Machine Learning-Based Reactive Power Optimization Control System and Method for Industrial Kilns

Abstract

This invention discloses a machine learning-based reactive power optimization control system and method for industrial kilns, relating to the fields of industrial automation and power control technology. The method includes the following steps: real-time monitoring of the adjustment process of the reactive power compensation device, acquiring reactive power output data of the compensation device during operation, thereby understanding its operating status and its impact on grid voltage stability. This invention, through real-time monitoring of power system data, combined with spectrum analysis and machine learning models, accurately identifies the matching risk between harmonic frequencies and the frequency of the reactive power compensation device, intelligently assesses and dynamically adjusts the number of capacitors connected, and optimizes the operating frequency of the compensation device. This method effectively avoids harmonic resonance, ensures grid voltage stability, reduces equipment failures and overheating, improves kiln operating efficiency, reduces energy waste and production downtime risks, ultimately saving equipment maintenance and energy costs, and improving the economic benefits of industrial production.

Description

Technical Field

[0001] This invention relates to the fields of industrial automation and power control technology, specifically to a reactive power optimization control system and method for industrial kilns based on machine learning. Background Technology

[0002] Machine learning-based reactive power optimization control for industrial kilns is a control method that utilizes machine learning technology to intelligently manage and optimize the power system in industrial kilns. Reactive power is the portion of the power system that cannot be directly converted into usable work, and it has a significant impact on the stability of the power grid and the operating efficiency of electrical equipment. During the operation of industrial kilns, by monitoring the kiln's electrical parameters and operating status in real time, machine learning models can analyze historical data and identify the variation patterns of reactive power under different operating conditions. Based on these analysis results, the model can intelligently predict and adjust reactive power demand, thereby optimizing the kiln's energy efficiency, reducing reactive power losses, improving equipment operating efficiency, reducing electricity consumption, and ensuring the stability of the power system. This method reduces human intervention through automation and intelligence, improving control accuracy and energy utilization efficiency.

[0003] Existing technologies have the following shortcomings: During the operation of industrial kilns, it is often necessary to provide capacitive reactive power through reactive power compensation devices (such as static var compensators, SVCs) to stabilize the grid voltage and ensure the normal operation of the kiln equipment. Reactive power compensation helps maintain the voltage level of the power system and avoids voltage fluctuations or overload problems in kiln heating equipment, motors, and other electrical equipment. However, when the operating frequency of the reactive power compensation device coincides with the harmonic frequency of the power system, it may cause resonance in the power system.

[0004] Resonance refers to the interaction between the capacitive reactive power of a compensation device and the harmonic frequencies in the power grid under specific conditions, leading to amplification of harmonic signals in the system. This amplification of harmonic frequencies exacerbates operating losses in electrical equipment, particularly in kilns, potentially causing abnormal temperature increases in heating elements, motors, and power control equipment, resulting in serious malfunctions such as overheating and burnout. Prolonged exposure to this unstable electrical environment not only significantly reduces kiln efficiency but can also lead to production line shutdowns and even disrupt the power supply of the entire industrial area, causing large-scale shutdowns and economic losses.

[0005] Therefore, in industrial kilns, if the resonance phenomenon between the reactive power compensation device and the harmonic frequency is not effectively avoided, it may seriously affect the stable operation of the kiln and the safety of the equipment, leading to unnecessary equipment maintenance costs, production stoppage risks and energy waste.

[0006] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0007] The purpose of this invention is to provide a machine learning-based reactive power optimization control system and method for industrial kilns. By monitoring power system data in real time, combined with spectrum analysis and machine learning models, the system accurately identifies the mismatch risk between harmonic frequencies and the frequency of reactive power compensation devices, intelligently assesses and dynamically adjusts the number of capacitors connected, and optimizes the operating frequency of the compensation devices. This method effectively avoids harmonic resonance, ensures grid voltage stability, reduces equipment failures and overheating, improves kiln operating efficiency, reduces energy waste and production downtime risks, and ultimately saves on equipment maintenance and energy costs, improving the economic benefits of industrial production, thereby solving the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a machine learning-based reactive power optimization control method for industrial kilns, comprising the following steps: Real-time monitoring of the adjustment process of the reactive power compensation device, obtaining reactive power output data of the compensation device during operation, thereby understanding its working status and its impact on grid voltage stability; Simultaneously, voltage and current waveform signals in the power system are collected, and harmonic frequency change data are obtained through spectrum analysis to identify the existence state of harmonics and their frequency fluctuation trend over time. The acquired raw data is preprocessed, and the preprocessed data is classified and organized to establish a structured dataset. Key indicators reflecting the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system are extracted from the dataset. The extracted key indicators are then comprehensively analyzed to quantify the degree of risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency. The analyzed key indicators are input into a pre-trained machine learning model, which intelligently assesses the risk of the reactive power compensation device's operating frequency coinciding with the harmonic frequency in the power system. When it is identified that the operating frequency of the reactive power compensation device may overlap with the harmonic frequency in the power system, the number of capacitors connected is dynamically reduced based on the assessment results to reduce the capacitive reactive power of the reactive power compensation device, thereby adjusting and changing the operating frequency of the compensation device to avoid resonance with the harmonic frequency.

[0009] Preferably, the acquired raw data is preprocessed and further categorized and organized to establish a structured dataset. The specific steps are as follows: First, the raw data is cleaned to remove noise, outliers, and invalid data, ensuring the accuracy and reliability of the data. Secondly, the data is synchronized and aligned in time to ensure consistency of different data sources under the same time base. Next, normalization or standardization methods are applied to unify different physical quantities to the same dimension or numerical range, which facilitates subsequent analysis and modeling. Subsequently, the data was categorized and organized according to its characteristics; Finally, the cleaned, aligned, standardized, and grouped data is transformed into a structured format, forming a dataset that can be directly used for model calls and analysis.

