Method for accurately detecting power of wind power generation system based on direct current offset estimation compensation

Abstract

The invention discloses a method for accurately detecting the power of a wind power generation system based on direct current offset estimation compensation. The method comprises the steps of 1, performing ADC sampling on voltage and current of a power grid to obtain sampling values; 2, sequentially extracting four continuous sampling points, and grouping by taking every four sampling points as one group; step 3, any group of data is taken, direct current is eliminated through difference, an intermediate quantity related to frequency is calculated, and input is provided for frequency estimation; 4, iteratively updating the frequency by using a normalized gradient method to obtain estimated values of the intermediate variable and the frequency of the sampling data; 5, estimating the direct current offset, the amplitude and the phase; step 6, performing power calculation based on the estimated direct current offset, amplitude and phase, and eliminating direct current interference; and step 7, circularly executing the step 4 to the step 6 until the output is stable. Under the complex working condition that the power grid voltage suffers from direct current offset interference, high-precision measurement of the success rate parameter can still be achieved, and the operation reliability of a power system is enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of measuring electrical variables, and in particular relates to a method for accurately detecting power of a wind power generation system based on DC offset estimation and compensation. Background Art

[0002] As a distributed power source, the continuous increase in the proportion of wind power access has led to an increasing number of DC offsets in the grid voltage and current signals, making the measurement of grid power more complex and difficult. As an important parameter reflecting the efficiency of power transmission and the operating status of the system, the accurate measurement of power is of great significance for energy management, electricity bill settlement, and system scheduling and control. Existing power detection methods are mostly based on ideal operating conditions during design, and insufficient attention is paid to the impact of DC offset. This may introduce large errors under complex operating conditions, affecting the power detection accuracy. Therefore, there is an urgent need for a high-precision power detection method that can effectively suppress the influence of DC offset to meet the actual application needs in complex power grid environments.

[0003] To mitigate the impact of DC offset, existing power detection methods typically use bandpass filters to eliminate the DC component or construct orthogonal signals to achieve signal decoupling. However, the use of filters often introduces significant issues such as phase delay and amplitude attenuation. This makes timely and accurate tracking difficult, especially when power changes rapidly or when high dynamic response requirements are required. Furthermore, calculation methods based on orthogonal signals are prone to introducing steady-state errors when the signal contains a DC component, affecting the accuracy of active and reactive power calculations.

[0004] In summary, existing methods have the following shortcomings in scenarios with DC offset:

[0005] First, when the frequency fluctuates or the load changes suddenly, the power calculation results lag or overshoot, and the dynamic tracking capability is poor; second, the coupling effect of DC offset, noise and harmonics leads to insufficient steady-state accuracy; third, some methods have high algorithm complexity and difficult parameter tuning. Summary of the Invention

[0006] In order to overcome the shortcomings of the above-mentioned existing technologies, the purpose of the present invention is to provide a method for accurately detecting the power of a wind power generation system based on DC offset estimation and compensation. Even under complex working conditions where the grid voltage is subject to DC offset interference, high-precision measurement of power parameters can still be achieved, thereby enhancing the operational reliability of the power system.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is:

[0008] A method for accurately detecting power of a wind power generation system based on DC offset estimation and compensation comprises the following steps:

[0009] Step 1: ADC samples the grid voltage and current to obtain sampling values;

[0010] Step 2: grouping four consecutive sampling points into a group;

[0011] Step 3: Take any set of sample values, eliminate DC by differential, and calculate the intermediate quantity related to frequency to provide input for frequency estimation;

[0012] Step 4: construct a frequency estimator, and based on the intermediate variable, use the normalized gradient method to iteratively update the frequency to obtain the estimated value of the sample value intermediate variable and frequency;

[0013] Step 5: Estimate the DC offset, amplitude, and phase based on the estimated values of the intermediate variables and frequency;

[0014] Step 6: Perform power calculation based on the estimated DC offset, amplitude, and phase to eliminate DC interference;

[0015] Step 7: Repeat steps 4 to 6 until the output is stable.

[0016] In the step 2: Current: [I k,0 ,I k,1 ,I k,2 ,I k,3 ], voltage: [U k,0 ,U k,1 ,U k,2 ,U k,3 ];

[0017] Wherein, the subscripts (k, 0), (k, 1), (k, 2), and (k, 3) are the first, second, third, and fourth sampling points of the kth group of sampling data, and k = 1, 2, 3, ...

