A remote state monitoring and fault diagnosis method based on LED explosion-proof lamp

Abstract

The application provides a remote state monitoring and fault diagnosis method based on an LED explosion-proof lamp and belongs to the technical field of state monitoring of lighting equipment. The application quantifies health state information contained in a thermal dynamic process sufficiently by extracting a temperature rise rate and a cooling time constant and fusing the same into a thermal characteristic response value, and can capture light decay and degradation of a heat dissipation path earlier than a scheme relying on electrical parameters or a steady-state temperature. Meanwhile, an adaptive baseline is established based on a historical sequence of the lamp and hierarchical diagnosis is performed, so that the defects of false alarms of a fixed threshold and failure of group comparison when lamps are rare are overcome.

Description

Technical Field

[0001] This invention belongs to the field of lighting equipment condition monitoring technology, specifically relating to a remote condition monitoring and fault diagnosis method based on LED explosion-proof lights. Background Technology

[0002] As core lighting equipment in flammable and explosive environments such as petroleum, chemical, and mining facilities, the operational reliability of LED explosion-proof lights directly affects production safety and the safety of personnel and property. Continuous monitoring and fault diagnosis of the thermal status of LED explosion-proof lights are crucial for ensuring safe operation, extending service life, and reducing maintenance costs. In recent years, the industry has conducted numerous studies on lighting condition monitoring and fault diagnosis, with the main technical approaches categorized into three types: first, monitoring methods based on electrical parameters (voltage, current, power factor, etc.), which determine whether the working status of the lights is abnormal by collecting electrical operating data; second, alarm methods based on temperature thresholds, which install temperature sensors inside the lights or in the housing, triggering an alarm when the temperature exceeds a preset threshold; and third, diagnostic methods based on group comparison, which compare multiple lights in the same area to identify individuals exhibiting abnormal behavior relative to other lights in the same group. While each of these technical approaches has played a role in its respective application scenario, each also has its own inherent limitations.

[0003] For example, CN121547924A discloses a data processing system and method for monitoring the status and diagnosing faults of lighting fixture groups. This system replaces fixed thresholds with grouped dynamic benchmarks, which to some extent solves the problem of misjudgment caused by inherent differences among different lighting fixtures. However, the core diagnostic basis of this technical solution is electrical parameters (voltage, current, power, etc.). These parameters reflect the electrical behavior of the lighting fixtures rather than their physical health status, which may create diagnostic blind spots. This solution relies on grouping lighting fixtures with similar operating characteristics into the same subgroup, requiring a sufficient number of similar lighting fixtures in the area as a reference. In scenarios where the number of lighting fixtures is small or the operating conditions of each fixture differ significantly, the rationality of the group division is difficult to guarantee, and the reliability of the benchmark value decreases accordingly. Furthermore, the solution can only identify lighting fixtures that exhibit abnormal behavior relative to other lighting fixtures in the same group, and cannot further locate the root cause or section of the abnormality. It lacks effective means of identifying and locating multi-lamp synchronous anomalies caused by grid-side factors.

[0004] CN118482359A discloses an explosion-proof lighting fixture with lifespan monitoring based on full-area temperature and humidity detection, and a lifespan monitoring method. It compensates for the shortcomings of single-point temperature measurement by acquiring temperature data across the entire area and correlates the temperature data with the degradation characteristics of LEDs to predict remaining lifespan, thus improving the accuracy of lifespan prediction for explosion-proof lighting fixtures to some extent. However, this technical solution also has significant shortcomings. For example, the temperature monitoring object is the temperature of the internal space of the lighting fixture, rather than the difference between the housing temperature and the ambient temperature, failing to utilize the thermal resistance relationship between the housing temperature and the junction temperature to characterize the overall thermal path state of the lighting fixture. This solution only focuses on the steady-state temperature of the lighting fixture under operating conditions, without considering the cooling process after power failure. The dynamic response of the cooling process is precisely a key physical quantity characterizing the thermal resistance-thermal capacity characteristics of the lighting fixture, which can more sensitively reflect the aging and degradation of the heat dissipation path. Furthermore, this solution uses absolute temperature values ​​or temperature-humidity combinations as the basis for lifespan prediction, essentially still belonging to a threshold-based judgment logic, lacking an adaptive baseline based on the lighting fixture's own historical state, and making it difficult to exclude the interference of normal factors such as seasonal changes in ambient temperature on the diagnostic results. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this section, the abstract and title of the invention. Such simplifications or omissions shall not be used to limit the scope of the present invention.

[0006] In view of the aforementioned existing problems, the present invention is proposed.

[0007] Therefore, the technical problem solved by this invention is that the diagnostic granularity is limited to the single lamp level, and it fails to aggregate the scattered lamp monitoring data into the fault perception capability of the power grid section, thus failing to achieve the functional leap from single lamp anomaly identification to power grid section location.

[0008] To address the aforementioned technical problems, the present invention provides the following technical solution: A remote status monitoring and fault diagnosis method based on LED explosion-proof lights includes: collecting the housing temperature and ambient temperature of the explosion-proof light, and recording the power-on and power-off times of the light fixture; obtaining the heating rate during the heating phase and the cooling time constant during the cooling phase based on the power-on and power-off times and the collected temperature data; generating and storing the thermal characteristic response value for the current cycle based on the heating rate and the cooling time constant; calculating a baseline value based on the stored historical thermal characteristic response values, and outputting the health status of the light fixture based on the deviation of the current thermal characteristic response value from the baseline value; and locating abnormal sections based on the health status of multiple light fixtures and the power grid topology, and outputting the location results.

[0009] As a preferred embodiment of the present invention, before collecting the housing temperature and ambient temperature of the explosion-proof lamp, the method further includes: using an independent backup power supply to power the controller; using a voltage detection circuit to monitor the voltage amplitude at the lamp's power supply terminal in real time; when the voltage amplitude jumps from a first preset percentage below the rated voltage to a second preset percentage above the rated voltage, marking the current moment as the power-on start moment; when the voltage amplitude drops from the second preset percentage above the rated voltage to the first preset percentage below the rated voltage, marking the current moment as the power-off moment.

[0010] As a preferred embodiment of the present invention, before obtaining the heating rate of the heating stage and the cooling time constant of the cooling stage, the method further includes: calculating the difference between the shell temperature and the ambient temperature as the initial temperature difference at the time of power failure; collecting shell temperature data and ambient temperature data during the cooling stage after power failure, and adopting a dual-condition termination strategy during the collection process, stopping recording when any of the following conditions are met: the collection time reaches the preset maximum safe time; the difference between the shell temperature and the ambient temperature calculated in real time decays to below a first preset percentage of the initial temperature difference.

[0011] As a preferred embodiment of the present invention, before obtaining the heating rate of the heating stage and the cooling time constant of the cooling stage, the method further includes an ambient temperature validity check: determining whether the ambient temperature is within a preset physical reasonable range; determining whether the initial temperature difference is greater than a preset minimum effective temperature difference threshold; if any check fails, the current diagnostic process is terminated and the ambient temperature reference failure information is reported.

[0012] In a preferred embodiment of the present invention, obtaining the cooling time constant of the cooling stage includes: counting the number of effective sampling points in the cooling stage; if the number of effective sampling points is less than a preset threshold, terminating the current diagnostic process; when the number of effective sampling points meets the requirements, using the initial temperature difference as a fixed benchmark, performing nonlinear fitting on the real-time temperature difference sequence of the cooling stage according to a first-order exponential decay model to obtain the cooling time constant; if the cooling process is terminated prematurely and the real-time temperature difference sequence has not yet decayed to a first preset percentage of the initial temperature difference, terminating the current diagnostic process.

[0013] As a preferred embodiment of the present invention, obtaining the heating rate of the heating stage includes: extracting an effective heating segment from the heating stage data, linearly fitting the temperature data within the effective heating segment relative to the time data, and using the fitted slope as the heating rate; generating and storing the thermal characteristic response value of the current period based on the heating rate and the cooling time constant; multiplying the heating rate by the cooling time constant, and using the product as the thermal characteristic response value of the current period; validating the thermal characteristic response value; the validity verification includes: the thermal characteristic response value is positive, the heating rate is greater than a preset minimum heating slope, and the cooling time constant is within a preset reasonable time interval; if the absolute deviation of the thermal characteristic response value from the median of the historical sequence exceeds a preset outlier multiple, it is determined to be an outlier, the current data is discarded and the current diagnostic process ends, and subsequent baseline calculation and health status output are not performed; if it is not an outlier, the current thermal characteristic response value is written into the historical sequence.

[0014] As a preferred embodiment of the present invention, when storing the thermal characteristic response value into the historical sequence, the first-in-first-out principle is used to maintain the sequence length, and the sequence length is limited to a preset maximum storage quantity. If the quantity is exceeded, the earliest data point is deleted.