[0010] Preferably, key indicators reflecting the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system are extracted from the dataset. The extracted indicators include the degree of mutual influence between the harmonic signals in the power system and the operating frequency of the reactive power compensation device, and the output stability of the reactive power compensation device. The degree of mutual influence between the harmonic signals in the power system and the operating frequency of the reactive power compensation device, and the output stability of the reactive power compensation device are comprehensively analyzed under the detection window to generate harmonic cross-response reference values ​​and reactive power compensation stability reference values, respectively. The degree of risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency is quantified by the harmonic cross-response reference values ​​and the reactive power compensation stability reference values.

[0011] Preferably, the specific steps for generating harmonic cross-response reference values ​​by comprehensively analyzing the mutual influence between harmonic signals in the power system and the operating frequency of the reactive power compensation device within a detection window are as follows: This paper analyzes the frequency distribution of harmonic signals in the power system and the operating frequency of the reactive power compensation device, evaluates the relationship between the harmonic frequency and the operating frequency of the compensation device, and calculates the frequency deviation index for each frequency component. The calculation expression is as follows: , In the formula, It is the frequency deviation index. It is the operating frequency of the reactive power compensation device. These are harmonic frequencies in a power system; Obtaining the frequency deviation index Then, the frequency deviation is combined with the amplitude response of the harmonic signal to generate a harmonic cross-response reference value. By quantifying the amplitude influence of the harmonic signal, the interaction strength between the operating frequency of the compensation device and the harmonic frequency is further determined. The formula is as follows: , In the formula, This is the reference value for harmonic cross-response. It is the first Frequency deviation index of a harmonic signal It is the first The amplitude of the harmonic signal It is the amplitude of the capacitive reactive power output by the reactive power compensation device.

[0012] Preferably, the specific steps for generating a reactive power compensation stability reference value through comprehensive analysis of the reactive power compensation device's output stability within the detection window are as follows: Within the detection window, the real-time data of the output power of the reactive power compensation device and the grid voltage are continuously monitored to obtain comprehensive fluctuation response data under disturbance conditions, and a fluctuation response index is generated. The generation formula is as follows: , In the formula, It is a volatility response index. It is a reactive power compensation device in time Real-time output reactive power at all times. This is the rated maximum output power of the reactive power compensation device. It is the power grid in time The actual voltage value at that moment. This is the reference voltage value. It is the total duration of the detection window. It refers to minute changes over an extremely short period of time; Based on volatility response index By combining the offset between the operating frequency of the compensation device and the harmonic frequency of the power system, a reference value for reactive power compensation stability is constructed. The specific calculation formula is as follows: , In the formula, This is a reference value for reactive power compensation stability. It is the frequency offset. It is the power system reference frequency. It is the maximum fluctuation response threshold.

[0013] Preferably, the analyzed harmonic cross-response reference value and reactive power compensation stability reference value are input into a pre-trained machine learning model. The machine learning model generates a frequency overlap risk coefficient, and the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system is intelligently assessed using the frequency overlap risk coefficient.

[0014] Preferably, the frequency overlap risk coefficient generated by the pre-trained machine learning model during the intelligent assessment of the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system is compared and analyzed with a pre-set reference threshold for the frequency overlap risk coefficient to determine whether there is a risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system. The judgment logic is as follows: If the frequency overlap risk coefficient is greater than the preset reference threshold for frequency overlap risk coefficient, it is determined that there is a risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system; if the frequency overlap risk coefficient is less than or equal to the preset reference threshold for frequency overlap risk coefficient, it is determined that there is no risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system.

[0015] Preferably, when it is identified that the operating frequency of the reactive power compensation device overlaps with the harmonic frequency in the power system, the specific steps for dynamically reducing the number of capacitors connected and lowering the capacitive reactive power of the reactive power compensation device based on the assessment results are as follows: When a frequency overlap risk is detected, capacitive reactive power is reduced by adjusting the number of capacitors connected, thereby avoiding resonance. The adjustment amount for the number of capacitors connected is calculated based on the difference between the frequency overlap risk coefficient and the reference threshold. The formula for calculating the adjustment amount is as follows: , In the formula, It refers to the number of capacitors that need to be reduced. This represents the total number of capacitors currently connected to the reactive power compensation device. It is to adjust the sensitivity coefficient. It is an exponential adjustment parameter. It is the frequency overlap risk coefficient. This is a reference threshold for the frequency overlap risk coefficient. It is a rounding function that ensures the number of capacitors is adjusted to an integer. After determining the number of capacitors that need to be reduced, the capacitive reactive power of the reactive power compensation device is adjusted according to the number of capacitors reduced, thereby changing the operating frequency of the compensation device to avoid resonance with harmonic frequencies. The formula for calculating the new operating frequency of the compensation device after adjustment is as follows: , In the formula, This refers to the adjusted operating frequency of the reactive power compensation device. It is the original operating frequency. This is the total number of capacitors connected.

[0016] The machine learning-based reactive power optimization control system for industrial kilns includes a reactive power compensation device monitoring module, a harmonic frequency acquisition and analysis module, a data preprocessing and structuring module, a key indicator extraction and risk quantification module, a machine learning evaluation module, and a dynamic adjustment module. The reactive power compensation device monitoring module monitors the adjustment process of the reactive power compensation device in real time, obtains the reactive power output data of the compensation device during operation, and thus understands its working status and its impact on the stability of the power grid voltage. The harmonic frequency acquisition and analysis module simultaneously acquires voltage and current waveform signals from the power system and obtains harmonic frequency change data through spectrum analysis, thereby identifying the existence state of harmonics and their frequency fluctuation trend over time. The data preprocessing and structuring module preprocesses the acquired raw data, classifies and organizes the preprocessed data, and establishes a structured data set. The key indicator extraction and risk quantification module extracts key indicators from the dataset that reflect the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system. It then performs a comprehensive analysis of the extracted key indicators to quantify the degree of risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency. The machine learning evaluation module inputs the analyzed key indicators into a pre-trained machine learning model, which then intelligently assesses the risk of the reactive power compensation device's operating frequency coinciding with the harmonic frequency in the power system. The dynamic adjustment module, when it identifies a risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system, dynamically reduces the number of capacitors connected based on the assessment results, thereby reducing the capacitive reactive power of the reactive power compensation device, and thus adjusting and changing the operating frequency of the compensation device to avoid resonance with the harmonic frequency.

