Water supply equipment electrical performance detection system based on artificial intelligence

Abstract

The invention relates to the technical field of electrical performance testing, in particular to a water supply equipment electrical performance detection system based on artificial intelligence, which comprises an electrical parameter acquisition module, an energy efficiency analysis module, an insulation performance evaluation module and a comprehensive diagnosis module. According to the invention, a three-phase voltage and current root-mean-square value is dynamically calculated through a broadband current transformer and a sliding window integration method, a 50Hz fundamental component amplitude is extracted to separate power frequency and high frequency harmonic waves, a continuous period energy efficiency ratio standard deviation triggers an insulation detection threshold value, and a parameter fluctuation and insulation degradation response mechanism is established. Harmonic interference correction synchronous processing energy efficiency ratio and leakage current integration, the detection error being lower than 5%, temperature rise rate linear regression and bearing current threshold value cooperative determination, winding aging cross validation, fault misjudgment rate reduction to 26%, closed-loop detection fusion transient capture, dynamic threshold value triggering and multi-parameter diagnosis, early warning success rate reaching 92%, and positioning response being reduced to 3 seconds.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical performance testing, and particularly to an electrical performance detection system for water supply equipment based on artificial intelligence. Background Art

[0002] The technical field of electrical performance testing includes the detection, analysis, and evaluation of electrical parameters of various electrical equipment and systems under different working conditions. The core content of this technical field is to obtain electrical performance indicators such as voltage, current, resistance, insulation performance, and power factor in the electrical system through special detection equipment and methods to judge the electrical state and safety performance of equipment operation. Generally speaking, the technical field of electrical performance testing covers power system monitoring, electrical equipment testing, insulation performance measurement, electrical safety detection, and related data acquisition and analysis methods, and is widely used in industries such as power, transportation, industrial control, and water conservancy to ensure the stability and reliability of electrical equipment.

[0003] Among them, the electrical performance detection system for water supply equipment refers to a device or combined structure that measures the performance parameters of electrical components and circuits in the water supply system and diagnoses the state when the identification state is false. The technical matters targeted by this patent theme cover the detection of the working states of key electrical components such as electric pumps, controllers, relays, and frequency converters in water supply equipment. Specifically, by configuring an electrical parameter acquisition circuit, setting voltage and current detection interfaces, integrating an impedance measurement unit, and establishing data analysis rules, the electrical characteristics of the power supply path and load are analyzed and identified. Such systems usually complete the detection work by means of analog sampling, DC impedance measurement, voltage sag monitoring, and control signal comparison to ensure the rationality and stability of the electrical performance of water supply equipment under normal working conditions.

[0004] The prior art uses fixed-period sampling and static impedance measurement, and cannot capture the transient voltage sag and current harmonic distortion caused by load mutations. For example, the instantaneous overcurrent during the start-stop stage of an electric pump is not covered by the detection window, resulting in an energy efficiency evaluation deviation exceeding 15%. The single-parameter threshold determination is not related to the ambient temperature rise factor, and the leakage current detection of cable joints frequently gives false alarms during high-temperature periods. The off-line insulation impedance detection is separated from the actual operating conditions of the equipment. In a certain case, the winding that passed the off-line detection broke down after 72 hours of continuous loading. The control signal comparison relies on preset rules and lacks the ability to dynamically suppress frequency converter harmonics. A water supply system misjudged and stopped due to third-harmonic interference. The accuracy rate of compound fault diagnosis of the existing system is less than 75%, and the insulation deterioration warning lags by more than 48 hours. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings existing in the prior art, and to propose an electrical performance detection system for water supply equipment based on artificial intelligence.

[0006] To achieve the above object, the present invention adopts the following technical solutions: An electrical performance detection system for water supply equipment based on artificial intelligence includes: An electrical parameter acquisition module, which is used to obtain the effective values of three-phase voltage and current through a broadband current transformer, calculate the root mean square value within a period by using the sliding window integration method, extract the amplitude of the 50Hz fundamental wave component in the current waveform, generate a sequence of effective voltage values and the amplitude of the fundamental current, and transmit them to the energy efficiency analysis module; An energy efficiency analysis module, which is used to calculate the instantaneous value of active power based on the sequence of effective voltage values and the amplitude of the fundamental current, generate a sequence of energy efficiency ratios, perform a standard deviation operation on the energy efficiency ratios of 10 consecutive periods, generate a fluctuation mark when the value exceeds the threshold of 0.05 determined by the confidence interval analysis based on 30 groups of samples, and transmit the sequence of energy efficiency ratios and the fluctuation mark to the insulation performance evaluation module; An insulation performance evaluation module, which is used to trigger the detection of the insulation impedance of the water pump winding based on the fluctuation mark, perform harmonic interference correction on the sequence of energy efficiency ratios, unify the unit of the leakage current amplitude of the cable joint to amperes, and synchronously perform a time-domain integration operation on the leakage current amplitude of the cable joint. When the insulation impedance value is lower than 1MΩ and the leakage current integral value is greater than 10mA·s, an insulation deterioration mark is generated, and the insulation impedance detection value and the insulation deterioration mark are transmitted to the comprehensive diagnosis module.

[0007] As a further solution of the present invention, the sequence of effective voltage values includes the root mean square voltage of the period, the change trend of the three-phase voltage amplitude, and the voltage fluctuation characteristics. The amplitude of the fundamental current includes the estimated result of the fundamental amplitude, the current symmetry index, and the current waveform stability index. The sequence of energy efficiency ratios includes the periodic power factor, the change trend of the average active power, and the load response characteristics. The fluctuation mark is specifically an energy efficiency ratio fluctuation warning mark, a fluctuation period calibration, and a fluctuation intensity level. The insulation impedance detection value includes the insulation attenuation degree of the water pump winding, the insulation response time index, and the impedance value time series. The insulation deterioration mark is specifically an insulation performance warning code, a deterioration severity level, and a timestamp record.

[0008] As a further solution of the present invention, the standard deviation determination threshold of 0.05 is determined by the confidence interval analysis of 30 groups of samples, and the confidence level is 95%; The leakage current integral value of 0.01A·s is obtained by performing a time-domain integration operation after unifying the unit of the leakage current amplitude of the cable joint to amperes.

