Load type identification method based on circuit breaker intelligent algorithm

Abstract

The invention discloses a load type identification method based on a circuit breaker intelligent algorithm, and the method comprises the steps: collecting various types of electrical parameters, such as transient harmonic signals and temperature drift data, of a load through a detection unit; and a transient harmonic attenuation coefficient, an electromagnetic coupling interference correction coefficient and a load characteristic comprehensive discrimination value are obtained through three-stage progressive characteristic calculation, and are compared with a load type characteristic library to output a result. According to the method, multi-dimensional key influence factors are fused, interference is cooperatively eliminated through a progressive algorithm, the problem that an existing method is inaccurate in identification is solved, accurate identification of the load type is achieved, and the method is suitable for industrial and civil electric power scenes.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of circuit breaker control, and particularly relates to a load type identification method based on a circuit breaker intelligent algorithm. BACKGROUND

[0002] In the operation process of a power system, load type identification is a core prerequisite for the circuit breaker to realize accurate control, fault early warning and optimal configuration of power resources. Existing load type identification methods mostly rely on conventional static electrical parameters such as voltage effective value, current effective value and power factor, and complete identification through simple threshold comparison or a single algorithm model. However, in actual power scenarios, transient harmonic signals are generated during the startup and running stage of the load, the signals contain load essential characteristic information, but are easily affected by temperature changes of the internal conductive loop of the circuit breaker, and temperature drift will cause distortion of the amplitude and attenuation characteristics of the harmonic signals. At the same time, there is an electromagnetic coupling effect between the load and the circuit breaker, and the coupling strength fluctuates with the change of the running state of the load, further interfering with the accuracy of the characteristic parameters. The existing technology does not fully consider the mutual influence among the transient harmonic signals, temperature drift and electromagnetic coupling strength, and only identifies based on conventional static parameters, resulting in insufficient identification accuracy under complex working conditions and failing to meet the high-precision requirements of the power system for load type identification.

[0003] Based on the above problems, there is an urgent need for a technical solution that can integrate multi-dimensional key influencing factors, eliminate interference and accurately identify. SUMMARY

[0004] The purpose of the application is to provide a load type identification method based on a circuit breaker intelligent algorithm, which comprises collecting electrical parameter data in the running process of a load through a detection unit of a circuit breaker, performing feature processing on the collected electrical parameter data, and completing load type identification based on the processed feature parameters. The feature processing comprises: S1: collecting transient harmonic signals, temperature drift data, electromagnetic coupling strength data, voltage effective value, current effective value, power factor and current distortion rate of the load through the detection unit; S2: performing denoising and normalization preprocessing operations on all the collected data; S3: performing three-level progressive feature calculation based on the preprocessed data to obtain a transient harmonic attenuation coefficient, an electromagnetic coupling interference correction coefficient and a load feature comprehensive discriminant value in sequence; S4: comparing the load feature comprehensive discriminant value with a preset load type feature library, and outputting a load type identification result.

[0005] Preferably, the detection unit comprises a harmonic sensor for collecting transient harmonic signals within 0.1s to 1s of the load starting stage, a temperature sensor for collecting temperature drift data of the internal conductive loop of the circuit breaker, and an electromagnetic coupling detector for collecting electromagnetic coupling strength data between the load and the circuit breaker.

[0006] Further preferably, the preprocessing operation comprises denoising the transient harmonic signals by using a wavelet threshold denoising algorithm, and normalizing the temperature drift data, the electromagnetic coupling strength data, the voltage effective value, the current effective value, the power factor, and the current distortion rate to the interval of 0 to 1 by using a maximum-minimum normalization method.

[0007] Further preferably, in the three-level progressive feature calculation, the first-level calculation obtains a transient harmonic attenuation coefficient based on the transient harmonic signals and the temperature drift data, the second-level calculation obtains an electromagnetic coupling interference correction coefficient based on the transient harmonic attenuation coefficient and the electromagnetic coupling strength data, and the third-level calculation obtains a load feature comprehensive discrimination value based on the electromagnetic coupling interference correction coefficient, the voltage effective value, the current effective value, the power factor, and the current distortion rate.

[0008] Further preferably, the transient harmonic attenuation coefficient is obtained by a transient harmonic attenuation coefficient calculation formula, which is: where λ is the transient harmonic attenuation coefficient, dimensionless; H n is the amplitude of the nth transient harmonic, in volts; N is the total number of transient harmonics, dimensionless; t n is the attenuation time of the nth harmonic, in seconds; τ is the harmonic attenuation time constant, in seconds; T d is the duration of the transient process, in seconds; α is the temperature drift influence coefficient, dimensionless; ΔT is the temperature drift data, in degrees Celsius.

[0009] Further preferably, the electromagnetic coupling interference correction coefficient is obtained by an electromagnetic coupling interference correction formula, which is: where μ is the electromagnetic coupling interference correction coefficient, dimensionless; λ is the transient harmonic attenuation coefficient, dimensionless; β is the electromagnetic coupling influence weight coefficient, dimensionless; C s is the electromagnetic coupling strength data, in microtesla; D is the installation distance of the load and the circuit breaker detection unit, in meters; γ is the frequency offset correction coefficient, dimensionless; Δf is the offset of the load operating frequency from the rated frequency, in hertz.