[0017] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention, through real-time monitoring and acquisition of key data in the power system, combined with spectrum analysis and machine learning models, accurately identifies harmonic frequency changes and their matching degree with the operating frequency of reactive power compensation devices. By intelligently assessing risks and dynamically adjusting the number of capacitors connected, the operating frequency of the reactive power compensation device can be adjusted in real time to avoid harmonic resonance, ensure grid voltage stability, reduce the risk of equipment failure, overheating, or damage, and significantly improve the operating efficiency of kilns and other equipment. Ultimately, this method not only improves the stability and safety of kilns but also reduces energy waste, lowers the risk of production downtime, and saves on equipment maintenance and energy costs, thereby bringing higher economic benefits to industrial production. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a flowchart of the reactive power optimization control method for industrial kilns based on machine learning, as described in this invention.

[0020] Figure 2 This is a schematic diagram of the module of the industrial kiln reactive power optimization control system based on machine learning of the present invention. Detailed Implementation

[0021] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0022] This invention provides, for example Figure 1 The machine learning-based reactive power optimization control method for industrial kilns, as shown, includes the following steps: Real-time monitoring of the adjustment process of the reactive power compensation device, obtaining reactive power output data of the compensation device during operation, thereby understanding its working status and its impact on grid voltage stability; Reactive power compensation devices (such as SVCs) adjust reactive power output according to changes in the power system's load to ensure grid voltage stability. Real-time monitoring of this adjustment process allows for the acquisition of data on the device's operating status and changes. By monitoring reactive power output, any abnormal fluctuations can be detected promptly, providing a data foundation for subsequent analysis and forecasting. Real-time data monitoring can provide rapid feedback for detecting potential harmonic risks, ensuring that every adjustment process during the operation of the compensation device can be effectively tracked.

[0023] Simultaneously, voltage and current waveform signals in the power system are collected, and harmonic frequency change data are obtained through spectrum analysis to identify the existence state of harmonics and their frequency fluctuation trend over time. Harmonics in power systems are generated by nonlinear loads, and their frequencies are typically dynamic, fluctuating with load changes and the operating status of power equipment. Real-time monitoring of harmonic frequencies can capture any frequency variations, especially when they approach the operating frequency of reactive power compensation devices. Acquiring this data helps in further analyzing potential resonance risks to the system and provides timely information for subsequent adjustment strategies.

[0024] By analyzing the reactive power output data of reactive power compensation devices and the real-time changes in harmonic frequencies in the power system, the risk of overlapping operating frequencies can be effectively identified. Specifically, the output behavior of the compensation device reflects its current operating status and regulation frequency, while harmonic frequency data reveals the frequency distribution and variation trend of non-fundamental components in the power grid. By performing correlation analysis on these two types of data, key indicators such as the degree of proximity between them in the frequency dimension (e.g., whether the frequency difference is close to the critical value) and dynamic coupling trend (e.g., consistent frequency change direction) can be calculated. When it is identified that the regulation frequency of the compensation device is approaching or overlapping a major harmonic frequency in the power grid, especially low-order harmonics (e.g., the 3rd, 5th, and 7th harmonics), it can be determined that there is a risk of resonance triggering, requiring dynamic avoidance or control in advance. Therefore, this two-way data fusion analysis is a key foundation for realizing intelligent early warning and active suppression of harmonic resonance.

[0025] The acquired raw data is preprocessed, and the preprocessed data is classified and organized to establish a structured dataset. The acquired raw data undergoes preprocessing, further classification, and organization to establish a structured dataset. The specific steps are as follows: First, the raw data is cleaned to remove noise, outliers, and invalid data, ensuring accuracy and reliability. Second, the data undergoes time synchronization and alignment to ensure consistency across different data sources (such as harmonic frequencies, voltage, and current) under the same time reference. Next, normalization or standardization methods are applied to unify different physical quantities to the same units or numerical ranges, facilitating subsequent analysis and modeling. Subsequently, the data is classified and organized according to its characteristics, such as grouping by time window, equipment type, or operating status. Finally, the cleaned, aligned, standardized, and grouped data is transformed into a structured format (such as time series tables and feature matrices), forming a dataset that can be directly used for model invocation and analysis. Through this process, the raw data is transformed into a high-quality data foundation with consistency, readability, and operability.

[0026] Raw data often contains noise or outliers, which may be due to factors such as sensor malfunctions or instantaneous fluctuations in the power system. Preprocessing is an essential step to ensure data accuracy. Preprocessing includes outlier removal, smoothing, and signal filtering, aiming to make the harmonic frequency data more accurate and ensure that subsequent analysis and prediction can rely on high-quality data input. The preprocessed dataset will become the basis for subsequent key indicator extraction.

[0027] Key indicators reflecting the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system are extracted from the dataset. The extracted key indicators are then comprehensively analyzed to quantify the degree of risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency. Based on the relationship between reactive power compensation devices and harmonic frequencies in the power system, key indicators reflecting the risk of overlap can be extracted. These include frequency difference and harmonic amplitude variations. These key indicators help the system identify when the operating frequency of the reactive power compensation device may be close to the harmonic frequency, causing resonance risk. Extracting these indicators provides actionable data for subsequent risk analysis and offers important guidance for adjustment strategies.

