A dynamic adjusting method of high-voltage power supply of electrostatic precipitator based on fuzzy control

Abstract

This invention discloses a dynamic adjustment method for high-voltage power supplies in electrostatic precipitators based on fuzzy control. The method includes: extracting impedance changes and dust concentration based on real-time operating parameters; determining tripping risks and adjusting the output voltage through fuzzy inference; optimizing pulse parameters using a fuzzy control algorithm based on dust distribution characteristics to generate waveforms matching particle motion; collecting environmental parameters to construct an operating condition model; switching control modes after scenario identification; predicting load change trends and calculating millisecond-level response strategies; integrating dust removal efficiency and energy consumption targets; dynamically allocating weights; calculating a comprehensive control strategy and verifying the balance state; iteratively adjusting weight coefficients to achieve dynamic adjustment of the high-voltage power supply. This invention significantly improves the stability, dust removal efficiency, and energy consumption balance of the high-voltage power supply, achieving intelligent control with multi-objective optimization.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of high-voltage power supply dynamic regulation, and particularly relates to a high-voltage power supply dynamic regulation method for electrostatic precipitators based on fuzzy control. BACKGROUND

[0002] As the core equipment for controlling dust emission in the industrial field, the performance of the high-voltage power supply of the electrostatic precipitator directly determines the dust removal efficiency and the stability of the equipment operation. With the continuous improvement of environmental protection requirements, the electrostatic precipitator needs to maintain efficient operation under complex and variable working conditions, while taking into account energy consumption and equipment safety. This makes the dynamic regulation of the high-voltage power supply a key research field. The traditional high-voltage power supply control technology is inadequate in dealing with variable environments and is difficult to meet the needs of modern industry for efficiency, flexibility and intelligence. Therefore, developing a dynamic regulation method that can adapt to complex working conditions is not only crucial for environmental protection, but also has far-reaching significance for the sustainability of industrial production.

[0003] Most existing high-voltage power supply control methods rely on fixed parameter settings or simple feedback mechanisms. These methods often show inadequate adaptability when faced with rapid changes in environmental factors such as dust concentration, humidity, and airflow. For example, traditional control systems usually operate according to pre-set voltage and current ranges. When the working conditions change dramatically, such as a sudden increase in dust concentration or the occurrence of turbulent airflow, the system is difficult to quickly adjust the output characteristics. This can lead to a mismatch in electric field strength, affecting dust capture efficiency, and even causing equipment tripping, increasing maintenance costs. More importantly, these methods lack real-time sensing ability for the operating environment and cannot dynamically optimize power supply parameters according to specific working conditions, limiting the overall performance of the precipitator.

[0004] In this field, one of the core technical difficulties is how to achieve rapid response of the high-voltage power supply to complex working conditions. The working environment of the electrostatic precipitator is complex and variable, and factors such as particle size distribution of dust particles, airflow speed, and humidity will affect the working state of the electric field. Rapid response means that the power supply needs to perceive these changes and adjust the output parameters within milliseconds, but traditional control systems often have slow calculation speed or parameter adjustment lag, causing the electric field strength to deviate from the optimal working interval. For example, in high-load working conditions, a sudden increase in dust concentration can cause impedance to drop sharply. If the power supply cannot lower the output voltage in time, it may cause arc discharge and even cause the equipment to trip.

[0005] Another key technical difficulty is the flexibility of the power output waveform. Traditional high-voltage power supplies usually output regular square waves or sine waves. Such waveforms are difficult to form a matching electric field distribution with the dust movement trajectory when facing uneven dust distribution or turbulent airflow. For example, when there is turbulence in the airflow, the movement path of the dust particles becomes disordered, and regular waveforms cannot effectively capture dust particles of different particle sizes, resulting in a decrease in capture efficiency. Flexible waveform adjustment requires the power supply to be able to adjust the pulse width, frequency and amplitude according to the dynamic characteristics of the dust, but existing technologies are often limited in achieving such dynamic matching due to the lack of precise environmental perception and complex calculation models.