[0010] Further preferably, the load feature comprehensive discriminant value is obtained by a load feature comprehensive discriminant formula, and the load feature comprehensive discriminant formula is: wherein Φ is the load feature comprehensive discriminant value, dimensionless; μ is an electromagnetic coupling interference correction coefficient, dimensionless; k1 is a voltage weight coefficient, dimensionless; U is a normalized voltage effective value, dimensionless; k2 is a current weight coefficient, dimensionless; I is a normalized current effective value, dimensionless; k3 is a power factor weight coefficient, dimensionless; cos φ is a power factor, dimensionless; k4 is a current distortion rate weight coefficient, dimensionless; and THD is a current distortion rate, dimensionless.

[0011] Further preferably, the load type feature library is constructed by collecting load feature comprehensive discriminant values of a plurality of standard load types under different operating conditions, and classifying and clustering all discriminant values by using a K-means clustering algorithm, and the standard load types include resistive load, inductive load, capacitive load and hybrid load.

[0012] Further preferably, the comparison operation calculates the similarity of the load feature comprehensive discriminant value and the center discriminant value of each category in the load type feature library by using a cosine similarity algorithm, and determines the category with the highest similarity as the final load type.

[0013] Further preferably, after outputting the load type identification result, the method further includes a step of verifying the identification result, and the verifying step is to collect the electrical parameter data in the load operation process again, repeat the operations of S2 to S4 to obtain a second load feature comprehensive discriminant value, and if the deviation between the second load feature comprehensive discriminant value and the first load feature comprehensive discriminant value is less than a preset threshold, it is confirmed that the identification result is valid, otherwise the entire identification process is re-executed.

[0014] Compared with the prior art, the present application has the following beneficial effects: The core creative technical point of the present application is the three-level progressive feature calculation and the synergistic fusion of multi-dimensional influencing factors. By collecting key parameters such as transient harmonic signals and temperature drift data, the first level calculation eliminates the interference of temperature drift on transient harmonics, the second level calculation corrects the deviation caused by electromagnetic coupling, and the third level calculation fuses the conventional electrical parameters to form a comprehensive discriminant value. This scheme solves the core problem that the prior art does not consider the mutual interference of multiple factors, realizes the accurate identification of the load type, and provides reliable support for intelligent control of circuit breakers and optimized configuration of power systems. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings needed to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary, and other drawings can be obtained by those skilled in the art without creative effort on the basis of the provided drawings.

[0016] Figure 1 The flow chart of the load type identification method based on the intelligent algorithm of the circuit breaker of the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort belong to the protection scope of the present application.

[0018] The concepts involved in the present application will be described below in combination with the drawings. It should be pointed out here that the descriptions of the concepts below are only for the purpose of making the content of the present application easier to understand, and do not represent the limitation of the protection scope of the present application. Meanwhile, the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the drawings and in combination with the embodiments.

[0019] The conventional load type identification method only relies on the conventional static electrical parameters such as the effective value of voltage and the effective value of current, and does not consider the mutual influence of key factors such as transient harmonic signals, temperature drift, and electromagnetic coupling strength, resulting in insufficient identification accuracy under complex working conditions.

[0020] Based on this, please refer to Figure 1 The present embodiment provides a load type identification method based on an intelligent algorithm of a circuit breaker. The electrical parameter data in the running process of the load is collected by a detection unit of the circuit breaker, the collected electrical parameter data is processed in features, the load type identification is completed based on the processed feature parameters, the feature processing includes four steps S1 to S4: S1 collects the transient harmonic signal, the temperature drift data, the electromagnetic coupling strength data, the effective value of voltage, the effective value of current, the power factor, and the current distortion rate of the load by the detection unit; S2 performs the denoising and normalization preprocessing operation on all the collected data; S3 performs the three-level progressive feature calculation based on the preprocessed data, and sequentially obtains the transient harmonic attenuation coefficient, the electromagnetic coupling interference correction coefficient, and the load feature comprehensive discrimination value; and S4 compares the load feature comprehensive discrimination value with the preset load type feature library, and outputs the load type identification result.

[0021] The core of the technical solution is to construct a complete logic of multi-dimensional data acquisition and three-level progressive feature calculation. In S1, key parameters in the load running process are comprehensively collected. Transient harmonic signals reflect the essential characteristics of the load start-up stage. Temperature drift data reflect the influence of the internal environment of the circuit breaker on the signal. Electromagnetic coupling strength data represent the interaction between the load and the circuit breaker. The conventional electrical parameters provide basic running information. In S2, the noise in the original data is eliminated and the data scale is unified through denoising and normalization processing, laying a foundation for subsequent calculation. The three-level progressive feature calculation in S3 forms the core innovation. The first level calculation combines transient harmonic signals and temperature drift data to eliminate the interference of temperature on harmonic features. The second level calculation integrates electromagnetic coupling strength data based on the results of the previous level to correct the coupling interference. The third level calculation integrates all effective parameters to form a comprehensive discriminant value that can fully reflect the load type. In S4, the precise identification of the load type is realized by comparing with the feature library. Through multi-parameter cooperation and progressive calculation, the defects of the prior art are systematically solved, and the precise identification of the load type is realized.

