Intelligent data acquisition method and system of distributed photovoltaic power station

Abstract

The invention provides an intelligent data acquisition method and system for a distributed photovoltaic power station, and relates to the technical field of photovoltaic power station data acquisition, and the method comprises the steps: obtaining spectral intensity data and sensor measurement data in the distributed photovoltaic power station; performing feature extraction according to the spectral intensity data to obtain spectral time sequence change features, and constructing a sensor linear drift model and a sensor nonlinear temperature drift model based on the spectral time sequence change features; calculating based on the sensor linear drift model and the sensor nonlinear temperature drift model to obtain a drift compensation coefficient; and performing zero drift calibration and range drift calibration in combination with the drift compensation coefficient and the sensor measurement data to obtain a calibrated data acquisition result. According to the method, a sensor drift compensation mechanism based on spectrum time sequence change characteristics is constructed, so that the problem of data acquisition accuracy reduction caused by sensor precision attenuation in long-term operation of the distributed photovoltaic power station is effectively solved.

Description

TECHNICAL FIELD

[0001] The present application relates to photovoltaic power station data acquisition technology, and particularly to an intelligent data acquisition method and system for a distributed photovoltaic power station. BACKGROUND

[0002] As an important part of renewable energy, the operation state monitoring and data acquisition technology of distributed photovoltaic power stations has experienced a development process from manual inspection to automatic monitoring. Early photovoltaic power station data acquisition mainly relied on manual periodic recording of electric meter readings and environmental parameters. This approach has the problems of low data acquisition frequency and high labor costs. With the development of sensor technology and communication technology, automatic data acquisition systems based on wired and wireless transmission have gradually been applied to photovoltaic power station monitoring. Modern distributed photovoltaic power stations generally use a distributed data acquisition architecture to achieve real-time monitoring of power generation unit operating parameters by deploying a large number of current sensors, voltage sensors, temperature sensors, and other devices. To improve the intelligent level of data acquisition, a remote data acquisition platform based on Internet of Things technology and a local data processing method based on edge computing have been proposed. However, as the operation time of photovoltaic power stations extends and environmental conditions change, maintaining the accuracy of sensor measurements becomes a key technical challenge for data acquisition systems.

[0003] Existing distributed photovoltaic power station data acquisition methods have limitations in maintaining sensor accuracy. Traditional data acquisition systems use fixed sampling frequencies and preset calibration parameters for data acquisition, which cannot adjust the acquisition strategy according to changes in environmental conditions, resulting in a decrease in data acquisition accuracy under real weather conditions. Existing sensor calibration methods mainly rely on periodic manual field calibration or automatic calibration programs based on fixed time intervals, which cannot respond promptly to sensor drift caused by environmental factors. Data acquisition optimization methods based on statistical models usually use static models trained on historical data, which lack real-time awareness of the current environmental state and cannot guarantee data acquisition accuracy when environmental parameters such as light, temperature, and humidity change rapidly. Existing intelligent data acquisition technologies mainly focus on data transmission efficiency and storage optimization, and do not pay enough attention to real-time calibration of sensor measurement accuracy. In particular, there is a lack of technical means to use photovoltaic power station environmental information to assess sensor state, which makes it difficult to ensure the measurement accuracy of sensors under real environmental conditions, and further causes the long-term operation accuracy of sensors to decrease, resulting in insufficient data acquisition accuracy. SUMMARY

[0004] Embodiments of the present application provide an intelligent data acquisition method and system for a distributed photovoltaic power station.

[0005] In a first aspect, the embodiments of the present application provide a method for intelligent data acquisition of a distributed photovoltaic power station, comprising: obtaining spectral intensity data and sensor measurement data in the distributed photovoltaic power station; performing feature extraction based on the spectral intensity data to obtain spectral time series variation features, and constructing a sensor linear drift model and a sensor nonlinear temperature drift model based on the spectral time series variation features; performing calculation based on the sensor linear drift model and the sensor nonlinear temperature drift model to obtain a drift compensation coefficient; and performing zero-point drift calibration and range drift calibration in combination with the drift compensation coefficient and the sensor measurement data to obtain a calibrated data acquisition result.

[0006] Optionally, in a possible implementation manner of the first aspect, the constructing of the sensor linear drift model based on the spectral time series variation features comprises: calculating a time derivative and a second-order derivative of the spectral intensity according to the spectral time series variation features; performing comparative analysis on the time derivative and the second-order derivative of the spectral intensity and a theoretical solar radiation model to generate spectral deviation; and fitting a linear relationship between the spectral deviation and a sensor measurement deviation by a least square method to establish the sensor linear drift model.

[0007] Optionally, in a possible implementation manner of the first aspect, the constructing of the sensor nonlinear temperature drift model based on the spectral time series variation features comprises: extracting a temperature-related spectral component in the spectral time series variation features; establishing a nonlinear mapping relationship between the temperature-related spectral component and an environmental temperature variation rate; and constructing the sensor nonlinear temperature drift model according to the nonlinear mapping relationship and a sensor temperature coefficient.

[0008] Optionally, in a possible implementation manner of the first aspect, the performing of the calculation based on the sensor linear drift model and the sensor nonlinear temperature drift model to obtain the drift compensation coefficient comprises: calculating a drift severity level of each sensor according to the sensor linear drift model and the sensor nonlinear temperature drift model; assigning a drift compensation weight to each sensor based on the drift severity level; and calculating a drift compensation coefficient in combination with the drift compensation weight by a weighted fusion algorithm, the drift compensation coefficient comprising a zero-point compensation coefficient and a range compensation coefficient.

