Automatic detection and intelligent classification method and system for storage materials of electrical equipment

Abstract

The invention provides a power equipment storage material automatic detection and intelligent classification method and system. Comprising the following steps: carrying out multi-dimensional parameter initialization acquisition on stored electric power materials, and synchronously obtaining the temperature and humidity gradient of a storage space, the initial oxide layer thickness of the materials, the electromagnetic coupling interference intensity of a stacking area and material parameters of the materials; dynamically calculating the thickness of the oxide layer, constructing a prediction model based on the temperature and humidity gradient and the storage time, and correcting a calculation result in combination with material parameters; performing electromagnetic signal attenuation and interference cooperative compensation, calculating signal attenuation according to the thickness of the oxide layer, and synchronously adjusting filtering parameters to suppress electromagnetic coupling interference; detection confidence calculation and intelligent classification are carried out, a dynamic weight coefficient is introduced, the regulation and control effect of the temperature and humidity gradient and the oxide layer thickness on the interference compensation effect is associated, the temperature and humidity gradient influence can be effectively associated, and the oxide layer thickness prediction precision is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent warehousing, in particular to a power equipment warehousing material automatic detection and intelligent classification method and system. BACKGROUND

[0002] In the management of power equipment warehousing, the oxidation state monitoring of metal power materials such as transformers and circuit breakers is a key link to ensure the quality of material storage and the safety of subsequent operation, and the accurate prediction of the thickness of the oxidation layer directly determines the effectiveness of material classification and control. The existing technology has a key defect in predicting the thickness of the oxidation layer of metal materials: it only relies on the temperature and humidity data collected at a single point in the warehouse environment, and uses a fixed constant temperature and humidity oxidation model for calculation, completely ignoring the key influencing factor of the temperature and humidity gradient existing in the warehouse space.

[0003] In actual warehouse scenarios, the temperature difference between the top and bottom layers of the shelves can reach 6-8℃ due to differences in ventilation conditions and sunlight, and the humidity difference is 12%-15%; the temperature and humidity change rate also differs significantly near the ventilation openings and in the corners, and this temperature and humidity gradient can cause the oxidation reaction rate of different parts of the same batch of materials to differ by more than 20%. The existing fixed model cannot reflect the nonlinear influence of this gradient on the growth of the oxidation layer, directly causing the prediction error of the oxidation layer thickness to exceed 30%, and further causing the analysis of electromagnetic detection signals based on the oxidation layer state to be distorted, ultimately leading to misclassification of the materials, and making it difficult to meet the actual needs of the power industry for accurate monitoring of the oxidation state of warehouse materials.

[0004] Based on the above problems, there is an urgent need for a technical solution that can effectively associate the influence of the temperature and humidity gradient and improve the prediction accuracy of the oxidation layer thickness. SUMMARY

[0005] The purpose of the present application is to solve the shortcomings in the prior art and to propose a power equipment warehousing material automatic detection and intelligent classification method, comprising:

[0006] S1: Multi-dimensional parameter initialization collection is performed on the warehoused power materials, and the temperature and humidity gradient of the warehouse space, the initial oxidation layer thickness of the materials, the electromagnetic coupling interference intensity of the stacking area, and the material quality parameters are simultaneously obtained;

[0007] S2: Dynamic calculation of the oxidation layer thickness is performed, a prediction model is constructed based on the temperature and humidity gradient and the storage time, and the calculation results are corrected in combination with the material quality parameters;

[0008] S3: Electromagnetic signal attenuation and interference are compensated, the signal attenuation amount is calculated according to the oxidation layer thickness, and the filtering parameters are adjusted to suppress electromagnetic coupling interference;

[0009] S4: Performs detection confidence calculation and intelligent classification, introduces dynamic weight coefficients, correlates the effect of temperature and humidity gradient and oxide layer thickness on interference compensation, divides material storage areas according to confidence threshold, and realizes automated detection and intelligent classification with multi-factor linkage.

[0010] Preferably, the multi-dimensional parameter initialization acquisition is performed by collecting three-dimensional temperature and humidity data of the storage space through distributed optical fiber sensors at a grid density of 1m×1m×1m, calculating the temperature and humidity gradient and the standard deviation of ambient humidity, wherein the temperature and humidity gradient is the vector composite value of the rate of change of temperature and humidity in the horizontal direction and the rate of change of temperature and humidity in the vertical direction; the initial oxide layer thickness is obtained by scanning the surface of the material using a laser interferometric thickness gauge, and the basic oxidation coefficient of the material is determined by combining the material type output by the material identification sensor; the electromagnetic coupling interference intensity at three points in the stacking area is collected by a broadband electromagnetic sensor and the average value is taken, and the initial value of the coupling interference compensation coefficient is set according to the number of stacking layers, wherein the coupling interference compensation coefficient is linearly adjusted by increasing by 0.04 for each additional stacking layer.

[0011] More preferably, the dynamic calculation of the oxide layer thickness is performed by the intelligent classification calculation module calling a preset prediction model, substituting temperature and humidity gradient, storage time, environmental humidity standard deviation and material oxidation basic coefficient to calculate the oxide layer thickness; when the difference between the calculated oxide layer thickness and the initial oxide layer thickness exceeds 20μm, the laser interferometric thickness gauge is triggered to perform a secondary scan of the material, and the oxide layer growth saturation correction coefficient is corrected according to the secondary scan result. The value range of the oxide layer growth saturation correction coefficient is determined according to the oxidation characteristics of the material.

[0012] In a further preferred embodiment, the electromagnetic signal attenuation and interference co-compensation is achieved by the electromagnetic coupling interference suppression module setting the initial detection signal frequency according to the material type and acquiring the initial signal strength; the intelligent classification calculation module inputs the oxide layer thickness, detection signal frequency, material permeability, oxide layer dielectric loss factor, and temperature and humidity gradient to calculate the signal attenuation; the adaptive filter bandwidth parameter is adjusted according to the signal attenuation, and the filter bandwidth is reduced by 10% for every 5dB increase in signal attenuation, and the electromagnetic coupling interference strength is reacquired after adjustment.

[0013] More preferably, the prediction model formula used for the dynamic calculation of the oxide layer thickness is:

[0014] ;

[0015] in, This represents the thickness of the oxide layer, in μm. The basic oxidation coefficient represents the material; the value is 0.85 for copper and 1.12 for iron. This represents the storage time of materials, expressed in days (d). represents the temperature and humidity gradient influence coefficient, with a value range of 0.04-0.06; represents the temperature and humidity gradient, with a unit of ℃ / m·%RH / m; represents the environmental humidity standard deviation, with a unit of %RH; represents the oxide layer growth saturation correction coefficient, with a value range of 0.02-0.05.