[0015] As a preferred embodiment of the present invention, the calculation of the baseline value based on the stored historical thermal characteristic response values ​​includes: reading the current total amount of data in the historical sequence; if the total amount of data is less than or equal to a preset minimum sample number, the system is determined to be in a data accumulation period, and deviation calculation is not performed, and the status information in the modeling process is directly output; if the total amount of data is greater than the preset minimum sample number, the baseline value is calculated; it is determined whether the thermal characteristic response value of the current period has been written into the historical sequence; if it has been written, the last data point written most recently is temporarily excluded from the historical sequence, and the most recent N data points are read backward; if the second to last data point points to the current value, it is skipped, ensuring that the data read does not contain the thermal characteristic response value of the current period; if it has not been written, the most recent preset number of historical thermal characteristic response values ​​are directly read from the historical sequence; the arithmetic mean of the read historical thermal characteristic response values ​​is calculated as the baseline value; if the baseline value is zero, the deviation is forced to be zero and a zero baseline alarm is triggered.

[0016] In a preferred embodiment of the present invention, the method of outputting the health status of the luminaire based on the degree of deviation between the current thermal characteristic response value and the baseline value includes: when the baseline value is not zero, calculating the absolute value of the difference between the current thermal characteristic response value and the baseline value, dividing the absolute value by the baseline value, and using the result as the deviation degree; and outputting the corresponding luminaire health status level according to a preset grading interval based on the deviation degree: when the deviation degree is less than or equal to a first preset threshold, outputting a health level; when the deviation degree is greater than the first preset threshold and less than or equal to a second preset threshold, outputting a warning level; when the deviation degree is greater than the second preset threshold and less than or equal to a third preset threshold, outputting an abnormal level; and when the deviation degree is greater than the third preset threshold, outputting a severe fault level.

[0017] As a preferred embodiment of the present invention, the following steps are taken: based on the health status of multiple lamps, and combined with the power grid topology, to locate abnormal sections and output the location results, the following steps are taken: collecting the deviation of each lamp on the same feeder, marking lamps with deviations greater than a preset abnormal threshold as abnormal nodes, and triggering a regional power grid abnormal event when the number of abnormal nodes exceeds a preset number; inferring the power grid abnormality type based on the direction of change of the thermal characteristic response value of each abnormal node relative to its respective baseline value: when the thermal characteristic response value of an abnormal node increases by more than a preset proportion, it is inferred to be a voltage high anomaly; when the thermal characteristic response value decreases by more than a preset proportion, it is inferred to be a voltage low anomaly; tracing the common upstream node of all abnormal nodes along the power supply path according to the power grid topology, locking the section between the common upstream node and the first lamp in the downstream direction that has not been marked as an abnormal node as a suspected abnormal section; and reporting the inferred abnormality type and the location of the suspected abnormal section to the power grid dispatch center or operation and maintenance master station system.

[0018] The beneficial effects of this invention are as follows: Compared with the prior art, the technical effects of this invention are as follows: By extracting the heating rate and cooling time constant and fusing them into thermal characteristic response values, this invention fully quantifies the health status information contained in the thermal dynamic process, which can capture light decay and heat dissipation path degradation earlier than schemes that rely on electrical parameters or steady-state temperature; at the same time, by establishing an adaptive baseline and hierarchical diagnosis based on the lamp's own historical sequence, it overcomes the defects of fixed thresholds being prone to false alarms and group comparison failing when there are few lamps; furthermore, it integrates single-lamp diagnosis into multi-lamp collaborative analysis, combines topology structure to locate abnormal sections, and transforms lamps into distributed sensing nodes of the power grid, providing section-level fault location information for the distribution network without the need for additional hardware. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the power-on and power-off detection process in an embodiment of the present invention.

[0020] Figure 2This is a schematic diagram of the process for determining thermal balance and selecting the effective heating segment in an embodiment of the present invention.

[0021] Figure 3 A schematic diagram of the structure of an electronic device for implementing the remote status monitoring and fault diagnosis method based on LED explosion-proof lights according to embodiments of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0023] like Figures 1-2 As shown, the remote status monitoring and fault diagnosis method based on LED explosion-proof lights of the present invention includes: S1: Collect the housing temperature and ambient temperature of the explosion-proof lamp, and record the power-on and power-off times of the lamp.

[0024] S1.1: Collect the housing temperature and ambient temperature of the explosion-proof lamp.

[0025] S1.1.1: A first temperature sensor is installed on the outer surface of the LED explosion-proof lamp housing to collect housing temperature data. Preferably, the first temperature sensor is an NTC negative temperature coefficient thermistor. The NTC thermistor is tightly attached to the root area of ​​the heat sink fins on the outer surface of the LED explosion-proof lamp housing using thermally conductive silicone grease, and is fixed with a clamping device or thermally conductive adhesive to ensure good thermal contact between the thermistor and the housing surface, thus accurately reflecting the temperature change of the LED explosion-proof lamp housing during operation. The housing temperature can effectively characterize the heat accumulation state of the LED light source and driving circuit—when the LED chip or driving power supply heats up, heat is transferred to the housing surface through the heat conduction path. There is a definite thermal resistance correlation between the housing temperature and the junction temperature; therefore, the housing temperature can serve as a reliable characterization parameter for the thermal state of the lamp.

[0026] In explosion-proof scenarios, the signal cable of the first temperature sensor must be equipped with an explosion-proof cable entry device to ensure that the explosion-proof performance of the sensor signal cable meets the requirements when passing through the housing.

[0027] Furthermore, a second temperature sensor is installed at a shaded location at least 10 cm away from the outer surface of the LED explosion-proof light and not directly affected by the heat radiation of the light fixture, to collect ambient temperature data. Among them, "not being directly affected by the heat radiation of the lamp" means that there is a physical shield between the location of the second temperature sensor and the lamp housing, such as the back of a sunshade installed above the lamp or the shaded area of ​​an adjacent structural component, and the ambient temperature at this location is not affected by the convective heat field caused by the heat generated by the lamp.

[0028] The second temperature sensor is also preferably an NTC negative temperature coefficient thermistor, with electrical parameters matching those of the first temperature sensor, so that the controller can process the signal using a unified signal conditioning circuit. The ambient temperature data collected by the second temperature sensor serves as a reference for calculating the temperature difference during the cooling phase—during the cooling process after the lamp is powered off, the difference between the housing temperature and the ambient temperature decays exponentially according to Newton's law of cooling, and accurate acquisition of the ambient temperature is a prerequisite for extracting the cooling time constant.

[0029] The first and second temperature sensors are connected to the analog-to-digital converter (ADC) input pins of the controller. The controller has a built-in temperature calculation module based on the Steinhart-Hart equation, which converts the voltage values ​​acquired by the ADC into temperature values. The coefficients in the Steinhart-Hart equation are pre-written into the controller's non-volatile memory according to the datasheet parameters of the selected NTC thermistor.

[0030] S1.1.2: In this embodiment, the controller is powered by a backup power supply independent of the LED explosion-proof lamp power supply circuit. Specifically, the controller's power supply system adopts a dual-redundancy design. The main power supply path is taken from the LED explosion-proof lamp power supply terminal and provides working power to the controller after AC-DC conversion. The backup power supply uses a rechargeable lithium battery pack or a supercapacitor module, which is connected in parallel with the main power supply path through a power switching circuit.

[0031] When the LED explosion-proof light is powered on, the main power supply path powers the controller while simultaneously charging the backup power supply. When the LED explosion-proof light is powered off, the power switching circuit automatically (without delay) switches to the backup power supply, ensuring that the controller can continue to collect temperature data during the cooling phase after the light is powered off. The power switching circuit can use a MOSFET ideal diode controller (such as the LTC4412 or a similar device) to achieve seamless switching between main and backup power supplies, with a switching time of less than 1 millisecond, ensuring that the controller does not reset or lose data during power switching.

[0032] The controller is powered by an independent backup power supply, which means that the acquisition of temperature data during the cooling phase is not affected by the power supply status of the lamps. This solves the fundamental defect of traditional solutions where the controller loses power immediately after the lamps are powered off, making it impossible to collect data during the cooling phase. This provides a complete data foundation for the extraction of the cooling time constant.

[0033] S1.1.3: The first temperature sensor and the second temperature sensor respectively collect housing temperature data and ambient temperature data at a fixed sampling frequency, and transmit the collected data to the controller in real time. In this embodiment, the fixed sampling frequency is set to 1Hz, that is, 1 data point is collected per second.

[0034] The 1Hz sampling frequency is sufficient to capture the temperature change characteristics during the thermal process of the LED explosion-proof lamp; the constant sampling interval ensures the consistency of the time base for subsequent digital signal processing and avoids the calculation error introduced by the variable sampling frequency.