[0009] As a further solution of the present invention, the electrical parameter acquisition module includes: The signal acquisition sub-module synchronously obtains the instantaneous analog quantities of three-phase voltage and current through a broadband current transformer, performs differential amplification on the analog signals to achieve noise elimination with a common-mode rejection ratio ≥ 80 dB, configures a band-pass filter to suppress high-frequency harmonic components with an attenuation coefficient of -40 dB / dec, and generates an instantaneous sequence of three-phase voltage and an instantaneous sequence of three-phase current; The effective value calculation sub-module calls the instantaneous sequence of three-phase voltage and the instantaneous sequence of three-phase current, sets the sliding window width to an integer multiple of the power frequency period of 20 ms, performs a square operation on each point of the instantaneous voltage value within the window, accumulates and calculates the mean value and then the square root to generate a sequence of voltage effective values; The fundamental wave extraction sub-module calls the current effective value data in the sequence of voltage effective values, decomposes the current spectrum using the discrete Fourier transform, extracts the real and imaginary components corresponding to the 50 Hz frequency point, calculates the complex modulus value and multiplies it by the spectral resolution coefficient Δf = 1 / T, where T is the total signal sampling duration, to generate the fundamental wave current amplitude.

[0010] As a further solution of the present invention, the energy efficiency analysis module includes: The power calculation sub-module obtains the sequence of voltage effective values and the fundamental wave current amplitude, performs phase alignment processing on the voltage values at each sampling point in the sequence, calculates the phase difference through an error compensation algorithm, multiplies it by the fundamental wave current amplitude and then superimposes a cosine phase difference correction amount, and performs point-by-point operations according to the instantaneous active power formula to generate an energy efficiency ratio sequence; The standard deviation analysis sub-module calls the energy efficiency ratio sequence, intercepts consecutive 10-cycle energy efficiency ratio data segments in chronological order, calculates the arithmetic mean value of the values within the data segment, performs a square operation on the difference between each value and the mean value and accumulates them, and calculates the dispersion through the standard deviation formula to obtain the standard deviation value; The fluctuation determination sub-module compares the standard deviation value with a preset fluctuation determination threshold of 0.05. When the determination exceeds the limit continuously for 3 times, it triggers dynamic threshold adjustment. If the former is greater than the threshold, it generates an identifier with a logical value of 1, otherwise it generates an identifier with a logical value of 0, encapsulates the energy efficiency ratio sequence and the logical identifier into a key-value pair structure, and outputs a fluctuation mark.

[0011] As a further solution of the present invention, the insulation performance evaluation module includes: The insulation detection trigger sub-module triggers the insulation impedance detection of the water pump winding based on the fluctuation mark, collects the voltage difference across the winding and the leakage current value, uses the four-wire measurement method to eliminate the influence of contact resistance, and applies Ohm's law to calculate the insulation impedance to generate an insulation impedance detection value; The harmonic correction sub-module calls the insulation impedance detection value, extracts the fundamental wave component and the harmonic component amplitude of the energy efficiency ratio sequence, and uses the formula: ; Perform harmonic interference correction on the energy efficiency ratio sequence to generate a harmonically corrected energy efficiency ratio; Wherein, represents the harmonically corrected energy efficiency ratio, represents the original energy efficiency ratio sequence, represents the th harmonic distortion rate, is the th harmonic weight factor, is the fundamental component ratio coefficient; The leakage current integration evaluation sub-module collects the time-domain waveform of the leakage current amplitude at the cable joint, calls the harmonically corrected energy efficiency ratio, and performs discrete integration operation on the leakage current amplitude using the trapezoidal method. The integration order is set to 2nd order accuracy. When the insulation impedance detection value is less than 1 MΩ and the integration result exceeds 10 mA·s, an insulation degradation flag is generated.

[0012] As a further aspect of the present invention, the system further includes: A comprehensive diagnosis module for judging whether the motor bearing current exceeds the 5 mA threshold collected by the Hall sensor at a frequency of 10 kHz based on the insulation impedance detection value, counting the number of times the insulation degradation flag is triggered, synchronously performing a linear regression calculation on the temperature rise rate of the junction box temperature, constructing a temperature difference sequence using a time window sliding mechanism, and generating a pump body winding aging fault code when the bearing current exceeds the limit or the temperature rise rate exceeds 2 °C / min.

[0013] As a further aspect of the present invention, the pump body winding aging fault code specifically refers to the motor bearing fault type flag, the fault trigger condition record, and the pump body temperature rise trend code.

[0014] As a further aspect of the present invention, the comprehensive diagnosis module includes: The current threshold judgment sub-module detects the motor bearing current data, extracts the effective value based on a 10 ms sampling interval, extracts the insulation impedance detection value, compares the detection value with the preset current threshold point by point, counts the time window length of continuous overlimit, and simultaneously judges whether the continuous overlimit condition is satisfied based on the time window length to generate a current overlimit flag; The flag statistics sub-module calls the trigger signal in the current overlimit flag, uses a fixed period as the statistical unit, accumulates the number of trigger signals, calculates the average value of the time interval between adjacent triggers and the standard deviation of the single trigger duration to generate a trigger frequency statistical value; The temperature rise calculation sub-module collects the time-series data of the junction box temperature, constructs a difference sequence based on the temperature difference between adjacent 10 seconds within the sliding time window, performs linear fitting on the sequence using the least squares method, extracts the slope as the temperature rise rate per unit time, and generates a temperature rise rate coefficient; The fault judgment submodule calls the authenticity of the current overlimit flag, the cumulative trend of the trigger frequency statistics and the temperature rise rate coefficient, and compares the temperature rise rate coefficient with the preset rate threshold. If the current overlimit flag is true or the temperature rise rate coefficient exceeds the limit, then combined with the continuous rising trend of the trigger frequency statistics, a pump winding aging fault code is generated.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, the three-phase voltage and current root mean square values ​​are dynamically calculated by combining a wide-band current transformer with a sliding window integration method to improve the accuracy of transient operating condition parameter acquisition. The 50Hz fundamental component amplitude extraction technology separates the power frequency signal from the high-frequency harmonics to reduce the error of the energy efficiency analysis base data. The standard deviation calculation of the continuous cycle energy efficiency ratio triggers the insulation impedance detection threshold, and establishes a dynamic response mechanism for parameter fluctuations and insulation degradation. The harmonic interference correction algorithm synchronously processes the energy efficiency ratio sequence and the cable leakage current time domain integral to suppress the impedance detection error to less than 5%. The temperature rise rate linear regression and the bearing current threshold are coordinated to determine the multi-dimensional cross-validation of winding aging, and the fault misjudgment rate is reduced by 26%. The closed-loop detection system integrates transient parameter capture, dynamic threshold triggering and multi-parameter fusion diagnosis, which increases the success rate of insulation degradation warning to 92% and shortens the fault location response time to 3 seconds. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a system flow chart of the present invention; Figure 2 This is a flow chart of the electrical parameter acquisition module of the present invention; Figure 3 This is a flow chart of the energy efficiency analysis module of the present invention; Figure 4 This is a flow chart of the insulation performance evaluation module of the present invention; Figure 5 This is a flow chart of the comprehensive diagnosis module of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more, unless otherwise specifically defined.

