A resident load data generation method based on dynamic harmonic admittance parameters

Abstract

The application discloses a kind of resident load data generation methods based on dynamic harmonic admittance parameter, state division and standardization pretreatment are carried out to original load current data according to sampling rule, then carry out fast Fourier transform calculation, obtain harmonic current vector data and define voltage vector data;According to harmonic current, voltage vector data, obtain admittance vector, utilize admittance transfer method to constrain admittance vector to one, four quadrant, obtain each period discrete harmonic admittance parameter, then obtain dynamic harmonic admittance parameter after polynomial fitting and introducing the random variable of simulation error;According to dynamic harmonic admittance parameter and load state switching time information, build complete load simulation model by power supply simulation model, load admittance equivalent simulation model under single operating state and load state switcher combination on Simulink platform, finally simulate and generate load current simulation waveform data.The application improves the randomness and richness of the generated data.

Description

Technical Field

[0001] This invention relates to non-intrusive load monitoring technology, and in particular to a method for generating residential load data based on dynamic harmonic admittance parameters. Background Technology

[0002] Non-intrusive load monitoring technology, as an emerging green and energy-saving technology, is an effective way to achieve green, low-carbon, energy-saving and emission-reduction in the energy sector, and has become one of the research hotspots for many scholars at home and abroad.

[0003] Feature database establishment and load identification are crucial components of non-intrusive load monitoring, directly impacting the accuracy and reliability of the overall monitoring results. However, the comprehensiveness of the feature database and the accuracy of the load identification process are limited by the breadth and realism of the datasets used. Existing measured and simulated datasets suffer from deficiencies in breadth, richness, and realism, thus limiting the accuracy of load identification. Summary of the Invention

[0004] The purpose of this invention is to provide a method for generating residential load data based on dynamic harmonic admittance parameters.

[0005] The technical solution for achieving the present invention is: a method for generating residential load data based on dynamic harmonic admittance parameters, comprising the following steps:

[0006] Step 1: Perform state division and normalization preprocessing on the raw load current data intercepted according to the sampling rules, and then perform fast Fourier transform calculation to obtain harmonic current vector data and define voltage vector data.

[0007] Step 2: Calculate the admittance vector based on the harmonic current and voltage vector data. Use the admittance transfer method to constrain the admittance vector to the first and fourth quadrants to obtain the discrete harmonic admittance parameters for each period. Then, obtain the dynamic harmonic admittance parameters by polynomial fitting and introducing random variables of simulation error.

[0008] Step 3: Based on the dynamic harmonic admittance parameters and load state switching time information, build a complete load simulation model on the Simulink platform, which consists of a power supply simulation model, a load admittance equivalent simulation model under single operating conditions, and a load state switcher. Finally, simulate and generate load current simulation waveform data.

[0009] A household appliance operation status monitoring system based on an online feature library is characterized by generating residential load data based on dynamic harmonic admittance parameters through the aforementioned method for generating residential load data based on dynamic harmonic admittance parameters.

[0010] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method for generating residential load data based on dynamic harmonic admittance parameters.

[0011] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for generating residential load data based on dynamic harmonic admittance parameters.

[0012] Compared with existing technologies, the significant advantages of this invention are: by performing harmonic decomposition, impedance transfer and other analysis on a small amount of measured load data, the dynamic harmonic admittance parameters of the load are calculated, and a residential load simulation model is established based on this to generate load current simulation data. This effectively solves the defects of existing data generation technologies, such as low goodness of fit, poor generalization ability and inability to fit reality, and improves the realism, breadth, richness, randomness and scientificity of the dataset. Attached Figure Description

[0013] Figure 1 This is a flowchart illustrating a method for generating residential load data based on dynamic harmonic admittance parameters according to the present invention.

[0014] Figure 2 This is a simulation diagram of a residential load data generation system based on dynamic harmonic admittance parameters.

[0015] Figure 3 A schematic diagram showing the sampling and segmentation results of the current waveform during the operation of a multi-state vertical inverter air conditioner;

[0016] Figure 4 A schematic diagram of the averaged current of a multi-state vertical inverter air conditioner in operating state 1.

[0017] Figure 5 A schematic diagram of the dynamic harmonic admittance parameters of a multi-state vertical inverter air conditioner in operating state 1.

[0018] Figure 6 The simulation results of a multi-state vertical inverter air conditioner are shown in the figure. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0020] like Figure 1As shown, a method for generating residential load data based on dynamic harmonic admittance parameters is presented. This method involves preprocessing a small amount of measured load data, performing harmonic decomposition, admittance transfer, polynomial fitting, and random number introduction to calculate the dynamic harmonic admittance parameters of the load. Based on these parameters, a residential load simulation model is built on the Simulink platform. Addressing the shortcomings of existing measured and simulation datasets in terms of breadth, richness, and realism, the method for generating residential load data includes the following steps:

[0021] Step 1: Perform state division and normalization preprocessing on the raw load current data intercepted according to the sampling rules, and then perform fast Fourier transform calculation to obtain harmonic current vector data and define voltage vector data.