[0016] Further preferably, the formula used for calculating the electromagnetic signal attenuation amount is:

[0017] ;

[0018] wherein, represents the electromagnetic detection signal attenuation amount, with a unit of dB; represents the material magnetic permeability, with a value of 4500 μH / m for transformer silicon steel and 1.05 μH / m for aluminum alloy; represents the detection signal frequency, with a unit of kHz and a value range of 50-200; represents the oxide layer dielectric loss factor, with a value range of 0.001-0.003; represents the oxide layer thickness, with a unit of μm; represents the initial signal intensity, with a unit of dBm; represents the additional attenuation coefficient of the temperature and humidity gradient on the signal, with a value of 0.015; represents the temperature and humidity gradient, with a unit of ℃ / m·%RH / m.

[0019] Further preferably, the formula used for calculating the detection confidence and classification weight is:

[0020] , ;

[0021] wherein, represents the detection confidence, with a value range of 0-1; represents the electromagnetic detection signal attenuation amount, with a unit of dB; represents the electromagnetic coupling interference intensity, with a unit of dBμV / m; represents the coupling interference compensation coefficient, with a value range of 0.8-1.2, dynamically adjusted according to the stacking layer number; represents the initial signal intensity, with a unit of dBm; represents the dynamic weight coefficient; represents the temperature and humidity gradient, with a unit of ℃ / m·%RH / m; represents the oxide layer thickness, with a unit of μm.

[0022] Further preferably, in the detection confidence calculation and intelligent classification step, when the detection confidence is greater than or equal to 0.95, it is determined that the material belongs to the normal storage area, and the classification weight is set to 1.0; when the detection confidence is greater than or equal to 0.9 and less than 0.95, it is determined that the material belongs to the maintenance area, the classification weight is set to 0.7, and a weekly re-inspection mechanism is triggered; when the detection confidence is less than 0.9, the detection signal frequency ±20 kHz and the coupling interference compensation coefficient ±0.1 are adjusted to recalculate the detection confidence, and if the recalculated detection confidence is still less than 0.9, it is determined that the material belongs to the priority out-of-stock area, and the classification weight is set to 0.3; after the classification execution module completes the material sorting according to the classification weight, the oxide layer thickness, signal attenuation and detection confidence are written into the RFID tag.

[0023] The power equipment warehouse material automatic detection and intelligent classification system is applied to the power equipment warehouse material automatic detection and intelligent classification method described in any one of the above, characterized in that it comprises a temperature and humidity gradient acquisition module, an oxide layer thickness detection module, an electromagnetic coupling interference suppression module, an intelligent classification calculation module and a classification execution module, wherein the temperature and humidity gradient acquisition module and the intelligent classification calculation module are electrically connected, used for acquiring warehouse space temperature and humidity data and transmitting them to the intelligent classification calculation module; the oxide layer thickness detection module and the intelligent classification calculation module are electrically connected, used for acquiring material oxide layer thickness and material information and transmitting them to the intelligent classification calculation module; the electromagnetic coupling interference suppression module and the intelligent classification calculation module are electrically connected, used for acquiring electromagnetic coupling interference intensity and receiving filter parameter adjustment instructions output by the intelligent classification calculation module; the classification execution module and the intelligent classification calculation module are electrically connected, used for receiving classification instructions output by the intelligent classification calculation module and executing sorting operations, and each module realizes real-time interaction of detection parameters and control instructions through electrical connection.

[0024] Further preferably, the temperature and humidity gradient acquisition module comprises a distributed optical fiber sensor and a data acquisition card, the distributed optical fiber sensor is used to acquire three-dimensional temperature and humidity data; the oxide layer thickness detection module comprises a laser interference thickness gauge and a material identification sensor, the laser interference thickness gauge is used to scan the surface of the material to acquire the oxide layer thickness, and the material identification sensor is used to identify the material type; the electromagnetic coupling interference suppression module comprises a wideband electromagnetic sensor and an adaptive filter, the wideband electromagnetic sensor is used to acquire electromagnetic coupling interference intensity, and the adaptive filter is used to suppress interference according to the filter parameter adjustment instruction; the intelligent classification calculation module comprises an edge computing unit, the edge computing unit is used to perform multi-factor linkage calculation; the classification execution module comprises a mechanical arm and an RFID tag writer, the mechanical arm is used to sort materials according to the classification instruction, and the RFID tag writer is used to write detection parameters into the RFID tag.

[0025] Technical effects:

[0026] The technical point of the present application is to construct a multi-factor linkage processing mechanism, synchronously acquire multi-dimensional parameters such as temperature and humidity gradient and initial oxide layer thickness, and correct the calculation of the oxide layer in combination with the temperature and humidity gradient and material parameters, associate the oxide layer thickness with the filtering parameters of the electromagnetic compensation, and introduce the dynamic weight associated with the temperature and humidity gradient and the oxide layer in the confidence calculation. This technical point solves the core problem of isolated processing of multiple factors in the background technology, reduces the prediction error of the oxide layer, improves the quantification accuracy of the electromagnetic signal attenuation and the stability of the interference suppression, reduces the misjudgment rate of the material classification, and meets the precise control demand of the power storage materials. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 The present application is an automatic detection and intelligent classification method for power equipment storage materials.

[0028] Figure 2 The present application is an automatic detection and intelligent classification method for power equipment storage materials. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0030] The traditional power equipment storage detection technology ignores the nonlinear influence of the storage space temperature and humidity gradient on the oxidation of metal parts, does not associate the coupling relationship between the oxide layer thickness and the electromagnetic detection signal attenuation, and lacks a cooperative suppression means for the electromagnetic coupling interference and the previous two factors, resulting in a large oxidation prediction error and a high classification misjudgment rate.

[0031] Based on this, please refer to Figure 1 The present application provides an automatic detection and intelligent classification method for power equipment storage materials, which comprises:

[0032] S1: multi-dimensional parameter initialization collection is performed on the storage power materials, and the storage space temperature and humidity gradient, the initial oxide layer thickness of the materials, the electromagnetic coupling interference intensity of the stacking area and the material quality parameters are synchronously acquired;

[0033] S2: dynamic calculation of the oxide layer thickness is performed, a prediction model is constructed based on the temperature and humidity gradient and the storage time, and the calculation result is corrected in combination with the material quality parameters;

[0034] S3: electromagnetic signal attenuation and interference cooperative compensation are performed, the signal attenuation amount is calculated according to the oxide layer thickness, and the filtering parameters are synchronously adjusted to suppress the electromagnetic coupling interference;

[0035] S4: Perform detection confidence calculation and intelligent classification, introduce dynamic weight coefficient, correlate temperature and humidity gradient and oxidation layer thickness to regulate the interference compensation effect, divide the material storage area according to the confidence threshold, and realize multi-factor linkage automatic detection and intelligent classification.