[0009] Optionally, in a possible implementation manner of the first aspect, the obtaining of the calibrated data acquisition result comprises: performing zero-point drift calibration in combination with the zero-point compensation coefficient and the sensor measurement data to obtain a zero-point drift calibration result; performing range drift calibration in combination with the range compensation coefficient and the sensor measurement data to obtain a range drift calibration result; and determining the calibrated data acquisition result in combination with the zero-point drift calibration result and the range drift calibration result.

[0010] Optionally, in a possible implementation manner of the first aspect, the zero-point drift calibration comprises: extracting a zero-point offset in the sensor measurement data; calculating the zero-point compensation coefficient and the zero-point offset to generate a zero-point calibration parameter; and adjusting a measurement reference value of the sensor according to the zero-point calibration parameter.

[0011] Optionally, in a possible implementation manner of the first aspect, the range drift calibration comprises: analyzing a range variation trend of the sensor measurement data to determine a range drift mode; applying the range compensation coefficient to the range drift mode to calculate a range calibration factor; and adjusting a gain characteristic of the sensor by the range calibration factor.

[0012] In a second aspect of the embodiments of the present application, a smart data acquisition system of a distributed photovoltaic power station is provided

[0013] Optionally, in a possible implementation manner of the second aspect, a data acquisition module is configured to acquire spectral intensity data and sensor measurement data in the distributed photovoltaic power station; a drift model construction module is configured to perform feature extraction according to the spectral intensity data to obtain spectral time series variation features, and construct a sensor linear drift model and a sensor nonlinear temperature drift model based on the spectral time series variation features; a drift compensation module is configured to calculate drift compensation coefficients based on the sensor linear drift model and the sensor nonlinear temperature drift model; and a drift calibration module is configured to perform zero-point drift calibration and range drift calibration in combination with the drift compensation coefficients and the sensor measurement data to obtain a calibrated data acquisition result.

[0014] In a third aspect of the embodiments of the present application, a computer device is provided, comprising a memory, a processor, and a computer program, the computer program is stored in the memory, and the processor executes the computer program to perform the smart data acquisition method of the distributed photovoltaic power station in the first aspect and various possible aspects related to the first aspect.

[0015] In a fourth aspect of the embodiments of the present application, a readable storage medium is provided, the readable storage medium stores a computer program, and the computer program is executed by a processor to implement the smart data acquisition method of the distributed photovoltaic power station in the first aspect and various possible aspects related to the first aspect.

[0016] The intelligent data acquisition method and system of the distributed photovoltaic power station provided in the application can realize real-time precision calibration of current, voltage, temperature and other key parameter sensors by constructing a sensor drift compensation mechanism based on spectral timing change characteristics, establishing a sensor state evaluation model by fully utilizing spectral information in the photovoltaic power station environment, effectively solving the problem of decreased data acquisition accuracy caused by sensor precision attenuation in the long-term operation of the distributed photovoltaic power station, reducing the operation and maintenance cost of the photovoltaic power station, improving the automation level and reliability of data acquisition, maintaining the long-term stability of sensor measurement precision in outdoor environmental conditions through the synergistic effect of the double drift model and the individualized compensation strategy, providing high-quality data support for the power generation statistics, fault diagnosis and performance evaluation of the photovoltaic power station, and ensuring the data acquisition precision and operation efficiency of the distributed photovoltaic power station. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the application or the related art, the drawings needed to be used in the description of the embodiments of the application or the related art will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other related drawings can be obtained by those skilled in the art without creative labor.

[0018] Figure 1 is the overall flowchart of the intelligent data acquisition method of the distributed photovoltaic power station involved in the application;

[0019] Figure 2 is the construction flowchart of the sensor linear drift model and the sensor nonlinear temperature drift model involved in the application;

[0020] Figure 3 is the calculation flowchart of the drift compensation coefficient involved in the application. DETAILED DESCRIPTION

[0021] In order to make the above-mentioned purposes, features and advantages of the application more obvious and easy to understand, the specific embodiments of the application will be described in detail in conjunction with the drawings of the specification.

[0022] In the following description, many specific details are set forth in order to provide a thorough understanding of the application, but the application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the application, therefore the application is not limited by the specific embodiments disclosed below.

[0023] Second, the "one embodiment" or "an embodiment" referred to herein means a specific feature, structure, characteristic, or combination of features and / or characteristics described in at least one implementation of the application. The various appearances of "in one embodiment" or "in an embodiment" in the specification do not all refer to the same embodiment.

[0024] Embodiment 1, Reference Figure 1 As a first embodiment of the present application, the embodiment provides an intelligent data acquisition method for a distributed photovoltaic power station, comprising:

[0025] S100: Obtain spectral intensity data and sensor measurement data in the distributed photovoltaic power station.

[0026] S200: Perform feature extraction according to the spectral intensity data to obtain spectral time series variation features, and construct a sensor linear drift model and a sensor nonlinear temperature drift model based on the spectral time series variation features.

[0027] S300: Perform calculation based on the sensor linear drift model and the sensor nonlinear temperature drift model to obtain a drift compensation coefficient.

[0028] S400: Perform zero-point drift calibration and range drift calibration in combination with the drift compensation coefficient and the sensor measurement data to obtain a calibrated data acquisition result.