[0019] Embodiment 1 Please refer to Figure 1 , an electrical performance detection system for water supply equipment based on artificial intelligence includes: An electrical parameter acquisition module, which is used to obtain the effective values of three-phase voltage and current through a broadband current transformer, calculate the root mean square value within a period by using the sliding window integration method, extract the amplitude of the 50Hz fundamental wave component in the current waveform, generate a sequence of effective voltage values and the amplitude of the fundamental current, and transmit them to the energy efficiency analysis module; An energy efficiency analysis module, which is used to calculate the instantaneous value of active power based on the sequence of effective voltage values and the amplitude of the fundamental current, generate a sequence of energy efficiency ratios, perform a standard deviation operation on the energy efficiency ratios of 10 consecutive periods, generate a fluctuation mark when the result exceeds the threshold of 0.05 determined by the confidence interval analysis based on 30 groups of samples, and transmit the sequence of energy efficiency ratios and the fluctuation mark to the insulation performance evaluation module; An insulation performance evaluation module, which is used to trigger the detection of the insulation impedance of the water pump winding based on the fluctuation mark, perform harmonic interference correction on the sequence of energy efficiency ratios, unify the unit of the leakage current amplitude of the cable joint to amperes, synchronously perform a time-domain integration operation on the leakage current amplitude of the cable joint, and generate an insulation degradation mark when the insulation impedance value is lower than 1MΩ and the leakage current integral value is greater than 10mA·s, and transmit the insulation impedance detection value and the insulation degradation mark to the comprehensive diagnosis module; A comprehensive diagnosis module, which is used to judge whether the motor bearing current exceeds the 5mA threshold collected by the Hall sensor at a frequency of 10kHz based on the insulation impedance detection value, count the number of times the insulation degradation mark is triggered, synchronously perform a linear regression calculation on the temperature rise rate of the junction box temperature, construct a temperature difference sequence by using the time window sliding mechanism, and generate a pump winding aging fault code when the bearing current exceeds the limit or the temperature rise rate exceeds 2℃ / min.

[0020] The effective voltage value sequence includes the periodic root mean square voltage, the variation trend of the three-phase voltage amplitude, and the voltage fluctuation characteristics. The fundamental wave current amplitude includes the estimated result of the fundamental wave amplitude, the current symmetry index, and the current waveform stability index. The energy efficiency ratio sequence includes the periodic power factor, the variation trend of the average active power, and the load response characteristics. The fluctuation mark specifically refers to the energy efficiency ratio fluctuation warning sign, the fluctuation period calibration, and the fluctuation intensity level. The insulation impedance detection value includes the insulation attenuation degree of the water pump winding, the insulation response time index, and the impedance value time series. The insulation deterioration mark specifically refers to the insulation performance warning code, the severe deterioration level, and the timestamp record. The pump body winding aging fault code specifically refers to the motor bearing fault type mark, the fault trigger condition record, and the pump body temperature rise trend code.

[0021] The standard deviation determination threshold of 0.05 is determined based on the confidence interval analysis of 30 groups of samples, and the confidence level is 95%; The leakage current integral value of 0.01 A·s is obtained by performing time-domain integration operation after unifying the unit of the leakage current amplitude of the cable joint to amperes.

[0022] Please refer to Figure 2 , the electrical parameter acquisition module includes: The signal acquisition sub-module synchronously obtains the three-phase voltage instantaneous analog quantity and the current instantaneous analog quantity through a broadband current transformer, performs differential amplification processing on the analog signal, realizes noise elimination with a common mode rejection ratio ≥ 80 dB, configures a band-pass filter to suppress high-frequency harmonic components with an attenuation coefficient of -40 dB / dec, and generates a three-phase voltage instantaneous sequence and a three-phase current instantaneous sequence; The signal acquisition sub-module synchronously obtains the three-phase voltage instantaneous analog quantity and the current instantaneous analog quantity through a broadband current transformer. Specifically, during implementation, for a specific model of centrifugal water pump motor in operation (rated voltage 380 V, rated current 25 A), the broadband current transformer (model LEMATO-10-B333-D10, bandwidth DC-1 MHz, accuracy 0.5%) is installed on the U, V, and W phases of the motor three-phase power supply cable. At the same time, a high-precision differential voltage probe (model TektronixP5200A, bandwidth 50 MHz, attenuation ratio can be set) is connected between the three-phase input terminals of the motor and the neutral point. The output analog signal of the sensor is connected to a multi-channel synchronous data acquisition card (model NIUSB-6363, 16-bit resolution). The sampling frequency of the data acquisition card is set to 10 kHz to ensure that the power frequency (50 Hz) and its high-order harmonic components can be captured. The data acquisition card synchronously performs analog-to-digital conversion on the analog signals of six channels (the voltage signals of Ua, Ub, Uc, and the current signals of Ia, Ib, Ic) at a sampling rate of 10 kHz to obtain the original voltage and current digital signal sequences.

[0023] To eliminate the common-mode noise interference introduced by line conduction and spatial electromagnetic field coupling, each phase voltage signal (such as the Ua phase) collected and its reference ground signal are fed into the front-end differential amplification and conditioning circuit built into the data acquisition card. This circuit configuration achieves a common-mode rejection ratio of 85 dB, significantly weakening the common-mode noise component. Specifically, if a common-mode interference voltage with an amplitude of 5 V and a frequency of 1 kHz is superimposed on the original signal, after differential amplification processing with a common-mode rejection ratio of 85 dB (i.e., times the suppression ability), the common-mode interference component in the output signal is suppressed to , which is much lower than the effective signal amplitude. Subsequently, the signal flows through a digital band-pass filter, which is set as a Butterworth fourth-order filter, and the passband range is set to 45 Hz to 2 kHz. The purpose is to retain the power frequency fundamental wave and the main low-order harmonic components, while filtering out high-frequency noises such as DC bias and switching frequencies. Outside the passband, especially in the high-frequency band (greater than 2 kHz), the attenuation slope of the filter is designed to be -40 dB / dec. For a noise component with an interference frequency of 5 kHz, its frequency is 2.5 times the upper limit of the passband, which is 2 kHz, , and the attenuation reaches , that is, the amplitude is attenuated by times. After the above processing, clear three-phase voltage instantaneous sequences and three-phase current instantaneous sequences are obtained. Table 1 shows some sampling data of phase A within one power frequency cycle (20 ms).

[0024] Table 1 Example of instantaneous sampling data of phase A voltage and current Sampling point serial number Time (ms) Instantaneous voltage value (V) Instantaneous current value (A) 1 0.1 15.71 0.82 2 0.2 31.40 1.64 100 10.0 310.55 15.11 199 19.9 -15.68 -0.81 200 20.0 0.00 0.00 As shown in Table 1, some instantaneous value data of phase A voltage and current at the start, near the peak, and end moments within one power frequency cycle (20 ms, corresponding to 200 sampling points) after signal acquisition and processing are listed. These data will be used for subsequent effective value calculation.

