Generator dynamic metering error compensation method and device

Abstract

The invention provides a generator dynamic metering error compensation method and device, relates to the technical field of generator metering errors, and is used for improving the technical problem of insufficient metering error compensation effect in the dynamic operation process of a generator in related technologies and improving the metering accuracy under a dynamic working condition. The method comprises the steps of obtaining generator operation parameters at the current moment; acquiring an original power measurement value of the generator at the current moment; according to the obtained operation parameters of the generator, an operation state influence factor at the current moment is calculated, and the operation state influence factor is used for representing the influence of the operation state of the generator on the metering error; estimating the dynamic metering error at the current moment according to the operation state influence factor and the estimated dynamic metering error at the previous moment; and obtaining a compensation power measurement value at the current moment based on the original power measurement value and the dynamic metering error at the current moment.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of generator metering errors, and in particular to a method and device for compensating for dynamic metering errors of generators. Background Art

[0002] Generators, as core equipment in power systems, require precise measurement of the electrical energy they generate. This not only impacts the economic profitability of power generation companies but also underpins power system dispatching, operational control, and market transactions. Traditional generator energy metering typically relies on voltage transformers (VTs) and current transformers (CTs) in conjunction with energy meters. These metering devices and methods offer high accuracy when the generator is operating in a stable state (e.g., when voltage and current waveforms are close to pure sine waves and the frequency is stable at the rated value).

[0003] However, during dynamic generator operation, such as startup, shutdown, load shedding, load addition, system faults (such as short circuits and ground faults), or grid disturbances (such as frequency fluctuations and voltage sags / swells), the voltage and current waveforms can become significantly distorted (including numerous harmonics and non-periodic transient components), while the frequency can rapidly deviate from the rated value. Under these dynamic conditions, existing metering equipment and methods face challenges. On the one hand, the response characteristics of traditional VTs and CTs deteriorate during non-fundamental frequencies or high-current transients. In particular, current transformers are prone to saturation under high current surges, resulting in nonlinear errors between the secondary output and the actual primary value. On the other hand, the metering algorithms of existing electricity meters are typically based on sampling and calculating the fundamental component or assuming integration within a standard power frequency cycle. This makes it difficult to accurately process non-steady-state signals containing numerous harmonics, transient components, and rapidly varying frequencies, thus introducing significant metering errors during dynamic operations.

[0004] The existence of these dynamic metering errors leads to a large deviation between the actual metering results and the true value when the generator changes load or experiences dynamic processes such as system disturbances. This not only affects the accuracy of the economic settlement of the power plant, but may also mislead the dispatching decisions and control strategies of the power system, and even affect the reliability of the protection functions that rely on accurate power calculations. Therefore, how to effectively compensate for the metering errors of the generator during dynamic operation has become a technical problem that needs to be solved urgently in the field of power metering. In the existing technology, although there have been some attempts to improve the dynamic measurement capabilities through hardware design or software algorithms, there are generally limited compensation effects, failure to fully consider the complex influencing factors in the dynamic process, or lack of description of the dynamic accumulation and saturation characteristics of the metering errors, resulting in insufficient compensation accuracy under complex and changeable dynamic working conditions. Summary of the Invention

[0005] The embodiments of the present application provide a method and device for compensating for dynamic metering errors of generators, which are used to improve the technical problem of insufficient metering error compensation effect during dynamic operation of generators in related technologies and to improve metering accuracy under dynamic working conditions.

[0006] To achieve the above objectives, the embodiments of the present application adopt the following technical solutions: In a first aspect, the present application provides a method for compensating a dynamic metering error of a generator, comprising: obtaining the operating parameters of the generator at a current moment; obtaining the original power measurement value of the generator at a current moment; calculating the operating state influence factor at a current moment based on the obtained operating parameters of the generator, wherein the operating state influence factor is used to characterize the influence of the operating state of the generator on the metering error; estimating the dynamic metering error at a current moment based on the operating state influence factor and the estimated dynamic metering error at a previous moment; and obtaining the compensated power measurement value at a current moment based on the original power measurement value and the dynamic metering error at a current moment.

[0007] In a possible implementation of the first aspect, the calculation of the operating status influencing factor includes: obtaining the deviation of at least one parameter of the generator operating parameters relative to its calibrated value; normalizing the deviation of the at least one parameter; performing a nonlinear transformation on the normalized deviation value; and combining the deviation values ​​after the nonlinear transformation to obtain the operating status influencing factor.