[0029] It should be noted that the distributed photovoltaic power station is an outdoor renewable energy facility, and its sensors are exposed to complex and variable environmental conditions for a long time. The dramatic changes in parameters such as light intensity, environmental temperature, and humidity cause the measurement accuracy of the sensors to drift, and the zero point and range characteristics of the sensors are also changed due to environmental factors. The existing data acquisition method uses fixed sampling frequency and preset calibration parameters to acquire data, which cannot adjust the acquisition strategy according to the changes in environmental conditions, resulting in a decrease in data acquisition accuracy under real weather conditions. At the same time, the traditional sensor calibration method mainly relies on periodic manual field calibration or automatic calibration program based on fixed time interval, which cannot respond to the drift changes of the sensors caused by environmental factors in time, and also cannot guarantee the accuracy of data acquisition when the environmental parameters such as light, temperature, and humidity change rapidly due to the lack of real-time sensing ability of the current environmental state.

[0030] Therefore, in order to solve the problems of sensor accuracy degradation and data acquisition accuracy reduction, the sensor drift compensation mechanism based on spectral time sequence change characteristics is constructed through the steps S100-S400, the sensor linear drift model and the nonlinear temperature drift model are obtained, and the accurate calculation of the drift compensation coefficient is realized; the real-time monitoring of the sensor drift is realized, and the zero drift calibration and the range drift calibration are performed, so as to realize the early warning of the sensor accuracy abnormal reduction; meanwhile, based on the synergistic effect of the double drift models and the individualized compensation strategy, the accurate guarantee of the data acquisition accuracy and the operation efficiency of the distributed photovoltaic power station is realized.

[0031] Embodiment 2, with reference to Figures 1 to 3 As a second embodiment of the present application, based on the above embodiment, an intelligent data acquisition method of a distributed photovoltaic power station is provided.

[0032] S100: Obtain spectral intensity data and sensor measurement data in the distributed photovoltaic power station.

[0033] In the embodiment of the present application, the spectral intensity data and the sensor measurement data are obtained through the data acquisition equipment deployed at each node of the distributed photovoltaic power station, which provides basic data support for subsequent sensor drift calibration.

[0034] S101: Obtain spectral intensity data in the distributed photovoltaic power station.

[0035] In the embodiment of the present application, the spectral intensity data refers to the solar spectrum radiation intensity information in the operating environment of the distributed photovoltaic power station, including the intensity data of ultraviolet spectrum, visible spectrum and near-infrared spectrum.

[0036] Specifically, obtaining the spectral intensity data in the distributed photovoltaic power station refers to collecting the environmental spectrum information in real time through the spectrometer equipment deployed in the distributed photovoltaic power station, obtaining the spectral intensity values reflecting the current lighting conditions and atmospheric transparency, including the following steps:

[0037] The environmental spectrum information is collected through the spectrometer equipment deployed in the distributed photovoltaic power station, and the spectrometer equipment includes one or more of ultraviolet spectrometer, visible spectrometer and near-infrared spectrometer.

[0038] The collected spectral signals are converted into digital signals, and continuous sampling is performed according to the preset time interval, and the time interval is set to 1 minute to 10 minutes.

[0039] The collected spectral intensity data is marked with a time stamp to ensure that the spectral data is synchronized with the sensor measurement data in time.

[0040] In an alternative embodiment, a multi-point distributed collection method can be adopted, multiple spectrum collection points are arranged in different areas of the distributed photovoltaic power station, and the spectrum data of each spectrum collection point is fused and processed through a spatial interpolation algorithm to obtain spectrum intensity data covering the entire power station area, further improving the spatial representativeness of data collection.

[0041] In another alternative embodiment, meteorological station data and satellite remote sensing data can also be combined to supplement and verify the spectrum intensity data. When the field spectrometer fails or the data is abnormal, the data is reconstructed using a meteorological model and historical spectrum data to ensure the continuity and integrity of the spectrum intensity data.

[0042] S102: Obtain sensor measurement data in the distributed photovoltaic power station.

[0043] In the embodiments of the present application, the sensor measurement data refers to real-time measurement information of various sensors during the operation of the photovoltaic power station, including current data, voltage data, ambient temperature data, and equipment temperature data.

[0044] Specifically, obtaining sensor measurement data in the distributed photovoltaic power station refers to collecting real-time operation parameters through sensors distributed in each power generation unit of the photovoltaic power station to obtain measurement values reflecting the actual operation state of the power station, including the following steps:

[0045] Current sensors, voltage sensors, and temperature sensors are deployed at each power generation unit of the distributed photovoltaic power station.

[0046] Current sensors are used to measure current data, voltage sensors are used to measure voltage data, and temperature sensors are used to measure ambient temperature data and equipment temperature data.

[0047] In an alternative embodiment, the acquisition of sensor measurement data can be combined with edge computing technology. Edge computing devices are deployed at each sensor node to pre-process and preliminarily screen the original measurement data, remove obviously abnormal data points, reduce data transmission volume, and improve data quality.

[0048] In another alternative embodiment, a redundant sensor configuration scheme can also be adopted. Backup sensors are configured at key monitoring points, and when the main sensor fails, the backup sensor is automatically switched to. At the same time, abnormal measurement values are identified through cross-validation of data between sensors to ensure the reliability and continuity of the sensor measurement data.

[0049] Exemplarily, in a typical 10MW distributed photovoltaic power station, 3-5 spectrum collection points can be arranged, each equipped with a full-spectrum spectrometer to collect spectral intensity data in the wavelength range of 280nm-2500nm; at the same time, current sensors and voltage sensors are arranged at each photovoltaic string, and temperature sensors are arranged at each inverter, to realize comprehensive monitoring of the operation state of the power station; all data are converged to the monitoring center through industrial Ethernet or wireless communication, and the sampling frequency is set to once per minute to ensure the timeliness and continuity of the data.