[0018] In the step three:

[0019] Use basic trigonometric formulas to calculate intermediate quantities in algebraic expressions and construct specific algebraic expressions between four consecutive samples;

[0020] The kth group of data is:

[0021] [I k,0 ,I k,1 ,I k,2 ,I k,3 ],[U k,0 ,U k,1 ,U k,2 ,U k,3 ]

[0022] [I k,0 ,I k,1 ,I k,2 ,Ik,3 ] are:

[0023]

[0024] [U k,0 ,U k,1 ,U k,2 ,U k,3 ] are:

[0025]

[0026] Where I and U represent the amplitude of current and voltage signals respectively, ω is the signal angular frequency, and f is the signal frequency. They are the current signals I k,0 , voltage signal U k,0 The phase angle, T s is the sampling time interval, f s is the sampling frequency, I dc , U dc are the DC offset contained in the current and voltage respectively.

[0027] The DC offset is explicitly eliminated by differential operation of four consecutive sampling points.

[0028] Calculate the intermediate quantities related to the current signal (the voltage signal is processed in the same way as the current signal, replacing I with U in steps 3 to 5):

[0029] Step 3-1: Define intermediate variables for

[0030]

[0031] Step 3-2: Define intermediate variables for

[0032]

[0033] Step 3-3: According to Calculate intermediate variables Construct a specific algebraic expression between four consecutive samples

[0034]

[0035] Therefore, there are

[0036]

[0037] Therefore, there are intermediate variables Thus, we can directly get the estimated value of frequency by traversing each set of sample data.

[0038] The step 4 is specifically as follows:

[0039] The normalized gradient method is used to construct a frequency estimator to achieve adaptive frequency estimation;

[0040] Step 4-1: Calculate the normalization coefficient m(k) of the kth group of data

[0041]

[0042] Intermediate variables obtained based on the kth group of sampling data;

[0043] Step 4-2: Calculate the estimated error e(k) for the kth group of data

[0044]

[0045] is the intermediate variable obtained based on the kth group of sampling data; To obtain the estimated value of the current intermediate variable γ when the kth group of sampling data is obtained, The initial value of The default value is 0.

[0046] Step 4-3: Calculate the estimated value of the intermediate variable γ when the k+1th group of data is obtained

[0047]

[0048] To obtain the estimated value of the current intermediate variable γ when the k+1th group of sampling data is obtained;

[0049] Step 4-4: Calculate the estimated value of the frequency when the k+1th group of sampling data is obtained

[0050]

[0051] The step five is specifically as follows:

[0052] Based on the estimated frequency obtained in step 4 and intermediate variables Calculate DC offset, amplitude and phase;

[0053] Step 5-1: Calculate the estimated value of the DC offset contained in the current when the k+1th set of sampling data is obtained (k+1):

[0054]

[0055] [I k+1,0 ,I k+1,1 ,I k+1,2 ,I k+1,3] is the k+1th group of data;

[0056] Step 5-2: Calculate the estimated value of the orthogonal signal of the current when the k+1th set of sampling data is obtained

[0057] Step 5-3: Calculate the estimated value of the current amplitude when the k+1th set of sampling data is obtained

[0058]

[0059] Step 5-4: Calculate the estimated value of the current phase when the k+1th set of sampling data is obtained

[0060]

[0061] The step six is specifically as follows:

[0062] Based on the above steps 3 to 5, the current and voltage amplitude and phase estimation of the k+1 group of sampling data are obtained.

[0063] remember This is the phase difference between voltage and current at this time;

[0064] The active power is calculated as follows:

[0065]

[0066] Reactive power is calculated as follows:

[0067]

[0068] The step seven is specifically as follows:

[0069] The frequency and DC offset parameters are updated using the latest data until the power estimation error for several consecutive times is less than a certain threshold. The estimated value is then considered to be the true value.

[0070] Beneficial effects of the present invention:

[0071] (1) The present invention proposes a power detection method for a power grid containing a DC offset. By introducing information based on the estimated frequency, the amplitude, phase and DC offset of the signal can be effectively extracted, thereby subtracting the DC offset from the original signal, ensuring that the power is determined only by the fundamental AC component, thereby improving the accuracy and stability of power detection.