[0028] Key indicators reflecting the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system are extracted from the dataset. The extracted indicators include the degree of mutual influence between the harmonic signals in the power system and the operating frequency of the reactive power compensation device, and the output stability of the reactive power compensation device. The degree of mutual influence between the harmonic signals in the power system and the operating frequency of the reactive power compensation device, and the output stability of the reactive power compensation device are comprehensively analyzed under the detection window, and reference values ​​for harmonic cross-response and reactive power compensation stability are generated respectively. The degree of risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency is quantified by the reference values ​​for harmonic cross-response and reactive power compensation stability.

[0029] The significant interaction between harmonic signals in the power system and the operating frequency of reactive power compensation devices indicates a risk of overlap between their operating frequencies. Harmonics, caused by nonlinear loads in the power system (such as motors and frequency converters), introduce specific frequency components that interact with the operating frequency of reactive power compensation devices. When the operating frequency of the compensation device approaches or overlaps with a harmonic frequency, resonance may occur, amplifying the harmonic signals in the system and increasing losses in electrical equipment. This is particularly problematic in high-power equipment like kilns, potentially leading to overheating and burnout. Furthermore, the amplification of harmonic frequencies further disrupts the voltage stability of the power system, affecting the normal operation of the reactive power compensation device and exacerbating grid voltage fluctuations. In this situation, the frequency overlap between the compensation device and the power system not only increases the risk of equipment failure but may also trigger broader power system instability issues. Therefore, the existence of overlapping frequencies signifies a high risk of resonance for both the equipment and the system, necessitating effective control measures.

[0030] The specific steps for generating harmonic cross-response reference values ​​by comprehensively analyzing the mutual influence between harmonic signals in the power system and the operating frequency of reactive power compensation devices within a detection window are as follows: This paper analyzes the frequency distribution of harmonic signals in the power system and the operating frequency of reactive power compensation devices, evaluates the relationship between harmonic frequencies and the operating frequency of the compensation devices, especially the degree of deviation between them, and calculates the frequency deviation index for each frequency component, representing the relative frequency deviation between the harmonic signal frequency and the operating frequency of the compensation devices. The calculation expression is as follows: , In the formula, It is the frequency deviation index, which indicates the degree of overlap between the operating frequency and harmonic frequency of the reactive power compensation device. It is the operating frequency of the reactive power compensation device. These are harmonic frequencies in a power system; By calculating the deviation between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system, the degree of frequency similarity between them is quantified, thereby assessing the potential resonance risk. This step helps identify the risk of the compensation device's frequency coinciding with the harmonic frequency, providing data for subsequent risk control and adjustments.

[0031] Obtaining the frequency deviation index Then, the frequency deviation is combined with the amplitude response of the harmonic signal to generate a harmonic cross-response reference value. By quantifying the amplitude influence of the harmonic signal, the interaction strength between the operating frequency of the compensation device and the harmonic frequency is further determined. The formula is as follows: , In the formula, This is the reference value for harmonic cross-response. It is the first Frequency deviation index of a harmonic signal It is the first The amplitude of the harmonic signal It is the amplitude of the capacitive reactive power output by the reactive power compensation device.

[0032] The above steps calculate the harmonic cross-response reference value by combining amplitude and frequency deviation. The amplitude ratio reflects the impact of the harmonic signal on the system, while the frequency deviation quantifies the proximity between the reactive power compensation device's frequency and the harmonic frequency. Through weighted interaction, the obtained harmonic cross-response reference value can effectively assess the risk of overlap between the reactive power compensation device's operating frequency and the harmonic frequency.

[0033] The larger the harmonic cross-response reference value generated after comprehensive analysis of the mutual influence between harmonic signals in the power system and the operating frequency of the reactive power compensation device within a detection window, the stronger the interaction between the operating frequency of the reactive power compensation device and the harmonic frequencies in the power system, indicating a higher risk of overlap. This usually means that the operating frequency of the compensation device is close to or resonates with the harmonic frequency, which may lead to the amplification of the harmonic signal in the system, causing problems such as equipment overheating, damage, or grid voltage instability. Conversely, when the reference value is small, it indicates a weaker mutual influence between the operating frequency of the reactive power compensation device and the harmonic frequency, usually meaning that the frequency difference between the two is large, there is no risk of resonance, and the equipment operates more stably.

[0034] A decrease in the output stability of a reactive power compensation device (VPC) can indeed indicate a risk of overlap between the VPC's operating frequency and the power system's harmonic frequencies. VPCs (such as SVCs) stabilize grid voltage by regulating capacitive or inductive reactive power, and their operating frequency typically depends on the power system's load demand and voltage conditions. When the VPC's operating frequency approaches or overlaps with the harmonic frequencies in the power system, resonance can occur, leading to system instability. In this case, the device's output may fluctuate significantly or fail to precisely regulate the required reactive power, resulting in decreased output stability. When the harmonic frequency is close to the VPC's operating frequency, the VPC's reactive power compensation effect is affected by the harmonics, increasing the system's response time and fluctuation amplitude, further exacerbating electrical equipment losses, and even causing equipment overheating or failure. Therefore, a decrease in the output stability of a VPC usually means that the system's harmonic frequencies and the VPC's operating frequency are approaching or overlapping, increasing the risk of resonance and affecting the normal operation of the power grid. This phenomenon is a crucial indicator of resonance risk and must be avoided by timely adjustment of the operating frequency or the use of filtering techniques.