[0022] Traditional detection units mostly use single or a few sensors, which cannot comprehensively collect multi-dimensional data such as transient harmonics, temperature drift, and electromagnetic coupling strength, resulting in a lack of effective data support for subsequent feature calculation.

[0023] Therefore, the detection unit includes a harmonic sensor, a temperature sensor, an electromagnetic coupling detector, a voltage transformer, and a current transformer. The harmonic sensor is used to collect transient harmonic signals within 0.1s to 1s during the load start-up stage. The temperature sensor is used to collect temperature drift data of the internal conductive loop of the circuit breaker. The electromagnetic coupling detector is used to collect electromagnetic coupling strength data between the load and the circuit breaker.

[0024] This technical solution realizes the precise collection of multi-dimensional key data by specifically configuring sensor types. The harmonic sensor is specially used to collect transient harmonic signals during the load start-up stage. The harmonic signals in this stage are not disturbed by steady-state operation and can more truly reflect the essential characteristics of the load. The collection time window is set within 0.1s to 1s, which ensures the capture of the complete transient process and avoids redundant data. The temperature sensor focuses on the temperature drift data of the internal conductive loop of the circuit breaker. The temperature change at this location directly affects the transmission and detection accuracy of the harmonic signals, and is a necessary interference factor data that must be collected. The electromagnetic coupling detector is specially used to collect electromagnetic coupling strength data between the load and the circuit breaker, accurately representing the degree of electromagnetic interaction between the two. The voltage transformer and the current transformer are used to collect voltage effective value and current effective value respectively, providing data for conventional electrical parameter analysis. Each type of sensor has its own function and cooperates with each other to comprehensively cover the key data dimensions required for load identification, providing sufficient data support for subsequent three-level progressive feature calculation.

[0025] The original data collected contains environmental noise, inconsistent data scales and other problems, which directly affect the accuracy and effectiveness of feature calculation. Traditional preprocessing methods can only solve a single problem and cannot meet the dual requirements of denoising and normalization.

[0026] Therefore, the preprocessing operation includes using a wavelet threshold denoising algorithm to denoise the transient harmonic signal and using a maximum-minimum normalization method to normalize the temperature drift data, electromagnetic coupling strength data, voltage effective value, current effective value, power factor and current distortion rate to the interval of 0 to 1.

[0027] The technical solution adopts targeted preprocessing methods according to the characteristics of different types of data. Transient harmonic signals are easily disturbed by electromagnetic noise in the power environment. The wavelet threshold denoising algorithm has good time-frequency localization characteristics, which can effectively filter out high-frequency noise while preserving the key features of the harmonic signal, ensuring the authenticity of the harmonic data. The dimensions and numerical ranges of temperature drift data and electromagnetic coupling strength data are quite different. If they are directly used for calculation, it will lead to weight imbalance. The maximum-minimum normalization method can map these data to the interval of 0 to 1, which not only preserves the relative differences between data, but also eliminates the dimensional influence, so that different types of data can be effectively fused in subsequent calculations. The cooperative use of the two preprocessing methods comprehensively solves the noise interference and scale inconsistency problems in the original data, providing high-quality data input for the three-level progressive feature calculation.

[0028] Traditional feature calculation methods are mostly single-dimensional calculations, which do not consider the mutual influence between different parameters, cannot effectively eliminate interference factors, and lead to insufficient recognition of feature parameters.

[0029] Therefore, in the three-level progressive feature calculation, the first level calculation obtains the transient harmonic attenuation coefficient based on the transient harmonic signal and the temperature drift data, the second level calculation obtains the electromagnetic coupling interference correction coefficient based on the transient harmonic attenuation coefficient and the electromagnetic coupling strength data, and the third level calculation obtains the load feature comprehensive discriminant value based on the electromagnetic coupling interference correction coefficient, the voltage effective value, the current effective value, the power factor and the current distortion rate.

[0030] The technical scheme constructs layer-by-layer progressive and interrelated feature calculation logic. The first level calculation focuses on the fusion of the transient harmonic signal and the temperature drift data. The transient harmonic signal is a core feature reflecting the load type, but the temperature drift will cause distortion. Through the combination calculation of the two, the transient harmonic attenuation coefficient eliminating the temperature interference is obtained. The second level calculation is based on the first level result, and the electromagnetic coupling strength data is integrated. Electromagnetic coupling is another key interference factor. Through the correction calculation, the electromagnetic coupling interference correction coefficient eliminating the electromagnetic coupling interference is obtained, realizing the secondary elimination of interference factors. The third level calculation integrates the corrected core feature and the conventional electrical parameters. The conventional electrical parameters provide basic information of the load operation, and cooperate with the corrected core feature to form a comprehensive discrimination value which can comprehensively and accurately reflect the load type. The three levels of calculation are layer-by-layer progressive, and the previous level result provides a basis for the next level calculation. Each level of calculation eliminates a type of interference factor, and finally realizes the accurate extraction of the feature parameter.