[0006] Therefore, how to achieve millisecond-level dynamic response of the high-voltage power supply and generate flexible waveform output matching the dust movement characteristics through precise environmental perception and fast control algorithm under complex and variable working conditions has become a key problem for efficient operation of electrostatic precipitators. SUMMARY

[0007] To solve the above technical problems, the present application provides a dynamic adjustment method for a high-voltage power supply of an electrostatic precipitator based on fuzzy control, comprising:

[0008] Obtaining real-time operating parameters of the high-voltage power supply of the electrostatic precipitator, extracting impedance changes and dust concentration as input variables based on the real-time operating parameters, determining the trip risk level through a fuzzy inference mechanism, and obtaining a risk prediction result;

[0009] Determining the adjustment requirement of the output characteristics of the high-voltage power supply according to the risk prediction result, and reducing the output voltage amplitude if the risk level is higher than a preset threshold to obtain a preliminary adjustment scheme;

[0010] Obtaining the non-uniformity characteristics of dust distribution from the preliminary adjustment scheme, adjusting the pulse width and frequency using a fuzzy control algorithm, generating an irregular waveform output matching the particle movement trajectory, and obtaining optimized waveform parameters;

[0011] Collecting environmental parameter data through a sensor network, constructing a working condition scenario model in combination with the optimized waveform parameters and the environmental parameter data, determining the current scenario type using a scenario recognition algorithm, and obtaining a scenario switching instruction;

[0012] Selecting the corresponding fuzzy rule set and membership function according to the scenario switching instruction, driving the high-voltage power supply to switch the control mode and incorporating the load change rate as an additional input to obtain a mode adjustment output;

[0013] Extracting the voltage deviation and current fluctuation index from the mode adjustment output, predicting the load change trend using a fuzzy relationship model, calculating a millisecond-level response strategy if the load change trend shows rapid fluctuations, and obtaining response optimization parameters;

[0014] By integrating dust removal efficiency and energy consumption targets with response optimization parameters, assigning dynamic weight coefficients and calculating a comprehensive control strategy, the final adjustment command is obtained.

[0015] Verify the multi-objective balance state according to the final adjustment command. If the balance state does not meet the target, iteratively adjust the weight coefficients to obtain the high-voltage power supply control parameters. Perform dynamic adjustment based on the high-voltage power supply control parameters to obtain the dynamic adjustment result.

[0016] Compared with the prior art, the present invention has the following advantages and technical effects:

[0017] This invention obtains the real-time operating parameters of the high-voltage power supply of the electrostatic precipitator, extracts impedance changes and dust concentration as input variables, and uses a fuzzy reasoning mechanism to determine the tripping risk level. This allows for early prediction of tripping risks, effectively preventing equipment damage and downtime caused by high-voltage power supply anomalies, and improving the stability and reliability of equipment operation.

[0018] Based on risk prediction results, this invention reduces the output voltage amplitude in a timely manner when the risk level is higher than a preset threshold, thereby obtaining a preliminary adjustment scheme. It responds quickly to changes in risk, initially reduces system risk, provides a basis for subsequent optimization and adjustment, and reduces potential losses caused by excessive risk.

[0019] This invention obtains the non-uniformity characteristics of dust distribution from the preliminary adjustment scheme, uses a fuzzy control algorithm to adjust the pulse width and frequency, generates an irregular waveform output to match the particle motion trajectory, obtains optimized waveform parameters, improves dust removal efficiency, enhances the adaptability of the electrostatic precipitator to different dust characteristics, and improves the dust removal effect.

[0020] This invention collects environmental parameter data through a sensor network, constructs a working condition scenario model by combining optimized waveform parameters, uses a scenario recognition algorithm to determine the current scenario type, obtains a scenario switching command, selects the corresponding fuzzy rule set and membership function according to the scenario switching command, and drives the high-voltage power supply to switch control modes. This realizes intelligent identification and adaptive adjustment of different working conditions, and improves the system's operating performance under complex working conditions.

[0021] This invention extracts voltage deviation and current fluctuation indicators from the mode regulation output, uses a fuzzy relation model to predict the load change trend, and calculates a millisecond-level response strategy if the load change trend shows rapid fluctuations to obtain response optimization parameters. This enables the system to respond quickly to load changes, maintain stable system operation, and improve the system's dynamic response capability and anti-interference capability.