[0036] The technical scheme takes multi-factor linkage as the core, first acquires key basic data through multi-dimensional parameter initialization collection, wherein the temperature and humidity gradient reflects the spatial change rate of the warehouse space temperature and humidity, the material quality parameter provides material characteristic basis for subsequent oxidation layer calculation, and the stacking electromagnetic coupling interference intensity is the basis input for interference suppression; then, the oxidation layer thickness dynamic calculation link does not use the traditional constant temperature and humidity model, but incorporates the temperature and humidity gradient and storage time into the prediction model, and simultaneously corrects using the material quality parameter to ensure the oxidation layer calculation adaptability of different material goods; the electromagnetic signal attenuation and interference collaborative compensation link breaks through the limitation of traditional fixed filtering parameters, adjusts the filtering parameters based on the signal attenuation calculated by the oxidation layer thickness, realizes the linkage of attenuation compensation and interference suppression; the detection confidence calculation link introduces a dynamic weight coefficient to quantify the influence of the temperature and humidity gradient and the oxidation layer thickness on the interference compensation, so that the confidence calculation can reflect the comprehensive effect of multiple factors, and finally divides the storage area according to the confidence threshold, forming a complete technical link from parameter collection to classification execution. Each step is developed around solving the isolated processing problem of traditional technology to ensure that each technical feature works together.

[0037] During traditional multi-dimensional parameter collection, the low gradient calculation accuracy caused by the unclear temperature and humidity collection grid density, the lack of coordination between oxidation layer thickness and material quality parameter collection, the few sampling points of stacking electromagnetic interference, and the lack of basis for compensation coefficient adjustment result in large initial parameter errors affecting subsequent calculation.

[0038] Therefore, the multi-dimensional parameter initialization collection collects the three-dimensional temperature and humidity data of the warehouse space through a distributed optical fiber sensor with a grid density of 1m x 1m x 1m, calculates the temperature and humidity gradient and the environmental humidity standard deviation, and the temperature and humidity gradient is the vector composition value of the horizontal temperature and humidity change rate and the vertical temperature and humidity change rate; the initial oxidation layer thickness is obtained by scanning the surface of the goods with a laser interferometer, and the material oxidation basic coefficient is determined in combination with the material quality type output by the material quality identification sensor; the electromagnetic coupling interference intensity of three points in the stacking area is collected by a broadband electromagnetic sensor, and the average value is taken, the initial value of the coupling interference compensation coefficient is set according to the stacking layer number, and the coupling interference compensation coefficient is linearly adjusted by 0.04 amplitude for each increase of 1 layer of stacking layer number.

[0039] The technical scheme focuses on the accuracy of initial parameter collection. In the temperature and humidity data collection link, a fixed grid density of 1m x 1m x 1m is used to deploy distributed optical fiber sensors. This density can balance the collection accuracy and efficiency, avoid gradient distortion caused by excessive sparseness, and increase costs caused by excessive density. When calculating the temperature and humidity gradient, the vector synthesis of the horizontal and vertical direction change rate is used instead of simple arithmetic superposition to ensure that the gradient can truly reflect the spatial three-dimensional temperature and humidity difference. At the same time, the environmental humidity standard deviation is collected synchronously to provide a quantitative basis for the subsequent oxidation layer calculation. In the oxidation layer thickness and material parameter collection link, the initial oxidation layer thickness obtained by the laser interference thickness gauge is directly associated with the material type identified by the material identification sensor. According to the oxidation characteristics of different materials, the material oxidation basic coefficient is determined to match the basic coefficient with the actual material quality and avoid errors caused by traditional uniform coefficients. In the pile-up electromagnetic interference collection link, three points are collected to obtain the average value, which can better reflect the overall interference level of the pile-up compared to single-point sampling. The coupling interference compensation coefficient is adjusted linearly by 0.04 for each layer increase to make the coefficient dynamically adapt to the change of the pile-up density and ensure the rationality of the initial compensation coefficient to provide accurate initial parameters for subsequent interference suppression. This scheme improves the initial parameter collection accuracy and provides reliable data basis for subsequent oxidation layer calculation signal processing, reducing source errors.

[0040] In traditional oxidation layer thickness calculation, the prediction model lacks error calibration mechanism after being called. The thickness difference threshold is not clear, which leads to improper triggering time of secondary detection. The oxidation layer growth saturation correction coefficient has no basis, which causes the oxidation layer calculation error to exceed the acceptable range.

[0041] Therefore, the oxidation layer thickness dynamic calculation calls the preset prediction model by the intelligent classification calculation module, and substitutes the temperature and humidity gradient, storage time, environmental humidity standard deviation, and material oxidation basic coefficient to calculate the oxidation layer thickness. When the difference between the calculated oxidation layer thickness and the initial oxidation layer thickness exceeds 20pm, the laser interference thickness gauge is triggered to perform secondary scanning on the materials. According to the secondary scanning result, the oxidation layer growth saturation correction coefficient is corrected. The value range of the oxidation layer growth saturation correction coefficient is determined according to the oxidation characteristics corresponding to the material quality to ensure that the calculation error of the corrected oxidation layer thickness is less than 5%.

[0042] The technical scheme constructs a prediction and calibration closed loop for the calculation of the thickness of the oxidation layer. First, when the intelligent classification calculation module calls the preset prediction model, the four key parameters of temperature and humidity gradient, storage time, environmental humidity standard deviation and material oxidation basic coefficient are clearly substituted. These four parameters correspond to the factors of spatial environment, time accumulation, humidity fluctuation and material characteristics that affect oxidation, ensuring comprehensive model input. Then, set the thickness difference threshold of 20 μm as the secondary detection trigger condition. This threshold is determined based on the oxidation risk assessment of metal parts of power equipment. When the difference exceeds this value, the thickness of the oxidation layer may have affected the performance of the material, and the real thickness needs to be obtained by secondary scanning with a laser interference thickness gauge. After secondary scanning, the calculated value is not simply replaced, but the oxidation layer growth saturation correction coefficient is corrected according to the scanning results. When correcting, the coefficient value range is determined according to the oxidation characteristics of the material quality, for example, the oxidation saturation rate of copper material is different from that of iron material, and the coefficient range is also different. Finally, the corrected oxidation layer thickness calculation error is less than 5%, which can meet the precision requirements of power equipment storage for oxidation layer monitoring, and avoid errors caused by excessive error in subsequent signal processing and classification judgment. The scheme realizes accurate calculation of the thickness of the oxidation layer, controls the error through a closed loop calibration mechanism, and provides accurate thickness parameters for electromagnetic signal attenuation calculation.

[0043] In traditional electromagnetic signal attenuation and interference compensation, the initial detection signal frequency setting has no material adaptability, the signal attenuation amount calculation filter bandwidth adjustment ratio is not clear, and the interference reduction amplitude after compensation has no standard, resulting in poor signal processing effect that cannot meet the detection requirements.

[0044] Therefore, the electromagnetic signal attenuation and interference compensation is set by the electromagnetic coupling interference suppression module according to the material quality type, and the initial detection signal frequency is set. The initial signal intensity is collected. The intelligent classification calculation module substitutes the oxidation layer thickness, detection signal frequency, material permeability, oxidation layer dielectric loss factor and temperature and humidity gradient to calculate the signal attenuation amount. According to the signal attenuation amount, the adaptive filter bandwidth parameters are adjusted. The filter bandwidth is reduced by 10% for every 5 dB increase in signal attenuation amount. After adjustment, the electromagnetic coupling interference intensity is collected again to ensure that the electromagnetic coupling interference intensity after compensation is reduced by 30%-40%.