[0008] In a possible implementation of the first aspect, the dynamic metering error is obtained based on a dynamic error model, and an update rule of the dynamic error model adopts a discrete time form, wherein the discrete time form indicates that the estimated dynamic metering error at the current moment is an incremental update of the estimated dynamic metering error at the previous moment based on the operating state influencing factor; wherein the incremental update is determined by an incremental update formula, and the incremental update formula includes a driving term that depends on the operating state influencing factor and a saturation term that depends on the estimated error at the previous moment and the maximum potential error.

[0009] In a possible implementation of the first aspect, the dynamic error model has a parameter characterizing the maximum potential error, and the incremental update formula is determined based on the estimated dynamic measurement error at the previous moment, the operating state influencing factor, and the maximum potential error parameter.

[0010] In a possible implementation manner of the first aspect, the method further includes a step of correcting the dynamic error model, and the correcting step includes: periodically obtaining a reference power measurement value of the generator under a specific operating state; Obtaining an error residual based on the reference power measurement value and the compensated power measurement value obtained under the corresponding specific operating state; A parameter representing a maximum potential error in the dynamic error model is adjusted according to the error residual.

[0011] In a possible implementation of the first aspect, adjusting a parameter representing a maximum potential error in the dynamic error model includes: Converting the error residual through a preset adjustment function to obtain a parameter adjustment amount, wherein the preset adjustment function is a proportional function, and the parameter adjustment amount is proportional to the error residual; The parameter adjustment amount is applied to the current parameter representing the maximum potential error to obtain an updated parameter representing the maximum potential error.

[0012] In a possible implementation of the first aspect, the generator operating parameter includes at least one parameter, and the at least one parameter is selected from one of voltage, current, frequency, active power, reactive power, power factor, speed and temperature.

[0013] In the second aspect, the present application also provides a generator dynamic metering error compensation device, including: a parameter acquisition module for acquiring the generator operating parameters at the current moment; a measurement acquisition module for acquiring the original power measurement value of the generator at the current moment; an influence factor calculation module for calculating the operating state influence factor at the current moment based on the acquired generator operating parameters, wherein the operating state influence factor characterizes the comprehensive influence of the generator operating state on the metering error; an error estimation module for estimating the dynamic metering error at the current moment based on the calculated operating state influence factor and the estimated dynamic metering error at the previous moment; a compensation module for obtaining the compensated power measurement value at the current moment, wherein the compensated power measurement value is based on the acquired original power measurement value and the estimated dynamic metering error.

[0014] In a possible implementation of the second aspect, the impact factor calculation module is used to obtain the deviation of at least one parameter of the generator operating parameters relative to its calibrated value, normalize the deviation of the at least one parameter, perform nonlinear transformation on the normalized deviation value, and combine the deviation values ​​after nonlinear transformation to obtain the operating status impact factor.

[0015] In a possible implementation of the second aspect, the dynamic metering error is obtained based on a dynamic error model, and an update rule of the dynamic error model adopts a discrete time form, wherein the discrete time form indicates that the estimated dynamic metering error at a current moment is an incremental update of the estimated dynamic metering error at a previous moment based on the operating state influencing factor; The incremental update is determined by an incremental update formula, which includes a driving term that depends on the operating state influencing factor and a saturation term that depends on the estimation error at the previous moment and the maximum potential error. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flowchart of a method for compensating for dynamic metering errors of a generator provided in some embodiments of the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments.

[0018] Hereinafter, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature identified with "first," "second," etc., may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0019] In addition, in this application, directional terms such as "up", "down", "left", and "right" may be defined including but not limited to the orientation relative to the schematic placement of the components in the drawings. It should be understood that these directional terms may be relative concepts. They are used for relative descriptions and clarifications, and they may change accordingly according to changes in the orientation of the components in the drawings.

[0020] In this application, unless otherwise specified or limited, the term "connection" should be understood broadly. For example, "connection" can mean fixed connection, detachable connection, or integration; it can mean direct connection or indirect connection through an intermediate medium. In addition, the term "electrical connection" can refer to the method of electrical connection that enables signal transmission.

[0021] As used herein, “about,” “substantially,” or “approximately” includes the stated value and reference values ​​that are within an acceptable range of deviation from the particular value, as determined by one of ordinary skill in the art taking into account the measurements in question and the errors associated with the measurement of the particular quantity (i.e., the limitations of the measurement method).

[0022] The present application provides a generator dynamic metering error compensation device and its use method, aiming to solve the problem of insufficient accuracy of existing generator metering methods when dealing with dynamic changes in operating parameters. By establishing a dynamic error model and introducing a correction feedback mechanism, the metering accuracy under various operating conditions is improved.