[0022] This invention integrates dust removal efficiency and energy consumption targets by responding to optimized parameters, assigns dynamic weight coefficients, calculates a comprehensive control strategy, obtains the final adjustment command, verifies the multi-objective balance state based on the final adjustment command, and iteratively adjusts the weight coefficients if the balance state is not met to obtain the high-voltage power supply control parameters. This achieves a balance between dust removal efficiency and energy consumption, improves the overall performance of the system, and reduces operating costs.

[0023] This invention dynamically adjusts the high-voltage power supply control parameters to obtain dynamic adjustment results. Through continuous dynamic adjustment and optimization, it ensures that the high-voltage power supply of the electrostatic precipitator is always in the best operating state under different working conditions, thereby improving the long-term operating efficiency and stability of the system and extending the service life of the equipment.

[0024] This invention significantly improves the stability, dust removal efficiency, and energy consumption balance of high-voltage power supplies, and achieves intelligent control with multi-objective optimization. Attached Figure Description

[0025] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0026] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention. Detailed Implementation

[0027] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0028] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0029] like Figure 1 As shown, this embodiment provides a method for dynamic adjustment of high-voltage power supply for electrostatic precipitators based on fuzzy control, including:

[0030] The real-time operating parameters of the high-voltage power supply of the electrostatic precipitator are obtained. Based on the real-time operating parameters, impedance change and dust concentration are extracted as input variables. The tripping risk level is determined through a fuzzy inference mechanism to obtain the risk prediction result.

[0031] Based on the risk prediction results, the adjustment requirements for the output characteristics of the high-voltage power supply are determined. If the risk level is higher than the preset threshold, the output voltage amplitude is reduced to obtain a preliminary adjustment scheme.

[0032] The non-uniformity of dust distribution is obtained from the preliminary adjustment scheme. Fuzzy control algorithm is used to adjust the pulse width and frequency, generate irregular waveform output to match the particle motion trajectory, and obtain optimized waveform parameters.

[0033] Environmental parameter data is collected through a sensor network. A working condition scenario model is constructed by combining the optimized waveform parameters with the environmental parameter data. A scenario recognition algorithm is used to determine the current scenario type and obtain a scenario switching command.

[0034] Based on the scenario switching command, the corresponding fuzzy rule set and membership function are selected to drive the high-voltage power supply to switch control modes and incorporate the load change rate as an additional input to obtain the mode adjustment output.

[0035] Voltage deviation and current fluctuation indicators are extracted from the mode regulation output. A fuzzy relation model is used to predict the load change trend. If the load change trend shows rapid fluctuation, a millisecond-level response strategy is calculated to obtain the response optimization parameters.

[0036] By integrating dust removal efficiency and energy consumption targets with response optimization parameters, assigning dynamic weight coefficients and calculating a comprehensive control strategy, the final adjustment command is obtained.

[0037] Verify the multi-objective balance state according to the final adjustment command. If the balance state does not meet the target, iteratively adjust the weight coefficients to obtain the high-voltage power supply control parameters. Perform dynamic adjustment based on the high-voltage power supply control parameters to obtain the dynamic adjustment result.

[0038] Furthermore, the process of determining the tripping risk level and obtaining risk prediction results through fuzzy reasoning mechanisms includes:

[0039] Based on the voltage, current and power data obtained from the high voltage power supply of the electrostatic precipitator, the impedance change is calculated and the impedance change value is obtained.

[0040] Based on the impedance change value and the dust concentration data collected by the dust concentration sensor, a set of input variables is constructed to obtain standardized input variables;

[0041] A fuzzy inference mechanism is used to fuzzify the standardized input variables, generate fuzzy membership functions, and obtain a fuzzy input set.

[0042] By using a pre-established fuzzy rule base, reasoning is performed on the fuzzy input set to determine the tripping risk level and obtain the risk level assessment value;

[0043] If the risk level assessment value exceeds the preset threshold, the risk level is classified using the support vector machine algorithm to obtain the classified risk prediction result.

[0044] Based on the risk prediction results after classification, time series analysis is used to predict the future tripping risk trend and obtain the risk trend prediction value.