[0045] The technical scheme forms a frequency adaptation, attenuation calculation, bandwidth adjustment and effect verification link of electromagnetic signal processing, an initial detection signal frequency setting link, an electromagnetic coupling interference suppression module that sets the frequency according to the material quality type, different electromagnetic characteristics of different materials, and the adaptive frequency can reduce the initial signal loss; the initial signal strength collection provides a reference value for subsequent attenuation calculation; when the intelligent classification calculation module calculates the signal attenuation, the thickness of the oxide layer, the detection signal frequency, the material magnetic permeability, the oxide layer dielectric loss factor and the temperature and humidity gradient are substituted into the five parameters, covering material characteristics, signal parameters, oxidation influence and environmental factors, to ensure comprehensive attenuation calculation; the filter bandwidth adjustment adopts a quantitative ratio of 10% bandwidth reduction per 5dB attenuation, avoiding the randomness of traditional experience adjustment, and the ratio is determined through a large number of electromagnetic simulations and experiments, which can effectively compensate for signals of different attenuation degrees; after adjustment, the interference strength is collected again and the reduction amplitude is ensured to be in the 30%-40% range, which can ensure signal integrity while achieving effective interference suppression, avoiding signal distortion caused by excessive filtering or interference caused by insufficient filtering. The scheme realizes accurate processing of electromagnetic signals, effectively suppresses coupling interference, and provides high-quality signal data for detection confidence calculation.

[0046] Traditional oxide layer thickness prediction has no precise quantitative model, and cannot coordinate the influence of temperature and humidity gradient and environmental humidity standard deviation on oxidation, resulting in large deviation between predicted results and actual thickness, and unable to provide reliable basis for subsequent detection.

[0047] Therefore, the prediction model formula used in the dynamic calculation of the oxide layer thickness is:

[0048] ;

[0049] Among them: represents the thickness of the oxide layer, with the unit of μm; represents the material oxidation basic coefficient, with the value of 0.85 for copper material and 1.12 for iron material; represents the material storage time, with the unit of d; represents the temperature and humidity gradient influence coefficient, with the value range of 0.04-0.06; represents the temperature and humidity gradient, with the unit of ℃ / m·%RH / m; represents the environmental humidity standard deviation, with the unit of %RH; represents the oxide layer growth saturation correction coefficient, with the value range of 0.02-0.05, and the formula reflects the influence of warehouse space heterogeneity on oxidation rate through the coordination of temperature and humidity gradient and environmental humidity standard deviation, reducing the prediction error of oxide layer thickness.

[0050] The technical scheme realizes accurate prediction of the thickness of the oxide layer by building a multi-parameter coupled prediction model, and each parameter in the formula has a clear physical meaning and dimension, ensuring the quantitative accuracy of the model. As the output quantity, the actual measurement unit of the oxide layer thickness directly corresponds to μm, which is convenient for comparison with the actual measurement value; According to the specific values of copper, iron and other materials, the differences in the inherent oxidation characteristics of the materials are reflected, and the errors of the traditional uniform coefficient are avoided; In units of d, it conforms to the statistical habit of storage time of warehouse materials, The exponential form reflects the nonlinear characteristics of the oxide layer with time growth, rather than a simple linear relationship; The product term of The product term of The unit of ℃ / m·%RH / m comprehensively reflects the change rate of temperature and humidity in space, The value of 0.04-0.06 is used to quantify the influence of the gradient on oxidation, The exponential form strengthens the nonlinear acceleration effect of the gradient on the oxidation rate; The term The unit of %RH reflects the humidity fluctuation, and the 0.03 coefficient is the influence weight of humidity fluctuation on oxidation determined by experiment, which reflects the additional effect of humidity fluctuation on oxidation; The term is used to correct the saturation trend in the later stage of the growth of the oxide layer, The value of 0.02-0.05 is suitable for the saturation characteristics of different materials, The form reflects the change of the saturation rate with time. The synergistic effect of various parameters enables the model to accurately reflect the oxidation law of different materials under spatial heterogeneity.

[0051] The formula simulates the whole process from the initial growth of the oxide layer to the later saturation trend through the structure of the basic oxidation term plus the saturation correction term. Among them, The basic oxidation term is used to quantify the dynamic process of the oxide layer growth in the early stage under the combined action of material, time, temperature and humidity gradient and humidity fluctuation; The saturation correction term is used to compensate for the saturation trend in the later stage of the growth of the oxide layer due to the slowing down of the oxidation reaction rate of the metal surface, avoiding the problem that the prediction value of the traditional linear model is far beyond the actual thickness in the later stage.

[0052] Specifically, each operation link is disassembled: The exponential design of is not randomly selected, but based on the nonlinear law obtained from a large number of metal oxidation experiments. The metal oxidation rate does not grow at a constant speed. In the early stage, the oxidation reaction is active and the rate is fast. In the later stage, the oxide layer hinders the contact between oxygen and metal, and the rate gradually slows down, but it does not reach complete saturation. The growth curve of 1.2 times can accurately match this fast-to-slow transition process. If 1 times is used, the oxidation thickness in the later stage will be overestimated, and if 1.5 times is used, the oxidation thickness in the early stage will be underestimated; The exponential term is a core improvement addressing the temperature and humidity gradient in storage spaces. A larger temperature and humidity gradient means more significant differences in the temperature and humidity environments of different parts of the material, leading to greater differences in local oxidation rates. Traditional models ignore this gradient effect, resulting in large prediction errors, while the exponential function reflects the accelerating and amplifying effect of the gradient on oxidation—when… When it increases, The value increases non-linearly, demonstrating the physical law that a larger gradient corresponds to a faster overall oxidation rate. The value of 0.04-0.06 was obtained by calibrating the oxidation of copper and iron, two typical electrical materials, within the common storage gradient range of 5℃ / m·%RH / m to 15℃ / m·%RH / m, to ensure that the coefficient can cover the degree of influence of gradient on oxidation in most storage scenarios. This section focuses on humidity fluctuations, an easily overlooked factor, and humidity standard deviation. This reflects the instability of humidity in the storage environment. Greater fluctuations lead to more frequent wetting-drying cycles on the metal surface, accelerating the peeling and regeneration of the oxide layer. The coefficient of 0.03 is derived from statistical analysis of different... The rate of change of the metal oxidation rate was derived from this, showing that for every 1% increase in the standard deviation of humidity (RH), the oxidation rate increases by an average of 0.03 times. This design allows the model to simultaneously cope with the dual effects of static temperature and humidity fluctuations and dynamic humidity fluctuations; in the saturation correction term... The design is based on the fact that in the later stages of oxidation, the oxide layer on the metal surface gradually becomes denser, making it more difficult for oxygen to penetrate, and the oxidation rate tends to stabilize. The growth rate is much slower than that of the basic oxidation term. It can effectively correct the potentially rapid growth deviation that may occur in the basic oxidation term. The difference of 0.02-0.05 corresponds to the saturation characteristics of different materials—iron oxidizes and saturates faster than copper, therefore the corresponding value for iron is... The higher value ensures that the saturation stage predictions for both mainstream materials can accurately match the actual situation.