[0023] Figure 1 The flow chart of the method for compensating the dynamic measurement error of the generator of the present application is shown. Figure 1 , the method comprising: S101. Obtain the current generator operating parameters. Obtain various key operating data of the generator at the current moment. It is understandable that these parameters reflect the generator's immediate operating status and are key factors affecting its measurement error.

[0024] For example, the generator operating parameters include but are not limited to the generator output voltage U, output current I, grid frequency f, generator internal temperature or ambient temperature T, output active power P raw , output reactive power Q raw , power factor PF and speed RPM, etc. The parameters can be collected in real time by various sensors, transmitters and smart meters installed at the generator or grid connection point.

[0025] Exemplarily, the above parameters are obtained in the form of discrete time series data, and the sampling period Δt is determined according to actual application requirements. The time interval Δt is a system configuration parameter used to determine the update frequency of the model, and its value can be determined during system design or deployment.

[0026] S102: Obtain the original power measurement value of the generator at the current moment.

[0027] For example, the raw power measurement value may be the output active power measurement value P of the generator. meas Or reactive power measurement value Q meas , which is usually output by conventional metering instruments connected to the generator output side. The data collected by these instruments are also sampled at a period similar to the operating parameters. In this embodiment, the raw power measurement value specifically refers to the raw active power measurement value P meas .

[0028] S103: Calculate the operating status influencing factor at the current moment according to the acquired generator operating parameters.

[0029] For example, this embodiment integrates multi-dimensional generator operating parameter information into a single parameter, the Operating State Influence Factor (OSIF), which quantifies the potential impact of the current operating state on metering error. The OSIF characterizes the degree to which the current generator operating state deviates from its nominal operating conditions and its overall impact on the generation and accumulation of metering error. It is understood that the OSIF is a unique, non-negative value.

[0030] Exemplarily, the calculation of the operating status influencing factor includes: Obtain the deviation of at least one of the generator operating parameters relative to its calibration value. For example, for the parameter pᵢ(k) at the current moment k (such as the voltage U k ), whose calibration value is pᵢ ，nom , then the deviation is Δpᵢ(k)=pᵢ(k)-pᵢ ，nom The calibration value pᵢ ，nom These are the design, rated, or calibrated operating point parameter values ​​for generators, instrument transformers, or meters. These values ​​are typically obtained from equipment specifications, factory documentation, or historical stable operating data. They are determined during system design or deployment and are preset system parameters.

[0031] Normalize the deviation of the at least one parameter. It is understood that normalization is used to eliminate the dimensional differences of different parameters and unify their influence ranges. For example, the absolute value of the relative deviation can be used for normalization: The scale factor sᵢ is used to scale the deviation to a similar numerical range, such as the rated value, full-scale value, or statistical characteristics of historical data based on the parameter (such as standard deviation or maximum absolute deviation). These values ​​can usually be obtained by performing offline statistical analysis on the system's historical operating data or setting them according to the equipment's technical indicators. They are preset system parameters. sᵢ is usually a constant greater than zero, and its specific value depends on the physical dimension and expected range of variation of the corresponding parameter. For example, for voltage U, sᵢ is, for example, U It can be taken as a value between 1% and 5% of the rated voltage of the generator, such as 3% of the rated voltage value; for frequency f, for example, s f It can be taken as a value between 0.1% and 0.5% of the nominal frequency of the power grid, such as 0.2 Hz; for temperature T, for example, s T It can be taken as a value within the typical range of expected ambient temperature variation (such as 10 degrees Celsius to 30 degrees Celsius), such as 20 degrees Celsius.

[0032] The normalized deviation value is subjected to a nonlinear transformation. This nonlinear transformation of the normalized deviation value can amplify the effect of deviations away from the calibration point, reflecting the characteristic that errors grow nonlinearly with parameters. For example, the error of a metering device may be related to the square of the current or voltage, and the effect of temperature increase on resistance may be exponential. The nonlinear transformation is to perform a power operation higher than the first power on the normalized deviation value: , where αᵢ is an exponent greater than 1, for example, it can be 2 (corresponding to the square relationship) or other real numbers greater than 1. For example, α U It can be 2, α f It can be 1.5, α TThe value is 2. The exponent αᵢ determines the strength of the nonlinearity, reflecting the nonlinear exacerbation of error by parameter deviation. These exponent values ​​are typically determined through offline analysis of the nonlinear relationship between parameter deviations and corresponding errors observed in historical operating data, or based on an understanding of the physical characteristics of the equipment. They are preset model parameters.