[0025] The effective value calculation sub-module calls the three-phase voltage instantaneous sequence and the three-phase current instantaneous sequence, sets the sliding window width as an integer multiple of the power frequency cycle of 20 ms, performs a square operation on each point of the voltage instantaneous value within the window, accumulates and then calculates the mean value and the square root to generate the voltage effective value sequence; The effective value calculation sub-module calls the three-phase voltage instantaneous sequence and the three-phase current instantaneous sequence. When executing, it reads data from the phase A voltage instantaneous sequence and the phase A current instantaneous sequence obtained from the previous module. A sliding window is set, and its width must be an integer multiple of the number of sampling points within the power frequency cycle . Select as the integer multiple coefficient, then the window width is determined to be sampling points. A single - cycle window ( ) is selected to quickly respond to the change of the effective value, while ensuring that the complete power - frequency cycle is covered to obtain an accurate effective value. Select the voltage instantaneous values within the current calculation window (set from the th sampling point to the th sampling point) .

[0026] For each voltage instantaneous value within the window perform a square operation to obtain . Accumulate the squared values of all points within the window, and calculate . In a specific calculation window, the accumulated result of the sum of the squared voltages is . Then divide the accumulated sum by the window width to obtain the mean value, getting . Finally, calculate the square root of this mean value to obtain the effective value of the phase - A voltage corresponding to this window . Slide the window forward by one sampling point to a new window starting from , and repeat the above operations of squaring, accumulating, averaging, and square - rooting. For the current instantaneous sequence perform exactly the same calculation steps. If the sum of the squared currents in the corresponding window is , then the mean value is , and the effective value of the current is . Continuously process to generate the voltage effective - value sequence and the current effective - value sequence.

[0027] The fundamental - wave extraction sub - module calls the current effective - value data in the voltage effective - value sequence, decomposes the current spectrum using the discrete Fourier transform, extracts the real and imaginary components corresponding to the 50 - Hz frequency point, calculates the complex - modulus value and multiplies it by the spectrum - resolution coefficient Δf = 1 / T, where T is the total signal - sampling duration, to generate the fundamental - wave current amplitude.

[0028] The fundamental - wave extraction sub - module calls the current instantaneous sequence. Specifically, in implementation, a segment is intercepted from the phase - A current instantaneous sequence obtained from the signal - acquisition sub - module for analysis. Set the total signal duration for spectrum analysis. Its selection needs to balance the frequency resolution and the calculation real - time performance. Here, is selected, corresponding to a data segment containing sampling points. Use the discrete Fourier transform (DFT) algorithm to decompose the current instantaneous sequence of 10,000 points (where ) for spectrum decomposition.

[0029] The DFT calculation will output a series of corresponding discrete frequency points The complex spectral components , where the spectral resolution is determined by the total sampling duration, . The calculation process is to perform for each frequency index ( ). It is necessary to extract the component corresponding to the power frequency of 50 Hz, and the index corresponding to this frequency is . The real part and the imaginary part of and are calculated. Through calculation, and are obtained. Then, the modulus value of this complex component

[0030] The modulus value of the DFT calculation result needs to be amplitude-corrected to obtain the current amplitude in the actual physical sense. The standard correction method is to multiply by to obtain the peak value. Therefore, the fundamental current peak value . The fundamental current amplitude (rms value) is the peak value divided by , that is . This step generates the fundamental current amplitude .

[0031] Please refer to Figure 3 , the energy efficiency analysis module includes: The power calculation sub-module obtains the voltage rms value sequence and the fundamental current amplitude, performs phase alignment processing on the voltage values at each sampling point in the sequence, calculates the phase difference through an error compensation algorithm, multiplies it by the fundamental current amplitude, and then superimposes the cosine phase difference correction amount, and performs point-by-point calculation according to the instantaneous active power formula to generate the energy efficiency ratio sequence; The power calculation sub-module obtains the voltage rms value sequence and the fundamental current amplitude. Obtain the rms value of phase A voltage (result taken from paragraph 2), and obtain the fundamental current amplitude of phase A (result taken from paragraph 3) from the fundamental extraction sub-module. To calculate the active power, it is necessary to determine the phase difference between the fundamental components of voltage and current. This can be achieved by analyzing the voltage instantaneous sequence and the current instantaneous sequence . One method is to apply DFT to the voltage and current sequences respectively, extract the complex components and corresponding to the 50 Hz frequency point, and the difference in their arguments is the phase difference . The phase difference is obtained through calculation.

[0032] Considering that fixed phase errors may be introduced by sensors, circuits, etc. in actual measurements, a pre-calibrated error compensation algorithm is applied for correction. The calibration process can be tested by inputting signals with known phase differences to establish a compensation lookup table or function. Let the compensated phase difference be , and the compensation value be , then . The active power is calculated according to the formula . Substitute the obtained values into: . Calculate . Then the active power of phase A is . The calculated value represents the average active power at that moment or within that window, and it is used as a point in a sequence of "energy efficiency ratio" (here the term "energy efficiency ratio" comes from the original text, and actually the active power is calculated). Repeat this process for each calculation window to generate a sequence of energy efficiency ratio (active power).

[0033] The standard deviation analysis sub-module calls the energy efficiency ratio sequence, intercepts a continuous 10-cycle energy efficiency ratio data segment in chronological order, calculates the arithmetic mean of the values within the data segment, performs a square operation on the difference between each value and the mean and accumulates them, and at the same time calculates the dispersion degree through the standard deviation formula to obtain the standard deviation value; The standard deviation analysis sub-module calls the energy efficiency ratio (active power) sequence. Obtain the generated active power sequence of phase A from the power calculation sub-module . Intercept a continuous data segment in chronological order for volatility analysis, and select a data segment with a length of 10 consecutive power frequency cycles. If the power calculation is performed cycle by cycle, then intercept the latest 10 power values. The obtained data segment is .

[0034] Calculate the arithmetic mean for this data segment containing values. The calculation is . Then, for each value in the data segment, calculate its difference from the mean . The difference for the first point is . Then perform a square operation on each difference, and the squared difference for the first point is . Accumulate all squared differences to obtain . Then apply the sample standard deviation formula to calculate the dispersion degree. The standard deviation value . The standard deviation value for this data segment is 48.37 W.

