exhibits periodic patterns. Aeration pumps may start and stop according to dissolved oxygen concentration setpoints, and stirring motors may operate intermittently at fixed time intervals. Identifying these periodic patterns helps distinguish between normal periodic fluctuations and abnormal random fluctuations. Autocorrelation analysis is performed on historical parameter data recorded in the operating parameter baseline table to calculate the correlation coefficient between the parameter time series and its delayed replicas. The delay amount at which the correlation coefficient peaks corresponds to the equipment's operation cycle. Fast Fourier Transform (FFT) is used to perform spectral analysis on the parameter time series. The main peak frequency in the spectrum corresponds to the fundamental frequency component of the load operation cycle. The presence of multiple significant peaks indicates multiple overlapping cycles in the equipment's operation. A stability index for the load operation cycle is calculated, defined as the ratio of the standard deviation of the cycle length of several adjacent cycles to the average cycle length. A stability index less than 5% indicates that the load operation cycle is stable and reliable. The detected load operation cycles are stored according to equipment identification, and the confidence level of the cycle detection is recorded. The confidence level is determined by the significance of the autocorrelation peak and the signal-to-noise ratio of the main peak in the spectrum. Some ecological restoration equipment uses variable frequency drive, and the operating frequency is dynamically adjusted according to process requirements. The load operation cycle of such equipment exhibits quasi-periodic characteristics, requiring time-frequency joint analysis using short-time Fourier transform or wavelet transform.

[0028] For example, the step of adaptively adjusting the boundaries of the normal parameter segment and the early warning parameter segment to generate a dynamic threshold boundary based on the load operation cycle includes: identifying parameter mutation characteristics of the normal parameter segment and the early warning parameter segment according to the load operation cycle to determine the detection range; tracking the parameter change process within the detection range to construct a parameter change map; identifying gradual deterioration features and mutation anomalies within the parameter change map and marking them to generate change classification labels; and jointly generating a dynamic threshold boundary based on the change classification labels and the parameter change map.

[0029] The detection range is determined by identifying parameter mutation characteristics in normal and warning parameter segments according to the load operation cycle. The parameter time series is segmented according to the load operation cycle, and parameter sample points are extracted near the boundaries of the normal and warning parameter segments within each cycle. The vicinity of the boundary is defined as the area within ±5% of the boundary value. First-order differencing is performed on the parameter samples near the boundary. The first-order differencing value reflects the instantaneous rate of change of the parameter. Points with an absolute difference value exceeding 3% of the parameter mean are marked as mutation candidate points. The phase distribution of each mutation candidate point within the load operation cycle is statistically analyzed. If mutation candidate points are concentrated in a specific phase interval, that interval corresponds to the state switching time of the equipment, such as start / stop time or load mutation time. The phase interval where mutation candidate points are concentrated is determined as the detection range. The starting and ending phases of the detection range are determined based on the 5th and 95th quantiles of the mutation candidate point distribution. Parameter changes are gradual in phase intervals outside the detection range, allowing for a more lenient threshold boundary. Parameter changes are drastic in phase intervals within the detection range, requiring fine division using time-varying threshold boundaries. The water circulation pump in the ecological restoration equipment exhibits a significant current surge during startup, jumping from zero current in a static state to approximately six times the rated current. The detection range of this surge typically covers a time window of 0 to 500 ms after startup.

[0030] A parameter change graph is constructed by tracking parameter changes within the detection range. Within the phase interval corresponding to the detection range, a parameter time series is acquired using an encrypted sampling rate, typically 5 to 10 times the normal sampling rate, to capture the details of rapid parameter changes. The parameter time series within the detection range is then smoothed and filtered using a moving average or exponentially weighted moving average filter. The filter window length is set to 1% to 2% of the load operating cycle to remove high-frequency noise while preserving the parameter change trend. The first and second derivatives of the filtered parameter series are calculated. The first derivative reflects the rate of parameter change, and the second derivative reflects the acceleration of parameter change. Zero-crossing points of the first derivative correspond to local extrema of the parameter, and zero-crossing points of the second derivative correspond to inflection points of the rate of parameter change. The original parameter values, filtered parameter values, first derivatives, and second derivatives within the detection range are aligned and organized according to sampling time to form a parameter change graph. The parameter change graph uses sampling time as the horizontal axis and various parameter values ​​as the vertical axis, presenting a cluster of curves showing synchronous changes across multiple channels. The current boundary lines of the normal parameter segment and the warning parameter segment are marked on the parameter change graph. The intersection of the boundary line and the parameter curve corresponds to the trigger time of the state switch. The parameter change graph also records the parameter change trajectory of historical operating cycles. By comparing the trajectory differences between the current cycle and the historical cycle, the gradual trend of equipment operating status can be identified.

[0031] Based on the parameter change graph, gradual deterioration features and abrupt anomalies are identified and labeled to generate change classification labels. The parameter change trajectories in the parameter change graph are classified by feature: gradual deterioration features are characterized by a slow drift of the parameter baseline, while abrupt anomalies are characterized by a step jump in parameters. When detecting gradual deterioration features, the mean parameter sequence for multiple consecutive load operation cycles in the parameter change graph is extracted. Linear regression analysis is performed on the mean sequence; if the regression slope is significantly non-zero and lasts for more than 10 cycles, gradual deterioration is identified, which may be caused by equipment wear, insulation aging, or increased contact resistance. When detecting abrupt anomalies, the absolute peak value of the first derivative in the parameter change graph is calculated. If the peak value exceeds 1.5 times the peak value during normal start-up and shutdown, abrupt anomalies are identified, which may be caused by load jamming, short circuits, or phase loss. Change classification labels are generated for the detected gradual deterioration and abrupt anomaly features. The gradual deterioration label includes the direction of deterioration, the rate of deterioration, and the cumulative amount of deterioration; the abrupt anomaly label includes the time of abrupt change, the magnitude of the abrupt change, and the duration of the abrupt change. A correlation index is established between the change classification markers and their corresponding positions in the parameter change graph, facilitating rapid location of abnormal features during subsequent analysis. Increased mechanical resistance caused by bearing wear in equipment can lead to a gradual increase in current of approximately 0.5% to 1% per month. If this gradual deterioration is not identified, overcurrent protection may be triggered suddenly after several months without warning.

[0032] A dynamic threshold boundary is generated based on the change classification markers and parameter change graphs. The threshold boundary in the parameter change graph is compensated and adjusted according to the gradual deterioration information in the change classification markers. If gradual deterioration characteristics exist, the threshold boundary in the parameter change graph is compensated and adjusted in the opposite direction to the deterioration direction. The compensation amount accumulates periodically according to the deterioration rate, with the upper limit set at 10% of the original boundary value. When the compensation upper limit is exceeded, an equipment maintenance reminder is generated instead of further widening the boundary. Abrupt change anomaly information in the change classification markers is read. If abrupt change anomaly records exist in historical data, the threshold boundary is tightened at the corresponding phase position to improve the detection sensitivity of subsequent anomalies. The boundary tightening amount is proportional to the average amplitude of historical abrupt changes. Combining the adjustments of gradual compensation and abrupt tightening, the boundaries of the normal parameter segment and the warning parameter segment in the parameter change graph are jointly corrected. The corrected boundary curve exhibits periodic changes and adaptively tracks the evolution of equipment status. The corrected boundary curve is discretized into a sequence of threshold points to form a dynamic threshold boundary. The spacing between the threshold points of the dynamic threshold boundary is consistent with the sampling interval of the load operation cycle, ensuring point-by-point comparison during real-time monitoring. The dynamic threshold boundary also includes a boundary update timestamp and update reason record. When the change classification label changes, the dynamic threshold boundary is triggered by incremental update. The update process adopts a gradual smoothing method to avoid false triggering caused by boundary step.