[0031] The traditional transient harmonic feature calculation does not consider the influence of temperature drift, so that the calculation result cannot truly reflect the harmonic characteristics of the load. Therefore, a precise calculation method capable of fusing the temperature drift factor is needed.

[0032] Therefore, the transient harmonic attenuation coefficient is obtained through a transient harmonic attenuation coefficient calculation formula, and the transient harmonic attenuation coefficient calculation formula is: ; Wherein, λ is the transient harmonic attenuation coefficient, dimensionless; Hn is the amplitude of the nth transient harmonic, unit: volt; N is the total number of transient harmonics, dimensionless; tn is the attenuation time of the nth harmonic, unit: second; τ is the harmonic attenuation time constant, unit: second; Td is the duration of the transient process, unit: second; α is the temperature drift influence coefficient, dimensionless; ΔT is the temperature drift data, unit: degree Celsius.

[0033] The core design goal of the formula is to quantify the attenuation characteristics of the transient harmonic in the load starting stage and eliminate the interference of the temperature drift on the harmonic feature. The theoretical basis is derived from the physical attenuation law of the transient harmonic in the power electronics field and the influence mechanism of the temperature on the conductive loop. The derivation process follows the logical chain of original feature extraction, interference factor quantization and normalization correction.

[0034] From a theoretical basis, the transient harmonic containing rich spectral information will be generated when the load is started, and the amplitude of such harmonic presents an exponential decay trend over time, which is determined by the equivalent impedance characteristics of the load itself - the inductive elements of inductive load and the capacitive elements of capacitive load will form energy storage and release process at the moment of power-on, resulting in the gradual trend of harmonic amplitude to steady state. The existing technology only simply collects the harmonic amplitude, without considering its decay dynamic characteristics, resulting in insufficient feature recognition, while the formula numerator part accurately captures this dynamic process through summation operation. In the summation term, represents the amplitude of the nth transient harmonic, and the difference in amplitude of different order harmonics directly reflects the spectral characteristics of the load, for example, the transient harmonic amplitude of resistive load decays quickly and the proportion of high-order harmonic is low, while the inductive load is the opposite; is the exponential decay term, as the harmonic decay time constant, determined by the ratio of equivalent inductance and resistance of the load, is the inherent property of the load, is the decay time of the nth harmonic, and the two together characterize the decay rate of a single harmonic; summing the product of the "amplitude-decay rate" of all harmonics can fully represent the overall decay profile of the transient harmonic, avoiding the one-sidedness of single harmonic features.

[0035] The design of the denominator part specifically solves the interference problem of temperature drift, which is one of the core defects of the existing technology. The temperature change of the internal conductive circuit of the circuit breaker will cause the resistance value to change, and then affect the transmission and detection accuracy of the harmonic signal. The greater the temperature drift, the more serious the distortion of the harmonic feature. is the duration of the transient process, which is used to normalize the summation result of the numerator in the time dimension, ensuring the comparability of the harmonic decay coefficients of loads with different transient durations - if not normalized, the summation result of the long transient process will naturally be larger, and cannot truly reflect the difference in decay characteristics. is the temperature drift correction factor, where directly collects the temperature change data of the internal conductive circuit of the circuit breaker, is the temperature drift influence coefficient, which is determined by a large number of experiments: under standard temperature environment, multiple transient harmonic collections are performed on different types of loads, the environmental temperature is gradually changed and the change amount of harmonic decay characteristics is recorded, the offset proportion of harmonic decay coefficient when the temperature changes by 1°C is determined through linear regression analysis, and the reasonable value range of is finally determined. The physical meaning of this correction factor is that the distortion degree of harmonic feature caused by temperature drift is approximately linearly related to the temperature change, and through this factor, this distortion can be inversely offset, so that the calculated is closer to the real transient harmonic decay characteristics of the load under standard temperature.

[0036] The derivation logic of the whole formula is progressive: first, the dynamic attenuation characteristics of transient harmonics are extracted through the numerator part, then the time dimension difference and temperature drift interference are eliminated through the denominator part, and finally the transient harmonic attenuation coefficient which can accurately reflect the essential characteristics of the load is obtained. This design not only conforms to the physical propagation law of transient harmonics, but also specifically solves the pain point of temperature interference which is not solved by the existing technology. Its calculation form has clear theoretical support and technical improvement rationality. The specific values of and can completely realize the calculation process through conventional experimental methods according to the type and use environment of the load.

[0037] The theoretical design basis of the formula is the attenuation characteristics of transient harmonics and the interference mechanism of temperature drift. The numerator part accurately represents the overall attenuation process of transient harmonics by summing the amplitude and attenuation index terms of each transient harmonic, where the index term represents the attenuation law of harmonic amplitude with time, and τ is the attenuation time constant determined by the characteristics of the load itself. The denominator part introduces a temperature drift correction term (1+α·ΔT), where α is the temperature drift influence coefficient obtained through experiments, and ΔT is the temperature drift data. This correction term can quantify the influence of temperature change on the attenuation characteristics of harmonics, and T d is the duration of the transient process, which is used to normalize the influence of time dimension. The whole formula represents the original harmonic attenuation characteristics through the numerator, eliminates the temperature drift interference through the denominator, and realizes the accurate calculation of the transient harmonic attenuation coefficient, providing a reliable foundation for subsequent feature calculation.