[0022] Step 1.1: Use a waveform recording device to sample the waveform. The rule is as follows: within the power frequency cycle, take the moment when the voltage waveform crosses zero upwards as the starting point to extract the original load current data. The sampling time is an integer multiple of the cycle to ensure that the current signal in each cycle can be used for harmonic decomposition calculation.

[0023] Step 1.2, introduce the Pearson similarity coefficient to characterize the difference in current waveform between adjacent cycles, denoted as the waveform similarity between adjacent cycles, as shown in equation (1):

[0024]

[0025] Where i is the dynamic index of the number of periodic current sequences, and x i Let x be the current sequence of the current period. i-1 For the current sequence of the previous adjacent period, N c x represents the number of periodic current sequences in the original load data, i.e., the number of cycles. imax / x i-1max They are current sequences x i / x i-1 The maximum element value in the array, max(x) imax ,xi -1max ) represents x i Maximum element value and x i-1 The larger of the two values ​​is the maximum element value.

[0026] Using equation (1) to monitor state changes in the raw load data, if the waveform similarity between adjacent cycles is f(x) i When x ≥ 0.9, the load operating status is determined to have not changed, and x is set to... i With x i-1 If f(x) is assigned to the same running state; i When the load operating status is less than 0.9, it is determined that the load operating status has changed, and the cycle sequence index of the status change occurrence is recorded. p and xi It is reassigned to a new operating state. The determination criteria are shown in equation (2):

[0027]

[0028] After monitoring is completed, if multiple operating states exist in the original load data, then the record index i will be used as the basis for determining the operating state. p The original data is segmented to obtain multiple running states as shown in equation (3):

[0029]

[0030] Where y(p) is the duration of the operating state, p is the dynamic index of the number of times the load operating state changes, and i p N is a dynamic index for the periodic sequence of load operating state changes. p This refers to the number of times the load operating status changes.

[0031] The start / stop times Ton / Toff for each operating state can be calculated from 2πf*i p-1 With 2πf*i p The calculation yielded the result.

[0032] Step 1.3: Perform normalization preprocessing on the load data for each operating state. Define the normalization preprocessing procedure for the original electrical load data, where the per-unit calculation procedure is shown in Equation (4), the averaging calculation procedure is shown in Equation (5), and the restoration procedure is shown in Equation (6):

[0033]

[0034]

[0035] x i-N =x i-r =max(x i |)×x i-avg (6)

[0036] Where, x i Given the original electrical load periodic current sequence, x i-pu For a normalized periodic current sequence, x i-avg For the averaged periodic current sequence, x i-r The restored periodic current sequence is denoted as the normalized periodic current sequence x. i-N , m is the dynamic index of the number of sampling periods, N c N represents the number of sampling periods. s This represents the number of sampling points within a single period.

[0037] Step 1.4, for the standard periodic current sequence x i-NPerform a Fast Fourier Transform (FFT) to obtain the harmonic current vectors and define the harmonic voltage vectors.

[0038] For a standard periodic current sequence x i-N By performing an FFT, the component X of its k-th harmonic can be obtained. i (k), let X i (k)=a(k)+jb(k), and its modulus and phase angles |X(k)| and arg[X(k)] are shown in equations (7)~(8), respectively;

[0039]

[0040]

[0041] For any term X i (k), and its corresponding k-th order time-domain signal expression is shown in equation (9):

[0042]

[0043] Where n is the sampling point number within the period, ω is the angular velocity, and A, Let the amplitude and phase angle of the time-domain signal be respectively, which can be expressed by equations (10) to (11):

[0044]

[0045]

[0046] Where, N s This represents the number of sampling points per cycle.

[0047] In this invention, the number of voltage and current signal sampling points N within a single steady-state cycle s =32, then according to equations (6) to (8), the amplitude and phase of each current component can be calculated as shown in equations (12) to (13):

[0048]

[0049]

[0050] Among them, I k , Let be the amplitude and phase angle of the kth harmonic current.

[0051] Therefore, the vectors of each harmonic current can be expressed as:

[0052]

[0053] Based on the waveform sampling rules, the amplitude and phase of each harmonic voltage are defined as shown in equations (15) to (16):

[0054]

[0055]

[0056] Among them, V k Let V be the amplitude of the kth harmonic voltage. N The standard voltage for residential electricity in my country is 220V. S is the phase angle of the k-th harmonic voltage. vk This is a voltage direction parameter; the default value is 1.