[0033] The optimal division position within the dynamic threshold boundary is located to generate the monitoring cycle configuration. The dynamic threshold boundary exhibits periodic changes, requiring the determination of the optimal time and parameter position for monitoring cycle switching on the boundary curve. The gradient of the dynamic threshold boundary is analyzed; locations with larger gradients correspond to periods of rapid parameter change, during which switching the monitoring cycle may lead to discontinuous sampling. Cycle switching points should be avoided at gradient peaks. The curvature of each point on the dynamic threshold boundary is calculated. Flat sections with smaller curvature are suitable as stable operation judgment intervals for the monitoring cycle, while inflection sections with larger curvature are suitable as trigger judgment intervals for state switching. A first-level cycle switching point is set at the boundary between the normal parameter segment and the warning parameter segment of the dynamic threshold boundary. Crossing this boundary triggers a switch from normal mode to encrypted mode. A second-level cycle switching point is set at the boundary between the warning parameter segment and the alarm interval. Crossing this boundary triggers a switch to continuous sampling mode. The position coordinates, switching direction, and switching delay parameters of each level of cycle switching point are integrated to form the monitoring cycle configuration. This monitoring cycle configuration is stored in association with the dynamic threshold boundary, forming a complete adaptive monitoring strategy. The monitoring cycle configuration also includes a hysteresis interval setting. When the parameter drops from the warning parameter range to the normal parameter range, the monitoring frequency needs to be reduced only after the parameter has been running stably for more than 3 load operation cycles within the normal parameter range, so as to avoid frequent switching of the monitoring cycle when the parameter fluctuates near the boundary.

[0034] The current monitoring cycle is determined based on the monitoring cycle configuration. Real-time operating parameters of the ecological restoration equipment are read, and the real-time parameter values ​​are compared with the boundary values ​​of each threshold interval in the monitoring cycle configuration to determine the current threshold interval. The corresponding monitoring frequency parameter is extracted from the monitoring cycle configuration based on the current threshold interval. The monitoring frequency parameter determines the sampling interval time, and the reciprocal of the sampling interval time is the sampling rate of the current monitoring cycle. When the real-time parameter crosses from one threshold interval to another, the current monitoring cycle is dynamically adjusted. The adjustment process uses a gradual approach to avoid data discontinuity caused by sudden changes in the sampling rate; the gradual adjustment time is set to 5 sampling cycles. The determination of the current monitoring cycle also needs to consider the joint judgment of multiple parameters. When the current parameter is in the normal range and the voltage parameter is in the warning range, the current monitoring cycle is executed according to the monitoring frequency corresponding to the higher risk level. The historical change trajectory of the current monitoring cycle is recorded. If the current monitoring cycle switches frequently within a short period, it indicates that the equipment's operating status is unstable, and a special diagnostic program needs to be initiated. The monitoring cycle configuration also includes a minimum monitoring cycle limit to prevent excessively high sampling rates under extreme operating conditions from overloading the processor. The minimum monitoring cycle is usually set to 1ms, corresponding to a maximum sampling rate of 1kHz.

[0035] A status determination command is generated based on the deviation between the current monitoring cycle and the equipment's rated parameters. Real-time operating parameters of the equipment are acquired at the sampling time determined in the current monitoring cycle. These real-time operating parameters are compared with the equipment's rated parameters to calculate the percentage deviation. The deviation represents the relative degree of deviation of the real-time parameters from the rated parameters. A corresponding status determination command is generated based on the magnitude and duration of the deviation: a normal operation command is generated when the deviation is less than 10% and the duration is unlimited; a status attention command is generated when the deviation is between 10% and 20% and the duration exceeds 30 seconds; a warning command is generated when the deviation is between 20% and 35% and the duration exceeds 10 seconds; and an alarm command is generated when the deviation exceeds 35% or the duration exceeds 5 seconds. Status determination commands use priority coding: the highest priority alarm command is coded as 0x01, the warning command as 0x02, the status attention command as 0x03, and the normal operation command as 0x00. After generation, the status determination command is sent to the protection execution module via the internal bus. The protection execution module determines whether to trigger a protection action based on the priority of the status determination command. The operating environment of ecological restoration equipment changes significantly during seasonal transitions. In the summer, the high temperature increases the resistance of the motor windings, which may cause the current deviation to remain in the range of 10% to 15% for a long time. If seasonal factors are not considered, status attention instructions will be generated frequently, causing fatigue of maintenance personnel. Therefore, the generation logic of status judgment instructions needs to be comprehensively judged in combination with the historical trend of the current monitoring cycle.

[0036] Step S130: Based on the status determination instruction, perform multi-channel synchronous detection to obtain fault characteristic parameters, identify the start-up surge characteristics of the load current data to generate the start-up current compensation amount, determine the risk assessment coefficient through edge computing based on the fault characteristic parameters and the start-up current compensation amount, and generate an abnormal triggering flag based on the risk assessment coefficient.

[0037] Specifically, multi-channel synchronous detection acquires fault characteristic parameters based on status determination commands. The status determination commands define the current monitoring mode and sampling frequency. Under the control of these commands, multi-channel synchronous detection simultaneously acquires multiple parameters such as current, voltage, power factor, and temperature. A multi-channel synchronous sampling circuit is configured, with each channel's sampling clock driven by a unified time base signal. The sampling time deviation between channels is controlled within 10μs, ensuring a strict time correspondence between the parameters of each channel. When the status determination command is a normal operation command, multi-channel synchronous detection executes at the normal frequency, and the sampled data is preprocessed and stored in a circular buffer for analysis. When the status determination command is a warning or alarm command, multi-channel synchronous detection switches to encrypted sampling mode, and the sampled data is sent to the fault analysis module for online processing in real time. Feature extraction is performed on the raw data acquired by multi-channel synchronous detection: RMS value, peak value, and harmonic content are extracted for the current channel; RMS value, fluctuation amplitude, and three-phase imbalance are extracted for the voltage channel; and active power, reactive power, and power factor are extracted for the power channel. The feature parameters extracted from each channel are aligned and integrated according to the timestamp to form fault feature parameters. The fault feature parameters are listed with parameter type as the column field and sampling time as the row index. When a single-phase grounding fault occurs in the ecological restoration equipment, the three-phase current will show obvious asymmetrical distribution. Detecting only the single-phase current cannot reliably identify this type of fault. The fault feature parameters obtained by multi-channel synchronous detection can quickly locate the fault type through the symmetrical component analysis of the three-phase current.

[0038] In some embodiments, the step of identifying the start-up surge characteristics of the load current data to generate the start-up current compensation amount includes: extracting the peak envelope of the load current data to generate the current peak boundary; performing a multiplier analysis based on the current peak boundary to identify the surge dominance range; evaluating the duration of the surge dominance range to generate the surge duration; and performing compensation mapping based on the surge duration to determine the start-up current compensation amount.