[0035] In the specific implementation, the controller internally sets a timer to generate a sampling trigger signal at a rate of 1Hz, driving the ADC (analog-to-digital converter) to simultaneously sample and convert the signals from two temperature sensors. The ADC's sampling resolution is preferably 12 bits or higher to ensure that the temperature measurement resolution is better than 0.1°C.

[0036] Upon power failure, the controller initiates continuous temperature data acquisition during the cooling phase. During acquisition, it calculates the real-time difference between the casing temperature and the ambient temperature and determines whether to stop recording based on pre-set acquisition termination logic. Before acquisition terminates, the controller continuously acquires and buffers temperature difference data at a frequency of 1Hz. When the acquisition termination condition is triggered, the controller submits the buffered complete temperature difference sequence to subsequent steps for processing.

[0037] S1.2: Record the power-on and power-off times of the lamps.

[0038] S1.2.1: The controller has a built-in voltage detection circuit that uses a resistor divider and ADC analog-to-digital conversion method to monitor the voltage amplitude at the lamp power supply terminal in real time. Specifically, the voltage detection circuit includes: a first voltage divider resistor connected in series between the lamp power supply terminal and the reference ground. Second voltage divider resistor And a voltage follower connected between the common node of the first and second voltage divider resistors and the controller ADC input pin. In one specific embodiment, if the lamp's rated voltage =220V, the controller ADC input range is 0~3.3V, then select the first voltage divider resistor. =200kΩ, second voltage divider resistor =3kΩ. At this point, under rated voltage, the voltage at the ADC input terminal is approximately 3.26V (220×(3 / (200+3))≈3.26V), leaving a safety margin.

[0039] The ADC samples and quantizes the voltage signal after voltage division, converting the analog voltage value into a digital value. The controller obtains the real-time voltage amplitude at the power supply end of the lamp by reading the ADC conversion result.

[0040] S1.2.2: The controller monitors the ADC sampling value of the voltage detection circuit in real time. When it detects that the power supply voltage amplitude is lower than the rated voltage... A 10% jump to above the rated voltage When the light fixture reaches 90% completion, it is determined that the power-on start-up has been completed, and the current time is marked as the power-on start time. Setting two thresholds, 10% and 90%, instead of a single threshold (such as 50%), avoids false triggering caused by voltage fluctuations or noise, thus creating a reliable hysteresis judgment range. When the voltage rises from zero but has not yet reached 90% of the rated value, it may be in a power grid transient process or soft start phase. At this time, the lamp has not yet entered a stable operating state and should not trigger the power-on marker; only when the voltage reliably rises above 90% of the rated value does it indicate that the lamp has been stably powered on.

[0041] S1.2.3: When the amplitude of the supply voltage is detected to be higher than the rated voltage... 90% of the voltage dropped below the rated voltage. When the power level reaches 10%, the light fixture is determined to be de-energized, and the current time is marked as the power outage time. Similarly, using a 90% to 10% falling edge hysteresis for judgment avoids misjudgment caused by instantaneous voltage drops (such as grid flicker).

[0042] S1.2.4: If no power outage is detected within a single monitoring cycle. If the LED explosion-proof light is in a long-term continuous operation state without power failure, the controller will terminate the subsequent diagnostic process of the current cycle, clear the temporary data, and wait for the next power failure event to trigger.

[0043] Furthermore, the controller will specify the power-on start time. Until the power outage The casing temperature data between these points is marked as heating phase data; the power outage time is... The continuously collected shell temperature data and corresponding ambient temperature data are then labeled as cooling stage data.

[0044] At the moment of power outage The controller immediately calculates the difference between the housing temperature and the ambient temperature, and records it as the initial temperature difference. .

[0045] The controller associates the collected temperature data with timestamp information and stores it in local memory (such as EEPROM or Flash memory) to form a temperature time series dataset with timestamps, which can be called in subsequent steps.

[0046] S2: Based on the power-on time, power-off time, and collected temperature data, obtain the heating rate during the heating phase and the cooling time constant during the cooling phase.

[0047] Before formally calculating the heating rate and extracting the cooling time constant, the controller first verifies the validity of the ambient temperature data. It's important to note that the extraction of the cooling time constant relies entirely on the exponential decay of the difference between the casing temperature and the ambient temperature. If the ambient temperature reference source fails, all subsequent calculations lose their physical meaning. Pre-verification avoids invalid calculations and outputs a clear fault indication. Specifically, the controller judges the validity of the collected ambient temperature data... Whether it is within a preset physically reasonable range. In a preferred embodiment, the physically reasonable range is set to -40°C to +85°C. This range covers the operating temperature range of industrial-grade electronic equipment, and also encompasses the temperature boundaries of most outdoor and industrial environments where LED explosion-proof lights may be deployed. If the temperature exceeds this range, the second temperature sensor is deemed to be faulty.

[0048] Furthermore, the controller determines the initial temperature difference. Whether it exceeds the preset minimum effective temperature difference threshold. In a preferred embodiment, the minimum effective temperature difference threshold is set to 0.5°C. When the difference between the shell temperature and the ambient temperature is less than 0.5°C, the amplitude of the temperature difference signal during the cooling stage is on the same order of magnitude as the quantization noise and measurement noise of the temperature sensor itself. The signal-to-noise ratio is too low, and the reliability of the cooling time constant fitting result cannot be guaranteed.

[0049] If the ambient temperature exceeds the reasonable range or the initial temperature difference is less than or equal to the minimum effective temperature difference threshold, the controller will immediately terminate the current diagnostic process, will not proceed to any subsequent calculation steps, and will report an ambient temperature reference failure through the communication interface. Please check the error code of the second sensor.

[0050] S2.1: The controller starts from the power-on time marked in step one. Start by calculating the casing temperature in real time. The absolute value of the rate of temperature change within a continuous sliding time window.

[0051] Specifically, the sliding time window width is set to 30 sampling points, corresponding to 30 seconds, and the sampling frequency is 1Hz.

[0052] The controller calculates the rate of temperature change within the sliding window in real time from the moment of power-on. However, to avoid misjudgment caused by rapid temperature changes during the initial transient phase (within the first 60 seconds), only the rate of change calculated after a power-on duration of 60 seconds or more is used for formal determination of the thermal equilibrium state. If the light fixture is de-energized within 60 seconds of power-on, the controller directly determines that the thermal equilibrium state has not been reached for that cycle and marks this cycle as a non-steady-state cycle. If the initial temperature difference meets the validity verification requirements in S2 at this time, the data from... to The entire heating data is used to fit the heating rate; otherwise, the diagnostic process is terminated.

[0053] The purpose of setting a 60-second delay is to allow the lamps to fully experience the initial rapid heating phase, avoiding misjudgments caused by the transient startup process. Subsequently, at each sampling time... The absolute value of the linear regression slope of temperature data relative to time within a continuous sliding window is calculated as the rate of temperature change at that moment. For the sampling point sequence within the window, the absolute value of the slope is calculated using the least squares method. When the absolute value of this rate of change is less than a preset threshold ε for 30 consecutive seconds, and the difference between the maximum and minimum shell temperature within that 30-second window does not exceed 0.3℃, the controller determines that the shell temperature has first entered a state of natural thermal equilibrium, and this moment is recorded as [the point where the temperature changes]. .

[0054] In a preferred embodiment, the preset threshold ε is set to 0.05°C / second. Under steady-state operating conditions, the temperature of the LED explosion-proof light casing naturally fluctuates between ±0.02°C / second and ±0.05°C / second due to factors such as ambient airflow fluctuations and slight fluctuations in grid voltage. Setting the threshold to 0.05°C / second can accurately distinguish between thermal equilibrium and non-equilibrium heating processes, while avoiding misjudgments caused by normal thermal fluctuations.

[0055] In determining Then, the controller makes a judgment. With the time of power outage Relationship: like < This indicates that the LED explosion-proof light had reached thermal equilibrium before the power was cut off, and the controller intercepted the signal from... to The data segment is taken as the effective heating segment, and this cycle is marked as the steady-state cycle; if ≥ This indicates that the lighting fixture was powered off before reaching thermal equilibrium (i.e., a power outage under abnormal operating conditions), and the controller intercepted the signal from... to The entire heating data segment is used as the effective heating segment, and this cycle is marked as a non-steady-state cycle. Simultaneously, an auxiliary warning message indicating power outage before reaching thermal equilibrium is reported via the communication interface. This auxiliary warning message is for operational reference only and does not participate in subsequent baseline calculations or health status diagnostic decisions, to avoid contaminating the normal diagnostic model with non-steady-state data.