[0050] It should be noted that, since the conventional sensor calibration method mainly relies on the output characteristics of the sensor itself or a preset mathematical model for calibration, it cannot perceive the real-time influence of the environmental conditions of the photovoltaic power station on the performance of the sensor, the present application introduces spectral intensity data as a new dimension of environmental perception, establishes a correlation bridge between the sensor state and the environmental spectrum information, realizes comprehensive perception of the environmental state of the entire power station area, instead of relying on single-point environmental data for speculation; at the same time, through the time stamp synchronization mechanism, the time consistency of the spectral data and the sensor data is ensured, providing a reliable data basis for subsequent correlation analysis, which can detect abnormalities in the early stage of sensor drift, thereby realizing preventive maintenance instead of passive repair.

[0051] S200: performing feature extraction according to the spectral intensity data to obtain spectral time series variation features, and constructing a sensor linear drift model and a sensor nonlinear temperature drift model based on the spectral time series variation features.

[0052] S201: performing feature extraction according to the spectral intensity data to obtain spectral time series variation features.

[0053] In the embodiments of the present application, the spectral time series variation features refer to feature parameters extracted by analyzing the variation law of the spectral intensity data in the time dimension, which are used to reflect the dynamic change trend and periodic characteristics of the environmental lighting conditions.

[0054] Specifically, the spectral time series variation features are obtained by performing feature extraction according to the spectral intensity data, including the following steps:

[0055] The spectral intensity data is preprocessed by filtering, wherein a Butterworth low-pass filter is used to remove high-frequency noise, and the filter cutoff frequency is set to 1 / 10 of the sampling frequency.

[0056] The filtered spectral intensity data is normalized, and the specific formula is as follows:

[0057]

[0058] wherein, I norm(t, λ) is the normalized spectral intensity data; I(t, λ) is the original spectral intensity at time t and wavelength λ; I max (λ) is the maximum spectral intensity at wavelength λ. min (λ) is the minimum spectral intensity at wavelength λ.

[0059] The time series of spectral intensity data is analyzed by a sliding window algorithm, where the sliding window length is set to 30-60 minutes and the step length is set to 5-10 minutes, the spectral time series variation characteristics are extracted, and the specific formula is as follows:

[0060]

[0061] Where, F spectrum (t) is the spectral time series variation characteristic at time t; W is the sliding window length. is the time derivative of the normalized spectral intensity.

[0062] In an optional embodiment, the multi-scale wavelet transform method can be used to further analyze the spectral time series variation characteristics, to identify short-term fluctuations and long-term trends by decomposing different frequency components, and to improve the accuracy and robustness of feature extraction.

[0063] In another optional embodiment, the principal component analysis method can also be combined to reduce the dimension of the multi-band spectral data, extract the main spectral variation pattern, and reduce the calculation complexity while retaining the key feature information.

[0064] S202: Constructing a sensor linear drift model based on spectral time series variation characteristics.

[0065] In the embodiments of the present application, as shown in Figure 2 The construction flowchart of the sensor linear drift model and the sensor nonlinear temperature drift model is shown in FIG. 1, the sensor linear drift model refers to a mathematical model established based on the linear relationship between the spectral time series variation characteristics and the sensor measurement deviation, which is used to predict and compensate the linear drift of the sensor.

[0066] Specifically, constructing a sensor linear drift model based on spectral time series variation characteristics refers to establishing a quantitative linear drift prediction relationship based on spectral time series variation characteristics by analyzing the correlation between the change rule of spectral time series variation characteristics and the output deviation of the sensor, including the following steps:

[0067] The time derivative and second derivative of the spectral intensity are calculated according to the spectral time series variation characteristics, and the specific formula is as follows:

[0068]

[0069] Where, is the first derivative of the spectral time series, reflecting the rate of spectral change; is the second derivative of the spectral time series, reflecting the acceleration of spectral change; and △t is the time interval.

[0070] The time derivative and the second derivative of the spectral intensity are compared with the theoretical solar radiation model to generate a spectral deviation, according to the following formula:

[0071]

[0072] where D spectrum (t) is the spectral deviation, indicating the difference between the actual spectral change and the predicted value of the theoretical solar radiation model; F theory (t) is the output value of the theoretical solar radiation model.

[0073] A linear relationship between the spectral deviation and the sensor measurement deviation is fitted by the least squares method to establish a sensor linear drift model, according to the following formula:

[0074] △S linear (t) = K linear · D spectrum (t) + b linear ;

[0075] where △S linear (t) is the linear drift of the sensor; K linear is the linear drift coefficient; and b linear is the linear drift bias term.

[0076] In an optional embodiment, the recursive least squares method can be used to realize online updating of the sensor linear drift model, and the linear drift coefficient K linear and the linear drift bias term b linear of the sensor linear drift model are adjusted according to newly collected data to improve the adaptive ability of the model.

[0077] In another optional embodiment, the multi-sensor data fusion technology can also be combined to cross-verify the accuracy of the spectral deviation by using the measurement information of different types of sensors to improve the reliability of the sensor linear drift model.

[0078] S203: Constructing a sensor nonlinear temperature drift model based on the spectral time series.

[0079] In the embodiments of the present application, the sensor nonlinear temperature drift model refers to a mathematical model established based on the nonlinear relationship between the temperature-related component in the spectral time series and the sensor temperature drift, which is used to predict and compensate for the nonlinear drift of the sensor caused by temperature change.

[0080] Specifically, the sensor nonlinear temperature drift model is constructed based on the spectral time series variation characteristics, which means that the temperature-related spectral components are identified and extracted from the spectral time series variation characteristics, and a nonlinear mapping relationship between the temperature-related spectral variation and the sensor temperature drift is established, including the following steps:

[0081] By analyzing the correlation between the spectral intensity at different wavelengths and the temperature, the temperature-related spectral components in the spectral time series variation characteristics are extracted, and the temperature-sensitive waveband is identified, and the specific formula is as follows:

[0082]

[0083] Where F temp (t) is the temperature-related spectral component; N is the number of temperature-sensitive wavelengths; w i is the weight coefficient of wavelength λ i ; ρ(λ i ) is the temperature sensitivity coefficient of wavelength λ i .