[0035] The fluctuation determination sub-module compares the standard deviation value with the preset fluctuation determination threshold of 0.05. When the limit is exceeded for three consecutive determinations, dynamic threshold adjustment is triggered. If the former is greater than the threshold, a flag with a logical value of 1 is generated; otherwise, a flag with a logical value of 0 is generated. The energy efficiency ratio sequence and the logical flag are encapsulated into a key-value pair structure, and the fluctuation mark is output.

[0036] The fluctuation determination sub-module compares the standard deviation value with the preset fluctuation determination threshold. Obtain the standard deviation value of the active power calculated by the previous module . Compare this value with the preset fluctuation determination threshold . This threshold is set with reference to the power fluctuation statistical characteristics of this type of water pump under a large number of normal operating conditions. By analyzing historical data, calculate the standard deviation of the power sequence with a normal operating duration of 10 cycles, and obtain a series of standard deviation values. Let the average value of these standard deviation values be , and the standard deviation of the standard deviation be . To cover the vast majority of normal fluctuation situations (such as a 95% confidence interval), the threshold is set to the mean plus twice the standard deviation, that is .

[0037] The standard deviation calculated this time is compared with the threshold . Because , the determination result this time is that the limit is not exceeded. If the standard deviation calculation and determination are carried out three times continuously, and the results are respectively , and all these three results are greater than the threshold , it is determined that the limit has been exceeded three times continuously. At this time, trigger the dynamic threshold adjustment mechanism: adjust the threshold according to the average value of the last three over-limit values, and the new threshold , and this new threshold will be used for subsequent comparisons. According to the comparison result, if , a flag with a logical value of 1 is generated; if , a flag with a logical value of 0 is generated. For this comparison , so the logical flag is generated. Finally, the active power sequence and the generated logical flag are encapsulated into a key-value pair structure, expressed as , and the fluctuation mark is output.

[0038] Please refer to Figure 4 , the insulation performance evaluation module includes: The insulation detection trigger sub-module triggers the insulation impedance detection of the water pump winding based on the fluctuation mark, collects the voltage difference across the winding and the leakage current value, uses the four-wire measurement method to eliminate the influence of contact resistance, applies Ohm's law to calculate the insulation impedance, and generates an insulation impedance detection value; The insulation detection trigger sub-module triggers the insulation impedance detection of the water pump winding based on the fluctuation mark. The system continuously monitors the fluctuation mark output by the previous module. When the logical identifier in the fluctuation mark becomes 1 (indicating abnormal power fluctuation, which may predict potential problems), the insulation impedance detection program of the water pump motor winding is automatically started. Before performing the detection, the control system safely disconnects the three-phase power supply of the water pump motor. Use an insulation resistance tester (model Fluke1507, test voltage options are 250V, 500V, 1000V), and connect it using the four-wire measurement method to eliminate the influence of lead resistance and contact resistance. Connect the L line of the current output terminal of the tester to a phase wire terminal (U phase) of the motor winding, and connect the E line of the current return terminal to the grounding terminal of the motor housing. At the same time, connect the Guard line of the voltage measurement terminal to the U phase terminal, and connect the other voltage measurement line to the grounding terminal of the motor housing.

[0039] Select the test voltage range to 500VDC. After starting the test, the tester applies a 500V DC voltage between the winding and the housing, and accurately measures the leakage current flowing through the insulating medium and the actual voltage difference applied across the insulating medium . The voltage difference obtained from this measurement , leakage current . It is necessary to convert the unit of the leakage current to amperes (A) for calculation. The conversion rule is 1A = 1000mA, so . Apply Ohm's law to calculate the insulation impedance. The insulation impedance detection value . Convert the result to the more commonly used unit of megaohms (MΩ). The conversion rule is 1MΩ = 1,000,000Ω, so . Generate the insulation impedance detection value .

[0040] The harmonic correction sub-module calls the insulation impedance detection value, extracts the fundamental wave component and the harmonic component amplitude of the energy efficiency ratio sequence, and uses the formula: ; Perform harmonic interference correction on the energy efficiency ratio sequence to generate the harmonic-corrected energy efficiency ratio; Among them, represents the harmonic-corrected energy efficiency ratio, represents the original energy efficiency ratio sequence, represents the th harmonic distortion rate, is the th harmonic weight factor, is the fundamental wave component ratio coefficient; The harmonic correction sub-module calls the insulation impedance detection value and calls the energy efficiency ratio sequence (modified to call the instantaneous current sequence for harmonic analysis). Obtain the insulation impedance detection value from the previous module . At the same time, for harmonic analysis, it is necessary to call the instantaneous current sequence obtained by the signal acquisition sub-module again , or obtain the calculated harmonic amplitudes of each order from the fundamental wave extraction sub-module. Here, use the harmonic information obtained by DFT analysis in Paragraph 3: the effective value of the fundamental wave current , and it is also necessary to calculate the effective values of the harmonic currents of the 2nd to 5th orders (n = 2, 3, 4, 5). By performing a similar amplitude correction calculation on the modulus value at the corresponding frequency , we get: Second harmonic ( ) ; Third harmonic ( ) Fourth harmonic ( ) Fifth harmonic ( ) These are the harmonic distortion rates of each order

[0041] Set the weighting factors of each harmonic . These weighting factors reflect the relative influence degrees of different harmonics on a specific evaluation target (here it is the correction of the energy efficiency ratio). The setting basis is the statistical analysis of a large amount of operation data of similar pumps, and the correlation between different harmonic contents and parameters such as the deviation of the energy efficiency index and the failure rate of the equipment is studied. The analysis shows that low-order harmonics (especially the 2nd and 3rd orders) have a relatively greater impact on the additional losses and temperature rise of the winding, so higher weights are given. The set values are determined as: . Calculate the proportion coefficient of the fundamental wave component . This coefficient represents the proportion of the fundamental wave current in the total current. First, calculate the effective value of the total current including the fundamental wave and the 2nd to 5th harmonics ; Then the fundamental wave proportion coefficient .

[0042] Call an original energy efficiency ratio index (here it is assumed to be the basic efficiency evaluation value without considering the influence of harmonics, set as ), and apply the harmonic correction formula for calculation. In the formula , represents the corrected energy efficiency ratio represents the original energy efficiency ratio, is the th harmonic distortion rate, is the th harmonic weight factor, is the fundamental component ratio coefficient. All terms involved in the summation are dimensionless ratios. Therefore, and have the same unit or are both dimensionless ratios. Substitute the values for calculation: ; ; ; ; The benefit of the formula is that by distinguishing the effects of different harmonics (through and ) and considering the dominance of the fundamental wave (through ), it can more precisely evaluate the comprehensive impact of harmonics on the system energy efficiency index, thus obtaining a more accurate evaluation result than relying solely on the original energy efficiency ratio or the total harmonic distortion rate . The calculated harmonic-corrected energy efficiency ratio . This result indicates that after considering the effects of key harmonics, the corrected energy efficiency evaluation index is significantly different from the original value, which will be used as the input for subsequent evaluations. Generate the harmonic-corrected energy efficiency ratio .