[0033] The deviation values ​​after the nonlinear conversion are combined to obtain the operating state influencing factor. The combination can be a weighted summation of the deviation values ​​after the nonlinear conversion. The weight wᵢ reflects the relative influence of the corresponding parameter pᵢ on the measurement error. For example, if the voltage deviation has a greater influence on the error than the temperature, the weight w of the voltage deviation is U will be greater than the weight w of the temperature deviation T It is understandable that the above weight values ​​can be determined by offline analysis of the correlation between the changes in various parameters and the corresponding error changes in historical operation data (such as multivariate regression) or based on expert experience, and are preset model parameters. wi is usually a non-negative constant, and its relative size reflects the importance of the influence of the corresponding parameter. For example, all w can be normalized to i The sum is 1, for example, we can take w U =0.5, w f =0.3, w T =0.2w U =0.5, w f =0.3, w T =0.2.

[0034] For example, assuming that only voltage (U), frequency (f), and temperature (T) parameters are considered, and the operating parameters at the current moment k are U(k), f(k), and T(k), then OSIF(k) can be calculated as: Among them, the weight w U , w f , w T , scale factor s U , s f , s T , and the exponent α U , α f , α T These are all system parameters determined in advance through historical data analysis, offline modeling or expert experience.

[0035] Step S104: Estimate the dynamic metering error at the current moment according to the operating state influencing factor and the estimated dynamic metering error at the previous moment.

[0036] Existing technologies, such as static calibration or simple linear models, struggle to capture the complex dynamic evolution of errors as they evolve with operating conditions. In some scenarios, metrology errors can dynamically generate and accumulate with operating conditions, saturating when they reach certain physical or system limits.

[0037] Exemplarily, the step of estimating the dynamic measurement error of the present application is based on a preset dynamic error model, and the updating rule of the model adopts its discrete time form. Let E(k) be the estimated dynamic measurement error at the current moment k, E(k-1) be the estimated dynamic measurement error at the previous moment k-1, OSIF(k) be the operating state influence factor calculated at the current moment k, and Δt be the sampling time interval. The dynamic error model has a parameter that characterizes the maximum potential error, denoted as E max This parameter defines the maximum steady-state value or saturation upper limit that the dynamic measurement error can theoretically reach under the current operating state. When iteratively calculating the error E(k) at the current moment k, the model uses the maximum potential error value determined at the end of the previous moment k-1 or the beginning of the current moment, which is recorded as E max (k-1).

[0038] Exemplarily, the step of estimating the dynamic metrology error uses the following iterative update formula to calculate the error variation ΔE: Where Rate is the instantaneous rate of change of the error, which is determined by the following formula: Where E(k) represents the estimated dynamic metrology error at the current time k. At the first time step, or after external high-precision calibration or system reset, the initial value of E(k-1) is typically set to zero or the known error value obtained through calibration at that time. E(k-1) represents the estimated dynamic metrology error at the previous time k-1. During the iteration process, this parameter is the value calculated and output in the previous step S104 during the previous time period.

[0039] Δt represents the time interval between the current moment and the previous moment, which is determined by the parameter acquisition period in step S101 and is used to convert the instantaneous change rate Rate into the total error change ΔE within the time interval.

[0040] Indicates the maximum potential error parameter used when calculating the error at the current moment, representing the limit value that the dynamic measurement error may approach under the current operating state. The initial value of can be determined by analyzing the maximum error observed in historical data offline or by estimating it based on the physical characteristics of the device.

[0041] A(OSIF(k)) and B(OSIF(k)) are functions that depend on the operating state influencing factor OSIF(k). Together, they determine the specific driving force and form of the error growth rate, Rate, for a given value of OSIF(k). These two functions map the external operating state influence OSIF(k) into internal driving force parameters for the dynamic changes in the error.

[0042] A(OSIF(k)) represents the portion of the error growth rate that is relatively independent of the current accumulated error level, E(k-1). It describes the fundamental component of the error generation rate that is directly caused by the current operating state and persists even when the current error is small. For example, sustained temperature deviations from the calibration value may cause a fundamental component drift rate that is not significantly dependent on the total amount of accumulated drift within a time step. Larger values ​​of A(OSIF(k)) indicate a stronger fundamental driving force of the current operating state on error generation.

[0043] B(OSIF(k)) represents the part of the error growth rate that is related to the current accumulated error level E(k-1). max (k-1)) to form a driving term that depends on the current error ratio B(OSIF(k)) describes the tendency for errors to self-accelerate or reinforce: the greater the current accumulated error (relative to its upper limit), the faster the rate of further error accumulation driven by unfavorable operating conditions, OSIF(k). This cumulative effect can arise from continued stress accumulation, increased secondary effects due to temperature rise, or accelerated component aging. Larger values ​​of B(OSIF(k)) indicate a more pronounced self-reinforcement or cumulative effect of errors under the current operating conditions.