[0084] A nonlinear mapping relationship between the temperature-related spectral component and the environmental temperature change rate is established, wherein a cubic polynomial function is used to describe the nonlinear characteristics, and the specific formula is as follows:

[0085]

[0086] Where, is the environmental temperature change rate; a3, a2, a1, a0 are polynomial coefficients, which are obtained by fitting historical data.

[0087] According to the nonlinear mapping relationship and the sensor temperature coefficient, a sensor nonlinear temperature drift model is constructed, and the specific formula is as follows:

[0088]

[0089] Where, △S temp (t) is the nonlinear temperature drift of the sensor; α temp is the temperature drift proportion coefficient; β temp is the nonlinear adjustment parameter; tanh(·) is the hyperbolic tangent function, which is used to limit the change range of the drift.

[0090] In an alternative embodiment, a more complex sensor nonlinear temperature drift model can be established using a neural network algorithm, which learns the nonlinear relationship between the temperature-related spectral component and the sensor drift through a multilayer perceptron, thereby improving the prediction accuracy of the model.

[0091] In another alternative embodiment, a personalized sensor nonlinear temperature drift model can also be trained using machine learning algorithms in combination with historical temperature response data of the sensor, to establish a specialized compensation strategy for different models and batches of sensors.

[0092] For example, in a typical photovoltaic power station application, by analyzing the near-infrared spectral data in the wavelength range of 800-1200 nm, the spectral features related to the change of ambient temperature are extracted, and the sensor nonlinear temperature drift model is established in combination with the measured data of the temperature sensor. When the ambient temperature rises from 25℃ to 45℃, the drift of the temperature sensor can be predicted by the sensor nonlinear temperature drift model, which is about ±0.1℃, and the drift of the current sensor is about ±0.02A, thereby providing an accurate prediction basis for subsequent drift compensation.

[0093] It should be noted that by constructing a dual drift model based on the spectral time series change characteristics, the problem of incomplete modeling of the sensor drift mechanism in the prior art is solved. The existing sensor drift compensation technology usually uses a single linear model or a simple temperature compensation algorithm, which cannot accurately describe the multiple influencing factors and nonlinear characteristics of the sensor drift in the photovoltaic power station environment. The present application realizes fine modeling of the sensor drift mechanism by constructing a linear drift model and a nonlinear temperature drift model respectively. The linear drift model captures the influence of environmental factors on the basic performance of the sensor by analyzing the deviation of the spectral time series change characteristics from the theoretical solar radiation model. The nonlinear temperature drift model is specifically modeled for the nonlinear effects caused by temperature changes, solving the problem of insufficient accuracy of traditional linear temperature compensation methods in large temperature difference environments. More importantly, the present application uses environmental spectral information as the core input for modeling, rather than relying only on temperature, humidity and other conventional environmental parameters, so that the model can perceive more environmental factors that affect the performance of the sensor, improving the accuracy and adaptability of drift prediction.

[0094] S300: Calculate the drift compensation coefficient based on the sensor linear drift model and the sensor nonlinear temperature drift model.

[0095] S301: Calculate the drift severity level of each sensor according to the sensor linear drift model and the sensor nonlinear temperature drift model.

[0096] In the embodiments of the present application, as shown in Figure 3 The drift compensation coefficient calculation flowchart is shown in the figure, and the drift severity level refers to a quantitative index that divides the overall drift condition of the sensor into different levels by comprehensively evaluating the influence degree of the sensor linear drift and the nonlinear temperature drift.

[0097] Specifically, the calculating the drift severity level of each sensor according to the sensor linear drift model and the sensor nonlinear temperature drift model refers to comprehensively analyzing the linear drift amount and the nonlinear temperature drift amount, quantitatively evaluating the drift severity of each sensor, and including the following steps:

[0098] The total drift amount of the sensor is calculated, and the linear drift and the nonlinear temperature drift are superimposed, and the specific formula is as follows:

[0099] △S total (t)=△S linear (t)+△S temp (t);

[0100] Wherein, △S total (t) is the total drift amount of the sensor; △S linear (t) is the linear drift amount of the sensor; and △S temp (t) is the nonlinear temperature drift amount of the sensor.

[0101] The total drift amount of the sensor is converted into an index in the range of 0-1 through normalization processing, to obtain a drift severity index, and the specific formula is as follows:

[0102]

[0103] Wherein, I drift (t) is the drift severity index; and S range is the measurement range of the sensor.

[0104] The drift severity level is determined according to the drift severity index.

[0105] For example, when determining the drift severity level, the drift severity index is compared with a preset threshold value, and then the level division is realized, and the specific grading standard is as follows:

[0106] When I drift (t)≤2%, the drift severity level is level 1, i.e. slight drift.

[0107] When 2%<I drift (t)≤5%, the drift severity level is level 2, i.e. moderate drift.

[0108] When 5%<I drift (t)≤10%, the drift severity level is level 3, i.e. high drift.

[0109] When I drift (t)>10%, the drift severity level is level 4, i.e. severe drift.

[0110] In an alternative embodiment, the drift severity level can be calculated by time window averaging method, by analyzing the drift data in the past N hours, calculating the average drift severity index, to avoid misjudgment of the level caused by instantaneous fluctuations.