[0039] Peak envelope extraction is performed on load current data to generate peak current boundaries. The load current data is segmented according to the power frequency cycle, and the instantaneous maximum current value within each cycle is extracted as the peak sampling point for that cycle. Peak sampling points from multiple consecutive cycles constitute a peak time series. An upper envelope extraction algorithm is applied to the peak time series. This algorithm combines local maximum detection with spline interpolation, with a detection window length set to 5 power frequency cycles. Local maximum points are identified within the window and connected using cubic spline interpolation to form a continuous upper envelope curve. This upper envelope curve represents the upper boundary of the current peak boundary, reflecting the maximum envelope trajectory of the current peak in the load current data. Similarly, a lower envelope extraction algorithm is applied to the peak time series to generate the lower boundary of the current peak boundary. The area between the upper and lower boundaries reflects the normal fluctuation range of the current peak. The distance between the upper and lower boundaries of the current peak boundary changes over time. A larger distance during the equipment startup phase reflects the uncertainty of the startup process, while a narrower distance during steady-state operation reflects the stability of the operating state. The upper boundary curve, lower boundary curve, and boundary distance sequence of the current peak boundary are stored as structured data for subsequent surge interval identification and analysis. During the speed regulation process, the peak current of the variable frequency drive water pump will fluctuate with the frequency. The current peak boundary can adaptively track this change and distinguish between normal speed regulation and abnormal impact.

[0040] Surge-dominant regions are identified through ratio analysis based on the current peak boundary. The ratio of each point on the upper boundary curve of the current peak boundary to the steady-state current peak is calculated. The steady-state current peak is the average value of the current peak boundary during the steady-state operation phase. The ratio characterizes the amplification degree of the instantaneous peak current relative to the steady-state peak current. A ratio threshold is set to delineate surge and non-surge regions. The ratio threshold is typically set between 1.5 and 2.0. When the current ratio exceeds the ratio threshold, the region is considered to be in a surge region; when the current ratio falls below the ratio threshold, the region is considered to be out of a surge region. The continuous period where the current ratio on the current peak boundary consistently exceeds the threshold is identified as the surge-dominant region. The start time of the surge-dominant region corresponds to the trigger point of the initiation event, and the end time corresponds to the moment when the surge current decays to a steady-state level. Further analysis of the temporal distribution characteristics of the surge dominance interval reveals that the surge dominance interval is subdivided into impact sub-intervals and attenuation sub-intervals based on the magnitude of the surge. The impact sub-interval, corresponding to periods with surge magnitudes exceeding 5, typically lasts 20 to 50 ms, while the attenuation sub-interval, corresponding to periods with surge magnitudes gradually decreasing from 5 to 1.5, typically lasts 200 to 2000 ms. Recording the duration and peak surge magnitude of each sub-interval forms a structured description of the surge dominance interval.

[0041] The surge duration is generated by assessing the duration of the dominant surge region. The start and end times of the dominant surge region are statistically analyzed, and the difference between these two times is calculated to obtain the total surge duration. Further analysis of the energy distribution within the dominant surge region is performed, calculating the duration of the impact sub-region and the attenuation sub-region separately; their sum equals the total surge duration. The reasonableness of the surge duration is evaluated by comparing the measured surge duration with the equipment's nominal start-up time, which can be obtained from the equipment nameplate or technical manual. If the measured duration exceeds 150% of the nominal duration, it is marked as an abnormal start-up. A correlation model between surge duration and load status is established. The surge duration for no-load start-up is typically 50% to 70% of that for full-load start-up; the current load level can be inferred from the surge duration. The various components of the surge duration are integrated to form a structured duration description, including total duration, impact duration, attenuation duration, and deviation rate relative to the nominal duration. The surge duration is also used to set the delay parameters of the protection logic during startup. Within the surge duration range, the overcurrent protection judgment is automatically shielded to avoid accidental cut-off during normal startup. If the surge duration of the aeration pump is abnormally prolonged due to pipeline blockage during startup, continuous monitoring can detect such mechanical faults in a timely manner.

[0042] The starting current compensation amount is determined by a compensation mapping based on the surge duration. A mapping function is established between the surge duration and the starting current compensation amount. The mapping function is a piecewise linear function. The first segment, corresponding to the impact sub-interval, uses a high compensation coefficient; the second segment, corresponding to the attenuation sub-interval, uses a decreasing compensation coefficient; and the third segment, corresponding to the steady-state interval, has a compensation coefficient that drops to zero. The starting current compensation amount for the impact sub-interval is calculated using the formula C_impact = K_impact × (I_peak - I_rated) / I_rated × T_impact / T_ref, where C_impact is the compensation amount for the impact segment, K_impact is the impact compensation coefficient ranging from 0.8 to 1.0, I_peak is the peak current in the dominant surge interval, I_rated is the rated current of the equipment, T_impact is the duration of the impact sub-interval, and T_ref is a reference time constant with a value of 100 ms. The compensation amount for the attenuation sub-interval is calculated by integrating the excess current over time. The integration interval is the attenuation sub-interval, and the compensation coefficient ranges from 0.3 to 0.5. The total starting current compensation is obtained by adding the two compensation values. This starting current compensation is a dimensionless value, representing the equivalent risk weight of excess current during startup. After being passed to the risk assessment module, the starting current compensation is used to correct current-related indicators in the fault characteristic parameters, eliminating the interference of startup surges on fault diagnosis. The statistical average of the starting current compensation after multiple equipment starts can serve as an indirect indicator of equipment health. A continuous increase in the compensation may indicate deterioration of the motor assembly insulation or wear of the mechanical transmission system.

[0043] In some embodiments, determining the risk assessment coefficient by edge computing based on the fault characteristic parameters and the starting current compensation includes: constructing a parameter time series record based on the fault characteristic parameters; extracting multiple types of parameter extreme points from the parameter time series record to construct multi-channel extreme value calibration points; performing cross-validation on the multi-channel extreme value calibration points to identify synchronous abnormal parameters and generate cross-validation results; and determining the risk assessment coefficient by edge computing based on the cross-validation results and the starting current compensation.

[0044] A parameter time series record is constructed based on the fault characteristic parameters. The parameters of each dimension in the fault characteristic parameters are expanded into a time series according to the sampling time, with each dimension forming an independent parameter channel. All channels share a unified time axis scale. A circular buffer is configured to store the parameter time series record, with a buffer depth set to 1000 to 5000 sampling points. Based on the current sampling frequency, it can cover 1 to 50 seconds of historical data. New data overwrites the oldest data point to achieve rolling updates. Preprocessing is performed on the data of each channel in the parameter time series record. The preprocessing steps include outlier removal, missing value imputation, and baseline drift correction. Outliers are removed using the 3σ criterion, missing values ​​are imputed using linear interpolation, and baseline drift is eliminated using high-pass filtering. A status label is added to each sampling point in the parameter time series record. The label content includes the equipment operation stage, the current threshold range, and a validity flag. The operation stage is divided into three categories: startup, steady state, and shutdown. The validity flag indicates whether the data point has passed quality inspection. The parameter time series records form the data foundation for fault analysis. Subsequent extreme value extraction and cross-validation are all based on the parameter time series records. The intermittent operation characteristics of ecological restoration equipment require that the parameter time series records can maintain data integrity across multiple start-stop cycles.

[0045] Multi-channel extreme value calibration points are constructed by extracting extreme value points of various parameters from the parameter time series records. Extreme value detection is performed on each parameter channel in the parameter time series records using a local comparison method. When the parameter value of a sampling point is greater than its N neighboring points, it is determined to be a local maximum point. N is typically 5 to 10, corresponding to a time window of 50ms to 100ms. Similarly, local minimum points are detected. The maximum and minimum points of each channel are sorted by timestamp to form an extreme value point sequence. A significance index is calculated for each extreme value point. The significance index is defined as the difference in amplitude between the extreme value point and its neighboring extreme values ​​divided by the standard deviation of the channel. Extreme value points with a significance index greater than 2 are marked as significant extreme value points, and those with a significance index less than 2 are marked as ordinary extreme value points. Significant extreme points are selected from the extreme point sequences of each channel. These significant extreme points from multiple channels are then aligned and integrated by timestamp to form multi-channel extreme value calibration points. The data structure of these calibration points is a two-dimensional matrix, where the row index represents the time of the extreme event, the column index represents the parameter channel type, and the matrix elements represent the extreme value amplitude and type of that channel at that time. These multi-channel extreme value calibration points record the key state changes of each parameter channel. Equipment failures are often accompanied by the simultaneous occurrence of significant extreme values ​​in multiple parameter channels; analyzing the synchronicity of the multi-channel extreme value calibration points can help identify the failure event.