[0038] The transient harmonic attenuation coefficient is still disturbed by electromagnetic coupling. The traditional calculation method does not consider the comprehensive influence of electromagnetic coupling strength, installation distance, frequency offset and other factors, resulting in deviation of the characteristic parameters.

[0039] Therefore, the electromagnetic coupling interference correction coefficient is obtained through the electromagnetic coupling interference correction formula, which is: where μ is the electromagnetic coupling interference correction coefficient, dimensionless; λ is the transient harmonic attenuation coefficient, dimensionless; β is the electromagnetic coupling influence weight coefficient, dimensionless; C s is the electromagnetic coupling strength data, unit: microtesla; D is the installation distance between the load and the circuit breaker detection unit, unit: meter; γ is the frequency offset correction coefficient, dimensionless; Δf is the offset of the load operating frequency and the rated frequency, unit: hertz.

[0040] The core function of this formula is to calculate the transient harmonic attenuation coefficient On the basis of the above, further eliminate the electromagnetic coupling interference on the characteristic parameters, its theoretical design basis from the electromagnetic radiation coupling principle and the frequency characteristics of the load operation in power system, the derivation process around the "interference factor identification-interference intensity quantification-correction factor construction", is the depth optimization of the previous level characteristic parameters.

[0041] From the theoretical basis analysis, there is an electromagnetic coupling effect between the circuit breaker and the load, which is caused by the alternating magnetic field generated by the current through the conductor of the two, the alternating magnetic field will penetrate each other and affect the electric signal detection of the other party, which is another key factor leading to the distortion of the load characteristic parameters. The existing technology does not consider this interference, resulting in limited identification accuracy. The formula is based on the load characteristic parameters that have eliminated temperature interference , through the correction term to quantify and offset the electromagnetic coupling interference, the core logic is "basic characteristic parameters x interference offset factor", to ensure that the corrected is closer to the real characteristics of the load.

[0042] The design inside the correction term fully reflects the coordinated consideration of multiple factors of electromagnetic coupling interference. First, The design basis of the term is the inverse square law of electromagnetic radiation-the electromagnetic coupling strength is inversely proportional to the square of the distance from the radiation source to the receiving end, which is a basic law of electromagnetism. Directly collect the electromagnetic coupling strength data between the load and the circuit breaker, the larger the value, the stronger the interference; is the installation distance of the load and the circuit breaker detection unit, the closer the distance, the higher the coupling strength, the more serious the interference to the harmonic characteristics, therefore, the reciprocal of the square of the distance is used to quantify this distance dependence. For example, when the installation distance increases from 0.5 meters to 1 meter, the square of the distance becomes 4 times the original, the influence of the coupling strength will decrease to 1 / 4, this quantitative relationship accurately reflects the propagation characteristics of electromagnetic coupling.

[0043] Secondly, The term is a supplementary correction for the interference of load operating frequency deviation. The load may deviate from the rated frequency due to the change of working condition in actual operation, and the resonance frequency of electromagnetic coupling is related to the load operating frequency, so the frequency deviation will cause the coupling resonance strength to change, and further aggravate the interference. is the deviation of the load operating frequency from the rated frequency, is the frequency deviation correction coefficient, its calibration logic is similar to : by collecting the relationship data of electromagnetic coupling strength and characteristic parameter distortion under different frequency deviation, the change ratio of interference strength when the frequency deviates 1 hertz is determined, to ensure that the factor can accurately quantify the influence of frequency deviation on coupling interference.

[0044] As the electromagnetic coupling influence weight coefficient, its role is to balance the comprehensive interference contribution of the three factors 、 、 Due to the different influence degrees of the three interference factors in different application scenarios, for example, the influence of load density is more significant in industrial scenarios, and frequency offset may be more common in civil scenarios, the value range of is calibrated through experiments, so that the correction term can adapt to the interference characteristics of different scenarios. The overall correction term adopts the form of "1-interference quantization value", because the larger the interference quantization value, the more serious the damage to the characteristic parameters, and the higher the proportion that needs to be offset. When the interference quantization value is 0, the correction term is 1, which conforms to the ideal situation without interference.

[0045] The theoretical design basis of the formula is the correlation characteristics of the interference law and influencing factors of electromagnetic coupling. Based on the transient harmonic attenuation coefficient λ, the electromagnetic coupling interference is eliminated through the correction term [1-β·C s / (D²)·(1+γ·Δf)]. C s represents the strength of electromagnetic coupling, and the larger the value, the stronger the interference; D is the installation distance, and the electromagnetic coupling strength is inversely proportional to the square of the distance, so the 1 / D² term is introduced; Δf is the frequency offset, which will exacerbate the electromagnetic coupling interference, and the (1+γ·Δf) term quantifies this influence; β and γ are weight coefficients, which are determined through experimental calibration. The correction term accurately calculates the interference degree of electromagnetic coupling on harmonic characteristics by comprehensively considering the three key factors of electromagnetic coupling strength, installation distance, and frequency offset, and modifies λ to obtain the electromagnetic coupling interference correction coefficient μ that eliminates electromagnetic coupling interference, further improving the accuracy of the characteristic parameters.