[0072] (2) After obtaining the frequency, amplitude, and phase estimation results, this method, combined with the conventional power calculation model, can better restore the power parameters. Even when the grid frequency fluctuates or the load changes greatly, it can maintain a relatively stable and reliable detection effect through the designed adaptive method. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 It is a flow chart of the method for estimating power of a power grid implemented and provided by the present invention.

[0074] Figure 2 It is a flow chart of the grid power estimation method implemented and provided by the present invention. DETAILED DESCRIPTION

[0075] The present invention will be described in further detail below with reference to the accompanying drawings.

[0076] like Figure 1 、 Figure 2 As shown, a method for accurately detecting power of a wind power generation system based on DC offset estimation and compensation includes the following steps:

[0077] Step 1: ADC sampling of grid voltage and current to provide data basis for subsequent processing;

[0078] Step 2: Based on the sampling values obtained in step 1, extract four consecutive sampling points in sequence and group them into groups of four:

[0079] Current: [I k,0 ,I k,1 ,I k,2 ,I k,3 ], voltage: [U k,0 ,U k,1 ,U k,2 ,U k,3 ],

[0080] The subscripts (k,0), (k,1), (k,2), and (k,3) are the first, second, third, and fourth sampling points of the kth group of sampled data. (k=1,2,3...)

[0081] Step 3: Take any set of data, use the basic formula of trigonometric functions to calculate the intermediate value of the algebraic expression, and construct a specific algebraic expression between four consecutive samples;

[0082] To simplify the expression, take the kth group of data as an example:

[0083] [I k,0 ,I k,1 ,I k,2 ,I k,3 ],[U k,0 ,Uk,1 ,U k,2 ,U k,3 ]

[0084] To illustrate the implementation process in detail, the angular frequency of current and voltage is ω I ,ω U (the two are equal under normal circumstances), the frequency is f I , f U .

[0085] [I k,0 ,I k,1 ,I k,2 ,I k,3 ] are:

[0086]

[0087] ω I =2πf I

[0088] [U k,0 ,U k,1 ,U k,2 ,U k,3 ] are:

[0089]

[0090] Among them, I and U represent the amplitude of current and voltage signals respectively. They are the current signals I k,0 , voltage signal U k,0 The phase angle, T s is the sampling time interval, f s is the sampling frequency, I dc , U dc are the DC offset contained in the current and voltage respectively.

[0091] The DC offset is then removed explicitly by performing a differential operation on four consecutive sampling points.

[0092] Calculate the intermediate quantities related to the current signal. The voltage signal is processed in the same way as the current signal:

[0093] Step 3-1: Define intermediate variables for

[0094]

[0095] Step 3-2: Define intermediate variables for

[0096]

[0097] The DC component is directly offset by differential operation, and the trigonometric function relationship related to the frequency ω is retained, providing the necessary conditions for subsequent frequency estimation and parameter calculation.

[0098] Step 3-3: According to and Calculate intermediate variables Construct a specific algebraic expression between four consecutive samples

[0099]

[0100]

[0101] Therefore, there are

[0102]

[0103] Similarly,

[0104]

[0105] Therefore, there is an intermediate variable of current and voltage intermediate variables Thus, each set of sampled data is traversed to directly obtain the estimated value of each frequency and

[0106] However, due to the possible presence of harmonics and DC offset interference in actual working conditions, the estimated frequency value may jump and deviate from the true value.

[0107] Therefore, the present invention uses the normalized gradient method to achieve adaptive estimation, so that the observed γ I , γ U Gradually approaching the true value.

[0108] Step 4: Use the normalized gradient method to build a frequency estimator to achieve adaptive frequency estimation

[0109] Step 4-1: Calculate the normalization coefficient m of the kth group of data I (k), m U (k)

[0110]

[0111] is the intermediate variable obtained based on the kth group of sampling data.

[0112] Step 4-2: Calculate the estimated error e of the kth group of data I (k), e U (k)

[0113]

[0114] is the intermediate variable obtained based on the kth group of sampling data; To obtain the estimated value of the current intermediate variable γ when the kth group of sampling data is obtained, The initial value of The default is 0.

[0115] Step 4-3: Calculate the estimated value of the intermediate variable γ when the k+1th group of data is obtained

[0116]

[0117] To obtain the current intermediate variable γ when the k+1th group of sampling data is obtained I , γ U estimated value.