[0045] By comparing the predicted risk trends with historical data and adjusting the parameters of the fuzzy rule base, an optimized fuzzy inference model is obtained.

[0046] In one possible implementation, the voltage, current, and power data of the high-voltage power supply of the electrostatic precipitator in this embodiment are acquired in real time by sensors. For example, the high-voltage power supply outputs a voltage of 50kV, a current of 100mA, and a power of 5kW. These data are used to calculate the impedance change, which is obtained by dividing the voltage by the current. Assuming the initial impedance is 500kΩ, if the current increases to 120mA, the impedance drops to 416.67kΩ, and the impedance change rate is 16.67%. This change reflects the dynamic adjustment of the electric field characteristics within the precipitator, which helps to determine the operational stability.

[0047] Specifically, this embodiment constructs a set of input variables by combining impedance change values ​​with dust concentration sensor data. Assume the dust concentration is 200 mg / m³. 3 Combined with the impedance change rate of 16.67%, standardized input variables were formed. Standardization normalizes the data to the 0-1 range; for example, the impedance change rate is mapped to 0.167, and dust concentration to 0.2. This standardization ensures consistent processing of data with different dimensions, improving model consistency.

[0048] In one embodiment, the fuzzy inference mechanism fuzzifies the standardized input variables to generate a membership function.

[0049] For example, the impedance change rate of 0.167 is divided into three fuzzy sets: "low," "medium," and "high," with membership degrees of 0.8, 0.2, and 0, respectively. The dust concentration of 0.2 is processed similarly to generate a fuzzy input set. The fuzzy rule base, based on expert experience (e.g., "if the impedance change rate is high and the dust concentration is high, then the tripping risk is high"), infers from the input set to derive a risk level assessment value, assumed to be 0.75 (range 0-1). This fuzzification effectively addresses data uncertainty and improves the robustness of risk assessment.

[0050] For example, if the risk assessment value of 0.75 exceeds the preset threshold of 0.6, a support vector machine (SVM) is used for risk classification. The SVM is trained based on historical data and categorizes risks into three classes: "low," "medium," and "high." The classification result is assumed to be "high risk." This classification method, by segmenting the data using a hyperplane, can efficiently distinguish complex risk patterns and improve prediction accuracy.

[0051] In one possible implementation, time series analysis is used to predict future tripping risk trends based on classification results. For example, using an ARIMA model to analyze risk data from the past 24 hours, a risk trend value of 0.8 is predicted for the next 12 hours, indicating a persistently high risk. Time series analysis captures the temporal patterns of risk, helping to provide early warnings of equipment failures.

[0052] Specifically, the fuzzy rule base is optimized by comparing the predicted risk trends with historical data.

[0053] For example, historical data shows that high dust concentrations are often accompanied by high impedance changes. Adjusting the rule weights makes the model pay more attention to the impact of dust concentration. The optimized fuzzy inference model can more accurately predict tripping risks, reduce false alarm rates, improve the operational stability of the dust collector, and extend equipment life.

[0054] It is understood that the integrated implementation of the above methods, from data acquisition to model optimization, forms a closed-loop feedback mechanism. Each link supports the others, ensuring the accuracy and practicality of risk prediction and providing technical support for the efficient operation of electrostatic precipitators.

[0055] Furthermore, the process of obtaining a preliminary adjustment plan includes:

[0056] Based on the risk prediction results, the current output voltage amplitude and operating status parameters are extracted from the high-voltage power supply system operation data to obtain the risk level data.

[0057] Using threshold comparison logic, if the risk level data is higher than the preset threshold, the voltage regulation mechanism is triggered to determine the output voltage amplitude that needs to be reduced.

[0058] Based on the determined output voltage amplitude that needs to be reduced, the target output characteristics of the high-voltage power supply system are calculated, and a preliminary adjustment scheme is generated.

[0059] By simulating the operating environment, the impact of the preliminary adjustment scheme on the high-voltage power supply system is verified, and the operating status of the adjusted power supply is obtained.

[0060] If the adjusted power supply operating state meets the preset safety range, the preliminary adjustment scheme will be converted into an executable control command to generate the final adjustment scheme.

[0061] Based on the final adjustment plan, update the control parameters of the high-voltage power supply system, obtain the updated operating data, and determine the system stability.