[0035] The specific steps for generating a reference value for reactive power compensation stability by comprehensively analyzing the output stability of the reactive power compensation device under the detection window are as follows: Within the detection window, the real-time data of the output power of the reactive power compensation device and the grid voltage are continuously monitored to obtain comprehensive fluctuation response data under disturbance conditions, and a fluctuation response index is generated. The generation formula is as follows: , In the formula, It is a volatility response index. It is a reactive power compensation device in time The real-time output reactive power at any given moment represents the actual reactive power value provided by the compensation device at any given time. This is the rated maximum output power of the reactive power compensation device. It is the power grid in time The actual voltage value at that moment. This is a reference voltage value, representing the standard voltage of the power grid under stable and normal operating conditions. It is the total duration of the detection window. It refers to minute changes over an extremely short period of time; This step integrates the relative amplitude of reactive power output with the real-time voltage deviation, comprehensively reflecting the response amplitude of the compensation device under different voltage disturbance levels. A large value indicates that the reactive power compensation device fluctuates violently and responds unstablely under disturbances, which may indicate interference with harmonic frequencies and is a potential unstable signal.

[0036] Based on volatility response index By combining the offset between the operating frequency of the compensation device and the harmonic frequency of the power system, a reference value for reactive power compensation stability is constructed. The specific calculation formula is as follows: , In the formula, This is a reference value for reactive power compensation stability. It is the frequency offset, representing the reactive power compensation device at any given time. The instantaneous offset difference between the operating frequency and the harmonic frequency of the power system. It is the power system reference frequency. It is the maximum fluctuation response threshold, representing the maximum fluctuation response value that the reactive power compensation device can tolerate under historical or simulated conditions, as... The normalization standard.

[0037] This step quantifies the relationship between the frequency regulation capability of the compensation device and the system stability by coupling the frequency offset and fluctuation response. When the reactive power compensation stability reference value... A significant increase indicates that the operating frequency of the current compensation device is approaching the harmonic frequency range, and its output response is unstable, signifying a significant increase in the risk of resonance. Therefore, the reference value for reactive power compensation stability... It can serve as a key decision indicator for determining whether there is a risk of overlap between the operating frequency of a reactive power compensation device and its harmonic frequency.

[0038] A higher reactive power compensation stability reference value, generated after comprehensive analysis of the reactive power compensation device's output stability within a detection window, indicates poorer output stability and a potential risk of overlap with power system harmonic frequencies. Specifically, when the operating frequency of the compensation device approaches the harmonic frequencies of the power system, resonance occurs, causing significant fluctuations in the device's output and resulting in decreased system stability. Therefore, an increase in the stability reference value usually means that the compensation device cannot effectively compensate for the grid's reactive power, and the system's unstable response may be due to the proximity of its operating frequency to harmonic frequencies. Conversely, a lower stability reference value indicates that the compensation device's output remains relatively stable and there is no resonance with harmonic frequencies. Therefore, it can be inferred that there is no risk of overlap between the reactive power compensation device's operating frequency and the power system's harmonic frequencies.

[0039] The analyzed key indicators are input into a pre-trained machine learning model, which intelligently assesses the risk of the reactive power compensation device's operating frequency coinciding with the harmonic frequency in the power system. The analyzed harmonic cross-response reference value and reactive power compensation stability reference value are input into a pre-trained machine learning model. The machine learning model generates a frequency overlap risk coefficient, which is used to intelligently assess the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system.

[0040] A pre-trained machine learning model refers to a model trained on a specific task or problem using historical and known sample data, capable of automatically identifying patterns, predicting outcomes, or making decisions. These machine learning models are typically trained on large historical datasets, learning the complex relationship between input features (such as harmonic cross-response reference values ​​and reactive power compensation stability reference values) and target outputs (such as frequency overlap risk coefficients) through feature engineering and algorithm optimization. In industrial applications, machine learning models can help systems automatically identify and predict potential risks, failures, or optimization opportunities. In this scenario, the pre-trained machine learning model intelligently evaluates data such as harmonic cross-response and reactive power compensation stability to determine whether the operating frequency of the reactive power compensation device coincides with the harmonic frequencies in the power system, thereby predicting and quantifying the degree of risk.

[0041] The process of training a machine learning model typically involves several key steps, including data collection, data cleaning, feature selection, and model training. First, a large amount of power system data, particularly data on the operating frequency, harmonic frequencies, and equipment operating status of reactive power compensation devices, is collected and organized as a training dataset. Next, feature engineering methods are used to extract key features reflecting the relationship between the operating frequency and harmonic frequencies, such as frequency offset and harmonic response parameters. Then, a suitable machine learning algorithm (such as support vector machines, decision trees, or neural networks) is selected to train the model on this data, establishing a model that can predict frequency overlap risk coefficients based on input data. During training, the algorithm automatically learns patterns and regularities in the data and optimizes model parameters until it achieves the desired prediction accuracy on test data. Once training is complete, the model can be used for real-time monitoring and intelligent evaluation, helping to dynamically adjust the operating frequency of reactive power compensation devices, thereby effectively avoiding the resonance risk caused by frequency overlap.

[0042] The machine learning model is not limited here, but it can achieve the function of obtaining the harmonic cross-response reference value. and reactive power compensation stability reference value A comprehensive analysis is conducted to generate a frequency overlap risk coefficient. Any machine learning model is acceptable; this invention provides a specific implementation method for realizing the technical solution of this invention. Frequency overlap risk coefficient The formula for generating the formula is as follows: In the formula, and These are the reference values ​​for harmonic cross-response. and reactive power compensation stability reference value The preset proportional coefficient, and and All are greater than 0.

[0043] "Preset ratio coefficient" refers to the factor used in calculating the risk of overlapping frequencies. At that time, pre-set coefficients are used to adjust the contribution of each parameter. In the formula, and These are the two preset scaling factors, which are respectively related to the harmonic cross-response reference value. and reactive power compensation stability reference value The parameters are multiplied to determine their influence in the final frequency overlap risk calculation.