[0111] In another alternative embodiment, the grading criteria can also be adjusted in combination with the particularity of the sensor type, for example, a voltage sensor with higher accuracy requirement adopts more stringent grading criteria, and a temperature sensor with relatively lower accuracy requirement adopts relatively relaxed grading criteria.

[0112] S302: Assign a drift compensation weight to each sensor based on the drift severity level.

[0113] In the embodiment of the present application, the drift compensation weight refers to a compensation intensity coefficient determined according to the drift severity level of the sensor, used to control the compensation strength of different level sensors.

[0114] Specifically, assigning a drift compensation weight to each sensor based on the drift severity level means assigning a corresponding compensation weight coefficient to each sensor according to the drift severity level, to ensure that the more serious the drift of a sensor, the stronger the compensation processing it obtains, including the following steps:

[0115] Assigning a basic compensation weight W based on the drift severity level base .

[0116] Considering the sensor type correction factor, the weight is adjusted according to the characteristics of different types of sensors, and the specific formula is as follows:

[0117] W type =W base ×γ sensor ;

[0118] Wherein, W type is the correction weight considering the sensor type; γ sensor is the sensor type correction factor, the correction factor of current sensor is 1.2, the correction factor of voltage sensor is 1.1γ voltage = 1.1, and the correction factor of temperature sensor is 0.9.

[0119] Introducing drift change trend correction, adjusting the weight according to the development trend of drift, and the specific formula is as follows:

[0120]

[0121] Wherein, W final is the final drift compensation weight; η is the trend correction coefficient, taking the value of 0.2-0.5; is the time derivative of the drift severity index, reflecting the drift deterioration trend.

[0122] In an alternative embodiment, a weight adjustment mechanism can be employed to adjust the weight distribution strategy according to the feedback information of compensation effect, and automatically increase the compensation weight of a sensor when its compensation effect is not ideal.

[0123] In another alternative embodiment, the historical performance data of the sensors can also be combined to employ a conservative weight distribution strategy for newly installed sensors and a relatively aggressive weight distribution strategy for long-term stable sensors.

[0124] S303: Calculate the drift compensation coefficient by the weighted fusion algorithm in combination with the drift compensation weight.

[0125] In the embodiments of the present application, the drift compensation coefficient refers to a numerical parameter used to correct the measurement data of the sensor, including a zero-point compensation coefficient and a range compensation coefficient, which are respectively used to correct the reference offset and gain offset of the sensor.

[0126] Specifically, calculating the drift compensation coefficient by the weighted fusion algorithm in combination with the drift compensation weight refers to calculating the drift compensation coefficient by the weighted fusion method using the drift compensation weight, realizing accurate compensation of the sensor drift, including the following steps:

[0127] Calculate the zero-point compensation coefficient, mainly compensating for the offset of the measurement reference value of the sensor, and the specific formula is as follows:

[0128]

[0129] Wherein, C zero is the zero-point compensation coefficient; S nominal is the nominal value of the sensor; cos(·) is the cosine function, which is used to reduce the compensation strength when the drift degree is high, avoiding overcompensation.

[0130] Calculate the range compensation coefficient, mainly compensating for the change of the gain characteristic of the sensor, and the specific formula is as follows:

[0131]

[0132] Wherein, C range is the range compensation coefficient; sin(·) is the sine function, which is used to realize progressive compensation, providing smaller compensation when the drift is slight and providing larger compensation when the drift is severe.

[0133] In an alternative embodiment, the moving average method can be employed to smooth the compensation coefficient, by calculating the average value of the compensation coefficient of the past N sampling periods, avoiding the sharp fluctuation of the compensation coefficient, and improving the stability of the compensation.

[0134] In another alternative embodiment, a compensation effect evaluation mechanism can also be introduced to adjust the calculation parameters of the compensation coefficient by comparing the sensor measurement accuracy before and after compensation, thereby realizing continuous optimization of the compensation strategy.

[0135] It should be noted that the present application realizes accurate evaluation and hierarchical management of the sensor state by calculating the drift severity level, ensures reasonable allocation of calibration resources, and especially establishes a personalized weight allocation strategy by introducing a sensor type correction factor and a drift change trend correction, so that sensors of different types and states can obtain the most suitable calibration intensity, thereby not only improving the targeting of calibration, but also avoiding the problems of over-calibration or insufficient calibration; at the same time, the drift compensation coefficient calculated by the weighted fusion algorithm comprehensively considers the combined effects of linear drift and nonlinear temperature drift, and realizes unified quantitative processing of multidimensional drift factors.

[0136] S400: Perform zero-point drift calibration and range drift calibration on the sensor measurement data combined with the drift compensation coefficient to obtain the calibrated data acquisition result.

[0137] S401: Perform zero-point drift calibration on the sensor measurement data combined with the drift compensation coefficient.

[0138] In the embodiments of the present application, zero-point drift calibration refers to correcting the reference offset of the sensor measurement data by using the zero-point compensation coefficient to eliminate the influence of sensor zero-point drift on measurement accuracy.

[0139] Specifically, zero-point drift calibration is performed on the sensor measurement data combined with the zero-point compensation coefficient to obtain a zero-point drift calibration result, including the following steps:

[0140] The zero-point offset is extracted from the sensor measurement data to determine the current zero-point offset degree; for a current sensor, the output value when there is no current input is the zero-point offset; for a voltage sensor, the output value when there is zero voltage input is the zero-point offset; for a temperature sensor, the zero-point offset is determined by comparison with a standard thermometer.

[0141] The zero-point compensation coefficient is calculated with the zero-point offset to generate a zero-point calibration parameter, and the specific formula is as follows:

[0142] P zero =S measured,zero -C zero,final ×S range ;

[0143] Wherein, P zero is the zero-point calibration parameter; S measured,zero is the zero-point offset measured by the sensor; C zero,final is the zero-point compensation coefficient; S rangeMeasuring range of the sensor.

