Multi-model and multi-mode combined photovoltaic power generation device and system

Abstract

The invention discloses a multi-model and multi-mode combined photovoltaic power generation device and system, and belongs to the technical field of photovoltaic power supplies and transmission, the multi-model and multi-mode combined photovoltaic power generation device comprises a photovoltaic panel, a micro power device is electrically connected to the photovoltaic panel, an AI algorithm module and a carrier communication module are arranged in the micro power device, and an energy storage device is arranged outside the micro power device. A monitoring unit is arranged outside the micro power equipment; according to the invention, through distributed control and AI algorithm optimization, each photovoltaic panel can operate in an optimal working state, so that the overall power generation efficiency is improved, and under the condition that other gateways or communication management equipment is not needed, the system can be directly connected with energy storage equipment or a battery, can adapt to the states of voltage, SOC and the like of the energy storage equipment or the battery for charging, and is flexible in configuration; free combination of photovoltaic panels of different models and different platforms is supported, the scale and power of the system can be flexibly configured according to actual requirements, the adaptability and expansibility of the system are improved, the system is safe and reliable, and the monitoring unit can monitor the states of the photovoltaic panels and the system in real time.

Description

Technical Field

[0001] The present invention belongs to the field of photovoltaic power supply and transmission technology, and in particular relates to a photovoltaic power generation device and system with multiple models and modes. Background Art

[0002] In the prior art, photovoltaic optimizers and photovoltaic controllers are used;

[0003] A photovoltaic optimizer is a device installed on the back of a photovoltaic module to optimize the output performance of each photovoltaic module individually. However, a gateway or controller must be installed separately to achieve centralized management and data transmission, which increases the complexity and cost of the system. The compatibility between solar panels of different models, powers and platforms is poor, and there may be efficiency losses when used in combination.

[0004] The photovoltaic controller is a device used to control the working status of the photovoltaic system. Its main function is to protect the battery and prevent overcharging and over-discharging. However, it has strict restrictions on the specifications and connection methods of solar panels, lacks flexibility, and cannot achieve independent optimization and monitoring of each photovoltaic module.

[0005] Centralized control limitations: Existing photovoltaic power generation systems mostly use centralized photovoltaic charge controllers, which limits the free combination of solar panels. Solar panels of different models and platforms are difficult to connect in series and parallel, and cannot be flexibly configured according to actual needs;

[0006] Inefficiency: Centralized control cannot optimize each solar panel independently, resulting in low overall power generation efficiency. When some solar panels perform poorly or fail, the power generation of the entire system will be affected.

[0007] Lack of monitoring and management: The existing system lacks monitoring and management capabilities for each solar panel and the entire system, making it impossible to detect faults and perform maintenance in a timely manner, nor to perform data analysis and performance evaluation.

[0008] Based on this, the present invention designs a photovoltaic power generation device and system with multiple models and modes combined to solve the above problems. Summary of the invention

[0009] The purpose of the present invention is to solve the problems in the above-mentioned background technology and to propose a photovoltaic power generation device and system with multiple models and modes.

[0010] In order to achieve the above object, the present invention adopts the following technical solutions:

[0011] A photovoltaic power generation device with multiple models and modes, including a photovoltaic panel, to which a micro power device is electrically connected, wherein an AI algorithm module and a carrier communication module are provided inside the micro power device, an energy storage device is provided outside the micro power device, and a monitoring unit is provided outside the micro power device.

[0012] A photovoltaic power generation system with multiple models and modes, including:

[0013] Photovoltaic panels, which are used to collect and convert light energy;

[0014] Micro power system (MCS), which is distributed and installed near each photovoltaic panel, is used to convert, control and regulate the output current and voltage of the photovoltaic panel;

[0015] The AI ​​algorithm module runs in the MCS and is responsible for analyzing the performance and system status of the photovoltaic panels, and dynamically adjusting the MCS parameters to achieve current and voltage balance;

[0016] Carrier communication module, to realize communication between MCSs, exchange information and instructions;

[0017] Energy storage devices, used to store excess electrical energy;

[0018] The monitoring unit is used to monitor the status of the photovoltaic panels and the system, and provide data analysis and fault diagnosis functions, and edit and write data to operate the MCS module.