[0057] Therefore, the voltage vectors of each harmonic can be expressed as:

[0058]

[0059] In addition, the proportion η of each current harmonic needs to be calculated. k =I k / I1, and define the set m of harmonic principal component orders to facilitate subsequent simulation operations. The criteria for its discrimination are shown in equation (18):

[0060]

[0061] Step 2: Calculate the admittance vector based on the harmonic current and voltage vector data. Use the admittance transfer method to constrain the admittance vector to the first and fourth quadrants to obtain the discrete harmonic admittance parameters for each period. Then, obtain the dynamic harmonic admittance parameters by polynomial fitting and introducing random variables of simulation error.

[0062] Step 2.1: Calculate the admittance parameters (including resistance, inductance, and capacitance parameters) of each harmonic based on the current and voltage vectors of each harmonic.

[0063] First, calculate the amplitude and phase angle of the k-th harmonic admittance within a single period, as shown in equations (19) to (20):

[0064] Y k =I k / V k (19)

[0065]

[0066] Among them, Y k , Let Y be the magnitude and phase angle of the k-th harmonic admittance vector. Then, the harmonic admittance vector can be expressed as Y. k form.

[0067] Based on the definition of admittance, the resistance and reactance parameters of each harmonic of the residential load are further calculated, as shown in equations (21) to (22);

[0068]

[0069]

[0070] Among them, R k / X k These are the resistance / reactance parameters for the k-th harmonic, G. k / B k These are the conductivity / susceptance parameters.

[0071] After FFT decomposition, observe the k-th order harmonic admittance vector. Distribution of: If Located in the second and third quadrants, i.e., G k <0, R k <0. Considering that there is no negative resistance in the Simulink simulation software, an admittance transfer method is proposed, introducing the power supply direction parameter S in equation (16). vk By adjusting the value of this parameter, the harmonic admittance vector is constrained to the first and fourth quadrants to facilitate simulation verification.

[0072] The specific method for admittance transfer is as follows:

[0073] ■When When G is in the first quadrant, k >0, R k >0, B k >0, X k When the value is less than 0, the resistance parameter is positive and no admittance transfer is required; the reactance parameter is negative and exhibits capacitive behavior.

[0074] ■When When G is in the second quadrant, k <0, R k <0, B k >0, X k <0, at this time the resistance parameter is negative and the reactance parameter is negative, exhibiting capacitive behavior; let the voltage direction parameter S in equation (16) be negative. vk If the value is 0, that is, the voltage source for that harmonic is inverted, then at this time... The resistor changes from 0 to π after inversion. Reactance parameters The reactance opposite to that before the phase reversal is positive, exhibiting inductive properties. Move from the second quadrant to the fourth quadrant;

[0075] ■When When G is in the third quadrant, k <0, R k <0, B k <0,X k>0, at this time the resistance parameter is negative and the reactance parameter is positive, exhibiting inductive behavior; let the voltage direction parameter S in equation (16) be 0. vk If the value is 0, that is, the voltage source for that harmonic is inverted, then at this time... The resistor changes from 0 to π after inversion. Reactance parameters The reactance is negative, opposite to the direction of the inverting reactance, and exhibits capacitive behavior. Move from the third quadrant to the first quadrant;

[0076] ■When When located in the fourth quadrant, G k >0, R k >0, B k <0,X k When the value is >0, the resistance parameter is positive, and no admittance transfer is required. The reactance parameter is positive, and the behavior is inductive.

[0077] Based on the admittance transfer steps described above, update the resistance and reactance parameters:

[0078]

[0079]

[0080] Among them, R tk X tk These are the updated resistance and reactance parameters, respectively.

[0081] The inductance / capacitance parameters of each harmonic are calculated as shown in equations (25) to (26):

[0082]

[0083]

[0084] Among them, L tk / C tk ω represents the inductor / capacitor parameters for the kth harmonic, f represents the angular velocity, and f represents the frequency of electricity used by Chinese residents, taken as 50Hz.

[0085] Step 2.2, obtain the harmonic admittance parameter sequence {R} for all periods. tk0 ,R tk1 ,…,R tki}, {L tk0 ,L tk1 ,…,L tki} and {C tk0 C tk1 ,…,C tki}, where i is the dynamic index of the number of periods. With time t as the independent variable (t = i / f) and the harmonic admittance parameters as the dependent variable, a discrete sequence of harmonic admittance parameters is converted into a continuous time-varying function R using polynomial fitting up to order 30. tk (t), L tk (t), C tk (t), which is the dynamic harmonic admittance parameter.