[0043] The leakage current integration evaluation sub-module collects the time-domain waveform of the leakage current amplitude at the cable joint, calls the harmonic-corrected energy efficiency ratio, and applies the trapezoidal method to perform discrete integration operations on the leakage current amplitude. Set the integration order to 2nd-order accuracy. When the insulation impedance detection value is less than 1 MΩ and the integration result exceeds 10 mA·s, generate an insulation deterioration flag.

[0044] The leakage current integration evaluation sub-module collects the time-domain waveform of the leakage current amplitude at the cable joint and calls the harmonic-corrected energy efficiency ratio. When the insulation impedance detection value (from paragraph 7) and the harmonic-corrected energy efficiency ratio (from paragraph 8) are available, the system starts monitoring the leakage current at the connection between the pump motor power cable and the junction box. Use a clamp-on leakage current sensor (model HantekCC-65, measurement range 1 mA - 65 ADC / AC), and clamp it on the three-phase cable bundle near the cable joint (or measure each phase separately and then synthesize). Set the sampling frequency of the sensor output signal to 1 kHz and continuously monitor for a period of time. Set the monitoring duration to . Obtain the leakage current time-series data , including sampling points.

[0045] Apply the trapezoidal integration rule to the measured leakage current amplitude for discrete integration operations to estimate the total charge flowing through the insulation defect path during the monitoring period. The trapezoidal integration formula is , where is the sampling time interval. Set the integration order to 2nd order accuracy, which is reflected in the selection of the trapezoidal rule (whose truncation error is ). By performing cumulative calculations on the leakage current data (unit: mA) of 60,000 sampling points, the integration result obtained is . It is necessary to convert the unit of the result to . The conversion rule is 1s = 1000ms. Therefore .

[0046] Compare the calculated integration result with the preset integration threshold . At the same time, check whether the measured insulation impedance value is less than the preset insulation threshold . The setting basis of the insulation threshold is the requirement of relevant national or industry electrical safety regulations for the insulation resistance of low-voltage rotating motors. It is usually stipulated that the insulation resistance of an operating motor should not be lower than 1MΩ / kV (operating voltage). For a 380V motor, this threshold is set to . The setting of the integration threshold is based on the risk assessment of heat accumulation and accelerated insulation aging caused by long-term leakage of cable joints. It is determined through experiments and experience that when the cumulative leakage charge exceeds a certain value, the failure risk increases significantly. Here, it is set to . In this example, , does not meet the (i.e., ) condition. Although the integration result meets the (i.e., ) condition, but since the insulation resistance condition is not met, no insulation deterioration mark is generated. In another scenario, if it is measured that , then condition is met, and condition is also met. At this time, both conditions are achieved, and an insulation deterioration mark is generated, and its logical value is 1.

[0047] Please refer to Figure 5 , the comprehensive diagnosis module includes: The current threshold judgment sub-module detects the bearing current data of the motor, extracts the effective value based on a 10ms sampling interval, extracts the insulation impedance detection value, compares the detection value with the preset current threshold point by point, counts the time window length of continuous overrun, and at the same time judges whether the continuous overrun condition is satisfied based on the time window length, and generates a current overrun flag; The current threshold judgment sub-module detects the bearing current data of the motor. The current is led out through an insulating bearing installed on the bearing seat at the non-driving end of the motor or a grounding brush installed on the shaft, and is connected to a dedicated shaft current sensor (such as a Hall effect sensor) to monitor the bearing current of the water pump motor in real time. The sampling interval of the sensor is set to 10ms (corresponding to the sampling frequency ), and the instantaneous bearing current data is collected . Calculate the effective value of the collected data. Using a method similar to that in paragraph 2, the window width for calculating the effective value is set to , including sampling points. Calculate the effective value of the bearing current for each window .

[0048] Obtain the sequence of effective values of the bearing current. At the same time, obtain the insulation impedance detection value (using the value in paragraph 7 ). Compare the calculated sequence of effective values of the bearing current with the preset bearing current threshold point by point The setting of needs to comprehensively consider motor design (such as whether there is internal asymmetry), bearing type, lubrication state, load size, and manufacturer's suggestions. Referring to the manual of this type of water pump motor and the operation experience of similar equipment, the effective value of the bearing current is usually below 0.5A under normal operating conditions. Therefore, the threshold is set to . Analyze a sequence of effective values of the bearing current: . Compare each value with the threshold . It is found that the 3rd value (0.58A) to the 6th value (0.53A) all exceed the threshold

[0049] Count the time window length of continuous overrun. In this sequence, from the 3rd window to the 6th window, there is continuous overrun, a total of 4 windows. The duration of each window is 100ms. Therefore, the total time of continuous overrun . Judge whether this duration meets the preset continuous overrun condition . Set to prevent misjudgment caused by short-term electrical transients or measurement noise. Its value is determined according to the cumulative effect of bearing electrical erosion damage and is set to . Because the calculated continuous overrun time is greater than the set continuous overrun condition , so the continuous overlimit condition is met. Accordingly, a current overlimit identifier is generated, and its status is set to true (logical value 1). If , the identifier status is false (logical value 0). A current overlimit identifier with a status of true is generated.

[0050] The identifier statistics sub-module calls the trigger signal in the current overlimit identifier, uses a fixed period as the statistical unit, accumulates the number of trigger signals, calculates the average value of the time intervals between adjacent triggers and the standard deviation of the duration of a single trigger, and generates a trigger frequency statistical value; The identifier statistics sub-module calls the trigger signal in the current overlimit identifier. The system monitors the status change of the current overlimit identifier generated by the previous module. When the identifier status changes from false 0 to true 1, the moment is recorded as a trigger event. Set a fixed statistical period for summarizing trigger information, The selection of which should adapt to the typical time scale of fault development and also take into account the timeliness of alarm. Here it is set to . In each 5-minute statistical period, perform the following statistics: 1. Accumulate the total number of times the current overlimit identifier is triggered . In a statistical period, if it is recorded that the identifier changes from 0 to 1 a total of 4 times, then .

[0051] 2. Record the start time and end time (k = 1, 2, 3, 4).

[0052] 3. Calculate the time interval between the start times of two adjacent trigger events . If the trigger times are the 10th s, 70th s, 150th s, and 250th s within the statistical period respectively, the time intervals are , , . Calculate the average value of these time intervals .

[0053] 4. Calculate the duration of a single trigger event . According to the record, the durations of the 4 triggers are respectively . Calculate the standard deviation of these durations . First calculate the average duration . Then calculate the standard deviation ; Take the number of triggers within the statistical period, the average time interval and the standard deviation of the duration as the output results. Generate a trigger frequency statistical value.