[0044] The specific forms of the functions A(OSIF) and B(OSIF) (for example, they can be , Or more complex nonlinear functions such as polynomials, exponential functions, etc.) and the fixed parameters contained therein (such as a0, a1, b0, b1, etc. in linear form) can be determined through an offline training process. The values ​​of these parameters depend on the fitting results of historical data and can be arbitrary real numbers. Their signs and sizes reflect how OSIF affects the "innovation" and "imitation" growth rates of the error. For example, if a1 is positive, it means that the larger the OSIF, the higher the basic error growth rate; if b1 is positive, it means that the larger the OSIF, the stronger the self-acceleration effect of the error. The values ​​of these parameters do not have a universal numerical range and must be fitted according to the specific generator system and historical data. Their dimensions match Rate / (error dimension). For example, assuming that OSIF is dimensionless, the error E dimension is kW, and the time dimension is s, then the dimension of Rate is kW / s. At this time, if , then a0 dimension is kW / s, a1 dimension is kW / s; if , due to its relationship with (E / E max ) should have the dimension of 1 / s, and E / E max Since B(OSIF) is dimensionless, it has a dimension of 1 / s, and thus b0 and b1 have a dimension of 1 / s. For example, assuming the OSIF calculation range is typically [0, 10], the maximum error E is 200 kW, and Δt is 0.01 s, example values ​​for the parameters after offline training might be: a0 = 0.1 kW / s, a1 = 0.05 kW / s, b0 = 0.021 / s, and b1 = 0.011 / s.

[0045] The offline training process includes: Collect historical operating parameter data for the generator under various operating conditions, along with actual metrological error data obtained through high-precision reference measurements or calibration at the same or similar time points corresponding to these operating conditions. This historical data should cover as wide a range of operating parameters, loads, and durations as possible to reflect the dynamic behavior and cumulative characteristics of the error under different operating conditions.

[0046] For each set of historical operating parameter data collected, the corresponding historical operating status impact factor OSIF is calculated using the formula and parameters determined in step S103. hist .

[0047] Using the collected historical OSIF hist Data and corresponding actual error data Error hist The dynamic error model (discrete time form) is applied to these historical data series. The specific form of functions A(OSIF) and B(OSIF) and their internal fixed parameters (such as a0, a1, b0, b1 in linear form) are determined by optimization algorithms, and E can also be optimized. max The optimization goal is to minimize the difference (e.g., mean square error, maximum absolute error, etc.) between the estimated error sequence obtained by iterative model calculation using a certain function and parameters and the historical actual error sequence. It can be understood that this is essentially a system identification or parameter estimation problem based on historical observation data. Common optimization methods include but are not limited to parameter optimization algorithms such as least squares, nonlinear regression, gradient descent, and genetic algorithms. These algorithms iteratively adjust the model parameters to make the model output error closest to the actual observation error.

[0048] After the function form and fixed parameters are determined, during the real-time operation process, the OSIF(k) value calculated at each moment will be substituted into the functions A and B determined through offline training, thereby obtaining the specific values ​​A(OSIF(k)) and B(OSIF(k)) required at the current moment, which are used for the iterative calculation of the error estimate for the current time step.

[0049] (E max (k-1)-E(k-1)) represents the distance between the current estimated error E(k-1) and the theoretical maximum potential error E max How much "growth space" is left when E(k-1) is much smaller than E max (k-1), the term is larger, allowing a larger growth rate; when E(k-1) gradually increases and approaches E max (k-1), the term approaches zero, which makes the entire error rate approach zero and the error growth stagnate. In this way, the model provided in this application accurately describes the physical phenomenon of error saturation, that is, the accumulation of error is not infinite, but has an upper limit under certain conditions. This factor forces the prediction error of the model to be limited to E max (k-1) to prevent the predicted value from diverging infinitely.

[0050] (E(k-1) / E max (k-1)) This term indicates that the estimated error E(k-1) at the previous moment accounts for the maximum potential error E max The ratio of (k-1). Its value is usually between 0 and 1. In the error change rate formula, this proportional factor works together with B(OSIF(k)) to enhance the cumulative effect of the error. The specific term is This term simulates the accelerating growth or self-reinforcing effect based on the current error level, i.e. the more the current accumulated error (relative to its potential upper limit E max (k-1)), the further error growth momentum (the part contributed by item B) is stronger under the driving force of the unfavorable operating state OSIF(k).