[0086] Step 2.3 involves further introducing random variables following a probability distribution into the dynamic harmonic admittance parameters to simulate the random errors of the actual load and generate electrical load range currents that take into account the actual errors. The specific method is as follows:

[0087] Based on the distribution of elements in the harmonic admittance parameter sequence in step 2.2, the mean and standard deviation μ and σ of each harmonic admittance parameter sequence are calculated. A random variable x is introduced into the RLC parameter of the mathematical model, and it is assumed to follow a normal distribution. Its probability density f(x) expression is shown in equation (27):

[0088]

[0089] Therefore, the expression for the dynamic harmonic admittance parameter is updated to:

[0090] R tk ′(t)=R tk (t)+x rk (28)

[0091] L tk ′(t)=L tk (t)+x lk (29)

[0092] C tk ′(t)=C tk (t)+x ck (30)

[0093] Among them, R tk / L tk / C tk With R tk ' / L tk ' / C tk 'These are the resistance / inductance / capacitor parameters before and after the update, x' rk / x lk / x ck These are the random variables corresponding to the resistance / inductance / capacitance parameters.

[0094] Step 3: Based on the dynamic harmonic admittance parameters and load state switching time information, build a complete load simulation model (CLM) on the Simulink platform, which consists of a power supply simulation model (SM1), a load admittance equivalent simulation model under single operating conditions (SM2), and a load state switcher (SM3). Finally, simulated load current waveform data is generated.

[0095] Step 3.1: Use DC Voltage Source, AC Voltage Source, and Ground components to establish a power supply simulation model (SM1).

[0096] Calculations show that the main harmonic components in most residential load currents are concentrated before the 15th harmonic. Therefore, this invention only considers the first 15 harmonic components and the DC component in the current waveform. Thus, a total of 1 pair of DC voltage sources in opposite directions and 15 pairs of AC voltage sources in opposite directions are needed to equivalently realize the forward / reverse voltage sources of the DC component and each harmonic component, respectively. The amplitude of the voltage sources is shown in Equation (11). All the ports on the same side (32) of the voltage sources are connected to the grounding element, and the ports on the other side (32) are used as output terminals.

[0097] Step 3.2: Use Breaker (single-pole single-throw switch), Clock, Variable Resistor, Variable Inductor, Variable Capacitor, and MATLAB Function components to build an equivalent simulation model of load admittance under single operating conditions (SM2).

[0098] To address these harmonics, firstly, two single-pole single-throw switches are combined into a single-pole double-throw switch as a voltage direction selector (SM21). Its input has two ports, and its output has one port for connecting to the input of the RLC module (SM22). The control parameter is the voltage direction parameter S. vk If S vk =1, the switch connected to the positive voltage source is on, and the switch connected to the reverse voltage source is off; if S vk =0, the switch connected to the reverse voltage source is on, and the switch connected to the forward voltage source is off.

[0099] Secondly, a parallel RLC module (SM22) is formed by connecting a variable resistor, a variable inductor, and a variable capacitor in parallel. Its input has one port to connect to the output of a voltage direction selector (SM21), and its output has one port to connect to the input of a harmonic principal component selector (SM23). The values ​​of its resistor, inductor, and capacitor are all output by a MATLAB Function component. The inputs to the MATLAB Function are the clock signal t and the load operation status start / stop time T. on / T off The output is the resistance / inductance / capacitance parameter R. tk ' / L tk ' / C tk Assume T = tT on For the duration of the load operation state, the functional relationships within the component are shown in equations (31) to (33):

[0100]

[0101]

[0102]

[0103] To reduce the computational load of the simulation software and improve the simulation speed, a single-pole single-throw switch is used as the harmonic principal component selector. Its input has one port to connect to the output of the parallel RLC module (SM22), and its output has one port as the output of the single-harmonic admittance equivalent model (SMk). The control parameter is the harmonic principal component judgment parameter S. mk If S mk =1, switch is on; if S mk =0, switch off. S mk The rules for determining the value are shown in equation (34):

[0104]

[0105] In summary, a single-harmonic admittance equivalent model (SMk) is constructed by connecting a voltage direction selector, a parallel RLC module, and a harmonic principal component selector in series. Finally, the input terminals (16×2=32 ports) of the admittance equivalent model for all harmonic orders (SMk) are used as the input terminals of the load admittance equivalent simulation model (SM2) under single-operation conditions, and the output terminals (16 ports) are connected together to serve as the output terminals of the load admittance equivalent simulation model (SM2) under single-operation conditions.

[0106] Step 3.3: A load state switch (SM3) is built using a Breaker (single-pole single-throw switch), a Clock, and a MATLAB Function component. It has one input port and one output port. The input to the MATLAB Function component is the clock signal t and the load on / off time T. on / T off The output is a status switch signal K. s The functional relationships within the components are shown in equation (35):

[0107]

[0108] The input of a single-pole single-throw switch is K. s If K s =1, switch is on; if K s =0, switch off.