[0046] For example, the step of performing cross-validation to identify synchronous abnormal parameters and generate cross-validation results for the multi-channel extreme value calibration points includes: detecting fluctuation amplitudes of the multi-channel extreme value calibration points to generate a fluctuation amplitude distribution; identifying abnormally increasing positions in the fluctuation amplitude distribution to construct an abnormal position set; assessing the degree of abnormality of the abnormal position set to generate a corrected risk quantity; and jointly generating cross-validation results based on the corrected risk quantity and the multi-channel extreme value calibration points.

[0047] A fluctuation amplitude distribution is generated by detecting fluctuation amplitudes at multi-channel extreme value calibration points. The deviation amplitude of each extreme point relative to the channel baseline value is calculated; this deviation amplitude characterizes the degree of relative deviation of the extreme point amplitude from the baseline value. The deviation amplitudes of each extreme point are arranged in chronological order to form a time series of deviation amplitudes. Statistical analysis is performed on the time series of deviation amplitudes to calculate the mean, standard deviation, and extreme value range of the deviation amplitudes. The range of deviation amplitude values ​​is divided into several intervals, and the proportion of extreme points in each interval is statistically analyzed to form a histogram distribution of the deviation amplitudes. The horizontal axis of the histogram represents the deviation amplitude interval, and the vertical axis represents the frequency of extreme points in that interval. The histograms of deviation amplitudes for each channel of the multi-channel extreme value calibration points are integrated to form a fluctuation amplitude distribution. The fluctuation amplitude distribution is in matrix form, with row indices representing parameter channel types, column indices representing deviation amplitude intervals, and element values ​​representing frequencies. The fluctuation amplitude distribution reveals the statistical regularity of extreme value fluctuations in each channel. Under normal operation, the fluctuation amplitude distribution of each channel should exhibit an approximately Gaussian distribution. Abnormal distribution patterns, such as bimodal or heavy-tailed distributions, indicate the presence of non-steady-state factors.

[0048] An abnormal location set is constructed by identifying anomalous surge locations in the fluctuation amplitude distribution. Anomaly interval detection is performed on the fluctuation amplitude distribution, calculating the deviation between the expected frequency and the actual frequency for each deviation interval. Intervals with significantly positive deviations correspond to anomalous surge intervals. An anomaly judgment threshold is set to three times the expected frequency. When the actual frequency of an interval exceeds three times the expected frequency, an anomalous surge is determined to exist in that interval, indicating an abnormally high frequency of extreme events within that amplitude range. Multi-channel extreme value calibration points are traced back to extract the timestamps and channel identifiers of all extreme points whose deviation amplitude falls into the anomalous surge interval. These extreme points correspond to potential anomalous events. The timestamps of the anomalous extreme points are sorted by occurrence time. After removing expected extreme points generated during normal start-stop processes, the location information of the remaining extreme points is used to construct an anomalous location set. The data structure of the anomalous location set is a list, with each element containing three fields: anomalous time, anomalous channel, and deviation amplitude. Adjacent anomalous locations in the anomalous location set with a time interval of less than 1 second are merged into the same anomalous event. After merging, the start time, end time, and a list of involved channels are recorded. An abnormal surge in the right tail of the fluctuation amplitude distribution usually corresponds to a severe overload or short circuit event, while an abnormal surge in the left tail may correspond to a phase loss or poor contact event.

[0049] Anomaly severity assessment is performed on the set of abnormal locations to generate a corrected risk quantity. The severity of each abnormal event in the set of abnormal locations is quantitatively assessed, with assessment indicators including deviation magnitude, number of channels involved, and duration. The deviation magnitude score represents the ratio of the maximum deviation magnitude of the abnormal event to the protection trigger threshold; a score greater than 1 indicates that the protection trigger condition has been met. The channel coverage score represents the ratio of the number of channels involved in the abnormal event to the total number of monitored channels; a higher score indicates a wider range of impact. The duration score represents the ratio of the duration of the abnormal event to the critical duration, typically taken as 100ms; a score greater than 1 indicates that the duration of the abnormal event poses a threat. The corrected risk quantity of the abnormal event is calculated by combining the three scores using the formula: R_mod = w_amp × S_amp + w_ch × S_ch + w_dur × S_dur, where R_mod is the corrected risk quantity, S_amp is the deviation magnitude score, S_ch is the channel coverage score, S_dur is the duration score, and w_amp, w_ch, and w_dur are the weights of each dimension, typically 0.5, 0.3, and 0.2, respectively. The corrected risk values ​​of all anomalies in the anomaly location set are aggregated to form a corrected risk value sequence, with each element of the sequence corresponding one-to-one with an event in the anomaly location set. Anomalies with a corrected risk value exceeding 1.0 are marked as high-risk events requiring immediate action, those between 0.5 and 1.0 are marked as medium-risk events requiring close monitoring, and those below 0.5 are marked as low-risk events and included in routine monitoring.

[0050] Cross-validation results are generated jointly based on the corrected risk quantity and multi-channel extreme value calibration points. The corrected risk quantity sequence is correlated with the original extreme value information in the multi-channel extreme value calibration points. A risk weight label is attached to each significant extreme value point, with the risk weight label value being the corrected risk quantity of the corresponding abnormal event. The risk weight label for extreme value points not covered by the abnormal location set is set to 0. Multi-channel extreme value calibration points are grouped and analyzed by channel, and the number and distribution density of high-risk extreme value points in each channel are statistically analyzed. When the density of high-risk extreme value points in a certain channel is significantly higher than that in other channels, the parameter corresponding to that channel may have a systematic anomaly. Sliding window analysis is performed on the multi-channel extreme value calibration points by time window, and the cumulative corrected risk quantity within each time window is statistically analyzed. An upward trend in the cumulative value indicates that the equipment condition is deteriorating. The channel-dimensional analysis results and time-dimensional analysis results are integrated to form the cross-validation results. The data structure of the cross-validation results includes three parts: risk channel ranking, risk trend judgment, and comprehensive anomaly score. The comprehensive anomaly score in the cross-validation results is calculated using a time-decay weighted method, with the formula: Score = Σ(R_mod_i × e^(-Δt_i / τ)), where Score is the comprehensive anomaly score, R_mod_i is the corrected risk amount for the i-th anomaly event, Δt_i is the time difference between the event and the current time, τ is the decay time constant (valued at 5 seconds), and e is the natural constant, with more recently occurring anomalies contributing more to the score. The cross-validation results are then passed to the risk assessment module and used in conjunction with the starting current compensation to calculate the final risk assessment coefficient.

[0051] The risk assessment coefficient is determined through edge computing based on the cross-validation results and the starting current compensation. The cross-validation contribution is determined based on the comprehensive anomaly score in the cross-validation results. It is determined whether the device is currently in the startup phase; if so, a compensation value is extracted from the calculated starting current compensation; otherwise, the compensation value is set to 0. The final calculation of the risk assessment coefficient is performed at the edge computing node. The formula is R_final = Score_cv × (1 - C_total × K_comp) + R_base, where R_final is the risk assessment coefficient, Score_cv is the comprehensive anomaly score, C_total is the starting current compensation, K_comp is the compensation conversion coefficient (0.5 to 0.8), and R_base is the base risk value (0.1) representing the inherent risk of device operation. The risk assessment coefficient ranges from 0 to 10, with 0 to 1 representing low risk, 1 to 3 representing medium risk, 3 to 6 representing high risk, and 6 to 10 representing extremely high risk. Edge computing nodes perform time-smoothing on the risk assessment coefficients, using an exponentially weighted moving average to filter out transient spikes, with a smoothing coefficient of 0.3. The smoothed risk assessment coefficients are then output to the protection logic module, while the risk assessment coefficients and intermediate variables from their calculation process are stored in a log for post-event analysis.