[0053] This approach uses a precise model to predict oxide layer thickness, reducing prediction errors and providing accurate basic data for subsequent electromagnetic signal processing.

[0054] Traditional electromagnetic signal attenuation calculations do not consider oxide layer thickness and temperature and humidity gradients. They only consider a single factor, which leads to inaccurate quantification of attenuation. This cannot reflect the true signal attenuation under the coupling of multiple factors, affecting the subsequent interference compensation effect.

[0055] Based on this, the formula used to calculate the electromagnetic signal attenuation is:

[0056] ;

[0057] in This represents the attenuation of the electromagnetic detection signal, measured in dB. The magnetic permeability of the material is represented by a value of 4500 μH / m for silicon steel and 1.05 μH / m for aluminum alloy. This represents the frequency of the detected signal, in kHz, and ranges from 50 to 200. This represents the dielectric loss factor of the oxide layer, with a value range of 0.001-0.003. This represents the thickness of the oxide layer, in μm. This represents the initial signal strength, measured in dBm. This represents the additional attenuation coefficient of the signal due to the temperature and humidity gradient, with a value of 0.015. Representing the temperature and humidity gradient, with units of ℃ / m·%RH / m, this formula improves the accuracy of attenuation calculation by linking the oxide layer thickness with the temperature and humidity gradient to quantify the signal and achieve a dual attenuation effect.

[0058] This technical solution constructs a multi-factor coupled signal attenuation quantization model, with each parameter clearly defined and dimensionally consistent, ensuring accurate attenuation calculation. The unit dB conforms to the conventional measurement method of electromagnetic signal attenuation, making it easy to compare with the actual measured attenuation value; Specific values ​​are set according to different materials such as silicon steel and aluminum alloy in transformers. The unit μH / m reflects the material's ability to conduct magnetic fields, which directly affects the signal propagation loss in the material. The unit is kHz and the value ranges from 50 to 200, covering the commonly used frequency range for electromagnetic testing of power equipment. The exponential form reflects the nonlinear effect of frequency on attenuation. Item The value ranges from 0.001 to 0.003, quantifying the effect of the change in the dielectric properties of the oxide layer with frequency on the attenuation. Item The unit is μm, and the 1.5 power form reflects the accelerating effect of the oxide layer thickness on the attenuation, which is consistent with the attenuation law of electromagnetic waves propagating in the oxide layer. Item The unit is dBm, and the coefficient 0.02 is the basic contribution weight of the initial signal strength to the attenuation determined by experiments, reflecting the influence of the initial signal itself on the attenuation. Item The value is 0.015. The unit is ℃ / m·%RH / m, quantifying the additional signal attenuation caused by spatial environmental differences due to temperature and humidity gradients, thus compensating for the shortcomings of traditional models that ignore the influence of environmental gradients. Various parameters are linked to comprehensively reflect the combined effects of material, frequency, oxide layer, initial signal, and environmental gradient on attenuation. This scheme accurately quantifies electromagnetic signal attenuation, providing accurate attenuation data for interference compensation and confidence calculation, thereby improving the accuracy of subsequent detection.

[0059] The formula as a whole adopts a structure that combines a core attenuation term and an additional attenuation term, wherein As the core attenuation factor, we focus on three factors that directly affect signal propagation: material composition, detection frequency, and oxide layer thickness. An additional attenuation term is added to supplement the indirect effects of initial signal strength and temperature / humidity gradient on attenuation, ensuring that no key factors are omitted in the attenuation calculation.

[0060] A detailed analysis of the design logic of each stage: )and The product term is based on the physical laws governing the propagation of electromagnetic waves in metallic materials. Materials with higher magnetic permeability have a stronger ability to attract magnetic fields, resulting in greater signal attenuation due to energy loss from the magnetic field during propagation. The detection frequency... The effect of attenuation is not linear. Although high-frequency signals can improve detection resolution, the higher the frequency, the faster the magnetic field alternation speed and the greater the eddy current loss inside the material. However, when the frequency exceeds a certain threshold, the loss growth rate will slow down. Therefore, 0.8 is used instead of higher powers, which reflects the trend of attenuation increase due to frequency increase and conforms to the actual law of slowing loss growth in the high-frequency band. This index is obtained by calibrating the signal propagation experiment on silicon steel and aluminum alloy in the range of 50kHz to 200kHz. This project is a key improvement on oxide layer thickness. The oxide layer is a non-metallic insulating layer. When electromagnetic waves pass through it, they will be reflected, refracted and absorbed. The greater the thickness, the more significant these losses are, and the loss growth rate accelerates with the increase of thickness. When the oxide layer thickness increases from 10μm to 20μm, the attenuation does not double, but increases by 1.8-2.2 times. The 1.5-power design can accurately match this accelerated attenuation characteristic and avoid the traditional linear thickness attenuation model underestimating the impact of thick oxide layer on signal. The term focuses on the dielectric loss characteristics of the oxide layer. The dielectric constant of the oxide layer increases with increasing frequency, and the dielectric loss also increases rapidly. An exponential function can reflect this law that the higher the frequency, the faster the dielectric loss increases. The value of 0.001-0.003 was obtained by measuring the change in dielectric loss in the frequency band from 50kHz to 200kHz under different oxide layer thicknesses, ensuring that this coefficient can cover the effect of dielectric loss on attenuation under common oxide layer thicknesses; additional attenuation terms are included. The design is based on the initial signal strength. Not only the detection benchmark, but also the signal itself will naturally attenuate during transmission. Through a large number of experiments and statistics, it has been found that this natural attenuation accounts for about 2% of the initial signal strength. Therefore, a coefficient of 0.02 is used to quantify this part of the attenuation to avoid underestimating the overall attenuation value due to ignoring natural attenuation. The section further explains the indirect effects of temperature and humidity gradients—the greater the temperature and humidity gradient, the greater the difference in dielectric constant of the air in different areas of the storage space, and the greater the scattering loss that occurs when electromagnetic waves propagate in a non-uniform dielectric medium. The value of 0.015 is obtained by measuring signal scattering loss under different gradients, ensuring that the coefficient can accurately quantify the additional attenuation of the signal by the environmental gradient, so that the attenuation calculation covers both material factors and environmental factors.

[0061] Traditional detection confidence calculations do not incorporate dynamic weights and cannot correlate the effects of temperature and humidity gradients and oxide layer thickness on interference compensation, resulting in low confidence levels and insufficient classification reliability in high-interference scenarios.

[0062] Based on this, the formula used to calculate the detection confidence and classification weight is:

[0063] , ;

[0064] in This represents the detection confidence level, with a value ranging from 0 to 1. This represents the attenuation of the electromagnetic detection signal, measured in dB. Represents the intensity of electromagnetic coupling interference, with units of dBμV / m; This represents the coupling interference compensation coefficient, with a value ranging from 0.8 to 1.2, which is dynamically adjusted according to the number of stacking layers. This represents the initial signal strength, measured in dBm. Represents dynamic weighting coefficients; Represents the temperature and humidity gradient, with units of ℃ / m·%RH / m; The value represents the oxide layer thickness in μm. This formula uses dynamic weighting coefficients to achieve coordinated control of the temperature and humidity gradient and the oxide layer thickness on interference compensation, ensuring that the detection confidence meets the requirements under high interference scenarios.