[0019] As a further description of the above technical solution:

[0020] The monitoring unit includes a data acquisition unit, a data transmission unit, a data processing and analysis unit, and a fault diagnosis and alarm unit;

[0021] The data acquisition unit is used to collect various data in real time, including the voltage, current, power of the photovoltaic panel, the ambient light intensity, temperature, humidity, the power status of the energy storage device, and the power consumption of the load;

[0022] A data transmission unit, which transmits the collected data to a data processing and analysis unit through a communication interface;

[0023] Data processing and analysis unit, used to filter, denoise and calibrate the raw data to eliminate sensor noise and measurement errors;

[0024] Analyze the processed data in real time to monitor the operating status of the photovoltaic system;

[0025] Calculate the power generation efficiency of each photovoltaic panel and identify inefficient or abnormal photovoltaic panels;

[0026] Utilize historical data and current status to predict the system's future power generation and load demand, providing a basis for optimized control.

[0027] As a further description of the above technical solution:

[0028] The fault diagnosis and alarm unit monitors various indicators in the system in real time and identifies potential faults through set thresholds and anomaly detection algorithms;

[0029] Once a fault or abnormality is detected, the system immediately triggers an alarm and notifies the operation and maintenance personnel by sending a text message;

[0030] All fault events and system status changes are recorded in logs to facilitate subsequent fault analysis and maintenance.

[0031] As a further description of the above technical solution:

[0032] The AI ​​algorithm module adopts a balanced voltage optimization model, the goal of which is to minimize the deviation between the output voltages of each photovoltaic panel, thereby achieving voltage balance;

[0033] Optimization algorithm,In order to minimize the objective function E, the gradient descent method is used as the optimization algorithm.,The gradient descent method calculates the gradient of the objective function with respect to the control parameters, and updates the parameters in the reverse direction of the gradient to approach the optimal solution;

[0034] Initialization, determine the initial MCS parameters;

[0035] Calculate the initial photovoltaic panel output voltage V pvi ;

[0036] Calculate the average voltage, Where V pvi is the output voltage of the i-th photovoltaic panel, n is the total number of photovoltaic panels, V avg is the average output voltage of all photovoltaic panels;

[0037] Calculate the voltage deviation,

[0038] Calculate the gradient. For each MCS parameter, calculate the partial derivative of the objective function E with respect to the parameter.

[0039] Update parameters, update MCS parameters according to the gradient descent formula,

[0040] Among them, α is the learning rate, which controls the size of the update step;

[0041] Iteration, repeat the average voltage calculation-parameter update step until the objective function E converges below a preset threshold.

[0042] As a further description of the above technical solution:

[0043] The AI ​​algorithm module also includes an optimization unit, and the optimization steps include:

[0044] Data collection and preprocessing require the collection of weather-related data such as ambient temperature, humidity, wind speed, light intensity, and cloud coverage;

[0045] Model training, introducing weather variables as input features of the model, and using data augmentation technology to enhance the data used for training;

[0046] Model evaluation: Use test data to evaluate the performance of the model and continuously iterate and improve the model based on the evaluation results.

[0047] As a further description of the above technical solution:

[0048] The collected data is cleaned to remove outliers and noise, and is labeled to facilitate learning and training of the AI ​​algorithm.

[0049] As a further description of the above technical solution:

[0050] The enhancement techniques include random rotation and scaling to improve the generalization ability of the model.

[0051] As a further description of the above technical solution:

[0052] The numerical values ​​used to evaluate model performance include accuracy, recall, and F1.

[0053] As a further description of the above technical solution:

[0054] In the process of removing the outliers, the data is judged and processed using the median filtering algorithm:

[0055] For a data point z i , select a certain number of data points before and after it to form a window containing 2h+1 points, where h is the radius of the window;

[0056] The window range formula is: i-h ,…,z i ,…,z i+h};

[0057] Sort the data in the window from small to large, and take the median of the sorting as the filter value f(o) of the point p ,g p );

[0058] Median formula, f(o p ,g p )=med(z i-h,…,z i ,…,z i+h )

[0059] Among them, z i-h ,…,z i ,…,z i+h are all the data points in the window;

[0060] Calculate the standard deviation σ of the data in the window for subsequent determination of outliers;

[0061] Standard deviation formula,

[0062] Where μ is the mean of the data in the window;

[0063] Compare data points z i and the median f(o p ,g p ), and judge whether the data point is an outlier according to the set threshold hσ:

[0064] Judgment formula,

[0065] Among them, the value of h is usually 2 or 3, and σ is the standard deviation of the data in the window;

[0066] If a data point is judged to be abnormal, the following processing is performed:

[0067] Replace it with the median f(o p ,g p ) replaces the abnormal data point z i ;

[0068] Ignore and directly raise the abnormal data points;

[0069] Mark it as abnormal data for subsequent processing.