[0052] An abnormal trigger flag is generated based on a risk assessment coefficient. The risk assessment coefficient is compared with a preset trigger threshold, which is divided into two levels: a warning threshold and an alarm threshold. The warning threshold is typically set to 2.0, and the alarm threshold is typically set to 4.0. When the risk assessment coefficient exceeds the warning threshold but does not reach the alarm threshold, a warning-level abnormal trigger flag is generated (coded 0x02). After triggering, the protection switch enters a warning state, increases the monitoring frequency, and sends an alarm message to the monitoring center but does not perform a disconnect action. When the risk assessment coefficient exceeds the alarm threshold, an alarm-level abnormal trigger flag is generated (coded 0x01). After triggering, the protection switch performs protection actions according to a preset strategy, including current-limiting operation and tripping disconnection. The generation of the abnormal trigger flag also requires a duration confirmation condition. The risk assessment coefficient must remain above the threshold for more than the confirmation time before an abnormal trigger flag is generated. The confirmation time for the warning level is 500ms, and the confirmation time for the alarm level is 100ms. This confirmation mechanism filters out transient interference to avoid false triggering. After an anomaly trigger flag is generated, it notifies the protection execution module via a hardware interrupt. The execution module completes the protection action response within 10ms of receiving the anomaly trigger flag. The anomaly trigger flag also records the risk assessment coefficient value, fault characteristic parameter snapshot, and cross-validation result summary at the trigger time, forming a complete fault event record, which facilitates subsequent fault cause analysis and protection strategy optimization.

[0053] Step S140: Based on the abnormal trigger flag, determine the protection level and generate protection control parameters. Based on the protection control parameters, perform threshold comparison to generate the fault limit. Based on the fault limit, perform circuit breaker response control to generate the tripping signal.

[0054] In some embodiments, the step of determining the protection level and generating protection control parameters based on the abnormal trigger flag includes: converting the abnormal trigger flag into an abnormal level mapping table; detecting the number of consecutive abnormalities based on the abnormal level mapping table to generate an abnormal cumulative count; determining the protection level upgrade based on the abnormal cumulative count to generate an upgraded level identifier; and performing level aggregation on the upgraded level identifier to generate protection control parameters.

[0055] Convert the abnormal trigger flags into an abnormality level mapping table. Parse the encoded fields of the abnormal trigger flags to extract the protection level code, fault type code, and trigger time information. Protection level code 0x01 corresponds to Level 1 severe abnormality, 0x02 to Level 2 moderate abnormality, and 0x03 to Level 3 minor abnormality. Extract the risk assessment coefficient value and fault characteristic parameter summary from the additional information of the abnormal trigger flags. The risk assessment coefficient value is used to refine the severity ranking within the same protection level. Establish conversion rules from abnormal trigger flags to the abnormality level mapping table. The conversion rules map discrete abnormal trigger flag codes to continuous abnormality level values. The formula for calculating the abnormality level value is L_exc = L_base + R_final × K_scale, where L_exc is the abnormality level value, L_base is the base level value corresponding to the protection level code (Level 1 is 3.0, Level 2 is 2.0, and Level 3 is 1.0), R_final is the risk assessment coefficient, and K_scale is a scaling factor with a value of 0.1. The converted anomaly level value, along with the trigger time, fault type, and original anomaly trigger flag code, is integrated to form a record in the anomaly level mapping table. The anomaly level mapping table is stored in a time-series structure, with new records appended to the end. The table retains the 100 most recent anomaly records for cumulative analysis. Ecological restoration equipment may frequently generate anomaly trigger flags during severe weather. The anomaly level mapping table can comprehensively record the level distribution of each anomaly trigger flag, aiding in determining whether it is a systemic fault or environmental interference.

[0056] An anomaly cumulative count is generated based on the number of consecutive anomalies detected using an anomaly level mapping table. The timestamp sequence in the anomaly level mapping table is scanned to detect the time interval between adjacent anomaly records. When the time interval is less than a set continuity judgment window (usually 60 seconds), the two records are determined to belong to the same consecutive anomaly sequence. The number of records belonging to the current consecutive anomaly sequence in the anomaly level mapping table is counted; this number is the base value of the anomaly cumulative count. A level-weighted correction is applied to the anomaly cumulative count, with higher-level anomaly records having a greater weight than lower-level anomaly records in the anomaly level mapping table. The formula for calculating the weighted anomaly cumulative count is N_acc = Σ(w_level × n_level), where N_acc is the anomaly cumulative count, w_level is the weight coefficient for each level (Level 1 weight 3.0, Level 2 weight 2.0, Level 3 weight 1.0), and n_level is the number of records at each level. The anomaly cumulative count also incorporates a time decay factor. The contribution of earlier anomalies to the cumulative count decays over time, with the decay formula being N_decay=N_acc×e^(-Δt / τ), where Δt is the time difference from the current moment, and τ is the decay time constant with a value of 300 seconds. The time-decayed anomaly cumulative count is used as the final output, reflecting the recent anomaly activity level of the equipment. Initial wear of the agitator motor bearing may manifest as intermittent, slight overcurrent, accumulating multiple Level 3 anomaly records in the anomaly level mapping table. As the anomaly cumulative count increases over time, it can trigger preventative maintenance reminders.

[0057] The protection level is upgraded based on the accumulated anomaly count, generating an upgraded level identifier. An accumulated threshold for protection level upgrades is set; the threshold for Level 1 is typically set to 5.0, and for Level 2 to 3.0. When the accumulated anomaly count exceeds the Level 2 upgrade threshold but does not reach the Level 1 upgrade threshold, the protection level is upgraded one level. When the accumulated anomaly count exceeds the Level 1 upgrade threshold, the protection level is directly upgraded to the highest level. The current protection level identifier is read. If the current protection level is Level 3, the upgraded level identifier changes to Level 2; if the current protection level is Level 2, the upgraded level identifier changes to Level 1; if the current protection level is already Level 1, the upgraded level identifier remains unchanged but an emergency handling flag is added. The upgraded level identifier also considers the trend of the accumulated anomaly count. If the accumulated anomaly count rises rapidly in a short period (rising rate exceeding 1.0 / minute), the highest level upgraded level identifier is generated directly without considering the current level. A rapidly rising accumulated anomaly count usually indicates that the fault is worsening. The upgraded level identifier contains three fields: the target protection level code, the upgrade reason code, and the upgrade timestamp. The upgrade reason code distinguishes between two modes: accumulated threshold triggering and trend triggering. The protection level upgrade adopts a one-way irreversible mechanism. Automatic downgrading is not allowed during the period when the upgraded level label is in effect. The protection level can only be reset after manual confirmation that the fault has been eliminated. When ecological restoration equipment is operating in the field environment, minor anomalies may recur in a short period of time. The protection level upgrade driven by the cumulative anomaly count can prevent minor faults from developing into major accidents.