[0051] Thus, by combining the above two factors, the dynamic behavior described by the error change rate Rate has a typical S-shaped curve characteristic (when OSIF and model parameters are constant): when E(k-1) starts to grow from zero, the initial growth rate is mainly determined by the A term and the remaining capacity; as E(k-1) increases, especially when E(k-1) is approximately equal to E max When E(k-1) / 2, the growth contributed by B reaches its maximum, causing the overall rate to reach its peak and the error growth to accelerate; then, as E(k-1) approaches E max (k-1), saturation factor (E maxThe influence of (k-1)-E(k-1)) gradually dominates, causing the growth rate to slow down again and eventually approach zero. The error saturates at E max (k-1). This S-shaped growth dynamics more accurately simulates the growth and saturation processes that occur over time or in accumulated quantities in many real-world generator systems, such as heat accumulation, component wear, or concentration changes in certain chemical reactions. This improves the ability to describe the evolution of dynamic metering errors in generators. Furthermore, because the driving coefficients A and B are functions of the OSIF, the model can adapt to changes in operating conditions and dynamically adjust the speed and form of error accumulation, making it more flexible and adaptable than fixed-parameter S-shaped models.

[0052] Thus, the estimated dynamic metrology error E(k) at the current moment is obtained by adding the estimated value at the previous moment to the calculated change: E(k) = E(k-1) + ΔE. It can be understood that the E(k) value will be used as E(k-1) for the new iterative calculation at the next time step.

[0053] Step S105: Obtain a compensated power measurement value at the current moment based on the original power measurement value and the dynamic metering error at the current moment.

[0054] For example, the compensation power measurement value P at the current time k is comp (k) The original power measurement value P at the current time k meas (k) is added to the estimated dynamic measurement error E(k) to obtain: P comp (k)=P meas (k)+E(k) If the estimated error is negative, it is equivalent to subtracting its absolute value from the original measurement value. The compensated power measurement value is the final output after processing by this method and is used for subsequent metering, settlement or grid scheduling.

[0055] The above steps constitute a dynamic error compensation method. In order to further improve the accuracy and adaptability of the model, especially to cope with the characteristic changes caused by long-term operation of the generator or the seasonal impact of the environment, the method of the present application also includes the step of correcting the dynamic error model.

[0056] The correction step is performed periodically to calibrate the key parameters of the model online, especially the parameters that characterize the error "capacity" or "upper limit" (E max This allows the upper bound of the model to be adjusted to accommodate observed long-term error trends or the effects of unmodeled factors, eliminating the need for frequent and expensive full calibrations.

[0057] Exemplarily, the correction step includes: Periodically obtain the reference power measurement value P of the generator under a specific operating state ref The specific operating state can be when the generator is in stable operation, when a high-precision reference instrument is connected for measurement during planned maintenance, or when measurements are taken at an operating point where the error is known to be very small. It is understood that the reference measurement value is considered to be the "true" value or a high-precision estimate under the current conditions.

[0058] Based on the reference measurement value and the compensated power measurement value obtained under the corresponding specific operating state, an error residual R is calculated. The residual reflects the difference between the above compensation result and the high-precision reference value.

[0059] R=P ref -P comp (After obtaining the reference measurement value P ref At the same or similar time, obtain the operating parameters, execute S103-S105 to obtain P comp ).

[0060] According to the calculated error residual, adjust the parameter E in the dynamic error model that represents the maximum potential error max For example, if the compensation value of this method is continuously higher than the reference value (R is negative), it may mean that under the current conditions, the potential error upper limit E of the model estimation is max is underestimated, and the actual error may be higher than the model expects; if the compensation value is continuously lower than the reference value (R is positive), it may mean that E max Overrated.

[0061] Thus, by adjusting E max , the error model is able to continuously correct its predictions about the error saturation range, thus improving long-term accuracy.

[0062] Exemplarily, adjusting the parameter representing the maximum potential error in the dynamic error model includes: Converting the error residual through a preset adjustment function to obtain a parameter adjustment amount; The parameter adjustment amount is applied to the current parameter representing the maximum potential error to obtain an updated parameter representing the maximum potential error.

[0063] The preset adjustment function may be a proportional function, for example, the parameter adjustment amount , where β is a learning rate or adjustment gain factor. The value of β determines the E maxThe amplitude of the adjustment is usually determined through system debugging, simulation or empirical trial and error. It is a constant greater than zero. Its size affects the response speed and stability of the adjustment. For example, it can be a value in the range of [0.001, 0.1], such as 0.05, to ensure the stability and convergence of the adjustment.