[0056] It should be noted that the determination of the above-mentioned thermal equilibrium state is performed and remembered in real time during the operation of the controller. That is, the controller continuously monitors the sliding window data and records it once the determination condition is met. And store it in a local register. At any given moment, the controller directly reads the stored data. The value can be obtained instantly without backtracking calculations after a power outage, thus ensuring that the thermal balance determination result can be obtained immediately at the moment of power failure.

[0057] S2.2: The controller processes the temperature-time data within the captured effective heating range {(t, A univariate linear regression model was used for fitting. The univariate linear regression model is expressed as: ; in, Time is the independent variable. The shell temperature is the dependent variable. To fit the slope of the straight line, To fit the intercept of the straight line, the controller uses the least squares method to estimate the parameters. and This minimizes the sum of squared residuals. The slope obtained from the fitting is... denoted as the average heating rate Its dimension is °C / s.

[0058] It should be noted that the least squares method in this embodiment utilizes information from all sampling points within the effective heating range, which can effectively suppress the influence of temperature sensor quantization noise and random measurement noise on slope estimation, resulting in higher repeatability and anti-interference capability of the calculated heating rate.

[0059] S2.3: During the cooling data acquisition process, the controller executes the following termination judgment logic in real time: if the acquisition time reaches the preset maximum safe time, the acquisition is forcibly terminated; if the real-time temperature difference has decayed to below 10% of the initial temperature difference, the acquisition is terminated early. Based on the above judgment, the controller records the final cooling data termination time. And count the number of valid sampling points during the cooling phase: ; like If the number of iterations is less than a preset threshold (30 in this embodiment), the controller first determines the reason for termination: if termination is triggered because the temperature difference has decayed to less than 10% of the initial temperature difference, and subsequent fitting converges normally, then the subsequent fitting calculation continues; only if the number of iterations is less than the preset threshold due to reaching the maximum safe time or due to an unknown error will the controller terminate the current diagnostic process and report an error. After entering the fitting calculation, if the nonlinear fitting algorithm cannot converge normally (e.g., it still does not converge after 1000 iterations or the coefficient of determination R² does not meet the requirements), then the current diagnostic process is terminated and an error is reported. If the number is greater than or equal to the preset threshold, then proceed directly to the subsequent fitting calculation.

[0060] In this embodiment, the maximum safe duration is preferably set to 600 seconds. It should be noted that a 600-second acquisition duration can cover more than twice the cooling time constant (5-300 seconds) of a typical LED explosion-proof lamp, which is sufficient to reduce the temperature difference to less than 10% of the initial value in most actual working conditions. For extreme cases where τ is close to 480 seconds, complete attenuation can be achieved by configuring the acquisition duration to more than 1100 seconds.

[0061] In a preferred embodiment, the preset threshold for the number of sampling points is set to 30. It should be noted that nonlinear least squares fitting requires at least three valid data points to determine the parameters of a single exponential decay model. However, considering the presence of noise in actual temperature signals, at least 30 sample points (corresponding to 30 seconds of cooling data) are needed to obtain a reliable time constant estimate in order to ensure that the fitting result has statistical significance and sufficient degrees of freedom. If there are fewer than 30 cooling data points, the uncertainty of the fitting result will increase significantly, and underdetermined problems may even occur, making it unsuitable for subsequent diagnosis.

[0062] The controller uses a nonlinear least squares fitting method to extract the cooling time constant. .

[0063] Specifically, the controller will use the initial temperature difference calculated and fixed in step S1. As a known constant, the real-time temperature difference sequence during the cooling phase is fitted using a first-order exponential decay model. The mathematical expression of the first-order exponential decay model is: ; Among them, only It is a single parameter to be fitted.

[0064] It should be noted that the above fitting model assumes that the ambient temperature remains constant during the cooling process.

[0065] The controller constructs the residual sum of squares: ; in, For the cooling stage The measured temperature difference at each sampling point For the first The timestamp of each sampling point.

[0066] The controller estimates by minimizing the objective function. The optimal value is obtained. In a preferred embodiment, the Levenberg-Marquardt algorithm is used to iteratively solve the above nonlinear least squares problem.

[0067] In the specific implementation, the algorithm parameter configuration includes: the initial value of the damping factor is set to 0.01; the average ratio of the first three sampling points in the cooling stage is calculated. If the average value is greater than 0.9, the initial iteration value is set to 120 seconds; if it is between 0.5 and 0.9, the initial iteration value is set to 60 seconds; if it is less than 0.5, the initial iteration value is set to 15 seconds. If the above rules cannot be used, the default initial iteration value is 60 seconds. After each iteration, if the sum of squared residuals increases compared to the previous iteration, the damping factor is multiplied by 10; if the sum of squared residuals decreases, the damping factor is divided by 10. The iteration convergence condition is set as follows: the rate of change of the sum of squared residuals between two adjacent iterations is less than 10%. -6 Or the change in the iteration parameter τ is less than 10. -4 or gradient norm less than 10 -6 The maximum number of iterations is limited to 1000. If no convergence condition is met after 1000 iterations, the nonlinear fitting is deemed nonconvergent, and the controller terminates the current nonlinear fitting process. The above parameter configuration has been tested and shows stable convergence under the data characteristics of this application scenario.

[0068] Furthermore, the controller uses the extracted cooling time constant Perform a physical validity check. In a preferred embodiment, the physical validity check includes the following three parallel conditions: The first term is the time constant. More than 5 seconds. If A time of 5 seconds or less indicates an abnormally rapid cooling process, which does not conform to the physical cooling characteristics of LED explosion-proof lights determined by their heat capacity and thermal resistance. This may be due to sensor detachment, measurement noise, or fitting errors.

[0069] The second term, time constant Within a preset reasonable time range [5s, 480s], this value is derived from statistical analysis of measured data from lamps of different power levels. If If the readings exceed this range, it indicates that the cooling process is abnormally slow, exceeding the physically reasonable range of the LED explosion-proof light, which may indicate abnormal data or fit divergence.

[0070] Thirdly, the coefficient of determination R² is greater than 0.85. The coefficient of determination R² is defined as R² = 1 − Sres / Stot, where Sres is the sum of squared residuals and Stot is the sum of squared total deviations. R² characterizes the goodness of fit of the exponential decay model to the measured data; R² > 0.85 indicates that the model can explain more than 85% of the data variation, demonstrating a good fit.

[0071] All three conditions above must be met to determine the current cooling time constant. Valid. If any condition is not met, the controller determines that the cooling data is invalid, terminates the subsequent process, and reports an error message about a physical anomaly in the time constant through the communication interface.

[0072] S3: Generate and store the thermal characteristic response value for the current cycle based on the heating rate and the cooling time constant.

[0073] S3.1: In this invention, the thermal characteristic response value is defined as a quantitative index used to comprehensively characterize the thermal state changes of the luminaire during the current start-stop cycle. It couples the heat accumulation capacity during the heating phase with the heat dissipation capacity during the cooling phase into a scalar. In this embodiment, the controller uses the average heating rate calculated in step S2... With cooling time constant A fusion calculation is performed, and the product of the two is taken as the thermal characteristic response value H of the current start-stop cycle. The larger this product value is, the more significant the heat generation of the luminaire or the more difficult the heat dissipation is under the same conditions.

[0074] As can be seen, the embodiments of the present invention transform the temperature change characteristics in a single direction (heating only or cooling only) into a comprehensive index coupled in two directions, thereby maximizing the amount of thermal characteristic information obtained in a single start-stop cycle and improving the sensitivity of characterizing the health status of the lamp.

[0075] S3.2: The controller performs multiple parallel validity checks on the synthesized thermal characteristic response value H. Only after all checks pass can the data proceed to the subsequent storage stage. It should be noted that setting multiple checks to remove invalid data generated at the source due to sensor anomalies, operating condition anomalies, or algorithm anomalies can prevent dirty data from entering the historical sequence, thereby ensuring the purity of the baseline model and the reliability of diagnostic decisions.

[0076] S3.2.1: The controller determines whether H is greater than zero. Based on the physical principles in step two, the heating rate... Physically, it is always positive (temperature increases with energizing time), and the cooling time constant is... Physically, the product H is always positive (the duration of the cooling process), and it is physically positive if and only if it is positive. If the sensor readings are reversed, such as the first and second temperature sensor channels being cross-connected, resulting in the housing temperature reading being lower than the ambient temperature during the heating phase, or incorrect data labeling, or if the data division between the heating and cooling phases is reversed, or if the fitting algorithm outputs a negative value under extreme abnormal conditions, causing H≤0, then the data is physically invalid. The controller will immediately discard the data, not store it in subsequent data, and end the current diagnostic process.