[0054] The temperature rise calculation sub-module collects the time-series data of the junction box temperature, constructs a difference sequence based on the temperature difference in the adjacent 10 seconds within the sliding time window, uses the least squares method to linearly fit the sequence, extracts the slope as the temperature rise rate per unit time, and generates a temperature rise rate coefficient. The temperature rise calculation sub-module collects the time-series data of the junction box temperature. Through the PT100 platinum thermal resistance temperature sensor pre-installed near the terminal block inside the water pump motor junction box, it is connected to the data acquisition system. The acquisition time interval is set to 1 second, and the internal temperature of the junction box is continuously recorded to obtain the time-series temperature data. , with the unit of degree Celsius (°C). To analyze the temperature change trend, the sliding window method is used to calculate the temperature rise rate. Select a sliding time window with a length set to .

[0055] At each time point, 60 temperature data points within the current window (the past 60 seconds) are extracted. , where is the relative time (such as 0s, 1s, …, 59s). The least squares method is used to linearly fit these 60 data points to find the best fitting straight line. The formula for calculating the slope (i.e., the temperature rise rate) by the least squares method is , where . By calculating the data of a specific window (the data is shown in Table 2), the slope of the fitting straight line is obtained.

[0056] Table 2 Fragment of the time-series data of the junction box temperature (partial data within the window) Time (relative seconds within the window) Temperature (°C) 0 55.21 1 55.25 … … 30 56.28 … … 59 57.35 As shown in Table 2, it shows partial temperature sampling data within a 60-second window used for calculating the temperature rise rate. The calculated slope is regarded as the temperature rise rate per unit time at the current moment. This slope value is output as the temperature rise rate coefficient. Continuously slide the window and repeat the calculation to generate a time series of the temperature rise rate coefficient .

[0057] The fault determination sub-module calls the state authenticity of the current overlimit flag, the cumulative trend of the trigger frequency statistics value, and the temperature rise rate coefficient, compares the temperature rise rate coefficient with the preset rate threshold. If the current overlimit flag is true or the temperature rise rate coefficient exceeds the limit, combined with the continuous upward trend of the trigger frequency statistics value, a pump body winding aging fault code is generated.

[0058] The fault determination sub-module calls the state authenticity of the current over-limit indicator, the cumulative trend of the trigger frequency statistical value, and the temperature rise rate coefficient. This module collects the output information of multiple previous modules for comprehensive diagnosis. 1. Obtain the current state of the current over-limit indicator from the current threshold judgment sub-module. In this case, the state is true (logical value 1), indicating that continuous bearing current over-limit is detected (from paragraph 10). 2. Obtain the trigger frequency statistical value from the indicator statistics sub-module and analyze its change trend over time. Examine the trigger counts in the last three statistical periods (each 5 minutes) , obtaining the sequence (from paragraph 11). This sequence shows that the trigger frequency increases from 2 to 4 and remains at a high level, presenting a continuous abnormal or rising trend. 3. Obtain the current junction box temperature rise rate coefficient from the temperature rise calculation sub-module (from paragraph 12).

[0059] Compare the obtained temperature rise rate coefficient with the preset temperature rise rate threshold . The setting of needs to be based on the insulation class of the motor (such as class F, allowing a temperature rise of 105K), the rated operating temperature, the heat dissipation conditions, and relevant safety standards. For this water pump motor, referring to its design and operating environment, if the temperature rise rate continuously exceeds , it is considered abnormal and may cause accelerated insulation aging. Therefore, set . Compare the current value with the threshold , and the result is , indicating that the temperature rise rate exceeds the limit.

[0060] Finally, perform the fault logic determination. The determination rule is: If (the current over-limit indicator is true) or (the temperature rise rate coefficient exceeds the limit), and (the trigger frequency statistical value shows a continuous high level or an upward trend), then it is determined that a fault exists. In this example: Is the current over-limit indicator true? Yes (status = 1).

[0061] Does the temperature rise rate coefficient exceed the limit? Yes ( ).

[0062] Does the trigger frequency show a continuous high level or an upward trend? Yes (the sequence [2, 4, 4] shows a high level). Since at least one of the main conditions (current over-limit or temperature rise over-limit, both are satisfied here) is met, and the auxiliary condition (trigger frequency trend) is also satisfied, the system determines that the fault condition of "pump body winding aging" is met. Generate the corresponding fault code, such as "PUMP_AGING_01", and output an alarm or record. If none of the main conditions are met, or if the main conditions are met but the trigger frequency trend is stable or decreasing, then this fault code is not generated. Generate the pump body winding aging fault code.

[0063] The above are only the preferred embodiments of the present invention, and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. An electrical performance detection system for water supply equipment based on artificial intelligence, characterized in that: The system comprises: The electrical parameter acquisition module is used to obtain the three-phase voltage effective value and current effective value through the wide-band current transformer, calculate the root mean square value within the cycle using the sliding window integration method, extract the 50Hz fundamental component amplitude in the current waveform, generate the voltage effective value sequence and fundamental current amplitude, and transmit them to the energy efficiency analysis module; An energy efficiency analysis module, used to calculate the instantaneous value of active power based on the voltage effective value sequence and the fundamental current amplitude, generate an energy efficiency ratio sequence, perform standard deviation operation on the energy efficiency ratio of 10 consecutive cycles, generate a fluctuation mark when it exceeds a threshold value of 0.05 determined based on the confidence interval analysis of 30 groups of samples, and transmit the energy efficiency ratio sequence and the fluctuation mark to the insulation performance evaluation module; The insulation performance evaluation module is used to trigger the insulation impedance detection of the water pump winding based on the fluctuation mark, perform harmonic interference correction on the energy efficiency ratio sequence, unify the unit of the cable joint leakage current amplitude into ampere, and synchronously perform time domain integration operation on the cable joint leakage current amplitude. When the insulation impedance value is lower than 1MΩ and the leakage current integral value is greater than 10mA·s, an insulation degradation mark is generated, and the insulation impedance detection value and the insulation degradation mark are transmitted to the comprehensive diagnosis module.

2. The electrical performance detection system for water supply equipment based on artificial intelligence according to claim 1 is characterized in that: The voltage effective value sequence includes the periodic root mean square voltage, the three-phase voltage amplitude change trend, and the voltage fluctuation characteristics; the fundamental current amplitude includes the fundamental amplitude estimation result, the current symmetry index, and the current waveform stability index; the energy efficiency ratio sequence includes the periodic power factor, the active power mean change trend, and the load response characteristics; the fluctuation mark specifically includes the energy efficiency ratio fluctuation warning mark, the fluctuation period calibration, and the fluctuation intensity level; the insulation impedance detection value includes the insulation attenuation degree of the pump winding, the insulation response time index, and the impedance value time series; the insulation degradation mark specifically includes the insulation performance warning code, the degradation severity level, and the timestamp record.