[0065] This technical solution constructs a confidence calculation model through the linkage of two formulas, taking into account both interference compensation effect and multi-factor control, and the dimensions and physical meanings of each parameter are clear. The value ranges from 0 to 1, making it easy to intuitively judge the reliability of the test; 1 represents complete reliability. In the first formula... Xiang Jiang The attenuation influence factor is standardized to the 0-1 range. Item This represents the actual interference intensity. According to the number of stacking layers, the compensation strength is dynamically adjusted, The initial signal is the reference, and the combination of the three quantifies the basic influence of interference and compensation on confidence, The first formula represents the basic confidence level; the second formula is the calculation formula of the dynamic weight The S-shaped function is used to make The value is between 0 and 1, which can smoothly reflect The comprehensive influence of , where Reflects the environmental oxidation risk, Standardizes 0.05 quantifies the influence of the difference between the two on the weight, when Large and Small, Tends to 1, enhancing the confidence weight of high-oxidation-risk materials, otherwise reducing the weight, achieving coordinated regulation of temperature and humidity gradient and oxidation layer thickness on confidence. The linkage of the two formulas makes Can comprehensively reflect attenuation, interference, compensation and environmental material factors, ensuring stable confidence in high-interference scenarios. This scheme can still maintain stable detection confidence in high-interference scenarios, improving the reliability of classification and judgment, and avoiding misclassification.

[0066] The formula is divided into two parts: basic confidence calculation and dynamic weight regulation. The two work together to ensure that confidence can reflect the direct influence of signal attenuation and interference, and can also dynamically adjust the evaluation weight in combination with oxidation risk. First look at the basic confidence part : 1 represents a completely reliable confidence state in theory; Is the standardization of signal attenuation , because the signal attenuation in power material electromagnetic detection is usually not more than 100 dB, dividing By 100 can convert the attenuation impact into a coefficient in the range of 0-1. The larger the attenuation, the larger the coefficient, the more the basic confidence is deducted, which intuitively reflects the negative impact of attenuation on detection reliability; Then quantify the influence of electromagnetic coupling interference, The electromagnetic coupling interference strength directly reflects the degree of interference of adjacent materials on the detection signal in the stacking environment, The coupling interference compensation coefficient is dynamically adjusted with the number of stacking layers - the more the number of layers, the stronger the electromagnetic coupling of adjacent materials, The larger the value, the greater the deduction weight of interference on confidence, and divided by The initial signal strength is because the stronger the initial signal, the stronger the anti-interference ability, and the smaller the influence of interference on the detection result. The design enables the quantification of interference influence to combine the interference strength, the stacking density and the signal strength, avoids the one-sidedness of the traditional consideration of only the interference strength, and ensures that the basic confidence can objectively reflect the comprehensive influence of signal quality and interference environment.

[0067] Dynamic weight Calculation formula The formula uses an S-shaped function design, and the core purpose is to dynamically adjust the weight of the basic confidence according to the oxidation risk, so as to avoid using the same confidence standard for different oxidation risk materials. Among them, is a key risk assessment factor, The temperature and humidity gradient reflects the potential oxidation risk faced by the material. The larger the gradient, the faster the potential oxidation rate, and the more easily the state of the material changes; The oxidation layer thickness reflects the actual oxidation state of the material - the greater the thickness, the higher the actual oxidation degree, and Divided by 100 is to standardize its unit to a value similar in magnitude to Avoid distortion of the risk assessment factor due to too large a difference in the values of the two; The 0.5 coefficient in is used to adjust the steepness of the S-shaped function to ensure smooth changes in weight. When is large and is small, the value is large, and the negative exponential term tends to 0, tends to 1, at which time the weight of the basic confidence increases, meaning that such materials with high potential risk but good current state require more accurate confidence assessment to avoid misjudgment; when is small and is large, the value is small, and the negative exponential term increases, tends to 0.5-0.6, at which time the weight decreases, because such materials are relatively stable in state and do not need to amplify the influence of confidence, this dynamic adjustment enables the confidence calculation to meet the actual risk needs of different materials, improving the rationality of classification determination.

[0068] The traditional detection confidence does not have clear storage area division, the re-inspection mechanism trigger condition and parameter adjustment have no standard, and the material detection data is not complete, resulting in chaotic and inefficient classification operation and inability to trace the detection history.

[0069] Based on this, in the detection confidence calculation and intelligent classification step, when the detection confidence is greater than or equal to 0.95, it is determined that the material belongs to the normal storage area, and the classification weight is set to 1.0; when the detection confidence is greater than or equal to 0.9 and less than 0.95, it is determined that the material belongs to the maintenance area, the classification weight is set to 0.7, and a weekly re-inspection mechanism is triggered; when the detection confidence is less than 0.9, the detection signal frequency ±20 kHz and the coupling interference compensation coefficient ±0.1 are adjusted to recalculate the detection confidence, and if the recalculated detection confidence is still less than 0.9, it is determined that the material belongs to the priority out-of-stock area, and the classification weight is set to 0.3; after the classification execution module completes the material sorting according to the classification weight, the oxide layer thickness, signal attenuation and detection confidence are written into the RFID tag, realizing the whole life cycle traceability of the material detection data from storage to out-of-stock.

[0070] The technical scheme constructs a complete classification system of confidence, classification, re-inspection and traceability, the confidence threshold is clear, and the two thresholds of 0.95 and 0.9 are determined based on the risk level of power equipment storage. When greater than or equal to 0.95, the material state is good and belongs to the normal storage area, and the weight 1.0 represents priority normal storage; between 0.9 and 0.95, the material has a slight risk and belongs to the maintenance area, and the weight 0.7 prompts attention, and the weekly re-inspection mechanism can track the material state change in time to avoid risk expansion; when less than 0.9, the material risk is high, and the parameters need to be adjusted for re-detection, and the adjustment range of ±20 kHz frequency and ±0.1 compensation coefficient is an effective adjustment range determined based on signal processing experiments, which can ensure detection efficiency while trying to improve confidence, and still less than 0.9, then belongs to the priority out-of-stock area, and the weight 0.3 prompts priority processing; after classification execution, the oxide layer thickness, signal attenuation and detection confidence are written into the RFID tag, instead of only recording the classification result, ensuring that key detection data can be queried at any link in the whole cycle from storage detection to out-of-stock, realizing data traceability, avoiding the problem of data loss leading to untraceability after traditional classification, and ensuring classification order and efficiency. The scheme standardizes the classification operation, realizes the whole life cycle traceability of the detection data, and improves the classification efficiency and traceability.

[0071] The connection relationship of each module of the traditional power equipment storage detection system is fuzzy, the data transmission and instruction interaction have no standard, the multi-module collaborative work cannot be realized, the system stability is poor, and the linkage detection classification cannot be completed.