[0070] The median filter algorithm is a nonlinear filter that replaces each data point with the median of its surrounding neighborhood data points. The principle is that the median can effectively resist the influence of outliers because it is not affected by extreme values. For example, if a set of data contains an abnormal maximum value, the median will be lower than the mean, thereby better reflecting the true distribution of the data.

[0071] Advantages of median filtering algorithm:

[0072] Simple and easy: The algorithm principle is simple and easy to implement.

[0073] Strong robustness: strong resistance to outliers and noise.

[0074] Preserve edge information: Does not blur the edges of the image like a low-pass filter.

[0075] The denoised data can better reflect the true distribution of the data, thereby improving the prediction accuracy of the model;

[0076] Improve model robustness: Denoised data can reduce the impact of outliers and noise on the model, thereby improving the robustness of the model;

[0077] Reduce model complexity: Denoised data can reduce the number of features that the model needs to learn, thereby reducing the complexity of the model.

[0078] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0079] 1. In the present invention, through distributed control and AI algorithm optimization, the present invention can make each photovoltaic panel operate in the best working state, thereby improving the overall power generation efficiency; flexible configuration: supports free combination of photovoltaic panels of different models and platforms, and can flexibly configure the system scale and power according to actual needs, thereby improving the adaptability and scalability of the system; safe and reliable: the monitoring unit can monitor the status of the photovoltaic panels and the system in real time, detect faults in time and perform maintenance, ensure the safe and reliable operation of the system, and reduce maintenance costs; intelligent management: the AI ​​algorithm module can dynamically adjust the system parameters according to environmental changes and load requirements, realize intelligent management, improve energy utilization efficiency, and reduce energy waste; reduce costs: by improving power generation efficiency, reducing maintenance costs, etc., the overall cost of the photovoltaic power generation system can be effectively reduced, and the economic benefits can be improved;

[0080] This product integrates the functions of optimizer and controller and overcomes the limitations of both through innovative technology. Compared with photovoltaic optimizers and controllers, it has higher functional integration, reduces the number and complexity of equipment, reduces system costs, and has more efficient and convenient communication and management methods. It can achieve decentralized free networking without additional controllers or equipment. It has stronger compatibility with solar panels, more flexible installation and configuration, and can make full use of various resources to improve the power generation efficiency and stability of the entire photovoltaic system.

[0081] 2. In the present invention, a balanced voltage optimization model is adopted to minimize the deviation between the output voltages of each photovoltaic panel, thereby achieving voltage balance and improving the overall power generation efficiency; Gradient descent method: As an optimization algorithm, the gradient descent method calculates the gradient of the objective function with respect to the control parameters, and updates the parameters in the reverse direction of the gradient to approach the optimal solution and ensure the realization of the voltage balance target; the system will collect weather-related data such as ambient temperature, humidity, wind speed, light intensity, cloud coverage, etc., and clean, denoise and annotate them to provide high-quality data for the learning and training of AI algorithms; Model training and evaluation: weather variables are introduced as input features of the model, and data enhancement technology is used to enhance the training data to improve the generalization ability of the model, and indicators such as accuracy, recall and F1 are used to evaluate the model performance to ensure the reliability and effectiveness of the model;

[0082] Improve power generation efficiency: Through voltage balancing and intelligent optimization, the power generation efficiency of each photovoltaic panel can be maximized, thereby increasing the overall power generation; Reduce operation and maintenance costs: The photovoltaic panels and system status can be monitored in real time, and fault diagnosis can be performed to promptly discover and solve problems and reduce operation and maintenance costs; Extend equipment life: Through optimized control, the workload of the equipment can be reduced and the service life of the equipment can be extended; Improve safety: Various indicators can be monitored in real time and alarms can be triggered to ensure the safe operation of the system, and they do not need to be concentrated on a certain gateway or control device. Each one can be used separately, and can be flexibly matched in series and parallel to charge the battery. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 A schematic diagram of the connection of a photovoltaic power generation device and system of multiple models and modes proposed by the present invention;