[0064] Updated E max (k)=E max (k-1)+ΔE max _ adj In order to enhance the smoothness and robustness of the adjustment, avoid drastic fluctuations caused by noise, or introduce forgetting of old information, exponential smoothing updates can be used, for example: in, It is a smoothing factor or forgetting factor (its definition and value are different from αᵢ in OSIF calculation), and its value is usually between 0 and 1. It can be determined by system debugging, such as 0.4. max The periodic adjustment of parameters enables the model to gradually adapt to the long-term characteristic changes of the generator, environmental influences or calibration drift, improving the long-term accuracy and robustness of the system.

[0065] The present application also provides a dynamic metrology error compensation device, which includes a parameter acquisition module, a measurement acquisition module, an influence factor calculation module, an error estimation module, and a compensation module.

[0066] The parameter acquisition module connects to the generator's operating parameter sensors, transmitters, or SCADA system interface to collect real-time operating parameter data such as voltage, current, frequency, and temperature. The measurement acquisition module connects to the generator's raw metering instrument (such as an watt-hour meter or power meter) to obtain raw power measurements.

[0067] The impact factor calculation module receives the operating parameters output by the parameter acquisition module, executes the calculation logic of step S103 above, calculates and outputs the operating status impact factor. The module is pre-set or stored with various parameters (pᵢ ，nom , sᵢ, αᵢ, wᵢ, and specific parameters such as U nom , s U , α U , w U The module can be a processor unit with a specific calculation circuit or programmed to implement the OSIF calculation formula.

[0068] The error estimation module receives the OSIF(k) output by the impact factor calculation module and the stored estimated dynamic measurement error E(k-1) at the previous moment and the maximum potential error E at the previous moment. max(k-1), execute the iterative update calculation of step S104 above, estimate and output the dynamic measurement error E(k) at the current moment. This module stores the current state parameters of the dynamic error model (including E(k-1) and the currently valid E max (k-1)) and fixed parameters of the model determined through offline training (such as the definition of functions A(OSIF) and B(OSIF) and their internal parameters, such as a0, a1, b0, b1, etc. in linear functions). The module is typically implemented by a processor unit, which maintains the model state and performs iterative calculations.

[0069] The compensation module receives the original power measurement value output by the measurement acquisition module and the estimated dynamic metering error output by the error estimation module, performs the combined calculation in step S105 above, and outputs the compensated power measurement value. This module can be a simple adder or software implementation.

[0070] In some embodiments, the apparatus may include a correction module for implementing the above correction step. The correction module may include a reference measurement acquisition unit, a residual calculation unit, and a parameter adjustment unit.

[0071] The reference measurement acquisition unit is used to obtain or input a high-precision reference power measurement value provided by an external source when correction is required. The residual calculation unit receives the P value provided by the reference measurement acquisition unit. ref and the P output by the compensation module at the corresponding time comp , calculate the residual R. The parameter adjustment unit receives the R output by the residual calculation unit and performs the above adjustment E max The calculation logic of model parameters such as β and α is used to store or use the control parameters for parameter adjustment. adj ). The control parameters (β, α adj ) can be determined through system debugging.

[0072] Each module of the device can be implemented by hardware circuits, programmable logic devices (FPGAs, ASICs), general-purpose processors executing software programs, or a combination thereof. The entire device can be integrated into a standalone device or implemented as part of a generator control system, SCADA system, or smart meter. The data storage unit is used to store historical operating parameters, error estimates, model parameters, etc. The model parameters include but are not limited to OSIF calculation parameters (pᵢ ，nom , sᵢ, αᵢ, wᵢ, etc.), fixed parameters of the dynamic error model (such as the definition of functions A(OSIF) and B(OSIF) and their internal parameters), and control parameters of the correction mechanism (E max The current value of , and β and optionally α used to update it adjThe OSIF calculation parameters and the fixed parameters of the dynamic error model are usually determined through the above-mentioned offline training process and loaded when the device is deployed; max Its adjustment control parameters are dynamically updated and used during the operation of the device, and their initial values ​​are also loaded during deployment.

[0073] Those skilled in the art will understand that the above embodiments are merely illustrative. Various modifications and variations may be made to the sequence of steps, specific calculation formulas, parameter values, module division, and the like of the methods and devices without departing from the principles and spirit of this application, and such modifications and variations remain within the scope of protection of this application. Determining the specific values ​​and functional forms of various parameters is an engineering practice that typically involves analysis of historical data, offline model training, and online system debugging.