[0058] The upgraded protection level identifiers are used to generate protection control parameters through a level-based aggregation process. The protection parameter templates corresponding to the upgraded level identifiers are retrieved from the protection parameter library. Level 1 protection corresponds to the most stringent protection parameter template, with the shortest action delay, lowest current limiting ratio, and ultra-fast tripping method. Level 2 protection corresponds to a moderately stringent template, and Level 3 protection corresponds to a standard template. If the upgraded level identifier includes an emergency handling flag, the threshold parameters are further tightened based on the Level 1 protection parameter template, with the current threshold tightened to 1.1 times the rated current and the tripping delay shortened to 5ms. Based on the fault type distribution recorded in the anomaly level mapping table, the fault type-related parameters in the protection parameter templates are adjusted accordingly. If historical anomalies are predominantly overcurrent faults, the current protection parameters are tightened; if undervoltage faults are predominant, the voltage protection parameters are tightened. The adjusted protection parameter templates are instantiated into the final protection control parameters, which include fields such as current protection threshold, voltage protection threshold, temperature protection threshold, action delay time, tripping speed level, reclosing permission flag, and current limiting ratio. Once the protection control parameters take effect, they override the original protection configuration. The protection switch then performs subsequent fault diagnosis and protection actions based on the new protection control parameters. The peristaltic pump motor in the dosing device has relatively low power but frequent start-stop cycles. For this type of equipment, the protection control parameters need to have a more relaxed start-up impact tolerance to avoid frequent malfunctions.

[0059] Fault over-limit values ​​are generated based on threshold comparisons performed according to protection control parameters. Protection criteria such as current threshold, voltage threshold, and temperature threshold are extracted from the protection control parameters. The current threshold is typically set to 1.2 to 1.5 times the rated current, the voltage threshold to 85% to 110% of the rated voltage, and the temperature threshold to 90% of the equipment's maximum allowable temperature. Real-time operating parameters are obtained, and the real-time current value is compared with the current threshold in the protection control parameters. The current over-limit value is calculated as (I_real - I_threshold) / I_rated × 100%, where a positive value indicates overcurrent over-limit and a negative value indicates undercurrent over-limit. Similarly, voltage and temperature over-limit values ​​are calculated. The sign of each parameter's over-limit value indicates the direction of the over-limit, and the absolute value indicates the degree of over-limit. The over-limit values ​​of each parameter are weighted and summed according to the weighting coefficients specified in the protection control parameters to obtain the comprehensive fault over-limit value. The formula for calculating the comprehensive fault over-limit value is Q_fault=Σ(w_i×|ΔP_i|), where Q_fault is the fault over-limit value, w_i is the weighting coefficient of the i-th parameter, and ΔP_i is the over-limit value of the i-th parameter. The fault over-limit value is a dimensionless normalized value, ranging from 0 to infinity. A fault over-limit value of 0 indicates that all parameters are within the threshold range specified by the protection control parameters, and a larger fault over-limit value indicates a more severe fault. When the current of a water circulating pump slowly increases due to foreign objects entangled in the impeller, the fault over-limit value will gradually increase from 0. By continuously monitoring the changing trend of the fault over-limit value, the development process of the fault can be predicted.

[0060] In some embodiments, the step of generating a tripping signal based on the fault exceeding the limit response control includes: extracting the deviation direction and deviation amplitude characteristics of the fault exceeding the limit; classifying the fault level according to the deviation direction and deviation amplitude characteristics to generate a fault level identifier; generating differentiated tripping parameters based on the tripping speed adapted to the fault level identifier; and executing the tripping action according to the differentiated tripping parameters to generate a tripping signal.

[0061] Extract the deviation direction and magnitude characteristics of fault over-limit quantities. Decompose and analyze the over-limit state of each component parameter in the fault over-limit quantity, extracting the sign of the current over-limit quantity as the current deviation direction; positive values ​​indicate overcurrent direction, and negative values ​​indicate undercurrent direction. Similarly, extract the deviation directions of voltage and temperature over-limit quantities, encoding the deviation direction of each parameter into a direction vector. The components of the direction vector take values ​​of +1, 0, or -1, corresponding to positive over-limit, no over-limit, and negative over-limit, respectively. Extract the absolute value of each parameter's over-limit quantity as the deviation magnitude characteristic. The deviation magnitude characteristic represents the relative degree of deviation of each parameter from the threshold, and its numerical range corresponds to each component of the fault over-limit quantity. Calculate the maximum value in the deviation magnitude characteristic and its corresponding parameter type. The parameter corresponding to the maximum deviation magnitude is the dominant fault parameter, which determines the focus of subsequent protection actions. Integrate the deviation direction vector and the deviation magnitude characteristic vector to form a feature description of the fault over-limit quantity. The feature description also includes the dominant fault parameter identifier and the comprehensive deviation magnitude (i.e., the fault over-limit quantity itself). Overcurrent faults and undervoltage faults deviate in opposite directions. Overcurrent faults correspond to positive current over-limits, while undervoltage faults correspond to negative voltage over-limits. By observing the direction of the deviation, the fault type can be quickly distinguished and a targeted protection strategy can be selected.

[0062] Fault level classification is performed based on the deviation direction and amplitude characteristics to generate fault level labels. A fault level classification matrix is ​​established, where the row index represents the combination of deviation directions, the column index represents the deviation amplitude range, and the matrix elements are the corresponding fault level labels. The deviation amplitude range is divided into four levels: 0 to 1.0 corresponds to a minor over-limit and a level 4 fault level label; 1.0 to 2.0 corresponds to a moderate over-limit and a level 3 fault level label; 2.0 to 5.0 corresponds to a severe over-limit and a level 2 fault level label; and above 5.0 corresponds to an extremely severe over-limit and a level 1 fault level label. The deviation direction modifies the fault level label. If the fault type corresponding to the deviation direction is a rapidly developing type (such as a short circuit fault, with current exceeding the limit in the positive direction and the amplitude increasing rapidly), the fault level label is increased by one level; if the fault type corresponding to the deviation direction is a slowly developing type (such as a minor overload), the fault level label remains unchanged. The distribution of each parameter in the deviation amplitude characteristics is used to determine whether there is a compound fault. When the deviation amplitude of two or more parameters exceeds 1.0, it is determined to be a compound fault, and the fault level label of the compound fault is automatically increased by one level. The determined fault level code, along with the deviation direction, deviation amplitude characteristics, and composite fault flag, are integrated to form a fault level identifier, which is then transmitted to the tripping parameter adaptation module. During a single-phase ground fault, the fault phase current increases while the voltage decreases, and the deviation direction exhibits a combination of positive current over-limit and negative voltage over-limit characteristics, allowing for accurate identification of the ground fault type.

[0063] Differentiated tripping parameters are generated based on the fault level identifier to adapt the tripping speed. A correspondence is established between fault level identifiers and tripping speeds: Level 1 fault level identifiers correspond to ultra-fast tripping mode with a target tripping time of less than 10ms; Level 2 fault level identifiers correspond to fast tripping mode with a target tripping time of 15 to 25ms; Level 3 fault level identifiers correspond to normal tripping mode with a target tripping time of 30 to 50ms; and Level 4 fault level identifiers correspond to delayed tripping mode with a target tripping time of 50 to 100ms. The tripping drive parameters are adjusted according to the fault type information in the fault level identifier. Short-circuit faults require maximum tripping force to quickly open the arc, while overload faults can use a smaller tripping force to reduce mechanical impact. The drive current of the electromagnetic trip unit is adjusted according to the required tripping force. The differentiated tripping parameters include four fields: tripping speed level, drive current amplitude, drive pulse width, and expected tripping time. The drive current amplitude ranges from 80% to 120% of the rated drive current, and the pulse width ranges from 10 to 100ms. If the fault level identifier includes a composite fault flag, the differentiated tripping parameters are configured with the highest tripping speed level to ensure timely disconnection of composite faults. The adapted differentiated tripping parameters are output to the tripping execution module and simultaneously recorded in the protection action log. High-speed tripping can complete contact separation before the short-circuit current reaches its peak value, which can be more than 20 times the rated current. Rapid tripping is of great significance to protection equipment and power supply systems.