[0044] Specifically, and The reference value of the harmonic cross-response for each parameter was controlled. and reactive power compensation stability reference value (Risk of frequency overlap) The weight given in the evaluation. For example, if and A larger value indicates a higher harmonic cross-response reference value. and reactive power compensation stability reference value These parameters are of significant importance in risk calculation; conversely, smaller values ​​indicate a smaller impact on risk calculation. These proportional coefficients are pre-set in the model and are typically adjusted based on the specific needs of the system or optimization objectives.

[0045] As can be seen from the frequency overlap risk coefficient, the larger the harmonic cross-response reference value generated after comprehensively analyzing the mutual influence between the harmonic signals in the power system and the operating frequency of the reactive power compensation device under the detection window, and the larger the reactive power compensation stability reference value generated after comprehensively analyzing the output stability of the reactive power compensation device under the detection window, the larger the frequency overlap risk coefficient generated when the pre-trained machine learning model intelligently assesses the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system. This indicates that the probability of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system is greater, and vice versa.

[0046] The frequency overlap risk coefficient generated by the pre-trained machine learning model during the intelligent assessment of the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system is compared and analyzed with a pre-set reference threshold for the frequency overlap risk coefficient to determine whether there is a risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system. The judgment logic is as follows: If the frequency overlap risk coefficient is greater than the preset reference threshold for frequency overlap risk coefficient, it is determined that there is a risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system; if the frequency overlap risk coefficient is less than or equal to the preset reference threshold for frequency overlap risk coefficient, it is determined that there is no risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system.

[0047] When it is identified that the operating frequency of the reactive power compensation device may overlap with the harmonic frequency in the power system, the number of capacitors connected is dynamically reduced based on the assessment results to reduce the capacitive reactive power of the reactive power compensation device, thereby adjusting and changing the operating frequency of the compensation device to avoid resonance with the harmonic frequency. When it is identified that the operating frequency of the reactive power compensation device may overlap with the harmonic frequency in the power supply, the number of capacitors connected is dynamically reduced based on the assessment results. This reduces the capacitive reactive power of the reactive power compensation device, thereby adjusting and changing the operating frequency of the compensation device. This is to effectively avoid the resonance phenomenon between the harmonic frequency and the operating frequency of the compensation device, thus ensuring the stability of the power supply and the safety of the equipment.

[0048] The core function of this step is to reduce the reactive power generated by capacitors by adjusting the operating state of the compensation device, thus preventing resonance with harmonic frequencies. When resonance occurs, the frequency of the reactive power compensation device interacts with the harmonic frequencies in the power supply, leading to amplification of harmonic amplitudes. This exacerbates equipment losses, overheating, and in severe cases, may even cause equipment burnout, affecting the normal operation of the kiln and the entire power system. Dynamically adjusting the number of capacitors connected, by reducing capacitive reactive power output, can effectively regulate the operating frequency of the compensation device, preventing it from coinciding with harmonic frequencies and thus reducing the occurrence of resonance.

[0049] Furthermore, this dynamic adjustment process allows for flexible responses to different loads and grid conditions, ensuring that the reactive power compensation device always operates at its optimal state and avoiding voltage fluctuations caused by over-compensation or under-compensation. This method not only improves grid voltage stability and optimizes energy efficiency but also extends the service life of electrical equipment, reduces failures and maintenance costs, and ultimately guarantees the efficient and safe operation of industrial equipment such as kilns.

[0050] When it is identified that the operating frequency of the reactive power compensation device may overlap with the harmonic frequencies in the power system, the specific steps to dynamically reduce the number of capacitors connected and lower the capacitive reactive power of the reactive power compensation device based on the assessment results are as follows: When a frequency overlap risk is detected, capacitive reactive power is reduced by adjusting the number of capacitors connected, thereby avoiding resonance. The adjustment amount for the number of capacitors connected is calculated based on the difference between the frequency overlap risk coefficient and the reference threshold. The formula for calculating the adjustment amount is as follows: , In the formula, This refers to the number of capacitors that need to be reduced. This represents the total number of capacitors currently connected to the reactive power compensation device. This is the sensitivity adjustment coefficient, representing the adjustment range, and is typically set between 0.2 and 0.5. This is an exponential adjustment parameter that controls the sensitivity of the adjustment response; its value is typically between 1 and 5. It is the frequency overlap risk coefficient. This is a reference threshold for the frequency overlap risk coefficient. It is a rounding function that ensures the number of capacitors is adjusted to an integer. This step dynamically adjusts the parameters based on the difference between the frequency overlap risk coefficient and the reference threshold using an exponential relationship. At that time, the adjustment range will be increased to ensure that risks are controlled in a timely manner.

[0051] After determining the number of capacitors that need to be reduced, the capacitive reactive power of the reactive power compensation device is adjusted according to the number of capacitors reduced, thereby changing the operating frequency of the compensation device to avoid resonance with harmonic frequencies. The formula for calculating the new operating frequency of the compensation device after adjustment is as follows: , In the formula, This refers to the adjusted operating frequency of the reactive power compensation device. It is the original operating frequency. It represents the total number of capacitors connected, indicating the initial configuration of the device.

[0052] This step separates the operating frequency of the compensation device from the harmonic frequencies of the power grid by analyzing the relationship between the number of capacitors and the frequency, ensuring that the operating frequency stays away from the harmonic frequencies and thus avoiding resonance. By adjusting the number of capacitors connected, the operating frequency of the compensation device can be precisely controlled, ensuring the safe and stable operation of the system.