[0118] Step 4-4: Calculate the estimated value of the frequency when the k+1th group of sampling data is obtained

[0119]

[0120] Step 5: Estimate the DC offset and then subtract the DC component from the original signal to estimate the amplitude and phase;

[0121] Based on the estimated frequency obtained in step 4 and intermediate variables Calculate the DC offset, magnitude, and phase.

[0122] Step 5-1: Calculate the estimated DC offset of the current and voltage in the k+1th set of sampling data

[0123]

[0124] [I k+1,0 ,I k+1,1 ,I k+1,2 ,I k+1,3 ]、[U k+1,0 ,U k+1,1 ,U k+1,2 ,U k+1,3 ] is the k+1th group of data;

[0125] Step 5-2: Calculate the estimated values of the orthogonal signals of current and voltage for the k+1th set of sampling data

[0126]

[0127] Step 5-3: Calculate the estimated values of the current and voltage amplitudes for the k+1th set of sampling data

[0128]

[0129] Step 5-4: Calculate the estimated values of the current and voltage phases for the k+1th set of sampling data

[0130]

[0131] Step 6: Calculate power based on the obtained parameters

[0132] Based on the above steps 3 to 5, the current and voltage amplitude and phase estimation of the k+1 group of sampling data are obtained.

[0133] remember This is the phase difference between voltage and current at this time.

[0134] The active power is calculated as follows:

[0135]

[0136] Reactive power is calculated as follows:

[0137]

[0138] Step 7: Loop through steps 4 to 6, using the latest data to update parameters such as frequency and DC offset, until the power estimation error for several consecutive times is less than a certain threshold. The estimated value is then considered to be the true value.

[0139] The above specific process is only a practical application routine of the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention and should not be limited to the contents of the present invention in the above routine.

[0140] Application examples:

[0141] Taking a grid-connected wind power generation simulation system as an example, the system operates under complex working conditions with ±2% DC offset and frequency fluctuation (49Hz-51Hz) in the power grid.

[0142] To verify the actual effectiveness of the method of the present invention, it is compared with the traditional bandpass filter method and the orthogonal signal decoupling method. The traditional method relies on the whole cycle data cache (such as 200 points / cycle for 50Hz power frequency) and complex floating point operations, resulting in a single power analysis time of up to 20ms and limited dynamic tracking capabilities; while the present invention uses a four-sampling point grouping strategy combined with the linear combination of intermediate variables to Directly derive frequency and DC offset, shortening the single calculation cycle to 8ms and using only 5% of the memory used by traditional methods.

[0143] In frequency mutation scenarios (such as 50Hz→49.5Hz), traditional bandpass filters require more than 30ms of phase delay to stabilize, and the overshoot error reaches 0.15Hz. The present invention uses the normalized gradient method to achieve adaptive frequency estimation, completes tracking within 12ms, and the steady-state error is less than 0.02Hz. For DC offset compensation, the traditional orthogonal decoupling method requires iterative calculations and matrix operations, which takes more than 20ms and has an amplitude error of 1.2%. The present invention directly estimates the offset through three-step integer addition and subtraction and one division, compresses the compensation time to 5ms, and improves the fundamental amplitude accuracy to 0.05%. Actual tests show that in the scenario of DC offset and frequency fluctuation, the power detection error of the present invention is optimized by one order of magnitude compared with the traditional method. At the same time, because no external high-precision filter is required, the hardware cost is greatly reduced. This method is suitable for scenarios dominated by power frequency fluctuations and DC offsets. If there is non-power frequency interference in the power grid, it needs to be combined with a pre-filtering module to expand its applicability. However, it has the advantages of high real-time performance, low resource occupation and cost in new energy grid-connected systems, providing an innovative solution for accurate power detection in complex power grid environments.

Claims

1. A method for accurately detecting power of a wind power generation system based on DC offset estimation and compensation, characterized in that: The following steps are included: Step 1: ADC samples the grid voltage and current to obtain sampling values; Step 2: grouping four consecutive sampling points into a group; Step 3: Take any set of sample values, eliminate DC by differential, and calculate the intermediate quantity related to frequency to provide input for frequency estimation; Step 4: Based on the intermediate variable, the frequency is iteratively updated using the normalized gradient method to obtain an estimated value of the sampled intermediate variable and the frequency; Step 5: Estimate the DC offset, amplitude, and phase based on the estimated values of the intermediate variables and frequency; Step 6: Perform power calculation based on the estimated DC offset, amplitude, and phase to eliminate DC interference; Step 7: Repeat steps 4 to 6 until the output is stable.