[0109] Step 3.4: Combine the power supply simulation model (SM1), the load admittance equivalent simulation model under single operating conditions (SM2), and the load state switch (SM3) to obtain the single-state load simulation model (SLM). The combination method is as follows:

[0110] ■ Connect the output of the power supply simulation model (SM1) to the input of the load admittance equivalent simulation model (SM2) under single operation conditions according to the harmonic order and voltage direction;

[0111] ■ Connect the output of the load admittance equivalent simulation model (SM2) under single operating conditions to the input of the load state switch (SM3);

[0112] ■ Use the output of the load state switch (SM3) as the output of the single-state load simulation model (SLM);

[0113] ■ Connect all the outputs of the single-state load simulation model (SLM) to each other and ground them to serve as the outputs of the complete load simulation model (CLM).

[0114] Finally, the load current simulation waveform data can be obtained by measuring the output of the complete load simulation model (CLM) using a current measurement device.

[0115] Example

[0116] To verify the effectiveness of the present invention, the following experiment was conducted.

[0117] First, according to the waveform sampling rules, the current waveform of a multi-state vertical inverter air conditioner during normal operation is sampled. Based on the waveform similarity, the data is divided into five independent states, such as... Figure 3 As shown in the figure, during the 45-second operation period, the current waveform can be roughly divided into five parts, corresponding to five different operating states of the air conditioner:

[0118] ■State 1 is standby mode, the current component is mainly line noise, and the duration is [0,5]s;

[0119] ■State 2 is the instantaneous power-on state, where the current component is mainly pulse current, and the duration is (5, 5.1]s;

[0120] ■States 3 / 4 / 5 represent the steady-state current during normal operation of the air conditioner, distinguished by their amplitude.

[0121] ■ The duration of state 3 is (5.1, 9.6] s;

[0122] ■ The duration of state 4 is (9.6, 39.6] s;

[0123] ■ The duration of state 5 is (39.6, 45] s.

[0124] Because there are sudden changes in current amplitude between states, if they are not separated, the Simulink simulation will fail.

[0125] Next, the current at each state's periodic level is preprocessed using normalization. Here, we take state 1 as an example for illustration: The average current waveform at the state 1 periodic level is as follows... Figure 4 As shown, the principal harmonic components are: DC component (3.51%), fundamental frequency (50.67%), 3rd harmonic (23.71%), 5th harmonic (11.73%), 9th harmonic (3.70%), 11th harmonic (2.00%), and 13th harmonic (2.26%). Therefore, the selector switch state for the principal harmonic components is:

[0126]

[0127] The specific parameters of the dynamic harmonic admittance are calculated, and the time-varying function curve is obtained by polynomial fitting, as shown in the figure. Figure 5 As shown in the figure, the solid lines represent the discrete harmonic admittance parameter sequence, and the dashed lines represent the dynamic harmonic admittance parameters obtained by polynomial fitting.

[0128] The calculation results show that the resistance parameter in the harmonic admittance parameters corresponding to the 1st, 3rd, 5th, 7th, and 11th harmonics in the current waveform is negative. Therefore, the power supply direction for this harmonic should be reversed. Hence, the power supply direction selection parameter S... vk As shown in Table 1.

[0129] Table 1 Status 1 Power Direction

[0130] <![CDATA[Power direction S vk > 1 0 0 0 0 0 1

[0131] The dynamic harmonic admittance parameters for all five states are obtained using the same operating steps. Combined with the time information for each state, a complete load simulation model is established, and simulated load current waveform data is generated as follows: Figure 6 As shown in the figure, the simulated waveform and the measured waveform basically overlap. The fitting effect is good for each state throughout the process. The fitting effect is poor during the transition between states 1 and 2 and between 2 and 3. This is because the current amplitude changes too abruptly during the state transition.

[0132] As can be seen from the above research materials, the method of the present invention has a good fitting effect in the simulation and generation of long-term multi-state load current data.