[0077] S3.2.2: The controller determines the average heating rate. Is it greater than the preset minimum temperature rise rate? In a preferred embodiment, the minimum heating slope The setting is 0.01°C / s. That is, if the heating rate is less than 0.01°C / s, the total temperature change of the casing within a complete heating cycle (typically several minutes to tens of minutes) will be less than a few degrees Celsius. This significantly deviates from the normal operating temperature characteristics of an LED explosion-proof light casing. Possible reasons for an extremely low heating rate include poor thermal contact due to improper temperature sensor installation, sensor malfunction, or the controller performing the next power-on diagnostic before the light fixture has fully cooled down (i.e., the casing temperature is close to equilibrium), resulting in an excessively short effective heating period or a small temperature difference. Less than or equal to The controller determines that the data is invalid and terminates the subsequent process.

[0078] S3.2.3: Controller determines cooling time constant Is it within a preset reasonable time range? Internally, extensive experimental testing has shown that the thermal time constants of LED explosion-proof lights with different power levels (50W to 300W) and different heat dissipation structures range from 5 seconds to 480 seconds. Therefore, in the preferred embodiment, Set to 5 seconds. Set to 480 seconds. If If the readings exceed this range, it indicates that the thermophysical characteristics of the luminaire deviate significantly from the normal range, which may be due to fitting failure, sensor failure, or extreme environmental conditions. This data should not be included in subsequent diagnostic models.

[0079] If any of the above three checks fails, the controller determines that the thermal characteristic response value for the current cycle is invalid, discards the data, and terminates the current diagnostic process without executing subsequent steps. The order of the checks is not strictly limited; in alternative embodiments, parallel checks followed by a summary of the results can be used.

[0080] S3.3: After passing the above physical validity verification, the controller responds to the current thermal characteristic value. Perform a statistical outlier test to determine whether the data significantly deviates from the normal distribution range of historical data.

[0081] Specifically, the controller reads all existing data from the historical thermal characteristic response value sequence in the memory. If the historical sequence is empty, the outlier detection is skipped, and the data is directly processed. Write the historical sequence. If the historical sequence is not empty, calculate the median M and the absolute deviation of the median. The absolute deviation of the median is defined as: MAD = median(| -M|), which is the median of the absolute values ​​of the differences between each historical data point and the median. MAD is a robust measure of dispersion. Compared to standard deviation, MAD is not sensitive to outliers and can still accurately reflect the true dispersion of the data even with a small number of outliers.

[0082] The controller calculates the current value. The absolute deviation from the historical median M is calculated, and it is determined whether this absolute deviation is greater than a preset outlier multiple multiplied by MAD. In a preferred embodiment, the preset outlier multiple is set to 5. The basis for setting the outlier multiple to 5 times MAD is that for data following a normal distribution, approximately 99.7% of the data points fall within the range of ±3 times MAD of the median. That is, under a normal distribution, MAD is approximately equal to 0.6745 times the standard deviation. Therefore, 5 times MAD is approximately equal to 3.37 times the standard deviation, corresponding to a cumulative probability exceeding 99.9%. The thermal characteristic response value of LED explosion-proof lights will have certain natural fluctuations in actual use due to factors such as changes in ambient temperature and fluctuations in power grid voltage. However, extreme values ​​exceeding 5 times MAD can be statistically determined with high confidence to be outliers. Using 5 times MAD as the judgment threshold is neither too sensitive to misjudging normal fluctuations as outliers nor too insensitive to missing real abnormal data.

[0083] When the controller determines When a value is an outlier, it indicates that the thermal state corresponding to the data deviates from the historical normal range. This data should not be included in the historical sequence to contaminate the baseline model, nor should it be used for baseline comparison or diagnostic output. In this case, the controller discards the current data, does not write it into the historical sequence, directly ends the current diagnostic process, and reports the data outlier warning message through the communication interface to notify maintenance personnel to pay attention to any sudden changes in the status of the lighting fixture.

[0084] If the controller determines If it is a non-outlier value, then... The queue management process submitted to step S3.4 is where S3.4 performs sequence length checks and write operations.

[0085] It should be noted that the advantage of using MAD as an outlier criterion compared to using mean and standard deviation criteria is that the MAD method does not require the assumption that the data follows a normal distribution, is robust to the data distribution pattern, and is not affected by outliers. Even when there are a few outliers in the historical sequence, it can still reliably determine whether new data is an outlier. Therefore, it is particularly suitable for engineering scenarios where abnormal data may be mixed in under long-term field operation conditions.

[0086] S3.4: The controller will verify the thermal characteristic response value. Write the historical thermal characteristic response value sequence stored in memory in chronological order to the end, forming an ordered time series data structure.

[0087] The historical sequences are stored in non-volatile memory (such as EEPROM or Flash memory) to ensure that the historical sequence data is not lost after the controller is powered off and restarted. The controller also maintains a sequence length counter to record the current total amount L of historical sequence data.

[0088] Before writing new data each time, the controller determines whether the current sequence length L has reached the preset maximum storage quantity Lmax. In a preferred embodiment, the maximum storage quantity Lmax is set to 200. First, 200 data points are sufficient to construct a statistically significant baseline model. For LED explosion-proof lights that are started and stopped once or several times a day, 200 data points correspond to approximately one to six months of historical data, which can cover the slow drift of thermal characteristic response values ​​caused by seasonal changes in ambient temperature.

[0089] If the current sequence length L has reached Lmax, the controller deletes the oldest data point in the sequence before writing new data, and then shifts the remaining data points forward in sequence. Write to the end of the sequence. The above operation ensures that the sequence length never exceeds Lmax, while retaining the most recent historical data and discarding outdated data.

[0090] Furthermore, the first-in-first-out queue management method and the strategy of taking the most recent N historical values ​​in the baseline calculation work together to form an adaptive time window mechanism. That is, the historical sequence retains data redundancy for up to 200 periods, while the baseline calculation only takes the values ​​of the most recent N periods (N is preferably 10). This ensures that the baseline is sensitive to short-term fluctuations and retains enough redundant data for statistical outlier determination.

[0091] S4: Calculate the baseline value based on the stored historical thermal characteristic response values, and output the health status of the lamp according to the degree of deviation between the current thermal characteristic response value and the baseline value.

[0092] S4.1: Before starting baseline calculation, the controller first reads the current total data volume L of the historical thermal characteristic response value sequence and determines whether L meets the minimum sample size requirement for baseline calculation. In a preferred embodiment, the minimum sample size N is set to 10. Specifically, according to statistical principles, when the sample size is greater than or equal to 10, the standard error of the sample mean as an estimate of the population mean has decreased to an acceptable level (approximately 0.32 times the population standard deviation), which can meet the accuracy requirements of engineering diagnosis. Meanwhile, the accumulation of data over 10 cycles corresponds to approximately a 10-day data acquisition cycle in a typical daily start-stop lighting scenario, a waiting period that is acceptable in engineering.

[0093] If L ≤ N, the controller determines that the system is currently in the data accumulation phase and does not yet have reliable baseline calculation conditions. At this time, the controller does not perform deviation calculations or output any lighting health status diagnostic levels. Instead, it generates status information for the modeling process (after collecting L / N cycles) and reports it to the remote maintenance terminal via the communication interface. This status information allows maintenance personnel to understand that the monitoring system is in the data accumulation phase and has not yet entered formal diagnostic mode, rather than indicating a system failure or data loss. It should be clearly noted that although no diagnostic results are output for this cycle, the thermal characteristic response value for the current cycle is... The validity verification has been completed in step S3 and written into the historical sequence (if the verification passes), which has successfully enriched the historical database and accumulated a data foundation for diagnosis in subsequent cycles.

[0094] If L>N, the controller determines that the system has reliable baseline calculation conditions and proceeds to the subsequent baseline calculation and deviation analysis process.

[0095] It should be noted that during the data accumulation period, if a thermal characteristic response value for a certain period is identified as an outlier and discarded, that period will not be included in the total data volume L. To shorten the modeling gap, when the amount of valid historical data L is greater than 3 but less than or equal to N, the controller still calculates a temporary baseline value, which is the arithmetic mean of all existing valid data, and outputs a health status level with a temporary modeling label for maintenance personnel to refer to, but the confidence level of this level is marked as low confidence. When L is greater than N, it automatically switches to standard diagnostic mode. If L is still less than or equal to N after continuous operation for more than the preset modeling timeout period (e.g., 30 days), a modeling timeout will be reported through the communication interface. Please check the alarm information of the sensors and data acquisition links.

[0096] The temporary baseline value is the arithmetic mean of all available valid historical data (L data points). The deviation calculation method is the same, but due to the high uncertainty of baseline estimation with small samples, the health status grading intervals in the temporary mode are broadened by 1.5 times compared to the standard thresholds: healthy ≤ 7.5%, attentive ≤ 22.5%, and abnormal ≤ 45%, to compensate for the uncertainty of baseline estimation with small samples. A low-confidence (in progress) label is added to the output health level. When L increases to greater than N, it automatically switches to the formal diagnostic mode, and the grading intervals revert to the standard thresholds.