3. The electrical performance detection system for water supply equipment based on artificial intelligence according to claim 2 is characterized in that: The standard deviation judgment threshold of 0.05 was determined based on the confidence interval analysis of 30 groups of samples, with a confidence level of 95%; The leakage current integral value 0.01A·s is obtained by unifying the cable joint leakage current amplitude unit into ampere and then performing time domain integration operation.

4. The electrical performance detection system for water supply equipment based on artificial intelligence according to claim 3 is characterized in that: The electrical parameter acquisition module includes: The signal acquisition submodule synchronously obtains the instantaneous analog quantity of three-phase voltage and current through a wide-band current transformer, performs differential amplification on the analog signal, achieves noise elimination with a common mode rejection ratio of ≥80dB, configures a bandpass filter to suppress high-frequency harmonic components with an attenuation coefficient of -40dB / dec, and generates three-phase voltage instantaneous sequence and three-phase current instantaneous sequence; The effective value calculation submodule calls the three-phase voltage instantaneous sequence and the three-phase current instantaneous sequence, sets the sliding window width to an integer multiple of the power frequency period 20ms, performs square operation on the instantaneous voltage value in the window point by point, calculates the average value after accumulation and calculates the square root, and generates a voltage effective value sequence; The fundamental wave extraction submodule calls the current effective value data in the voltage effective value sequence, decomposes the current spectrum using discrete Fourier transform, extracts the real and imaginary components corresponding to the 50 Hz frequency point, calculates the complex modulus value and multiplies it by the spectrum resolution coefficient Δf=1 / T, where T is the total signal sampling time, to generate the fundamental wave current amplitude.

5. The electrical performance detection system for water supply equipment based on artificial intelligence according to claim 4 is characterized in that: The energy efficiency analysis module comprises: The power calculation submodule obtains the voltage effective value sequence and the fundamental current amplitude, performs phase alignment processing on the voltage value of each sampling point in the sequence, calculates the phase difference through the error compensation algorithm, multiplies it with the fundamental current amplitude and superimposes the cosine phase difference correction amount, performs point-by-point calculation according to the active power instantaneous value formula, and generates an energy efficiency ratio sequence; The standard deviation analysis submodule calls the energy efficiency ratio sequence, intercepts 10 consecutive cycles of energy efficiency ratio data segments in chronological order, calculates the arithmetic mean of the values ​​in the data segments, performs square operations on the differences between each value and the mean and accumulates them, and calculates the dispersion through the standard deviation formula to obtain the standard deviation value; The fluctuation determination submodule compares the standard deviation value with the preset fluctuation determination threshold value of 0.

05. When the limit is exceeded for three consecutive times, the dynamic threshold adjustment is triggered. If the former is greater than the threshold, an identifier with a logical value of 1 is generated, otherwise an identifier with a logical value of 0 is generated. The energy efficiency ratio sequence and the logical identifier are encapsulated into a key-value pair structure, and the fluctuation mark is output.

6. The electrical performance detection system for water supply equipment based on artificial intelligence according to claim 5 is characterized in that: The insulation performance evaluation module includes: The insulation detection trigger submodule triggers the insulation impedance detection of the water pump winding based on the fluctuation mark, collects the voltage difference and leakage current value at both ends of the winding, adopts the four-wire measurement method to eliminate the influence of contact resistance, calculates the insulation impedance by applying Ohm's law, and generates the insulation impedance detection value; The harmonic correction submodule calls the insulation impedance detection value to extract the amplitude of the fundamental component and the harmonic component of the energy efficiency ratio sequence, using the formula: ； Perform harmonic interference correction on the energy efficiency ratio sequence to generate a harmonic corrected energy efficiency ratio; in, stands for harmonic corrected energy efficiency ratio, represents the original energy efficiency ratio series, Representative Subharmonic distortion rate, For the Subharmonic weighting factor, is the fundamental wave component ratio; The leakage current integral evaluation submodule collects the time domain waveform of the leakage current amplitude of the cable joint, calls the harmonic corrected energy efficiency ratio, applies the trapezoidal method to perform discrete integration operation on the leakage current amplitude, sets the integration order to 2nd order accuracy, and generates an insulation degradation mark when the insulation impedance detection value is less than 1MΩ and the integration result exceeds 10mA·s.

7. The electrical performance detection system for water supply equipment based on artificial intelligence according to claim 6 is characterized in that: The system further comprises: The comprehensive diagnostic module is used to determine whether the motor bearing current exceeds the 5mA threshold collected by the Hall sensor at a frequency of 10kHz based on the insulation impedance detection value, count the number of times the insulation degradation mark is triggered, and simultaneously perform a linear regression calculation of the temperature rise rate on the junction box temperature. A time window sliding mechanism is used to construct a temperature difference sequence, and a pump winding aging fault code is generated when the bearing current exceeds the limit or the temperature rise rate exceeds 2°C / min.

8. The artificial intelligence-based water supply equipment electrical performance detection system according to claim 7 is characterized in that: The pump winding aging fault code specifically refers to the motor bearing fault type identification, fault trigger condition record, and pump body temperature rise trend code.

9. The artificial intelligence-based water supply equipment electrical performance detection system according to claim 8, characterized in that: The comprehensive diagnosis module comprises: The current threshold judgment submodule detects the motor bearing current data, extracts the effective value based on the 10ms sampling interval, extracts the insulation impedance detection value, compares the detection value with the preset current threshold point by point, counts the time window length of continuous over-limit, and judges whether the continuous over-limit condition is met based on the time window length, and generates a current over-limit mark; The identification statistics submodule calls the trigger signal in the current overlimit identification, takes a fixed period as a statistical unit, accumulates the number of trigger signals, calculates the mean time interval between adjacent triggers and the standard deviation of the duration of a single trigger, and generates a trigger frequency statistical value; The temperature rise calculation submodule collects the junction box temperature time series data, constructs a difference sequence based on the temperature difference of adjacent 10 seconds in the sliding time window, uses the least squares method to perform linear fitting on the sequence, extracts the slope as the temperature rise rate per unit time, and generates the temperature rise rate coefficient; The fault judgment submodule calls the authenticity of the current overlimit flag, the cumulative trend of the trigger frequency statistics and the temperature rise rate coefficient, and compares the temperature rise rate coefficient with the preset rate threshold. If the current overlimit flag is true or the temperature rise rate coefficient exceeds the limit, then combined with the continuous rising trend of the trigger frequency statistics, a pump winding aging fault code is generated.

Images