[0053] This invention, through real-time monitoring and acquisition of key data in the power system, combined with spectrum analysis and machine learning models, accurately identifies harmonic frequency changes and their matching degree with the operating frequency of reactive power compensation devices. By intelligently assessing risks and dynamically adjusting the number of capacitors connected, the operating frequency of the reactive power compensation device can be adjusted in real time to avoid harmonic resonance, ensure grid voltage stability, reduce the risk of equipment failure, overheating, or damage, and significantly improve the operating efficiency of kilns and other equipment. Ultimately, this method not only improves the stability and safety of kilns but also reduces energy waste, lowers the risk of production downtime, and saves on equipment maintenance and energy costs, thereby bringing higher economic benefits to industrial production.

[0054] This invention provides, for example Figure 2 The machine learning-based reactive power optimization control system for industrial kilns shown includes a reactive power compensation device monitoring module, a harmonic frequency acquisition and analysis module, a data preprocessing and structuring module, a key indicator extraction and risk quantification module, a machine learning evaluation module, and a dynamic adjustment module. The reactive power compensation device monitoring module monitors the adjustment process of the reactive power compensation device in real time, obtains the reactive power output data of the compensation device during operation, and thus understands its working status and its impact on the stability of the power grid voltage. The harmonic frequency acquisition and analysis module simultaneously acquires voltage and current waveform signals from the power system and obtains harmonic frequency change data through spectrum analysis, thereby identifying the existence state of harmonics and their frequency fluctuation trend over time. The data preprocessing and structuring module preprocesses the acquired raw data, classifies and organizes the preprocessed data, and establishes a structured data set. The key indicator extraction and risk quantification module extracts key indicators from the dataset that reflect the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system. It then performs a comprehensive analysis of the extracted key indicators to quantify the degree of risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency. The machine learning evaluation module inputs the analyzed key indicators into a pre-trained machine learning model, which then intelligently assesses the risk of the reactive power compensation device's operating frequency coinciding with the harmonic frequency in the power system. The dynamic adjustment module, when it identifies a risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system, dynamically reduces the number of capacitors connected based on the assessment results, thereby reducing the capacitive reactive power of the reactive power compensation device, and thus adjusting and changing the operating frequency of the compensation device to avoid resonance with the harmonic frequency.

[0055] The reactive power optimization control method for industrial kilns based on machine learning provided in this invention is implemented through the aforementioned reactive power optimization control system for industrial kilns based on machine learning. For details of the specific methods and processes of the reactive power optimization control system for industrial kilns based on machine learning, please refer to the embodiments of the aforementioned reactive power optimization control method for industrial kilns based on machine learning, which will not be repeated here.

[0056] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0057] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0058] It should be noted that, in this document, the use of relational terms such as "first" and "second" is merely for distinguishing one entity or operation from another, and does not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0059] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0060] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0061] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0062] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0063] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0064] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0065] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A machine learning-based method for optimizing reactive power control in industrial kilns, characterized in that, Includes the following steps: Real-time monitoring of the adjustment process of the reactive power compensation device, obtaining reactive power output data of the compensation device during operation, thereby understanding its working status and its impact on grid voltage stability; Simultaneously, voltage and current waveform signals in the power system are collected, and harmonic frequency change data are obtained through spectrum analysis to identify the existence state of harmonics and their frequency fluctuation trend over time. The acquired raw data is preprocessed, and the preprocessed data is classified and organized to establish a structured dataset. Key indicators reflecting the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system are extracted from the dataset. The extracted key indicators are then comprehensively analyzed to quantify the degree of risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency. The analyzed key indicators are input into a pre-trained machine learning model, which intelligently assesses the risk of the reactive power compensation device's operating frequency coinciding with the harmonic frequency in the power system. When it is identified that the operating frequency of the reactive power compensation device may overlap with the harmonic frequency in the power system, the number of capacitors connected is dynamically reduced based on the assessment results to reduce the capacitive reactive power of the reactive power compensation device, thereby adjusting and changing the operating frequency of the compensation device to avoid resonance with the harmonic frequency.

2. The reactive power optimization control method for industrial kilns based on machine learning according to claim 1, characterized in that, The acquired raw data is preprocessed and further categorized and organized to establish a structured dataset. The specific steps are as follows: First, the raw data is cleaned to remove noise, outliers, and invalid data, ensuring the accuracy and reliability of the data. Secondly, the data is synchronized and aligned in time to ensure consistency between different data sources under the same time base. Next, normalization or standardization methods are applied to unify different physical quantities to the same dimension or numerical range, which facilitates subsequent analysis and modeling. Subsequently, the data was categorized and organized according to its characteristics; Finally, the cleaned, aligned, standardized, and grouped data is transformed into a structured format, forming a dataset that can be directly used for model calls and analysis.

3. The method for optimizing and controlling the reactive power of industrial kilns based on machine learning according to claim 1, characterized in that, Key indicators reflecting the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system are extracted from the dataset. The extracted indicators include the degree of mutual influence between the harmonic signals in the power system and the operating frequency of the reactive power compensation device, and the output stability of the reactive power compensation device. The degree of mutual influence between the harmonic signals in the power system and the operating frequency of the reactive power compensation device, and the output stability of the reactive power compensation device are comprehensively analyzed under the detection window, and reference values ​​for harmonic cross-response and reactive power compensation stability are generated respectively. The degree of risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency is quantified by the reference values ​​for harmonic cross-response and reactive power compensation stability.