[0046] A single characteristic parameter or simple parameter superposition cannot fully reflect the load type, and a calculation method that can fuse multi-dimensional effective parameters and realize comprehensive characterization of load characteristics is needed.

[0047] Therefore, the load characteristic comprehensive discriminant value is obtained through a load characteristic comprehensive discriminant formula, and the load characteristic comprehensive discriminant formula is: wherein Φ is the load characteristic comprehensive discriminant value, dimensionless; μ is the electromagnetic coupling interference correction coefficient, dimensionless; k1 is the voltage weight coefficient, dimensionless; U is the normalized voltage effective value, dimensionless; k2 is the current weight coefficient, dimensionless; I is the normalized current effective value, dimensionless; k3 is the power factor weight coefficient, dimensionless; cosφ is the power factor, dimensionless; k4 is the current distortion rate weight coefficient, dimensionless; and THD is the current distortion rate, dimensionless.

[0048] The core objective of the formula is to integrate multi-dimensional effective feature parameters to form a comprehensive discriminant index that can accurately distinguish different load types. The theoretical basis comes from the sensitivity differences of different load types to various electrical parameters. The derivation process follows the logic of core feature dominance, auxiliary feature supplementation, and weighted fusion optimization, solving the technical problem that a single feature parameter cannot fully represent the load type.

[0049] From a theoretical basis, the essential differences between different load types are reflected in multiple electrical parameter dimensions: resistive loads have a power factor close to 1, low current distortion rate, and fast transient harmonic decay; inductive loads have a low power factor and slow transient harmonic decay; capacitive loads have a capacitive power factor and moderate current distortion rate; and hybrid loads exhibit mixed characteristics of multiple parameters. Existing technologies rely only on a single or a few parameters, which cannot cover these differences. However, the formula achieves comprehensive representation of load types through multi-parameter weighted fusion.

[0050] First part of the formula It is the core basis of the comprehensive discriminant value. As the core feature corrected by temperature and electromagnetic coupling interference, it can reflect the essential harmonic characteristics of the load. By taking it as the dominant factor, the core discriminability of the comprehensive discriminant value is ensured. The term is used to integrate the synergistic characteristics of voltage and current effective values. Voltage and current are the basic electrical parameters of load operation, but their individual effects cannot effectively distinguish load types. For example, resistive loads and inductive loads may have the same voltage effective value, but the current variation is different. The form of square sum and square root is used because the contribution of voltage and current to load power is in a square relationship. This form can accurately represent the comprehensive power-related characteristics of both; , are the weight coefficients of voltage and current. Their calibration is based on the sensitivity differences of different load types to voltage and current: by collecting a large amount of voltage and current data of standard load types, the contribution of both to load type identification is analyzed. For example, the current of a resistive load is linearly related to the voltage, and can be set to be approximately equal. The current of an inductive load lags behind the voltage, and the contribution of current to load characteristics is more significant, so the value of can be appropriately increased to ensure that this part can highlight the voltage and current synergistic characteristic differences of different load types.

[0051] Second part of the formula Focuses on the distinguishing role of power factor . Power factor is a key parameter that reflects the impedance properties of the load. The of a resistive load is close to 1, and the For inductive loads with a capacitance value less than 1, It is capacitive and less than 1, which is the core criterion for distinguishing the three types of basic loads. This is the power factor weighting coefficient. Its calibration logic is based on experimental analysis of the impact of the power factor on load type misjudgment. For example, when the power factor is close to 1, the probability of misjudging it as a non-resistive load is extremely low, therefore it is assigned a weighting factor of 1. Appropriate weighting ensures that this parameter effectively dominates the differentiation of basic load types.

[0052] Formula Part 3 For current distortion rate The supplementary role of current distortion rate. Current distortion rate reflects the degree of nonlinearity of the load; for purely resistive loads... Extremely low for inductive and capacitive loads. Medium-duty, mixed-type loads (such as loads containing nonlinear elements) The higher the value, the more important it is to distinguish between mixed loads and single-type loads. This is the current distortion rate weighting coefficient, and its calibration is based on different load types. Distribution differences, such as mixed loads Typically more than 50% higher than a single type of load, by assigning Appropriate weighting ensures that this parameter can effectively identify mixed loads.

[0053] The theoretical basis of this formula is the correlation between load type and multi-dimensional electrical parameters. The first part of the formula... The formula integrates the corrected core feature μ with the effective values ​​of voltage and current. Voltage and current are fundamental parameters for load operation, and their combined effect is represented by the sum of squares and the square root. k1 and k2 are weighting coefficients, calibrated according to the characteristics of different load types. The second part, k3·cosφ, represents the power factor, a key parameter for distinguishing resistive, inductive, and capacitive loads, and its influence is assigned an appropriate weight by the weighting coefficient k3. The third part, k4·THD, reflects the nonlinear characteristics of the load and is of great significance for identifying mixed loads; its role is reflected by the weighting coefficient k4. The entire formula comprehensively covers the essential characteristics of the load by weighted integration of the corrected core feature and conventional electrical parameters, forming a comprehensive discrimination value Φ that can accurately distinguish different load types.