[0133] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0134] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these modifications and improvements all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for generating residential load data based on dynamic harmonic admittance parameters, characterized in that, Includes the following steps: Step 1: Perform state division and normalization preprocessing on the raw load current data intercepted according to the sampling rules, and then perform fast Fourier transform calculation to obtain harmonic current vector data and define voltage vector data. Step 2: Calculate the admittance vector based on the harmonic current and voltage vector data. Use the admittance transfer method to constrain the admittance vector to the first and fourth quadrants to obtain the discrete harmonic admittance parameters for each period. Then, obtain the dynamic harmonic admittance parameters by polynomial fitting and introducing random variables of simulation error. Step 3: Based on the dynamic harmonic admittance parameters and load state switching time information, build a complete load simulation model on the Simulink platform, consisting of a power supply simulation model, a load admittance equivalent simulation model under single operating conditions, and a load state switch. Finally, simulated load current waveform data is generated. The specific method is as follows: Step 3.1: Establish a power supply simulation model SM1 using a DC voltage source, an AC voltage source, and a grounding element; Considering only the first 15 harmonic components and DC components in the current waveform, a total of 1 pair of DC voltage sources in opposite directions and 15 pairs of AC voltage sources in opposite directions are needed to equivalently realize the positive / reverse voltage sources of the DC components and each harmonic component. The same side port of all voltage sources is connected to the grounding element, and the other side port is used as the output terminal. Step 3.2: Use a single-pole single-throw switch, clock, variable resistor, variable inductor, variable capacitor, and MATLAB Function components to establish an equivalent simulation model SM2 of the load admittance under a single operating state; To address each harmonic, firstly, two single-pole single-throw switches are combined into a single-pole double-throw switch as a voltage direction selector. This selector has two input ports and one output port for connecting to the input of the RLC module. The control parameter is the voltage direction parameter S. vk If S vk =1, the switch connected to the positive voltage source is on, and the switch connected to the reverse voltage source is off; if S vk =0, the switch connected to the reverse voltage source is turned on, and the switch connected to the forward voltage source is turned off; next, a variable resistor, a variable inductor and a variable capacitor are connected in parallel to form a parallel RLC module. Its input terminal has one port to connect to the output terminal of the voltage direction selector, and its output terminal has one port to connect to the input terminal of the harmonic principal component selector. Its resistance, inductance and capacitance values ​​are all output by MATLAB Function components. Step 3.3: Build a load state switch SM3 using a single-pole single-throw switch, a clock, and a MATLAB Function component. SM3 has one input port and one output port. The inputs to the MATLAB Function component are the clock signal t and the load on / off time T. on / T off The output is a status switch signal K. s The functional relationships within the components are shown in equation (35): ； The input of a single-pole single-throw switch is K. s If K s =1, switch is on; if K =1, switch is on; s =0, switch off; Step 3.4: Combine the power supply simulation model SM1, the load admittance equivalent simulation model SM2 under single operating conditions, and the load state switch SM3 to obtain the single-state load simulation model SLM.