2. The method for accurate power detection of a wind power generation system based on DC offset estimation and compensation according to claim 1, characterized in that: Walkthrough: Electric current: [I k,0 ,I k,1 ,I k,2 ,I k,3 ],Electronic:[U k,0 ,U k,1 ,U k,2 ,U k,3 ]; Wherein, the subscripts (k, 0), (k, 1), (k, 2), and (k, 3) are the first, second, third, and fourth sampling points of the kth group of sampling data, and k = 1, 2, 3, ...

3. The method for accurate power detection of a wind power generation system based on DC offset estimation and compensation according to claim 2, characterized in that: In the step three: Use basic trigonometric formulas to calculate intermediate quantities in algebraic expressions and construct specific algebraic expressions between four consecutive samples; The kth group of data is: [AND k,0 ,AND k,1 ,AND k,2 ,AND k,3 ],[IN k,0 ,IN k,1 ,IN k,2 ,IN k,3 ] [I k,0 ,I k,1 ,I k,2 ,I k,3 ] are: [U k,0 ,U k,1 ,U k,2 ,U k,3 ] are: Where I and U represent the amplitude of current and voltage signals respectively, ω is the signal angular frequency, and f is the signal frequency. They are the current signals I k,0 , voltage signal U k,0 The phase angle, T s is the sampling time interval, f s is the sampling frequency, I dc , U dc are the DC offset contained in the current and voltage respectively.

4. The method for accurate power detection of a wind power generation system based on DC offset estimation and compensation according to claim 3, characterized in that: Explicitly eliminate DC offset through differential operation of four consecutive sampling points; Calculate the intermediate quantities related to the current signal; the voltage signal is processed in the same way as the current signal; Step 3-1: Define intermediate variables for Step 3-2: Define intermediate variables for Step 3-3: According to Calculate intermediate variables Construct a specific algebraic expression between four consecutive samples Therefore, there are Therefore, there are intermediate variables Thus, we can directly get the estimated value of frequency by traversing each set of sample data.

5. The method for accurate power detection of a wind power generation system based on DC offset estimation and compensation according to claim 4, characterized in that: The step 4 is specifically as follows: The normalized gradient method is used to construct a frequency estimator to achieve adaptive frequency estimation; Step 4-1: Calculate the normalization coefficient m(k) of the kth group of data Intermediate variables obtained based on the kth group of sampling data; Step 4-2: Calculate the estimated error e(k) for the kth group of data is the intermediate variable obtained based on the kth group of sampling data; To obtain the estimated value of the current intermediate variable γ when the kth group of sampling data is obtained, The initial value of The default value is 0. Step 4-3: Calculate the estimated value of the intermediate variable γ when the k+1th group of data is obtained To obtain the estimated value of the current intermediate variable γ when the k+1th group of sampling data is obtained; Step 4-4: Calculate the estimated value of the frequency when the k+1th group of sampling data is obtained 6. The method for accurate power detection of a wind power generation system based on DC offset estimation and compensation according to claim 5, characterized in that: The step five is specifically as follows: Based on the estimated frequency obtained in step 4 and intermediate variables Calculate DC offset, amplitude and phase; Step 5-1: Calculate the estimated value of the DC offset contained in the current when the k+1th set of sampling data is obtained [I k+1,0 ,I k+1,1 ,I k+1,2 ,I k+1,3 ] is the k+1th group of data; Step 5-2: Calculate the estimated value of the orthogonal signal of the current when the k+1th set of sampling data is obtained Step 5-3: Calculate the estimated value of the current amplitude when the k+1th set of sampling data is obtained Step 5-4: Calculate the estimated value of the current phase when the k+1th set of sampling data is obtained 7. The method for accurate power detection of a wind power generation system based on DC offset estimation and compensation according to claim 6, characterized in that: The step six is specifically as follows: Based on the above steps 3 to 5, the current and voltage amplitude and phase estimation of the k+1 group of sampling data are obtained. remember This is the phase difference between voltage and current at this time; The active power is calculated as follows: Reactive power is calculated as follows:

8. The method for accurate power detection of a wind power generation system based on DC offset estimation and compensation according to claim 7, characterized in that: The step seven is specifically as follows: The frequency and DC offset parameters are updated using the latest data until the power estimation error for several consecutive times is less than a certain threshold. The estimated value is then considered to be the true value.

Images