[0074] Through the description of the above implementation methods, technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0075] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0076] The units described as separate components may or may not be physically separate, and the components shown as units may be one physical unit or multiple physical units, that is, they may be located in one place or distributed in multiple places. Some or all of the units shown in this embodiment may be selected according to actual needs to achieve the purpose of this embodiment.

[0077] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware.

[0078] The above content is only a specific embodiment of this application, but the scope of protection of this application is not limited to this. Any changes or replacements within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A method for compensating a dynamic measurement error of a generator, characterized in that: include: Get the generator operating parameters at the current moment; Get the raw power measurement value of the generator at the current moment; Calculating an operating state influence factor at a current moment based on the acquired generator operating parameters, wherein the operating state influence factor is used to characterize the influence of the generator operating state on the measurement error; estimating the dynamic metering error at the current moment based on the operating state influencing factor and the estimated dynamic metering error at the previous moment; A compensated power measurement value at the current moment is obtained based on the original power measurement value and the dynamic metering error at the current moment.

2. The method according to claim 1, characterized in that The calculating the operating status influencing factor includes: Obtaining a deviation of at least one of the generator operating parameters relative to a calibrated value thereof; normalizing the deviation of the at least one parameter; Performing a nonlinear transformation on the normalized deviation value; The deviation values ​​after the nonlinear conversion are combined to obtain the operating state influencing factor.

3. The method according to claim 2, characterized in that The dynamic measurement error is obtained based on a dynamic error model. The update rule of the dynamic error model adopts a discrete time form. The discrete time form indicates that the estimated dynamic measurement error at the current moment is an incremental update of the estimated dynamic measurement error at the previous moment based on the operating state influencing factor. The incremental update is determined by an incremental update formula, which includes a driving term that depends on the operating state influencing factor and a saturation term that depends on the estimation error at the previous moment and the maximum potential error.

4. The method according to claim 3, characterized in that The dynamic error model has a parameter that characterizes the maximum potential error, and the incremental update formula is determined according to the estimated dynamic measurement error at the previous moment, the operating state influencing factor, and the maximum potential error parameter.

5. The method according to claim 4, characterized in that The method further comprises the step of correcting the dynamic error model, wherein the correcting step comprises: periodically obtaining a reference power measurement value of the generator under a specific operating state; Obtaining an error residual based on the reference power measurement value and the compensated power measurement value obtained under the corresponding specific operating state; A parameter representing a maximum potential error in the dynamic error model is adjusted according to the error residual.

6. The method according to claim 5, characterized in that The adjusting of the parameter representing the maximum potential error in the dynamic error model includes: Converting the error residual through a preset adjustment function to obtain a parameter adjustment amount, wherein the preset adjustment function is a proportional function, and the parameter adjustment amount is proportional to the error residual; The parameter adjustment amount is applied to the current parameter representing the maximum potential error to obtain an updated parameter representing the maximum potential error.

7. The method according to claim 1, characterized in that The generator operating parameters include at least one parameter selected from the group consisting of voltage, current, frequency, active power, reactive power, power factor, rotation speed, and temperature.

8. A generator dynamic measurement error compensation device, characterized in that: include: Parameter acquisition module, used to obtain the generator operating parameters at the current moment; The measurement acquisition module is used to obtain the original power measurement value of the generator at the current moment; An influence factor calculation module, configured to calculate an operating state influence factor at a current moment based on the acquired generator operating parameters, wherein the operating state influence factor represents a comprehensive influence of the generator operating state on the metering error; an error estimation module, configured to estimate the dynamic metering error at a current moment based on the calculated operating state influencing factor and the estimated dynamic metering error at a previous moment; The compensation module is configured to obtain a compensated power measurement value at a current moment, where the compensated power measurement value is based on the obtained original power measurement value and the estimated dynamic metering error.

9. The device according to claim 8, characterized in that The impact factor calculation module is used to obtain the deviation of at least one parameter of the generator operating parameters relative to its calibrated value, normalize the deviation of the at least one parameter, perform nonlinear transformation on the normalized deviation value, and combine the deviation values ​​after the nonlinear transformation to obtain the operating status impact factor.

10. The device according to claim 9, characterized in that The dynamic measurement error is obtained based on a dynamic error model. The update rule of the dynamic error model adopts a discrete time form. The discrete time form indicates that the estimated dynamic measurement error at the current moment is an incremental update of the estimated dynamic measurement error at the previous moment based on the operating state influencing factor. The incremental update is determined by an incremental update formula, which includes a driving term that depends on the operating state influencing factor and a saturation term that depends on the estimation error at the previous moment and the maximum potential error.

Images