2. The method for generating residential load data based on dynamic harmonic admittance parameters according to claim 1, characterized in that, Step 1: Perform state division and normalization preprocessing on the raw load current data extracted according to the sampling rules, and then perform Fast Fourier Transform calculation to obtain harmonic current vector data and define voltage vector data. The specific method is as follows: Step 1.1: Use a waveform recording device to sample the waveform. The rule is as follows: within the power frequency cycle, take the moment when the voltage waveform crosses zero upward as the starting point, and extract the original load current data. The sampling time is an integer multiple of the cycle to ensure that the current signal in each cycle can be used for harmonic decomposition calculation. Step 1.2, introduce the Pearson similarity coefficient to characterize the difference in current waveform between adjacent cycles, denoted as the waveform similarity between adjacent cycles, as shown in equation (1): ； Where i is the dynamic index of the number of periodic current sequences, and x i Let x be the current sequence of the current period. i-1 For the current sequence of the previous adjacent period, N c x represents the number of periodic current sequences in the original load data, i.e., the number of cycles. imax / x i-1max They are current sequences x i / x i-1 The maximum element value in the array, max(x) imax ,xi -1max ) represents x i Maximum element value and x i-1 The larger of the two values ​​is the maximum element value. Using equation (1) to monitor state changes in the raw load data, if the waveform similarity between adjacent cycles is f(x) i When x ≥ 0.9, the load operating status is determined to have not changed, and x is set to... i With x i-1 If f(x) is assigned to the same running state; i When the load operating status is less than 0.9, it is determined that the load operating status has changed, and the cycle sequence index of the status change occurrence is recorded. p and x i The criteria for classifying a new operating state are as shown in equation (2): ； After monitoring is completed, if multiple operating states exist in the original load data, then the record index i will be used. p The original data is segmented to obtain multiple running states as shown in equation (3): ； Where y(p) is the duration of the operating state, p is the dynamic index of the number of times the load operating state changes, and i p N is a dynamic index for the periodic sequence of load operating state changes. p This refers to the number of times the load operating status changes. The start / stop times Ton / Toff for each operating state are determined by 2πf*i p-1 With 2πf*i p Calculated; Step 1.3: Perform normalization preprocessing on the load data in each operating state; The normalization preprocessing procedure for the original electrical load data is defined, wherein the per-unit calculation process is shown in Equation (4), the averaging calculation process is shown in Equation (5), and the restoration process is shown in Equation (6): ； Where, x i-pu For a normalized periodic current sequence, x i-avg For the averaged periodic current sequence, x i-r The restored periodic current sequence is denoted as the normalized periodic current sequence x. i-N ; Step 1.4, for the standard periodic current sequence x i-N Perform a fast Fourier transform to obtain the harmonic current vectors and define the harmonic voltage vectors. For a standard periodic current sequence x i-N Performing an FFT yields the component X of its k-th harmonic. i (k), let X i (k) = a(k) + jb(k), and its modulus and phase angles |X(k)| and arg[X(k)] are shown in equations (7) to (8), respectively; ； For any term X i (k), and its corresponding k-th order time-domain signal expression is shown in equation (9): ； Where n is the sampling point number within the period, ω is the angular velocity, and A and φ are the amplitude and phase angle of the time-domain signal, respectively, as expressed by equations (10)~(11): ； Where, N s This represents the number of sampling points per cycle. The number of voltage and current signal sampling points in a single steady-state cycle is N s The magnitude and phase of each current component are calculated as shown in equations (12) to (13): ； Among them, I k φ ik The amplitude and phase angle of the k-th harmonic current; Therefore, the vectors of each harmonic current are expressed as: ； Based on the waveform sampling rules, the amplitude and phase of each harmonic voltage are defined as shown in equations (15) and (16): ； Among them, V k Let V be the amplitude of the kth harmonic voltage. N Taking the standard voltage for residential electricity in my country as 220V, φ vk S is the phase angle of the k-th harmonic voltage. vk This is a voltage direction parameter; its default value is 1. Therefore, the voltage vectors of each harmonic are expressed as: ； In addition, the proportion η of each current harmonic needs to be calculated. k =I k / I1, and define the set m of harmonic principal component orders to facilitate subsequent simulation operations. The criterion for the determination is shown in equation (18): 。 3. The method for generating residential load data based on dynamic harmonic admittance parameters according to claim 2, characterized in that, Step two: Calculate the admittance vector based on the harmonic current and voltage vector data. Use the admittance transfer method to constrain the admittance vector to the first and fourth quadrants to obtain the discrete harmonic admittance parameters for each period. Then, obtain the dynamic harmonic admittance parameters by polynomial fitting and introducing random variables of simulation error. The specific method is as follows: Step 2.1: Based on the harmonic current and voltage vectors, calculate the admittance parameters of each harmonic, including resistance, inductance, and capacitance parameters. First, calculate the amplitude and phase angle of the k-th harmonic admittance within a single period, as shown in equations (19)~(20): ； Among them, Y k φ vk Let Y be the magnitude and phase angle of the k-th harmonic admittance vector. Then, the admittance vector of this harmonic is expressed as Y. k ∠φ yk form; Based on the definition of admittance, the resistance and reactance parameters of each harmonic of the residential load are further calculated, as shown in equations (21) to (22). ； Among them, R k / X k These are the resistance / reactance parameters for the k-th harmonic, G. k / B k These are the conductivity / susceptance parameters; After FFT decomposition, observe the k-th harmonic admittance vector Y. k ∠φ yk Distribution of Y: k ∠φ yk Located in the second and third quadrants, i.e., G k <0, R k <0, considering that there is no negative resistance in the Simulink simulation software, the voltage direction parameter S in equation (16) is introduced based on the admittance transfer method. vk By adjusting the value of this parameter, the harmonic admittance vector is constrained to the first and fourth quadrants to facilitate simulation verification. The specific method for admittance transfer is as follows: When Y k ∠φ yk When G is in the first quadrant, k >0, R k >0, B k >0, X k When the value is less than 0, the resistance parameter is positive and no admittance transfer is required; the reactance parameter is negative and exhibits capacitive behavior. When Y k ∠φ yk When G is in the second quadrant, k <0, R k <0, B k >0, X k <0, at this time the resistance parameter is negative and the reactance parameter is negative, exhibiting capacitive behavior; let the voltage direction parameter S in equation (16) be negative. vk If the value is 0, that is, the voltage source for this harmonic is inverted, then φ will be zero. vk The resistor R changes from 0 to π after inversion. k =1 / Y k cos(φ yk -π)=-1 / Y k cosφ yk >0, reactance parameter X k =-1 / Y k sin(φ yk -π)=1 / Y k cosφ yk The reactance opposite to that before the phase reversal is positive, exhibiting inductive properties; therefore, Y... k ∠φ yk Move from the second quadrant to the fourth quadrant; When Y k ∠φ yk When located in the third quadrant, G k <0, R k <0, B k <0,X k >0, at this time the resistance parameter is negative and the reactance parameter is positive, exhibiting inductive behavior; let the voltage direction parameter S in equation (16) be 0. vk If the value is 0, that is, the voltage source for this harmonic is inverted, then φ will be zero. vk The resistor R changes from 0 to π after inversion. k =1 / Y k cos(φ yk -π)=-1 / Y k cosφ yk >0, reactance parameter X k =-1 / Y k sin(φ yk -π)=1 / Y k cosφ yk Since the reactance is negative (opposite to the direction of the inverting current), it exhibits capacitive behavior; therefore, Y... k ∠φ yk Move from the third quadrant to the first quadrant; When Y k ∠φ yk When located in the fourth quadrant, G k >0, R k >0, B k <0,X k When the value is >0, the resistance parameter is positive, and no admittance transfer is required. The reactance parameter is positive, and the behavior is inductive. Based on the admittance transfer steps described above, update the resistance and reactance parameters: ； Among them, R tk X tk These are the updated resistance and reactance parameters, respectively. The inductance / capacitance parameters of each harmonic are calculated as shown in equations (25) and (26): ； Among them, L tk / C tk The parameters are the inductance / capacitance of the kth harmonic, ω is the angular velocity, and f is the frequency of electricity consumption in my country, taken as 50Hz. Step 2.2, obtain the harmonic admittance parameter sequence {R} for all periods. tk0 ,R tk1 ,…,R tki }, { L tk0 ,L tk1 ,…,L tki } and { C tk0 C tk1 ,…,C tki }, where i is the dynamic index of the number of periods, with time t=i / f as the independent variable and the harmonic admittance parameters as the dependent variable. Polynomial fitting is used to convert the discrete harmonic admittance parameter sequence into a continuous time-varying function R. tk (t), L tk (t), C tk (t), which is the dynamic harmonic admittance parameter; Step 2.3 involves further introducing random variables following a probability distribution into the dynamic harmonic admittance parameters to simulate the random errors of the actual load and generate electrical load range currents that take into account the actual errors. The specific method is as follows: Based on the distribution of elements in the harmonic admittance parameter sequence in step 2.2, the mean and standard deviation μ and σ of each harmonic admittance parameter sequence are calculated. A random variable x is introduced into the RLC parameter of the mathematical model, and it is assumed to follow a normal distribution. Its probability density f(x) expression is shown in equation (27): ； Therefore, the expression for the dynamic harmonic admittance parameter is updated to: ； Among them, R tk / L tk / C tk With R tk ' / L tk ' / C tk 'These are the resistance / inductance / capacitor parameters before and after the update, x' rk / x lk / x ck These are the random variables corresponding to the resistance / inductance / capacitance parameters.