[0054] Traditional load type feature libraries are mostly built based on empirical thresholds, lacking coverage of different operating conditions, resulting in insufficient adaptability and accuracy of the feature libraries.

[0055] Based on this, the load type feature library is constructed by collecting comprehensive discrimination values ​​of load characteristics of various standard load types under different operating conditions, and then classifying and clustering all discrimination values ​​using the K-means clustering algorithm. The standard load types include resistive loads, capacitive loads, inductive loads, and mixed loads.

[0056] This technical solution constructs a widely adaptable load type feature library. First, data is collected for four standard load types: resistive loads, inductive loads, capacitive loads, and mixed loads, covering common load types in power systems. Data for each standard load type is collected under different operating conditions, including varying load power, ambient temperature, and operating frequency, ensuring that the collected comprehensive load characteristic discrimination values ​​fully reflect the load's characteristics in actual operation. The K-means clustering algorithm is used to classify and cluster all collected discrimination values. This algorithm automatically identifies cluster centers in the data, grouping discrimination values ​​belonging to the same load type together to form feature intervals for each category. The feature library constructed using this method not only covers common load types but also adapts to changes in different operating conditions, providing an accurate and reliable reference for subsequent comparison and identification.

[0057] Traditional comparison methods often rely on simple threshold judgments, which cannot effectively distinguish the feature differences of similar load types, resulting in low recognition accuracy.

[0058] Based on this, the comparison operation uses the cosine similarity algorithm to calculate the similarity between the comprehensive discriminant value of the load features and the discriminant values ​​of each category center in the load type feature library, and determines the category with the highest similarity as the final load type.

[0059] This technical solution achieves accurate comparison using a cosine similarity algorithm. The cosine similarity algorithm quantifies the cosine of the angle between two vectors, reflecting their degree of similarity. The value ranges from -1 to 1, with values ​​closer to 1 indicating higher similarity. During the comparison process, the comprehensive discriminant value of the load features is used as the vector to be identified, and the center discriminant values ​​of each category in the load type feature library are used as reference vectors. The cosine similarity between the vector to be identified and each reference vector is calculated. Since the center discriminant values ​​of different load types differ significantly, similarity calculation can accurately determine the degree of matching between the load to be identified and various standard loads, identifying the category with the highest similarity as the final load type. Compared to traditional threshold judgment, this comparison method can more accurately identify differences between similar load types, effectively improving the accuracy of load type identification.

[0060] A single identification process may lead to deviations in identification results due to accidental factors, and the lack of an effective verification mechanism makes it impossible to guarantee the reliability of the identification results.

[0061] Based on this, after outputting the load type identification result, the method further includes a step of verifying the identification result. The verification step involves collecting electrical parameter data during the load operation process again, repeating the operations from S2 to S4 to obtain a secondary load feature comprehensive discrimination value. If the deviation between the secondary load feature comprehensive discrimination value and the first obtained load feature comprehensive discrimination value is less than a preset threshold, the identification result is confirmed to be valid; otherwise, the entire identification process is re-executed.

[0062] This technical solution ensures the reliability of the identification results through a secondary verification mechanism. After the initial identification output, the electrical parameter data of the load is collected again, and the preprocessing and three-level progressive feature calculation steps are repeated to obtain a secondary comprehensive discriminant value of the load characteristics. The deviation between the two discriminant values ​​is calculated. If the deviation is less than a preset threshold, it indicates that the two identification results are consistent, the initial identification result is less affected by accidental factors, and the identification result is confirmed to be valid. If the deviation is greater than or equal to the preset threshold, it indicates that the initial identification result may have errors, and the entire identification process needs to be repeated until the deviation between the two identification results meets the requirements. This verification mechanism can effectively filter out identification errors caused by accidental factors, significantly improve the reliability and stability of load type identification results, and provide a reliable decision-making basis for subsequent circuit breaker control and power system optimization.

[0063] The embodiments and / or implementation methods described above are merely preferred embodiments and / or implementation methods for implementing the technology of the present invention, and are not intended to limit the implementation methods of the technology of the present invention in any way. Any person skilled in the art can make some modifications or alterations to other equivalent embodiments without departing from the scope of the technical means disclosed in the content of the present invention, but they should still be regarded as the technology or embodiments that are substantially the same as the present invention.

[0064] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are only preferred embodiments of this application. It should be noted that due to the limitations of written expression, while there are objectively infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner. These improvements, modifications, changes, or combinations, or the direct application of the inventive concept and technical solution to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. A load type identification method based on a circuit breaker intelligent algorithm, comprising collecting electric parameter data in a load operation process through a detection unit of a circuit breaker, performing feature processing on the collected electric parameter data, and completing load type identification based on the processed feature parameters, characterized in that, The feature processing comprises: S1: collecting transient harmonic signals, temperature drift data, electromagnetic coupling strength data, voltage effective value, current effective value, power factor and current distortion rate of the load by the detection unit; S2: performing denoising and normalization preprocessing operation on all collected data; S3: performing three-level progressive feature calculation based on the preprocessed data to obtain transient harmonic attenuation coefficient, electromagnetic coupling interference correction coefficient and load feature comprehensive discriminant value in sequence; S4: comparing the load feature comprehensive discriminant value with a preset load type feature library to output a load type recognition result.