[0072] Based on this, please refer to Figure 2The embodiment provides an electric power equipment warehouse material automatic detection and intelligent classification system, which is applied to the electric power equipment warehouse material automatic detection and intelligent classification method in any of the foregoing embodiments. The system comprises a temperature and humidity gradient acquisition module, an oxide layer thickness detection module, an electromagnetic coupling interference suppression module, an intelligent classification calculation module and a classification execution module. The temperature and humidity gradient acquisition module is electrically connected with the intelligent classification calculation module and is used to acquire warehouse space temperature and humidity data and transmit the data to the intelligent classification calculation module. The oxide layer thickness detection module is electrically connected with the intelligent classification calculation module and is used to acquire material oxide layer thickness and material information and transmit the information to the intelligent classification calculation module. The electromagnetic coupling interference suppression module is electrically connected with the intelligent classification calculation module and is used to acquire electromagnetic coupling interference intensity and receive a filtering parameter adjustment instruction output by the intelligent classification calculation module. The classification execution module is electrically connected with the intelligent classification calculation module and is used to receive a classification instruction output by the intelligent classification calculation module and execute a sorting operation. Each module realizes real-time interaction of detection parameters and control instructions through electrical connection.

[0073] The technical scheme clearly defines the system module composition and connection relationship, and builds a collaborative architecture with the intelligent classification calculation module as the core. In terms of module composition, the temperature and humidity gradient acquisition module, the oxide layer thickness detection module and the electromagnetic coupling interference suppression module serve as data acquisition and preprocessing modules, the classification execution module serves as an execution module, and the intelligent classification calculation module serves as a core processing module, with clear division of labor. In terms of connection relationship, all data acquisition and preprocessing modules are electrically connected with the intelligent classification calculation module, rather than directly connected with each other, ensuring that data is uniformly collected to the core module for processing and avoiding processing confusion caused by scattered data. The data and instruction flow is clear, the acquisition module only transmits data to the intelligent classification calculation module, the intelligent classification calculation module only outputs filtering parameter adjustment instructions to the interference suppression module and classification instructions to the classification execution module, forming a one-way instruction flow and a two-way data flow of acquisition, processing, control and execution. Each module realizes real-time interaction through electrical connection, rather than delayed transmission, ensuring that when the acquisition parameters change, the intelligent classification calculation module can process and output new instructions in a timely manner, and the interference suppression module and the classification execution module can respond in a timely manner, avoiding collaborative failure caused by chaotic connection of traditional modules, and ensuring stable operation and linkage function of the system.

[0074] The scheme clearly defines the module connection and interaction logic, realizes the collaborative work of multiple modules, and improves the stability and linkage detection and classification ability of the system.

[0075] The core components of the modules of the traditional detection system are not clear, the functional parameters have no specific standards, and the multi-factor linkage calculation and accurate sorting operation cannot be supported, which affects the overall detection and classification effect due to insufficient module performance.

[0076] Based on this, the temperature and humidity gradient acquisition module includes a distributed optical fiber sensor and a data acquisition card, the acquisition interval of the distributed optical fiber sensor is set to 1 m, which is used to obtain three-dimensional temperature and humidity data; the oxide layer thickness detection module includes a laser interference thickness gauge and a material identification sensor, the laser interference thickness gauge is used to scan the surface of the material to obtain the thickness of the oxide layer, and the material identification sensor is used to identify the material type of the material; the electromagnetic coupling interference suppression module includes a wideband electromagnetic sensor and an adaptive filter, the wideband electromagnetic sensor is used to collect electromagnetic coupling interference intensity covering the frequency band of 50 kHz-200 kHz, and the adaptive filter is used to suppress interference according to the filter parameter adjustment instruction; the intelligent classification calculation module includes an edge computing unit, the edge computing unit carries a TensorFlowLite framework, and is used to perform multi-factor linkage calculation; the classification execution module includes a mechanical arm and an RFID tag writer, the mechanical arm is used to sort materials according to the classification instruction, and the RFID tag writer is used to write detection parameters into the RFID tag.

[0077] The technical scheme clearly defines the core components and function parameters of each module, ensuring that the module performance is adapted to the overall technical requirements. In the temperature and humidity gradient acquisition module, the 1 m acquisition interval of the distributed optical fiber sensor can balance the accuracy and efficiency, and the data acquisition card can ensure the effective conversion and transmission of data, which together support the acquisition of three-dimensional temperature and humidity data; in the oxide layer thickness detection module, the laser interference thickness gauge provides high-precision thickness scanning, and the material identification sensor accurately identifies the material, which together provide basic data for oxide layer calculation; in the electromagnetic coupling interference suppression module, the wideband electromagnetic sensor covers the frequency band of 50 kHz-200 kHz, which matches the frequency range of the detection signal, ensuring that the interference intensity is fully collected, and the adaptive filter can respond to the parameter adjustment instruction to realize dynamic filtering; in the intelligent classification calculation module, the edge computing unit meets the real-time requirements of multi-factor linkage calculation, the TensorFlowLite framework adapts to the operation of lightweight algorithms, ensuring efficient calculation; in the classification execution module, the mechanical arm realizes accurate sorting, and the RFID tag writer ensures effective recording of detection parameters, the function parameters of each component are clear and mutually adapted, avoiding the performance not meeting the standard caused by the ambiguity of traditional module components, and providing hardware support for stable and efficient operation of the overall system. The scheme ensures that the performance of each module meets the standard, provides a reliable hardware foundation for multi-factor linkage detection and accurate classification, and guarantees the landing of the overall technical scheme.

[0078] The above merely describes the preferred embodiments of the present application, and is not intended to limit the present application in other forms. Any skilled person in the art can modify or change the disclosed technical content into equivalent embodiments with equivalent changes, and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present application, without departing from the technical solution content of the present application, still falls within the protection scope of the present application.

Claims

1. A method for automatic detection and intelligent classification of power equipment warehouse materials, characterized in that, The method comprises the following steps: S1: multi-dimensional parameter initialization collection of warehouse power materials, synchronous acquisition of warehouse space temperature and humidity gradient, initial oxidation layer thickness of materials, electromagnetic coupling interference intensity of stacking area and material quality parameters; S2: dynamic calculation of oxidation layer thickness, construction of prediction model based on temperature and humidity gradient and storage time, combination of material quality parameter correction calculation result; S3: electromagnetic signal attenuation and interference collaborative compensation, signal attenuation amount is calculated according to oxidation layer thickness, and electromagnetic coupling interference is suppressed by synchronous adjustment of filtering parameters; S4: detection confidence calculation and intelligent classification, introduction of dynamic weight coefficient, correlation of temperature and humidity gradient and oxidation layer thickness to interference compensation effect regulation, division of material storage area according to confidence threshold, realization of multi-factor linkage automatic detection and intelligent classification.