[0097] S4.2: Based on the acceptance condition in step S4.1 (i.e., L>N), the controller performs baseline value calculation. The core principle of baseline calculation is that the historical data used to calculate the baseline value strictly excludes the thermal characteristic response value of the current cycle. It is important to note that the baseline value serves as a reference point for the normal state. The observation value of the current period should not be used to define its own normal range; otherwise, a logical loop of evaluating itself based on its own baseline will occur, weakening the independence and effectiveness of the deviation diagnosis. This principle is consistent with the basic idea of ​​cross-validation (leave-one-out) in statistics, ensuring the rigor of the diagnosis.

[0098] The controller forces the use of the leave-one-out method to calculate the baseline, first determining the thermal characteristic response value of the current cycle. Does it already exist in the historical sequence? (1) If If the entry at index L has already been written into the historical sequence, then temporarily exclude it from the historical sequence. It reads the nearest N data points (indexed from LN to L-1) forward and calculates their arithmetic mean as the baseline value B.

[0099] (2) If If no historical data points are written to the historical sequence, the most recent N data points (indices L-N+1 to L) are read backwards from the end of the historical sequence, and their arithmetic mean is calculated as the baseline value B. In both of these cases, the data used to calculate the baseline strictly excludes the thermal characteristic response value of the current period. .

[0100] Furthermore, in this invention, the baseline value is defined as the expected level of the thermal characteristic response value of the luminaire under historical normal conditions, serving as a reference benchmark for determining whether the current state has deviated. The controller calculates the arithmetic mean of the N historical thermal characteristic response values ​​read, and records the calculation result as the baseline value B.

[0101] Furthermore, if the calculated baseline value B equals zero, the controller forces the deviation D to be zero and triggers a zero baseline alarm via the communication interface. A baseline value of zero indicates that all historical thermal characteristic response values ​​used to calculate the baseline are zero, which is physically impossible. Therefore, a zero baseline necessarily indicates a serious anomaly in the historical data, such as corrupted memory data, accidental zeroing of historical data, or long-term sensor failure resulting in only invalid zero values ​​being written. In this case, any deviation calculation with zero as the denominator will result in a division-by-zero error, and the system will be unable to output meaningful diagnostic results. Forcing D to zero and triggering an alarm is a necessary self-protection measure for the system.

[0102] S4.3: After completing the calculation of the baseline value B (and B≠0), the controller calculates the thermal characteristic response value for the current cycle. Deviation from baseline value B.

[0103] The formula for calculating the deviation is as follows: ; in, The deviation is expressed as a dimensionless ratio.

[0104] The sign information of the deviation, i.e. Whether the value is greater than or less than B, it is not used to distinguish the fault type in the single lamp health status diagnosis. However, this symbol information and the absolute value of the deviation are stored locally on the controller for subsequent steps of multi-lamp collaborative power grid anomaly location.

[0105] S4.4: The controller compares the calculated deviation D with the preset grading intervals, and outputs the corresponding lamp health status diagnostic level based on the interval D falls into. The grading intervals are preset within the controller as firmware parameters, specifically including four discrete grading intervals. The physical meaning of each level and its corresponding operation and maintenance strategy are as follows: (i) Health Level: When the deviation D is less than or equal to the first preset threshold, the controller determines that the lamp is in a healthy state. In a preferred embodiment, the first preset threshold is set to 5% (i.e., 0.05). Under the same ambient temperature, the periodic fluctuation of the thermal characteristic response value of a normally operating LED explosion-proof lamp is usually within the range of ±3% to ±5% due to normal factors such as normal fluctuations in the mains voltage (within ±5% of the rated voltage) and daily variations in ambient temperature. Setting the first preset threshold to 5% can cover the normal fluctuation range and avoid frequent false alarms. When outputting the health level, the controller generates a prompt message indicating that the lamp is in a normal state.

[0106] (ii) Attention Level: When the deviation D is greater than the first preset threshold and less than or equal to the second preset threshold, the controller determines that the luminaire is at the attention level. In a preferred embodiment, the second preset threshold is set to 15% (i.e., 0.15). When the deviation exceeds 5% but does not exceed 15%, it indicates that the thermal characteristics of the luminaire have undergone a detectable but not yet serious change. Possible causes include: slight dust accumulation on the luminaire's heat sink fins leading to a decrease in heat dissipation capacity, slight aging of the thermal grease, or significant seasonal changes in ambient temperature. This level does not constitute an imminent fault threat, but it is recommended that maintenance personnel arrange a regular inspection plan to monitor its changing trends. When outputting the attention level, the controller generates a prompt message indicating that the luminaire status requires attention and that an inspection is recommended.

[0107] (III) Abnormal Level: When the deviation D is greater than the second preset threshold and less than or equal to the third preset threshold, the controller determines that the luminaire is in an abnormal level. In a preferred embodiment, the third preset threshold is set to 30% (i.e., 0.30). A deviation exceeding 15% indicates that the thermal characteristics of the luminaire have changed significantly. Possible causes include: severe light decay of the LED light source leading to a large amount of electrical energy being converted into heat, cooling fan failure (for forced air-cooled luminaires), and significant degradation of the internal heat conduction path of the luminaire. Although the luminaire can still work at this level, it is in a sub-healthy state. If it continues to deteriorate, it may lead to luminaire failure or a sharp shortening of its lifespan. When outputting an abnormal level, the controller generates a prompt message indicating that the luminaire is in an abnormal state and recommends maintenance or replacement as soon as possible.

[0108] (iv) Severe Fault Level: When the deviation D exceeds the third preset threshold, the controller determines that the luminaire is in a severe fault level. In a preferred embodiment, the third preset threshold is set to 30%. At the severe fault level, the luminaire's thermal characteristic response value has significantly deviated from the normal range. Possible causes include overheating due to a damaged driver power supply, thermal runaway caused by severe light decay of the LED chip, or complete failure of the temperature sensor. When outputting a severe fault level, the controller generates a prompt message indicating a severe fault in the luminaire status and recommending immediate shutdown for repair.

[0109] The values ​​of the first preset threshold (5%), the second preset threshold (15%), and the third preset threshold (30%) are all determined based on empirical data accumulated from a large number of LED explosion-proof lamps in accelerated aging tests in the laboratory and long-term operation monitoring in the field, and were determined after statistical analysis. These thresholds can sensitively capture the early degradation of the lamp's thermal performance while effectively suppressing false alarms caused by normal fluctuations, achieving a balance between sensitivity and specificity.

[0110] S4.5: The controller combines the diagnostic data of the current period into a data frame through the preset communication module and sends it to the regional aggregation node or cloud operation and maintenance platform as the data source for the subsequent step S5 multi-lamp collaborative power grid anomaly location.

[0111] The communication module selects either wired or wireless communication methods based on the specific application scenario. In a preferred embodiment, for industrial sites with existing fieldbus infrastructure, an RS485 bus interface is used, and the Modbus-RTU protocol is selected as the communication protocol. The Modbus-RTU protocol is a widely used standard protocol in industrial sites, with advantages such as strong anti-interference capability and good device compatibility. For distributed deployment scenarios lacking wired communication infrastructure, a 4G wireless communication module is used, and the MQTT protocol is selected as the communication protocol. The MQTT protocol is a lightweight IoT message transmission protocol with low data packet overhead, support for reconnection after disconnection, and message persistence, making it suitable for remote monitoring scenarios.

[0112] The data frame contains the current timestamp (used to record the generation time of diagnostic data) and the current periodic thermal characteristic response value. (Raw data, for secondary analysis and data mining in the cloud), baseline value B (the benchmark for assessing deviation), deviation D (the core diagnostic criterion), generated steady-state / non-steady-state cycle markers (used to distinguish whether this diagnosis is based on a complete thermal equilibrium process; the confidence level of the diagnosis results for non-steady-state cycles is slightly lower, and the cloud can perform weighted processing accordingly), and the corresponding health status diagnosis level.

[0113] S4.6: The baseline value B is dynamically updated using a moving average of the most recent N historical values. This method has good adaptability to normal, slowly changing factors such as ambient temperature. However, it should be noted that when the luminaire itself undergoes slow aging, the thermal characteristic response value may drift at an extremely slow rate, and the moving baseline may follow this drift, thus delaying the rate at which the deviation D reaches the abnormal threshold. To overcome this limitation, the controller calculates the long-term deviation provided that the total length of the historical sequence L ≥ 150. The earliest G data points in the historical sequence are taken. Calculate the arithmetic mean as the long-term baseline value. Long-term deviation is This information is also reported to the maintenance platform as an auxiliary diagnostic indicator, allowing maintenance personnel to comprehensively assess the long-term performance degradation trend of the lighting fixtures. If If the value is zero, the long-term deviation calculation is skipped and no zero-value alarm is reported.