[0144] In an alternative embodiment, a segmented calibration method can be used, in which different zero-point calibration parameters are set for different intervals of the sensor measurement range, to improve the accuracy of calibration and be suitable for sensors with obvious non-linear characteristics.

[0145] In another alternative embodiment, a temperature compensation technique can also be combined to adjust the zero-point calibration parameters according to changes in the ambient temperature, to further improve the calibration effect in a temperature-varying environment.

[0146] S402: Range drift calibration is performed on the sensor measurement data in combination with the range drift compensation coefficient.

[0147] In the embodiments of the present application, the range drift calibration refers to correcting the gain offset of the sensor measurement data using the range compensation coefficient, to eliminate the influence of sensor range drift on the measurement accuracy.

[0148] Specifically, the range drift calibration is performed on the sensor measurement data in combination with the range compensation coefficient, to obtain a range drift calibration result, including the following steps:

[0149] The trend of the range variation of the sensor measurement data is analyzed to determine the range drift pattern; wherein the type and degree of the range drift are identified by comparing the difference between the output response and the theoretical response of the sensor under different input signals, and the range drift pattern includes three types of linear gain drift, non-linear gain drift, and mixed gain drift.

[0150] The range compensation coefficient is applied to the range drift pattern to calculate the range calibration factor, and the specific formula is as follows:

[0151]

[0152] Wherein, F range is the range calibration factor; C range,final is the range compensation coefficient; a is the range non-linear correction coefficient, and the value is 0.1-0.3; S measured is the current measurement value of the sensor; S nominal is the nominal value of the sensor.

[0153] In an alternative embodiment, a multi-point calibration method can be used, in which multiple calibration points are selected within the sensor measurement range, the corresponding range calibration factors are calculated respectively, and the calibration factor of any measurement point is determined through an interpolation algorithm.

[0154] S403: The calibrated data acquisition result is obtained.

[0155] In the embodiments of the present application, the calibrated data acquisition result refers to the final sensor measurement data obtained after zero drift calibration and range drift calibration, which has higher measurement accuracy and reliability.

[0156] Specifically, the calibrated data acquisition result refers to the sensor data after comprehensive calibration obtained by comprehensively processing the results of zero drift calibration and range drift calibration. The calculation formula of the calibrated data acquisition result is as follows:

[0157] S calibrated = (S raw -P zero ) × F range ;

[0158] Wherein, S calibrated is the calibrated sensor measurement value; S raw is the original sensor measurement value; P zero is the zero calibration parameter; and F range is the range calibration factor.

[0159] For example, for a current sensor with an original measurement value of 25.8A, after zero calibration to remove a zero point offset of 0.3A, and after range calibration by applying a calibration factor of 1.02, the final calibrated measurement value is 26.01A. Compared with the standard ammeter, the error before calibration is 3.2%, and the error after calibration is reduced to 0.8%, which improves the measurement accuracy.

[0160] It should be noted that, compared with the traditional sensor calibration which usually adopts an overall calibration method and cannot distinguish and independently process the two different error sources of zero drift and range drift, the calibration effect is limited. In the present application, the sensor calibration is divided into two independent processes of zero drift calibration and range drift calibration, and the reference offset and gain offset of the sensor are respectively corrected, so that more refined calibration control is realized. The zero drift calibration is specially used for processing the reference value offset of the sensor, and the range drift calibration is specially used for compensating the change of the sensitivity of the sensor, so that each kind of drift can be calibrated by the most suitable method.

[0161] In summary, the intelligent data acquisition method and system of the distributed photovoltaic power station provided by the application can realize real-time precision calibration of key parameter sensors such as current, voltage and temperature by constructing a sensor drift compensation mechanism based on spectral time sequence variation characteristics, establishing a sensor state evaluation model by fully utilizing spectral information in the photovoltaic power station environment, effectively solving the problem of decreased data acquisition accuracy caused by sensor precision decay in the long-term operation of the distributed photovoltaic power station, reducing the operation and maintenance cost of the photovoltaic power station, improving the automation level and reliability of data acquisition, maintaining the long-term stability of sensor measurement precision in outdoor environmental conditions through the synergistic effect of the double drift model and the individualized compensation strategy, providing high-quality data support for the power generation statistics, fault diagnosis and performance evaluation of the photovoltaic power station, and ensuring the data acquisition precision and operation efficiency of the distributed photovoltaic power station.

[0162] Embodiment 3 is a third embodiment of the application, which provides an intelligent data acquisition system of a distributed photovoltaic power station, comprising: a data acquisition module configured to acquire spectral intensity data and sensor measurement data in the distributed photovoltaic power station; a drift model construction module configured to perform feature extraction based on the spectral intensity data to obtain spectral time sequence variation characteristics, and construct a sensor linear drift model and a sensor nonlinear temperature drift model based on the spectral time sequence variation characteristics; a drift compensation module configured to calculate drift compensation coefficients based on the sensor linear drift model and the sensor nonlinear temperature drift model; and a drift calibration module configured to perform zero-point drift calibration and range drift calibration on the sensor measurement data in combination with the drift compensation coefficients to obtain calibrated data acquisition results.