4. The machine learning-based reactive power optimization control method for industrial kilns according to claim 3, characterized in that, The specific steps for generating harmonic cross-response reference values ​​by comprehensively analyzing the mutual influence between harmonic signals in the power system and the operating frequency of reactive power compensation devices within a detection window are as follows: This paper analyzes the frequency distribution of harmonic signals in the power system and the operating frequency of the reactive power compensation device, evaluates the relationship between the harmonic frequency and the operating frequency of the compensation device, and calculates the frequency deviation index for each frequency component. The calculation expression is as follows: ， In the formula, It is the frequency deviation index. It is the operating frequency of the reactive power compensation device. These are harmonic frequencies in a power system; Obtaining the frequency deviation index Then, the frequency deviation is combined with the amplitude response of the harmonic signal to generate a harmonic cross-response reference value. By quantifying the amplitude influence of the harmonic signal, the interaction strength between the operating frequency of the compensation device and the harmonic frequency is further determined. The formula is as follows: ， In the formula, This is the reference value for harmonic cross-response. It is the first Frequency deviation index of a harmonic signal It is the first The amplitude of the harmonic signal It is the amplitude of the capacitive reactive power output by the reactive power compensation device.

5. The machine learning-based reactive power optimization control method for industrial kilns according to claim 3, characterized in that, The specific steps for generating a reference value for reactive power compensation stability by comprehensively analyzing the output stability of the reactive power compensation device under the detection window are as follows: Within the detection window, the real-time data of the output power of the reactive power compensation device and the grid voltage are continuously monitored to obtain comprehensive fluctuation response data under disturbance conditions, and a fluctuation response index is generated. The generation formula is as follows: ， In the formula, It is a volatility response index. It is a reactive power compensation device in time Real-time output reactive power at all times. This is the rated maximum output power of the reactive power compensation device. It is the power grid in time The actual voltage value at that moment. This is the reference voltage value. It is the total duration of the detection window. It refers to minute changes over an extremely short period of time; Based on volatility response index By combining the offset between the operating frequency of the compensation device and the harmonic frequency of the power system, a reference value for reactive power compensation stability is constructed. The specific calculation formula is as follows: ， In the formula, This is a reference value for reactive power compensation stability. It is the frequency offset. It is the power system reference frequency. It is the maximum fluctuation response threshold.

6. The reactive power optimization control method for industrial kilns based on machine learning according to claim 3, characterized in that, The analyzed harmonic cross-response reference value and reactive power compensation stability reference value are input into a pre-trained machine learning model. The machine learning model generates a frequency overlap risk coefficient, which is used to intelligently assess the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system.

7. The machine learning-based reactive power optimization control method for industrial kilns according to claim 6, characterized in that, The frequency overlap risk coefficient generated by the pre-trained machine learning model during the intelligent assessment of the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system is compared and analyzed with a pre-set reference threshold for the frequency overlap risk coefficient to determine whether there is a risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system. The judgment logic is as follows: If the frequency overlap risk coefficient is greater than the preset reference threshold for frequency overlap risk coefficient, it is determined that there is a risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system; if the frequency overlap risk coefficient is less than or equal to the preset reference threshold for frequency overlap risk coefficient, it is determined that there is no risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system.

8. The machine learning-based reactive power optimization control method for industrial kilns according to claim 7, characterized in that, When it is identified that the operating frequency of the reactive power compensation device may overlap with the harmonic frequencies in the power system, the specific steps to dynamically reduce the number of capacitors connected and lower the capacitive reactive power of the reactive power compensation device based on the assessment results are as follows: When a frequency overlap risk is detected, capacitive reactive power is reduced by adjusting the number of capacitors connected, thereby avoiding resonance. The adjustment amount for the number of capacitors connected is calculated based on the difference between the frequency overlap risk coefficient and the reference threshold. The formula for calculating the adjustment amount is as follows: ， In the formula, It refers to the number of capacitors that need to be reduced. This represents the total number of capacitors currently connected to the reactive power compensation device. It is to adjust the sensitivity coefficient. It is an exponential adjustment parameter. It is the frequency overlap risk coefficient. This is a reference threshold for the frequency overlap risk coefficient. It is a rounding function that ensures the number of capacitors is adjusted to an integer. After determining the number of capacitors that need to be reduced, the capacitive reactive power of the reactive power compensation device is adjusted according to the number of capacitors reduced, thereby changing the operating frequency of the compensation device to avoid resonance with harmonic frequencies. The formula for calculating the new operating frequency of the compensation device after adjustment is as follows: ， In the formula, This refers to the adjusted operating frequency of the reactive power compensation device. It is the original operating frequency. This is the total number of capacitors connected.

9. A machine learning-based reactive power optimization control system for industrial kilns, used to implement the machine learning-based reactive power optimization control method for industrial kilns as described in any one of claims 1-8, characterized in that, It includes a reactive power compensation device monitoring module, a harmonic frequency acquisition and analysis module, a data preprocessing and structuring module, a key indicator extraction and risk quantification module, a machine learning evaluation module, and a dynamic adjustment module. The reactive power compensation device monitoring module monitors the adjustment process of the reactive power compensation device in real time, obtains the reactive power output data of the compensation device during operation, and thus understands its working status and its impact on the stability of the power grid voltage. The harmonic frequency acquisition and analysis module simultaneously acquires voltage and current waveform signals from the power system and obtains harmonic frequency change data through spectrum analysis, thereby identifying the existence state of harmonics and their frequency fluctuation trend over time. The data preprocessing and structuring module preprocesses the acquired raw data, classifies and organizes the preprocessed data, and establishes a structured data set. The key indicator extraction and risk quantification module extracts key indicators from the dataset that reflect the risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency of the power system. It then performs a comprehensive analysis of the extracted key indicators to quantify the degree of risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency. The machine learning evaluation module inputs the analyzed key indicators into a pre-trained machine learning model, which then intelligently assesses the risk of the reactive power compensation device's operating frequency coinciding with the harmonic frequency in the power system. The dynamic adjustment module, when it identifies a risk of overlap between the operating frequency of the reactive power compensation device and the harmonic frequency in the power system, dynamically reduces the number of capacitors connected based on the assessment results, thereby reducing the capacitive reactive power of the reactive power compensation device, and thus adjusting and changing the operating frequency of the compensation device to avoid resonance with the harmonic frequency.

Images