[0114] S5: Based on the health status of multiple lamps and the power grid topology, locate abnormal sections and output the location results.

[0115] S5.1: The controller of the regional aggregation node or cloud platform collects the deviation of the thermal characteristic response value of all LED explosion-proof lights connected to the same power supply feeder (or the same distribution substation) within the current time window, with a preset time window width as the period. ( Indicates the first (Each lamp) and the corresponding current periodic thermal characteristic response value of each lamp. and their respective baseline values .

[0116] Among them, the same power supply feeder specifically refers to the set of physical power supply paths in the distribution network that supply power from the low-voltage side bus outgoing switch of the same 10kV / 0.4kV distribution transformer and do not pass through on-load tap changers or independent reactive power compensation devices.

[0117] In a preferred embodiment, the time window width is set to 5 minutes. It should be noted that a 5-minute time window is sufficient to capture multi-lamp synchronization or quasi-synchronization deviation events caused by power grid anomalies.

[0118] Deviation of the controller from each lamp Perform a step-by-step comparison to determine if the value exceeds a preset abnormal threshold. In a preferred embodiment, the abnormal threshold The threshold is set to 15%, consistent with the initial threshold for the anomaly level in step S4; that is, a deviation greater than 15% is considered an anomaly. Using the same threshold as for single-lamp diagnosis ensures consistency in anomaly judgment criteria between single-lamp diagnosis and multi-lamp collaborative diagnosis.

[0119] when > When this happens, the controller marks the light fixture as a suspected abnormal node. The controller then counts the total number of light fixtures marked as suspected abnormal nodes on the current feeder. and the preset event trigger quantity threshold. A comparison is made. In a preferred embodiment, the event trigger quantity threshold is used. Setting it to 3, the simultaneous occurrence of thermal characteristic deviations in three or more lamps is statistically significant. If... ≥ If the controller determines that a regional power grid anomaly has occurred, it will proceed to the subsequent anomaly pattern recognition and section location process; otherwise, the regional power grid anomaly will not be triggered, and the anomaly of each lamp will be treated as a single lamp fault, outputting the fault diagnosis result of the individual lamp without reporting to the power grid location layer. If the total number of valid monitoring lamps connected to the same feeder is less than 3, a regional event will be triggered when more than two-thirds (rounded up) of the lamps are abnormal at the same time; if it is only 1, a single lamp anomaly will directly trigger a regional power grid anomaly, but the anomaly type confidence level will be marked as low confidence.

[0120] S5.2: After a regional power grid anomaly event is triggered, the controller identifies the anomaly pattern and infers the corresponding power grid anomaly type based on the direction of change of the thermal characteristic response values ​​of the lamps marked as suspected anomaly nodes relative to their respective baseline values. Here, the direction of change refers to the thermal characteristic response value of the current cycle. Compared with baseline value The comparison results—if > If it is less than 1, it is considered a positive deviation (increase); if it is less than 1, it is considered a negative deviation (decrease).

[0121] Furthermore, for LED explosion-proof lights using constant current driving, when the supply voltage increases, the input power of the driving power supply increases. Under the premise of constant current output, both the power supply's own losses and the forward voltage drop of the LED chip increase, leading to an increase in total heat generation and an upward trend in the heating rate v. Simultaneously, due to the increased steady-state temperature rise, the initial temperature difference after power failure increases, but the cooling time constant mainly depends on the thermal resistance-capacity structure of the lamp itself and is less affected by ambient temperature. Experiments show that within a voltage deviation range of ±10%, the change in v is much greater than the reverse change in τ, and the product H is monotonically positively correlated with the supply voltage. Therefore, a simultaneous increase in H for multiple lamps can be inferred as a high grid voltage, and conversely, a decrease can be inferred as a low grid voltage. Based on the above physical mechanism, the following judgment rule is set: When the thermal characteristic response values ​​of the vast majority (more than two-thirds) of the abnormal nodes show positive deviation, the controller infers that the power grid anomaly type is high supply voltage; when the thermal characteristic response values ​​of the vast majority (more than two-thirds) of the abnormal nodes show negative deviation, the controller infers that the power grid anomaly type is low supply voltage or undervoltage; when the abnormal nodes show a clustered distribution from upstream to downstream in the topology, that is, when the abnormal nodes are all located on the same continuous branch of a certain feeder, it is inferred that the anomaly may be caused by common factors on that branch, such as abnormal branch line impedance, neutral line open circuit, etc.

[0122] The physical basis of the above three abnormal modes all originates from the basic principles of power system analysis, and those skilled in the art can reproduce the inference logic based on the above disclosure.

[0123] S5.3: After completing the abnormal pattern recognition and power grid abnormality type inference, the controller, in conjunction with the power grid topology, traces the common upstream node of all abnormal nodes along the power supply path and locks the power distribution section between the common upstream node and the first normal lamp in its downstream direction as the suspected abnormal section.

[0124] Specifically, the controller reads the current feeder's power supply path topology. This topology data is pre-stored in the local non-volatile memory of the regional aggregation node in the IEC 61968 standard Common Information Model (CIM) format, or imported via a configuration file during system deployment. When a line change occurs in the distribution network, the operation and maintenance master station actively sends the updated topology model to the aggregation node via the IEC 61850 MMS protocol. The aggregation node performs incremental updates based on the model version number to ensure that the topology data remains consistent with the actual power grid structure.

[0125] In a preferred embodiment, the topology is described using a tree-like data structure based on an adjacency list, with the power source as the root node, each terminal device as a leaf node, and switches and branch boxes at all levels as intermediate nodes. Each node records a list of its parent nodes (upstream nodes) and child nodes (downstream nodes). The distribution network topology can be described and exchanged using the System Configuration Language (SCL) model in the IEC 61850 standard.

[0126] The controller spatially labels all lights marked as abnormal nodes in the topology and traces the parent node upstream of each abnormal node along the power supply path until it finds the first intersection of all abnormal node paths. This intersection is the smallest upstream node common to all abnormal nodes, meaning that at least one abnormal node exists in all branches supplying power downstream from this node, while not all branches supplying power downstream from this node's parent node contain abnormal nodes.

[0127] In a preferred embodiment, the upward tracing is implemented using the following algorithm: Initialize the common ancestor node set CA as the complete upstream path of the first anomalous node (all nodes from that node to the root node on the power supply side); traverse each remaining anomalous node, calculate the intersection of the complete upstream path of that node with the current CA, and update the CA to this intersection; after traversal, the node in CA closest to the anomalous node (i.e., the most downstream) is the minimum common upstream node of all anomalous nodes. The time complexity of this algorithm is O(K·R), where K is the number of anomalous nodes and R is the topology depth. For the scale of the distribution network (usually no more than several thousand nodes), the computational efficiency fully meets the real-time requirements.

[0128] Furthermore, after determining the minimum common upstream node, the controller searches downstream along the power supply path of that node for the first normal luminaire that is not marked as abnormal. If a normal luminaire exists that is not marked as abnormal, the power distribution section between the minimum common upstream node and the first normal luminaire is identified as a suspected abnormal section. If all luminaires downstream of the minimum common upstream node are marked as abnormal nodes (i.e., no normal luminaires exist), the entire power supply path from the minimum common upstream node to its downstream terminus is identified as a suspected abnormal section, and the confidence level is marked as low confidence. It is recommended to confirm this through on-site inspection.

[0129] It should be noted that the above-mentioned segment location method is based on a radial distribution network topology. To address potential boundary conditions during the location process, this method also includes the following processing logic: If the smallest common upstream node traced upwards is the root node on the power supply side, the segment from the outgoing switch of that root node to the first normal lamp downstream is locked as a suspected abnormal segment, and auxiliary information indicating that the abnormal source may be located on the bus or outgoing side is simultaneously reported; if all lamps downstream of the smallest common upstream node are marked as abnormal nodes, all segments from that common upstream node to the end of the power supply path are locked as suspected abnormal segments, and the confidence level of the location result is marked as low confidence, with confirmation recommended in conjunction with on-site inspection; for ring network structures, in application, the network topology is first analyzed to be equivalent to multiple radial subnets, and then segment locking is performed separately according to the above-mentioned radial topology tracing logic.

[0130] S5.4: The controller reports the complete power grid anomaly location results to the power grid dispatch center or operation and maintenance master station system in the form of standardized data frames. The data frame includes the event timestamp, anomaly type, location of the locked section, confidence level, list of lamp IDs involved in the judgment, deviation degree and direction of change of each anomaly node.

[0131] In a preferred embodiment, the data frames are encapsulated and transmitted using the General Object-Oriented Substation Event (GOOSE) message format or the Manufacturing Message Specification (MMS) protocol defined in the IEC 61850 standard.