[0064] The tripping action is executed according to the differentiated tripping parameters, generating a tripping completion signal. The output parameters of the trip unit drive circuit are configured based on the drive current amplitude and drive pulse width in the differentiated tripping parameters. The drive circuit uses a power MOSFET or IGBT to control the coil current. A drive pulse is applied to the trip unit coil; the rising edge of the drive pulse triggers the trip unit to operate. The armature of the electromagnetic trip unit overcomes the spring preload under electromagnetic force and begins to move, pushing the contact separation mechanism to perform the tripping action. The state changes during the contact separation process are monitored. The contact position sensor provides real-time feedback on the contact gap, and the auxiliary contacts switch states during contact separation. When the contact gap reaches the rated opening distance and the auxiliary contact state has switched, the tripping action is considered complete, and a tripping completion signal is generated to confirm successful tripping. The tripping completion signal includes four fields: tripping completion time, actual tripping time, final contact position, and auxiliary contact state. The actual tripping time is compared with the expected tripping time in the differentiated tripping parameters. If the actual opening time exceeds 150% of the expected opening time, it is considered a delay in opening. This delay may be caused by mechanical jamming or a deterioration in the performance of the trip unit, and needs to be recorded and a maintenance reminder triggered. If no opening signal is detected within the maximum permissible opening time (usually 200ms), it is considered a failure to open. In the event of a failure, a trip request must be immediately sent to the upstream circuit breaker to implement backup protection. An electric arc is generated during contact separation. The arc is elongated and cooled in the arc-extinguishing chamber until it is extinguished. The opening signal should be confirmed only after the arc is completely extinguished to ensure complete opening.

[0065] Step S150: The blocking action is triggered by the tripping signal to generate a blocking feedback state. An alarm completion flag is generated based on the blocking feedback state and the alarm reporting sequence. A protection control command is generated based on the alarm completion flag and the tripping signal.

[0066] Specifically, a blocking action is triggered by the tripping signal to generate a blocking feedback state. The tripping signal carries the execution result information of the tripping action, and the blocking action is used to lock the tripping state to prevent accidental closing. The validity of the tripping signal is checked, and it needs to meet the dual confirmation conditions of contact position and auxiliary contact state. After the tripping signal is validly confirmed, the blocking action sequence is initiated. The blocking action includes two levels: electrical blocking and mechanical blocking. Electrical blocking is achieved by cutting off the power supply to the closing circuit, and mechanical blocking is achieved by locking the release rod of the closing mechanism. The electrical blocking action sets the enable signal of the closing control circuit to an invalid state, with a response time of less than 5ms. The mechanical blocking action drives the blocking electromagnet to insert the locking pin into the locking hole of the closing release rod, with a response time of approximately 20 to 50ms. When both electrical and mechanical blocking are confirmed to be complete, a blocking feedback state is generated, which includes three fields: blocking time, blocking type, and blocking confirmation status. If the fault type is a severe fault such as a short circuit or grounding, a forced interlocking flag is added to the interlocking feedback status. The forced interlocking status can only be released through manual operation on-site. When ecological restoration equipment is installed in unattended outdoor locations, the interlocking feedback status prevents the automatic reclosing function from repeatedly attempting to close the circuit breaker while the fault is still present, thus preventing further damage to the equipment.

[0067] An alarm completion flag is generated based on the alarm reporting sequence triggered by the interlocking feedback state. The interlocking feedback state confirms that the device has entered the protection interlocking state, and the alarm reporting sequence sends the protection event information to the monitoring center. The alarm level is determined according to the severity of the interlocking feedback state: forced interlocking corresponds to a Level 1 alarm, and normal interlocking corresponds to a Level 2 alarm. An alarm data message is assembled, including the device identifier, interlocking feedback state details, a snapshot of fault characteristic parameters, and a protection action timeline. The alarm message is sent to the monitoring center via a communication module, supporting multiple communication methods such as 4G / 5G wireless networks, Ethernet, and LoRa. An acknowledgment and retransmission mechanism is used for alarm reporting. After sending, an acknowledgment waiting timer is started, with a waiting time set to 5 seconds. If an alarm acknowledgment response is received from the monitoring center, the alarm reporting is considered successful, and an alarm completion flag is generated. If the waiting timeout occurs, an alarm retransmission is triggered, with a maximum of 3 retransmissions. If all 3 retransmissions time out, the alarm information is cached in local storage and retransmitted after communication is restored. The alarm completion flag includes four fields: alarm sequence number, reporting time, acknowledgment time, and alarm level. A faulty aeration pump tripping can affect the aeration and oxygen supply function of the water. Reporting an alarm can enable maintenance personnel to be informed of the fault in a timely manner and arrange for emergency repairs.

[0068] The protection control command is generated based on the alarm completion flag and the tripping signal. The alarm completion flag confirms that the fault information has been successfully reported, and the tripping signal confirms that the protection action has been successfully executed; both trigger the generation of the protection control command. The timing consistency of the alarm completion flag and the tripping signal is checked; the tripping signal should be earlier than the alarm completion flag. The protection control command is generated by combining the information from the alarm completion flag and the tripping signal. The type of protection control command is determined based on the fault severity and the blocking status. If the blocking feedback status is forced blocking, the protection control command is a command to prohibit remote closing; if it is normal blocking and the alarm completion flag indicates that the alarm has been successfully delivered, the protection control command is a command to wait for remote authorization to close; if it is preventative blocking, the protection control command is a command to allow timed automatic recovery. The protection control command also includes parameters for subsequent monitoring requirements. Typically, an encrypted monitoring mode is used in the first operating cycle after recovery to promptly detect fault recurrence. The protection control command is sent to the main control module via the internal bus and simultaneously written to the protection action record memory to form a complete protection event file. The generation of protection and control commands marks the completion of the closed loop of this protection and control process, and the equipment enters a state of waiting for recovery or manual intervention.

[0069] To implement the above-described method embodiments, a smart protection switch control method for ecological restoration equipment is provided to achieve the corresponding functions and technical effects. See also... Figure 3 , Figure 3 This diagram illustrates a structural block diagram of an intelligent protection switch control system 300 for ecological restoration equipment according to an embodiment of this application. For ease of explanation, only the parts relevant to this embodiment are shown. The intelligent protection switch control system 300 for ecological restoration equipment provided in this embodiment includes: Data acquisition module 301 is used to collect load current data and voltage fluctuation data of ecological restoration equipment, and construct an operating parameter benchmark table based on the load current data and voltage fluctuation data; The monitoring and scheduling module 302 is used to divide the operating parameter benchmark table into threshold intervals to generate a monitoring cycle configuration, determine the current monitoring cycle based on the monitoring cycle configuration, and generate a status judgment instruction according to the deviation between the current monitoring cycle and the rated parameters of the equipment. Risk assessment module 303 is used to perform multi-channel synchronous detection to obtain fault characteristic parameters according to the state judgment instruction, perform start-up surge characteristic identification on the load current data to generate start-up current compensation amount, determine risk assessment coefficient by edge computing according to the fault characteristic parameters and the start-up current compensation amount, and generate an abnormal triggering flag according to the risk assessment coefficient. Protection execution module 304 is used to determine the protection level according to the abnormal trigger flag and generate protection control parameters, perform threshold comparison according to the protection control parameters to generate fault over-limit quantity, and perform circuit breaker response control according to the fault over-limit quantity to generate a tripping signal. The instruction generation module 305 is used to trigger a blocking action via the tripping signal to generate a blocking feedback state, trigger an alarm reporting sequence based on the blocking feedback state to generate an alarm completion flag, and generate a protection control instruction based on the alarm completion flag and the tripping signal.