[0084] Figure 2 A schematic diagram of the parallel connection of photovoltaic panels and micro power devices in a photovoltaic power generation device and system of multiple models and modes proposed by the present invention;

[0085] Figure 3 A schematic diagram of a photovoltaic panel and a micro power device in series connection in a photovoltaic power generation device and system of multiple models and modes proposed by the present invention;

[0086] Figure 4 A schematic diagram of system connection of a photovoltaic power generation device and system with multiple models and modes of combination proposed by the present invention;

[0087] Figure 5 This is a schematic diagram of the connection of a photovoltaic power generation device and an energy storage device in a system of multiple models and modes proposed by the present invention. DETAILED DESCRIPTION

[0088] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0089] Please see attached Figure 1 -Attached Figure 5 The present invention provides a technical solution: a photovoltaic power generation device with multiple models and modes, including a photovoltaic panel, a micro power device electrically connected to the photovoltaic panel, an AI algorithm module and a carrier communication module inside the micro power device, an energy storage device outside the micro power device, and a monitoring unit outside the micro power device.

[0090] A photovoltaic power generation system with multiple models and modes, including:

[0091] Photovoltaic panels, which are used to collect and convert light energy;

[0092] Micro power system (MCS), which is distributed and installed near each photovoltaic panel, is used to convert, control and regulate the output current and voltage of the photovoltaic panel;

[0093] The AI ​​algorithm module runs in the MCS and is responsible for analyzing the performance and system status of the photovoltaic panels, and dynamically adjusting the MCS parameters to achieve current and voltage balance;

[0094] Carrier communication module, to realize communication between MCSs, exchange information and instructions;

[0095] Energy storage devices, used to store excess electrical energy;

[0096] The monitoring unit is used to monitor the status of the photovoltaic panels and the system, and provide data analysis and fault diagnosis functions, and edit and write data to operate the MCS module.

[0097] The monitoring unit includes a data acquisition unit, a data transmission unit, a data processing and analysis unit, and a fault diagnosis and alarm unit;

[0098] The data acquisition unit is used to collect various data in real time, including the voltage, current, power of the photovoltaic panel, the ambient light intensity, temperature, humidity, the power status of the energy storage device, and the power consumption of the load;

[0099] A data transmission unit, which transmits the collected data to a data processing and analysis unit through a communication interface;

[0100] Data processing and analysis unit, used to filter, denoise and calibrate the raw data to eliminate sensor noise and measurement errors;

[0101] Analyze the processed data in real time to monitor the operating status of the photovoltaic system;

[0102] Calculate the power generation efficiency of each photovoltaic panel and identify inefficient or abnormal photovoltaic panels;

[0103] Utilize historical data and current status to predict the system's future power generation and load demand, providing a basis for optimized control.

[0104] The fault diagnosis and alarm unit monitors various indicators in the system in real time and identifies potential faults through set thresholds and anomaly detection algorithms;

[0105] Once a fault or abnormality is detected, the system immediately triggers an alarm and notifies the operation and maintenance personnel by sending a text message;

[0106] All fault events and system status changes are recorded in logs to facilitate subsequent fault analysis and maintenance.

[0107] The AI ​​algorithm module adopts a balanced voltage optimization model, the goal of which is to minimize the deviation between the output voltages of each photovoltaic panel, thereby achieving voltage balance;

[0108] Optimization algorithm,In order to minimize the objective function E, the gradient descent method is used as the optimization algorithm.,The gradient descent method calculates the gradient of the objective function with respect to the control parameters, and updates the parameters in the reverse direction of the gradient to approach the optimal solution;

[0109] Initialization, determine the initial MCS parameters;

[0110] Calculate the initial photovoltaic panel output voltage V pvi ;

[0111] Calculate the average voltage, Where V pvi is the output voltage of the i-th photovoltaic panel, n is the total number of photovoltaic panels, V avg is the average output voltage of all photovoltaic panels;

[0112] Calculate the voltage deviation,

[0113] Calculate the gradient. For each MCS parameter, calculate the partial derivative of the objective function E with respect to the parameter.