[0132] It can be seen that by aggregating the diagnostic results of a single lamp into multi-lamp collaborative analysis, the fault perception capability has been upgraded from the single lamp level to the feeder level and section level, transforming LED explosion-proof lamps from simple electrical equipment into distributed sensing nodes of the power grid status, and greatly expanding the application value of the condition monitoring system.

[0133] The remote status monitoring and fault diagnosis device based on LED explosion-proof lamps provided in this embodiment of the invention can execute the remote status monitoring and fault diagnosis method based on LED explosion-proof lamps provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0134] Figure 3This is a schematic diagram of the structure of an electronic device for implementing the remote status monitoring and fault diagnosis method based on LED explosion-proof lights, as described in this embodiment of the invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0135] like Figure 3 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded into the RAM 13 from storage unit 18. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0136] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0137] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, central processing unit (CPU), graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as remote condition monitoring and fault diagnosis methods based on LED explosion-proof lights.

[0138] In some embodiments, the remote condition monitoring and fault diagnosis method based on LED explosion-proof lights can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the remote condition monitoring and fault diagnosis method based on LED explosion-proof lights described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the remote condition monitoring and fault diagnosis method based on LED explosion-proof lights by any other suitable means (e.g., by means of firmware).

[0139] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0140] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0141] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0142] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0143] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0144] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0145] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0146] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for remote condition monitoring and fault diagnosis of LED-based explosion-proof lamps, characterized in that, include: Collect the housing temperature and ambient temperature of the explosion-proof lamp, and record the power-on and power-off times of the lamp. Based on the power-on time, power-off time, and collected temperature data, the heating rate during the heating phase and the cooling time constant during the cooling phase are obtained. Generate and store the thermal characteristic response value for the current cycle based on the heating rate and the cooling time constant; The baseline value is calculated based on the stored historical thermal characteristic response values, and the health status of the luminaire is output according to the degree of deviation between the current thermal characteristic response value and the baseline value. Based on the health status of multiple lamps and the power grid topology, abnormal sections are located and the location results are output.

2. The method for remote condition monitoring and fault diagnosis of LED-based explosion-proof lamp according to claim 1, characterized in that, Before collecting the housing temperature and ambient temperature of the explosion-proof lamp, the following steps are also taken: using an independent backup power supply to power the controller; The voltage amplitude at the power supply terminal of the lamp is monitored in real time using a voltage detection circuit. When the voltage amplitude jumps from a first preset percentage below the rated voltage to a second preset percentage above the rated voltage, the current moment is marked as the power-on start moment. When the voltage amplitude drops from a second preset percentage above the rated voltage to a first preset percentage below the rated voltage, the current moment is marked as the power outage moment.

3. The method for remote condition monitoring and fault diagnosis of LED-based explosion-proof lamp according to claim 2, characterized in that, Before obtaining the heating rate during the heating phase and the cooling time constant during the cooling phase, the following steps are also included: The difference between the casing temperature and the ambient temperature is calculated at the moment of power failure as the initial temperature difference; After power failure, the casing temperature and ambient temperature data are collected during the cooling phase. The data collection process employs a dual-condition termination strategy, stopping recording when any of the following conditions are met: The data collection time has reached the preset maximum safe duration; The difference between the real-time calculated shell temperature and the ambient temperature decays to below a first preset percentage of the initial temperature difference.

4. The method for remote condition monitoring and fault diagnosis of LED-based explosion-proof lamp according to claim 3, characterized in that, Before obtaining the heating rate during the heating phase and the cooling time constant during the cooling phase, the validity of the ambient temperature is also verified. Determine whether the ambient temperature is within a preset physically reasonable range; Determine whether the initial temperature difference is greater than a preset minimum effective temperature difference threshold; If any verification fails, the diagnostic process will be terminated and the ambient temperature reference failure information will be reported.

5. The method for remote condition monitoring and fault diagnosis of LED-based explosion-proof lamp according to claim 4, characterized in that, The cooling time constant for the cooling phase includes: The number of valid sampling points during the cooling phase is counted. If the number of valid sampling points is less than a preset threshold, the diagnostic process is terminated. When the number of effective sampling points meets the requirements, the initial temperature difference is used as a fixed benchmark, and the real-time temperature difference sequence of the cooling stage is nonlinearly fitted according to the first-order exponential decay model to obtain the cooling time constant. If the cooling process is terminated prematurely and the real-time temperature difference sequence has not yet decayed to the first preset percentage of the initial temperature difference, the diagnostic process will be terminated.

6. The remote status monitoring and fault diagnosis method based on LED explosion-proof lights according to claim 5, characterized in that, The heating rate during the heating phase includes: The effective heating segment is extracted from the heating stage data, and the temperature data within the effective heating segment is linearly fitted relative to the time data. The slope obtained from the fitting is used as the heating rate. Generate and store the thermal characteristic response value for the current period based on the heating rate and the cooling time constant: multiply the heating rate by the cooling time constant, and use the product as the thermal characteristic response value for the current period; The validity of the thermal characteristic response value is verified; The validity verification includes: the thermal characteristic response value is positive, the heating rate is greater than the preset minimum heating slope, and the cooling time constant is within the preset reasonable time range; If the absolute deviation of the thermal characteristic response value from the median of the historical sequence exceeds a preset outlier multiple, it is determined to be an outlier value. The current data is discarded and the current diagnostic process ends. Subsequent baseline calculations and health status outputs are not performed. If it is not an outlier, the current thermal characteristic response value is written into the historical sequence.

7. The method for remote condition monitoring and fault diagnosis of LED-based explosion-proof lamp according to claim 6, characterized in that, When storing the thermal characteristic response value into the historical sequence, the first-in-first-out principle is used to maintain the sequence length, limiting the sequence length to a preset maximum storage quantity. If the quantity is exceeded, the earliest data point is deleted.

8. The method for remote condition monitoring and fault diagnosis of LED-based explosion-proof lamp according to claim 7, characterized in that, The baseline value is calculated based on the stored historical thermal characteristic response values, including: Read the current total amount of data in the historical sequence. If the total amount of data is less than or equal to the preset minimum number of samples, the system is determined to be in the data accumulation period. No deviation calculation is performed, and the state information in the modeling process is directly output. If the total amount of data is greater than the preset minimum sample size, then calculate the baseline value: Determine whether the thermal characteristic response value of the current period has been written into the historical sequence; If it has already been written, temporarily exclude the last data point that was recently written from the historical sequence, and read the most recent N data points backward. If the second to last data point points to the current value, skip it to ensure that the data read does not contain the thermal characteristic response value of the current period. If not written, the most recent preset number of historical thermal feature response values ​​are read directly from the historical sequence; The arithmetic mean of the historical thermal characteristic response values ​​read is used as the baseline value; If the baseline value is zero, then the deviation is forced to zero and a zero baseline alarm is triggered.

9. The method for remote condition monitoring and fault diagnosis of LED-based explosion-proof lamp according to claim 8, characterized in that, The health status of the luminaire is output based on the degree of deviation between the current thermal characteristic response value and the baseline value, including: When the baseline value is not zero, calculate the absolute value of the difference between the current thermal characteristic response value and the baseline value, divide the absolute value by the baseline value, and use the result as the deviation. Based on the deviation, a preset grading interval is matched to output the corresponding lamp health status level: When the deviation is less than or equal to the first preset threshold, the health level is output; When the deviation is greater than the first preset threshold and less than or equal to the second preset threshold, the attention level is output. When the deviation is greater than the second preset threshold and less than or equal to the third preset threshold, the abnormality level is output. When the deviation exceeds the third preset threshold, a severe fault level is output.

10. The method for remote condition monitoring and fault diagnosis of LED-based explosion-proof lamp according to claim 9, characterized in that: Based on the health status of multiple light fixtures and the power grid topology, abnormal sections are located and the location results are output, including: Collect the deviation of each lamp on the same feeder, mark lamps with deviation greater than a preset abnormal threshold as abnormal nodes, and trigger a regional power grid abnormal event when the number of abnormal nodes exceeds a preset number. The type of power grid anomaly is inferred based on the direction of change of the thermal characteristic response value of each abnormal node relative to its respective baseline value: when the thermal characteristic response value of an abnormal node increases by more than a preset proportion, it is inferred to be a voltage-to-high anomaly; when the thermal characteristic response value decreases by more than a preset proportion, it is inferred to be a voltage-to-low anomaly. According to the power grid topology, tracing all common upstream nodes of the abnormal nodes along the power supply path upstream, locking a section between the common upstream nodes and the first luminaire in the downstream direction which is not marked as an abnormal node as a suspected abnormal section; reporting the inferred abnormal type and the location of the suspected abnormal section to the power grid dispatch center or the operation and maintenance master station system.

Images