2. The load type identification method based on intelligent algorithm of circuit breaker according to claim 1, characterized in that, The detection unit comprises a harmonic sensor, a temperature sensor, an electromagnetic coupling detector, a voltage transformer and a current transformer, the harmonic sensor is used to collect transient harmonic signals within 0.1s to 1s in the load starting stage, the temperature sensor is used to collect temperature drift data of the internal conductive loop of the circuit breaker, and the electromagnetic coupling detector is used to collect electromagnetic coupling strength data between the load and the circuit breaker.

3. The load type identification method based on intelligent algorithm of circuit breaker as claimed in claim 1, wherein, The preprocessing operation comprises denoising the transient harmonic signals by using a wavelet threshold denoising algorithm and normalizing the temperature drift data, electromagnetic coupling strength data, voltage effective value, current effective value, power factor and current distortion rate to the interval of 0 to 1 by using a maximum-minimum normalization method.

4. The load type identification method based on intelligent algorithm of circuit breaker as claimed in claim 1, wherein, In the three-level progressive feature calculation, the first-level calculation obtains the transient harmonic attenuation coefficient based on the transient harmonic signals and the temperature drift data, the second-level calculation obtains the electromagnetic coupling interference correction coefficient based on the transient harmonic attenuation coefficient and the electromagnetic coupling strength data, and the third-level calculation obtains the load feature comprehensive discriminant value based on the electromagnetic coupling interference correction coefficient, the voltage effective value, the current effective value, the power factor and the current distortion rate.

5. The load type identification method based on intelligent algorithm of circuit breaker as claimed in claim 4, wherein, The transient harmonic attenuation coefficient is obtained by a transient harmonic attenuation coefficient calculation formula, the transient harmonic attenuation coefficient calculation formula is: where λ is the transient harmonic decay coefficient, dimensionless; H n is the amplitude of the nth transient harmonic, in volts; N is the total number of transient harmonics, dimensionless; t n is the decay time of the nth harmonic, in seconds; τ is the harmonic decay time constant, in seconds; T d is the duration of the transient, in seconds; α is the temperature drift influence coefficient, dimensionless; ΔT is the temperature drift data, in degrees Celsius.

6. The load type identification method based on intelligent algorithm of circuit breaker as claimed in claim 4, wherein, The electromagnetic coupling interference correction coefficient is obtained by an electromagnetic coupling interference correction formula, the electromagnetic coupling interference correction formula is: wherein μ is an electromagnetic coupling interference correction coefficient, dimensionless; λ is a transient harmonic attenuation coefficient, dimensionless; β is an electromagnetic coupling influence weight coefficient, dimensionless; C s is the electromagnetic coupling strength data, with the unit of microtesla; D is the installation distance of the load and circuit breaker detection unit, with the unit of meters; γ is a frequency offset correction coefficient, dimensionless; Δf is the offset of the load operating frequency and the rated frequency, with the unit of hertz.

7. The load type identification method based on intelligent algorithm of circuit breaker as claimed in claim 4, wherein, The load feature comprehensive discriminant value is obtained by a load feature comprehensive discriminant formula, the load feature comprehensive discriminant formula is: wherein Φ is the load feature comprehensive discriminant value, dimensionless; μ is the electromagnetic coupling interference correction coefficient, dimensionless; k1 is the voltage weight coefficient, dimensionless; U is the normalized voltage effective value, dimensionless; k2 is the current weight coefficient, dimensionless; I is the normalized current effective value, dimensionless; k3 is the power factor weight coefficient, dimensionless; cosφ is the power factor, dimensionless; k4 is the current distortion rate weight coefficient, dimensionless; and THD is the current distortion rate, dimensionless.

8. The load type identification method based on intelligent algorithm of circuit breaker as claimed in claim 1, wherein, The load type feature library is constructed by collecting load feature comprehensive discriminant values of multiple standard load types under different operating conditions, and classifying and clustering all discriminant values by using a K-means clustering algorithm, the standard load types comprise resistive load, inductive load, capacitive load and hybrid load.

9. The load type identification method based on intelligent algorithm of circuit breaker as claimed in claim 1, wherein, The comparison operation calculates the similarity of the load feature comprehensive discriminant value and the center discriminant value of each category in the load type feature library by using a cosine similarity algorithm, and determines the category with the highest similarity as the final load type.

10. The load type identification method based on intelligent algorithm of circuit breaker as claimed in claim 1, wherein, After the load type identification result is output, the method further includes a verification step of verifying the identification result. The verification step is to collect the electrical parameter data during the load operation again, repeat the operations of S2 to S4 to obtain a secondary load feature comprehensive discriminant value, and if the deviation between the secondary load feature comprehensive discriminant value and the load feature comprehensive discriminant value obtained for the first time is less than a preset threshold, it is confirmed that the identification result is valid, otherwise the entire identification process is re-executed.

Images