4. The method for generating residential load data based on dynamic harmonic admittance parameters according to claim 3, characterized in that, Step 3.2 also includes: The input to the MATLAB Function is the clock signal t and the load operation status start / stop time T. on / T off The output is the resistance / inductance / capacitance parameter R. tk ' / L tk ' / C tk Assume T = tT on For the duration of the load operation state, the functional relationships within the component are shown in equations (31) to (33): ； To reduce the computational load of the simulation software and improve the simulation speed, a single-pole single-throw switch is used as the harmonic principal component selector. Its input has one port to connect to the output of a parallel RLC module, and its output has one port as the output of the single-harmonic admittance equivalent model. The control parameter is the harmonic principal component judgment parameter S. mk If S mk =1, switch is on; if S =1, switch is on; mk =0, switch off, S mk The rules for determining the value are shown in equation (34): ； In summary, the voltage direction selector, the parallel RLC module, and the harmonic principal component selector are connected in series to construct a single harmonic admittance equivalent model. Finally, the input of the admittance equivalent model of all harmonic orders is used as the input of the load admittance equivalent simulation model SM2 under single operating conditions, and the output is connected as the output of the load admittance equivalent simulation model SM2 under single operating conditions. Step 3.4 also includes the following combination methods: The output terminal of the power supply simulation model SM1 is connected to the input terminal of the load admittance equivalent simulation model SM2 under single operation state according to the harmonic order and voltage direction. Connect the output of the load admittance equivalent simulation model SM2 under single operating conditions to the input of the load state switch SM3; Use the output of the load state switcher SM3 as the output of the single-state load simulation model SLM. Connect all the outputs of the single-state load simulation model (SLM) to each other and ground them to serve as the outputs of the complete load simulation model (CLM). Therefore, the load current simulation waveform data can be obtained by measuring the output of the complete load simulation model (CLM) using an ammeter element.

5. A home appliance operation status monitoring system based on an online feature library, characterized in that, The method for generating residential load data based on dynamic harmonic admittance parameters, as described in any one of claims 1-4, enables the generation of residential load data based on dynamic harmonic admittance parameters.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the method for generating residential load data based on dynamic harmonic admittance parameters as described in any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for generating residential load data based on dynamic harmonic admittance parameters as described in any one of claims 1-4.

Images