[0114] Update parameters, update MCS parameters according to the gradient descent formula,

[0115] Among them, α is the learning rate, which controls the size of the update step;

[0116] Iteration, repeat the average voltage calculation-parameter update step until the objective function E converges below a preset threshold.

[0117] The AI ​​algorithm module also includes an optimization unit, and the optimization steps include:

[0118] Data collection and preprocessing require the collection of weather-related data such as ambient temperature, humidity, wind speed, light intensity, and cloud coverage;

[0119] Model training, introducing weather variables as input features of the model, and using data augmentation technology to enhance the data used for training;

[0120] Model evaluation: Use test data to evaluate the performance of the model and continuously iterate and improve the model based on the evaluation results.

[0121] The collected data is cleaned to remove outliers and noise, and is labeled to facilitate learning and training of AI algorithms.

[0122] Augmentation techniques include random rotation and scaling to improve the generalization ability of the model.

[0123] The values ​​used to evaluate model performance include precision, recall, and F1.

[0124] In the process of removing outliers, the data is judged and processed using the median filtering algorithm:

[0125] For a data point z i , select a certain number of data points before and after it to form a window containing 2h+1 points, where h is the radius of the window;

[0126] The window range formula is: i-h ,…,z i ,…,z i+h};

[0127] Sort the data in the window from small to large, and take the median of the sorting as the filter value f(o) of the point p ,g p );

[0128] Median formula, f(o p ,g p )=med(z i-h ,…,z i ,…,z i+h )

[0129] Among them, z i-h ,…,z i ,…,z i+h are all the data points in the window;

[0130] Calculate the standard deviation σ of the data in the window for subsequent determination of outliers;

[0131] Standard deviation formula,

[0132] Where μ is the mean of the data in the window;

[0133] Compare data points z i and the median f(o p ,g p ), and judge whether the data point is an outlier according to the set threshold hσ:

[0134] Judgment formula,

[0135] Among them, the value of h is usually 2 or 3, and σ is the standard deviation of the data in the window;

[0136] If a data point is judged to be abnormal, the following processing is performed:

[0137] Replace it with the median f(o p ,g p ) replaces the abnormal data point z i ;

[0138] Ignore and directly raise the abnormal data points;

[0139] Mark it as abnormal data for subsequent processing.

[0140] like Figure 2 and Figure 3 As shown, the photovoltaic panel and the micro power device can be connected in series or in parallel.

[0141] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A photovoltaic power generation device with multiple models and modes, including photovoltaic panels, characterized in that: A micro power device is electrically connected to the photovoltaic panel, an AI algorithm module and a carrier communication module are arranged inside the micro power device, an energy storage device is arranged outside the micro power device, and a monitoring unit is arranged outside the micro power device.

2. A photovoltaic power generation system with multiple models and modes, characterized in that: A photovoltaic power generation device with multiple models and modes according to claim 1, comprising: Photovoltaic panels, which are used to collect and convert light energy; Micro power system (MCS), which is distributed and installed near each photovoltaic panel, is used to convert, control and regulate the output current and voltage of the photovoltaic panel; The AI ​​algorithm module runs in the MCS and is responsible for analyzing the performance and system status of the photovoltaic panels, and dynamically adjusting the MCS parameters to achieve current and voltage balance; Carrier communication module, to realize communication between MCSs, exchange information and instructions; Energy storage devices, used to store excess electrical energy; The monitoring unit is used to monitor the status of the photovoltaic panels and the system, and provide data analysis and fault diagnosis functions, and edit and write data to operate the MCS module.

3. A photovoltaic power generation system with multiple models and modes according to claim 2, characterized in that: The monitoring unit includes a data acquisition unit, a data transmission unit, a data processing and analysis unit, and a fault diagnosis and alarm unit; The data acquisition unit is used to collect various data in real time, including the voltage, current, power of the photovoltaic panel, the ambient light intensity, temperature, humidity, the power status of the energy storage device, and the power consumption of the load; A data transmission unit, which transmits the collected data to a data processing and analysis unit through a communication interface; Data processing and analysis unit, used to filter, denoise and calibrate the raw data to eliminate sensor noise and measurement errors; Analyze the processed data in real time to monitor the operating status of the photovoltaic system; Calculate the power generation efficiency of each photovoltaic panel and identify inefficient or abnormal photovoltaic panels; Utilize historical data and current status to predict the system's future power generation and load demand, providing a basis for optimized control.