2. The method for automatic detection and intelligent classification of power equipment warehouse materials according to claim 1, characterized in that, The multi-dimensional parameter initialization collection collects three-dimensional temperature and humidity data of the warehouse space with a grid density of 1m×1m×1m through a distributed optical fiber sensor, calculates the temperature and humidity gradient and the environmental humidity standard deviation, the temperature and humidity gradient is the vector composition value of the horizontal temperature and humidity change rate and the vertical temperature and humidity change rate; the initial oxidation layer thickness is obtained by scanning the surface of the material with a laser interference thickness gauge, and the material oxidation basic coefficient is determined according to the material quality type output by the material quality identification sensor; the electromagnetic coupling interference intensity of three points in the stacking area is collected by a wideband electromagnetic sensor, and the average value is taken, the coupling interference compensation coefficient initial value is set according to the stacking layer number, and the coupling interference compensation coefficient is linearly adjusted by 0.04 amplitude for each increase of 1 layer of stacking layer number.

3. The method of claim 1, wherein the method further comprises: The dynamic calculation of the oxidation layer thickness calls the preset prediction model by the intelligent classification calculation module, and calculates the oxidation layer thickness by substituting the temperature and humidity gradient, storage time, environmental humidity standard deviation and material oxidation basic coefficient; when the difference between the calculated oxidation layer thickness and the initial oxidation layer thickness exceeds 20μm, the laser interference thickness gauge is triggered to scan the material again, and the oxidation layer growth saturation correction coefficient is corrected according to the secondary scanning result, and the value range of the oxidation layer growth saturation correction coefficient is determined according to the oxidation characteristics corresponding to the material quality.

4. The method of claim 1, wherein the method further comprises: The electromagnetic signal attenuation and interference collaborative compensation sets the initial detection signal frequency according to the material quality type by the electromagnetic coupling interference suppression module, and collects the initial signal intensity; the intelligent classification calculation module calculates the signal attenuation amount by substituting the oxidation layer thickness, detection signal frequency, material permeability, oxidation layer dielectric loss factor and temperature and humidity gradient; the filter bandwidth parameter is adjusted according to the signal attenuation amount, and the filter bandwidth is reduced by 10% for each increase of 5dB of the signal attenuation amount, and the electromagnetic coupling interference intensity is collected again after adjustment.

5. The method for power equipment warehouse material automatic detection and intelligent classification according to claim 1, characterized in that, The prediction model formula used in the dynamic calculation of the oxidation layer thickness is: ； wherein, represents the thickness of the oxide layer, in units of pm; represents the oxidation base coefficient of the material, with a value of 0.85 for copper and a value of 1.12 for iron; represents the storage time of the material, in units of days; represents the temperature and humidity gradient influence coefficient, with a value range of 0.04-0.06; represents the temperature and humidity gradient, in units of ℃ / m·%RH / m; represents the standard deviation of the environmental humidity, in units of %RH; represents the saturation correction coefficient of the oxide layer growth, with a value range of 0.02-0.

05.

6. The method for power equipment warehouse material automatic detection and intelligent classification according to claim 1, characterized in that, The formula used in the calculation of the electromagnetic signal attenuation amount is: ； Wherein, representing the electromagnetic detection signal attenuation, unit: dB; representing the material magnetic permeability, transformer silicon steel value 4500 μH / m, aluminum alloy value 1.05 μH / m; representing the detection signal frequency, unit: kHz, value range: 50-200; representing the oxide layer dielectric loss factor, value range: 0.001-0.003; representing the oxide layer thickness, unit: μm; representing the initial signal strength, unit: dBm; representing the additional attenuation coefficient of temperature and humidity gradient on the signal, value: 0.015; representing the temperature and humidity gradient, unit: ℃ / m·%RH / m.

7. The method for power equipment warehouse material automatic detection and intelligent classification according to claim 1, characterized in that, The formula used in the calculation of the detection confidence and classification weight is: 、 ； wherein, represents detection confidence, with a value range of 0-1; represents electromagnetic detection signal attenuation, with a unit of dB; represents electromagnetic coupling interference intensity, with a unit of dBμV / m; represents coupling interference compensation coefficient, with a value range of 0.8-1.2, dynamically adjusted with the number of stacking layers; represents initial signal intensity, with a unit of dBm; represents dynamic weight coefficient; represents temperature and humidity gradient, with a unit of ℃ / m·%RH / m; represents oxide layer thickness, with a unit of μm.

8. The method for power equipment warehouse material automatic detection and intelligent classification according to claim 1, characterized in that, The detection confidence calculation and intelligent classification step, when the detection confidence is greater than or equal to 0.95, it is determined that the material belongs to the normal storage area, and the classification weight is set to 1.0; when the detection confidence is greater than or equal to 0.9 and less than 0.95, it is determined that the material belongs to the maintenance area, and the classification weight is set to 0.7, and a weekly re-inspection mechanism is triggered; when the detection confidence is less than 0.9, the detection signal frequency ±20kHz and the coupling interference compensation coefficient ±0.1 are adjusted to recalculate the detection confidence, and if the recalculated detection confidence is still less than 0.9, it is determined that the material belongs to the priority out-of-stock area, and the classification weight is set to 0.3; After the classification execution module completes the material sorting according to the classification weight, the oxide layer thickness, signal attenuation and detection confidence are written into the RFID tag.

9. The power equipment warehouse material automatic detection and intelligent classification system is applied to the power equipment warehouse material automatic detection and intelligent classification method in any one of claims 1-8, characterized in that, The temperature and humidity gradient acquisition module and the intelligent classification calculation module are electrically connected, used for acquiring warehouse space temperature and humidity data and transmitting to the intelligent classification calculation module; the oxide layer thickness detection module and the intelligent classification calculation module are electrically connected, used for acquiring material oxide layer thickness and material information and transmitting to the intelligent classification calculation module; the electromagnetic coupling interference suppression module and the intelligent classification calculation module are electrically connected, used for acquiring electromagnetic coupling interference intensity and receiving the filter parameter adjustment instruction output by the intelligent classification calculation module; the classification execution module and the intelligent classification calculation module are electrically connected, used for receiving the classification instruction output by the intelligent classification calculation module and executing the sorting operation, and each module realizes real-time interaction of detection parameters and control instructions through electrical connection.

10. The power equipment warehouse material automatic detection and intelligent classification system according to claim 9, characterized in that, The temperature and humidity gradient acquisition module includes a distributed optical fiber sensor and a data acquisition card, the distributed optical fiber sensor is used to obtain three-dimensional temperature and humidity data; the oxide layer thickness detection module includes a laser interference thickness gauge and a material identification sensor, the laser interference thickness gauge is used to scan the surface of the material to obtain the oxide layer thickness, and the material identification sensor is used to identify the material type; the electromagnetic coupling interference suppression module includes a wideband electromagnetic sensor and an adaptive filter, the wideband electromagnetic sensor is used to acquire electromagnetic coupling interference intensity, and the adaptive filter is used to suppress interference according to the filter parameter adjustment instruction; the intelligent classification calculation module includes an edge computing unit, the edge computing unit is used to perform multi-factor linkage calculation; the classification execution module includes a mechanical arm and an RFID tag writer, the mechanical arm is used to sort materials according to the classification instruction, and the RFID tag writer is used to write detection parameters into the RFID tag.

Images