[0163] Embodiment 4 is a fourth embodiment of the application, which is different from the first three embodiments in that: if the function is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the application can be embodied in the form of a software product in essence or in the part that contributes to the prior art or part of the technical solutions. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device) to execute all or part of the steps of the method described in the embodiments of the application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0164] The logic and / or steps represented in the flowcharts and / or otherwise described herein, for example, can be considered as a sequence of executable instructions stored in a computer readable medium, which can be executed by an instruction execution system, apparatus or device, such as a computer-based system, processor, or other system that can fetch the instructions from the instruction execution system, apparatus or device and execute the instructions. For purposes of this specification, a "computer readable medium" can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus or device. The computer readable medium can be, for example but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device, propagation medium, or computer memory.

[0165] More specific examples (a non-exhaustive list) of the computer readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber (optical), and a portable compact disc read-only memory (CDROM). Additionally, the computer readable medium can also be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example via an optical scanner, then compiled, interpreted, or otherwise processed, and stored in a computer memory in order to be executed.

[0166] It should be understood that aspects of the application can be implemented in hardware, software, firmware or a combination thereof. In the above embodiments, various steps or methods can be implemented in software or firmware that is stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any of the following technologies, or combination thereof, can be used: a discrete logic circuit having logic gates for implementing logic functions upon data signals, an application specific integrated circuit having appropriate combinational logic gates, a programmable gate array (PGA), a field programmable gate array (FPGA), and / or the like.

[0167] It should be noted that the above-mentioned embodiments are merely used to illustrate the technical solutions of the present application but not to limit the present application, and although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalent replaced without departing from the spirit and scope of the present application, and all of them should be covered in the scope of the claims of the present application.

Claims

1. An intelligent data collection method for a distributed photovoltaic power station, characterized in that: include: Obtain spectral intensity data and sensor measurement data in distributed photovoltaic power plants; Extracting features based on the spectral intensity data to obtain spectral temporal variation features, and constructing a sensor linear drift model and a sensor nonlinear temperature drift model based on the spectral temporal variation features; Calculating based on the sensor linear drift model and the sensor nonlinear temperature drift model to obtain a drift compensation coefficient; The zero drift calibration and the span drift calibration are performed in combination with the drift compensation coefficient and the sensor measurement data to obtain a calibrated data acquisition result.

2. The intelligent data collection method for distributed photovoltaic power stations according to claim 1, characterized in that: A sensor linear drift model is constructed based on the spectral time series variation characteristics, including: Calculating the time derivative and second-order derivative of the spectrum intensity according to the time series variation characteristics of the spectrum; Comparing and analyzing the time derivative and the second derivative of the spectral intensity with a theoretical solar radiation model to generate a spectral deviation; The linear relationship between the spectral deviation and the sensor measurement deviation is fitted by the least square method to establish a sensor linear drift model.

3. The intelligent data collection method for distributed photovoltaic power stations according to claim 2, characterized in that: A nonlinear temperature drift model of the sensor is constructed based on the spectral time series variation characteristics, including: extracting a temperature-related spectral component from the spectral temporal variation characteristics; Establishing a nonlinear mapping relationship between the temperature-related spectral component and the ambient temperature change rate; A nonlinear temperature drift model of the sensor is constructed according to the nonlinear mapping relationship and the temperature coefficient of the sensor.

4. The intelligent data collection method for a distributed photovoltaic power station according to claim 3, characterized in that: Calculation is performed based on the sensor linear drift model and the sensor nonlinear temperature drift model to obtain a drift compensation coefficient, including: Calculating a drift severity level for each sensor based on the sensor linear drift model and the sensor nonlinear temperature drift model; assigning a drift compensation weight to each sensor based on the drift severity level; The drift compensation coefficient is calculated by combining the drift compensation weight through a weighted fusion algorithm, and the drift compensation coefficient includes a zero point compensation coefficient and a range compensation coefficient.

5. The intelligent data collection method for distributed photovoltaic power stations according to claim 4, characterized in that: The calibrated data acquisition results include: Performing zero drift calibration in combination with the zero point compensation coefficient and the sensor measurement data to obtain a zero drift calibration result; Performing a range drift calibration by combining the range compensation coefficient with the sensor measurement data to obtain a range drift calibration result; The calibrated data acquisition result is determined by combining the zero drift calibration result and the span drift calibration result.

6. The intelligent data collection method for a distributed photovoltaic power station according to claim 5, characterized in that: The zero drift calibration includes: Extracting a zero point offset from the sensor measurement data; The zero point compensation coefficient and the zero point offset are calculated to generate a zero point calibration parameter.

7. The intelligent data collection method for a distributed photovoltaic power station according to claim 6, characterized in that: The span drift calibration includes: Analyzing the range variation trend of the sensor measurement data to determine the range drift mode; The range compensation coefficient is applied to the range drift mode to calculate a range calibration factor.

8. An intelligent data acquisition system for a distributed photovoltaic power station, according to the intelligent data acquisition method for a distributed photovoltaic power station according to any one of claims 1 to 7, characterized in that: include: Data acquisition module, used to obtain spectral intensity data and sensor measurement data in distributed photovoltaic power stations; The drift model construction module is used to extract features based on spectral intensity data, obtain spectral time series variation features, and construct a sensor linear drift model and a sensor nonlinear temperature drift model based on the spectral time series variation features; A drift compensation module is used to calculate the drift compensation coefficient based on the sensor linear drift model and the sensor nonlinear temperature drift model; The drift calibration module is used to perform zero drift calibration and span drift calibration by combining the drift compensation coefficient with the sensor measurement data to obtain the calibrated data acquisition results.

9. A computer device, characterized in that: include: A memory, a processor, and a computer program, wherein the computer program is stored in the memory, and the processor runs the computer program to execute the intelligent data collection method for a distributed photovoltaic power station according to any one of claims 1 to 7.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which, when executed by a processor, is used to implement the intelligent data collection method for a distributed photovoltaic power station according to any one of claims 1 to 7.

Images