[0070] The aforementioned intelligent protection switch control system 300 for ecological restoration equipment can implement an intelligent protection switch control method for ecological restoration equipment as described in the above method embodiments. The options in the above method embodiments are also applicable to this embodiment, and will not be detailed here. The remaining content of this application embodiment can be referred to the content of the above method embodiments, and will not be repeated in this embodiment.

[0071] The above embodiments are not an exhaustive list based on the present invention, and there may be many other embodiments not listed. Any substitutions and improvements made without departing from the concept of the present invention are within the protection scope of the present invention.

Claims

1. A method for controlling an intelligent protection switch in ecological restoration equipment, characterized in that, include: Collect load current data and voltage fluctuation data of the ecological restoration equipment, and construct an operating parameter benchmark table based on the load current data and voltage fluctuation data; The monitoring cycle configuration is generated by dividing the threshold interval based on the operating parameter benchmark table, the current monitoring cycle is determined based on the monitoring cycle configuration, and a status judgment instruction is generated according to the deviation between the current monitoring cycle and the rated parameters of the equipment. Based on the state determination instruction, multi-channel synchronous detection is performed to obtain fault characteristic parameters. Start-up surge characteristics are identified for the load current data to generate a start-up current compensation amount. Based on the fault characteristic parameters and the start-up current compensation amount, a risk assessment coefficient is determined by edge computing. An abnormal triggering flag is generated based on the risk assessment coefficient. Based on the abnormal trigger flag, the protection level is determined and protection control parameters are generated. Based on the protection control parameters, threshold comparison is performed to generate the fault limit exceedance. Based on the fault limit exceedance, circuit breaker response control is performed to generate the tripping signal. The blocking action is triggered by the tripping signal to generate a blocking feedback state. An alarm reporting sequence is triggered based on the blocking feedback state to generate an alarm completion flag. A protection control command is generated based on the alarm completion flag and the tripping signal.

2. The method according to claim 1, characterized in that, The step of constructing an operating parameter benchmark table based on the load current data and the voltage fluctuation data includes: Based on the load current data, current amplitude features are extracted to construct a current feature set; The current feature set is correlated with the voltage fluctuation data to construct a current-voltage mapping relationship; Based on ambient temperature data, temperature correlation characteristics are analyzed within the current-voltage mapping relationship to generate a temperature compensation coefficient. The current-voltage mapping relationship is corrected based on the temperature compensation coefficient to generate a reference table of operating parameters.

3. The method according to claim 1, characterized in that, The step of dividing the operating parameter benchmark table into threshold intervals to generate the monitoring cycle configuration includes: The operating parameter benchmark table is divided into a normal parameter segment and an early warning parameter segment; Periodic feature detection is performed on the aforementioned operating parameter benchmark table to generate a load operating cycle; Based on the load operation cycle, the normal parameter segment and the early warning parameter segment are adaptively adjusted to generate dynamic threshold boundaries; The optimal division position is located within the dynamic threshold boundary to generate the monitoring cycle configuration.

4. The method according to claim 1, characterized in that, The step of generating a startup current compensation amount by identifying startup surge characteristics based on the load current data includes: Peak envelope extraction is performed on the load current data to generate current peak boundaries; Surge-dominant regions are identified by performing a rate analysis based on the current peak boundary. The duration of the surge is evaluated to generate the surge duration. The starting current compensation amount is determined by compensation mapping based on the surge duration.

5. The method according to claim 1, characterized in that, The step of determining the risk assessment coefficient through edge computing based on the fault characteristic parameters and the starting current compensation includes: Construct a parameter timing record based on the aforementioned fault characteristic parameters; Multiple types of parameter extreme points are extracted from the parameter time series records to construct multi-channel extreme value calibration points; Cross-validation is performed on the multi-channel extreme value calibration points to identify synchronous abnormal parameters and generate cross-validation results. The risk assessment coefficient is determined by edge computing based on the cross-validation results and the starting current compensation amount.

6. The method according to claim 1, characterized in that, The step of determining the protection level and generating protection control parameters based on the abnormal trigger flag includes: Convert the aforementioned exception trigger flag into an exception level mapping table; An anomaly cumulative count is generated based on the number of consecutive anomalies detected using the anomaly level mapping table. Based on the accumulated anomaly count, a protection level upgrade determination is made to generate an upgraded level identifier; The upgraded level identifier is used to generate protection control parameters through level aggregation.

7. The method according to claim 1, characterized in that, The step of generating a tripping signal based on the fault exceeding the limit response control includes: Extract the deviation direction and deviation amplitude characteristics of the fault exceeding the limit; Fault level identifiers are generated by classifying fault levels based on the deviation direction and the deviation amplitude characteristics. Based on the fault level identifier, differentiated tripping parameters are generated to adapt the tripping speed. The tripping action is executed according to the differentiated tripping parameters to generate a tripping completion signal.

8. The method according to claim 3, characterized in that, The step of adaptively adjusting the boundaries of the normal parameter segment and the early warning parameter segment based on the load operation cycle to generate dynamic threshold boundaries includes: Based on the load operation cycle, the normal parameter segment and the early warning parameter segment are identified to determine the detection range by identifying parameter mutation characteristics. A parameter change graph is constructed by tracking parameter changes within the detection range. Based on the parameter change map, gradual deterioration features and abrupt anomalies are identified and labeled to generate change classification labels; A dynamic threshold boundary is generated by combining the change classification label with the parameter change map.

9. The method according to claim 5, characterized in that, The step of cross-validating the multi-channel extreme value calibration points to identify synchronous abnormal parameters and generate cross-validation results includes: The fluctuation amplitude is detected and a fluctuation amplitude distribution is generated by performing fluctuation amplitude detection on the multi-channel extreme value calibration points; Identify locations of abnormal spikes in the fluctuation amplitude distribution to construct an abnormal location set; An anomaly severity assessment is performed on the aforementioned set of abnormal locations to generate a corrected risk level. Cross-validation results are generated based on the corrected risk amount and the multi-channel extreme value calibration points.

10. An intelligent protection switch control system for ecological restoration equipment, characterized in that, include: The data acquisition module is used to collect load current data and voltage fluctuation data of the ecological restoration equipment, and to construct an operating parameter benchmark table based on the load current data and voltage fluctuation data. The monitoring and scheduling module is used to divide the operating parameter benchmark table into threshold intervals to generate a monitoring cycle configuration, determine the current monitoring cycle based on the monitoring cycle configuration, and generate a status judgment instruction according to the deviation between the current monitoring cycle and the rated parameters of the equipment. The risk assessment module is used to perform multi-channel synchronous detection to obtain fault characteristic parameters based on the state determination instruction, perform start-up surge characteristic identification on the load current data to generate start-up current compensation amount, determine the risk assessment coefficient through edge computing based on the fault characteristic parameters and the start-up current compensation amount, and generate an abnormal triggering flag based on the risk assessment coefficient. The protection execution module is used to determine the protection level according to the abnormal trigger flag, generate protection control parameters, perform threshold comparison according to the protection control parameters to generate fault over-limit quantity, and perform circuit breaker response control according to the fault over-limit quantity to generate a tripping signal. The instruction generation module is used to trigger a blocking action via the tripping signal to generate a blocking feedback state, trigger an alarm reporting sequence based on the blocking feedback state to generate an alarm completion flag, and generate a protection control instruction based on the alarm completion flag and the tripping signal.

Images