4. A photovoltaic power generation system with multiple models and modes according to claim 3, characterized in that: The fault diagnosis and alarm unit monitors various indicators in the system in real time and identifies potential faults through set thresholds and anomaly detection algorithms; Once a fault or abnormality is detected, the system immediately triggers an alarm and notifies the operation and maintenance personnel by sending a text message; All fault events and system status changes are recorded in logs to facilitate subsequent fault analysis and maintenance.

5. A photovoltaic power generation system with multiple models and modes according to claim 2, characterized in that: The AI ​​algorithm module adopts a balanced voltage optimization model, the goal of which is to minimize the deviation between the output voltages of each photovoltaic panel, thereby achieving voltage balance; Optimization algorithm,In order to minimize the objective function E, the gradient descent method is used as the optimization algorithm.,The gradient descent method calculates the gradient of the objective function with respect to the control parameters, and updates the parameters in the reverse direction of the gradient to approach the optimal solution; Initialization, determine the initial MCS parameters; Calculate the initial photovoltaic panel output voltage V pvi ; Calculate the average voltage, Where V pvi is the output voltage of the i-th photovoltaic panel, n is the total number of photovoltaic panels, V avg is the average output voltage of all photovoltaic panels; Calculate the voltage deviation, Calculate the gradient. For each MCS parameter, calculate the partial derivative of the objective function E with respect to the parameter. Update parameters, update MCS parameters according to the gradient descent formula, Among them, α is the learning rate, which controls the size of the update step; Iteration, repeat the average voltage calculation-parameter update step until the objective function E converges below a preset threshold.

6. A photovoltaic power generation system with multiple models and modes according to claim 2, characterized in that: The AI ​​algorithm module also includes an optimization unit, and the optimization steps include: Data collection and preprocessing require the collection of weather-related data such as ambient temperature, humidity, wind speed, light intensity, and cloud coverage; Model training, introducing weather variables as input features of the model, and using data augmentation technology to enhance the data used for training; Model evaluation: Use test data to evaluate the performance of the model and continuously iterate and improve the model based on the evaluation results.

7. A photovoltaic power generation system with multiple models and modes according to claim 6, characterized in that: The collected data is cleaned to remove outliers and noise, and is labeled to facilitate learning and training of the AI ​​algorithm.

8. A photovoltaic power generation device and system with multiple models and modes according to claim 6, characterized in that: The enhancement techniques include random rotation and scaling to improve the generalization ability of the model.

9. A photovoltaic power generation system with multiple models and modes according to claim 6, characterized in that: The numerical values ​​used to evaluate model performance include accuracy, recall, and F1.

10. A photovoltaic power generation system with multiple models and modes according to claim 7, characterized in that: In the process of removing the outliers, the data is judged and processed using the median filtering algorithm: For a data point z i , select a certain number of data points before and after it to form a window containing 2h+1 points, where h is the radius of the window; The window range formula is: i-h ,…,z i ,…,z i+h }; Sort the data in the window from small to large, and take the median of the sorting as the filter value f(o) of the point p ,g p ); Median formula, f(o p ,g p )=med(z i-h ,…,z i ,…,z i+h ) Among them, z i-h ,…,z i ,…,z i+h are all the data points in the window; Calculate the standard deviation σ of the data in the window for subsequent determination of outliers; Standard deviation formula, Where μ is the mean of the data in the window; Compare data points z i and the median f(o p ,g p ), and judge whether the data point is an outlier according to the set threshold hσ: Judgment formula, Among them, the value of h is usually 2 or 3, and σ is the standard deviation of the data in the window; If a data point is judged to be abnormal, the following processing is performed: Replace it with the median f(o p ,g p ) replaces the abnormal data point z i ; Ignore and directly raise the abnormal data points; Mark it as abnormal data for subsequent processing.

Images