[0062] By analyzing the degree of matching between the updated operational data and the risk prediction results, the threshold comparison logic is optimized, resulting in an improved adjustment mechanism.

[0063] For example, when extracting the current output voltage amplitude and operating status parameters from the operating data of a high-voltage power supply system, specific data is obtained through real-time monitoring equipment. Assume that the high-voltage power supply system of an electrostatic precipitator is operating with an output voltage amplitude of 50kV, and operating status parameters including a current intensity of 200mA and a temperature of 45℃. This data is collected in real time by sensors to generate risk level data. The generation of risk level data is based on historical data analysis, calculating a risk index through voltage and current fluctuations.

[0064] For example, voltage fluctuations exceeding ±5% or abnormal current increases result in a risk level assessment of "medium." This data extraction method ensures the accuracy of subsequent risk assessments.

[0065] In one possible implementation, threshold comparison logic is used to determine whether the risk level exceeds a preset threshold. For example, if the preset threshold risk index is 0.6 and the currently calculated risk index is 0.75, exceeding the threshold, a voltage regulation mechanism is triggered. This mechanism can then reduce the output voltage to 48kV to alleviate system stress. The threshold comparison logic achieves fine-grained management by setting multiple threshold levels (such as 0.4, 0.6, and 0.8), facilitating rapid response to potential risks.

[0066] Specifically, when calculating the target output characteristics of a high-voltage power supply system, power and impedance are recalculated based on the reduced voltage amplitude. For example, after the voltage drops from 50kV to 48kV, the power can be reduced from 10kW to 9.6kW, generating a preliminary regulation scheme. This scheme considers the system's load capacity and stability to ensure that the system can still operate efficiently after regulation. The generation of the preliminary regulation scheme relies on a mathematical model of the system's operating characteristics, combined with optimized output based on actual operating conditions.

[0067] For example, when verifying the initial adjustment scheme, the effect of the adjustment is tested by simulating the operating environment. In the simulated environment, a 48kV voltage and a 200mA current are input, and the system is observed to see if overheating or tripping occurs. If the simulation results show that the temperature drops to 40℃ without abnormal fluctuations, the adjusted operating state is considered to meet the safety range. This verification method ensures the feasibility of the scheme by simulating real-world operating conditions.

[0068] In one possible implementation, specific control signals are generated when the initial adjustment scheme is transformed into control instructions to be executed.

[0069] For example, if the regulation scheme requires a voltage reduction of 2kV, the control command will be sent through the PLC system to the transformer module of the high-voltage power supply to adjust the output voltage. This conversion process ensures accurate execution of the command and avoids errors caused by human intervention.

[0070] Specifically, after updating the control parameters of the high-voltage power supply system, the updated operating data is obtained, such as the voltage stabilizing at 48kV and the current dropping to 190mA. When assessing system stability, the voltage and current fluctuation curves are analyzed to confirm whether the system remains within the safe range of ±3%. Stable operating data indicates that system risks are effectively controlled.

[0071] For example, when optimizing the threshold comparison logic, this embodiment analyzes the degree of matching between the updated operating data and the risk prediction results. Suppose the prediction result shows the risk index should be 0.5, but the actual value is 0.55, then the threshold is adjusted (e.g., reduced from 0.6 to 0.58) to improve the logic's sensitivity. This optimization continuously improves the adjustment mechanism through data feedback, enhancing the system's ability to cope with risks.

[0072] In one possible implementation, the logical progression from the core solution to the extended solution can further enrich the regulation mechanism.

[0073] For example, the core solution relies solely on voltage and current adjustments, while the extended solution incorporates dust concentration data as an auxiliary variable. If an increase in dust concentration leads to a change in impedance, the threshold is dynamically adjusted, generating a more flexible control scheme. This multi-variable collaborative approach enhances the system's adaptability, ensuring stable operation even under complex conditions.

[0074] Furthermore, the process of obtaining optimized waveform parameters includes:

[0075] Non-uniformity characteristics are obtained from dust distribution data, and statistical analysis methods are used to calculate the variance and skewness of the distribution to obtain the feature vector.

[0076] Based on the feature vector, a fuzzy control algorithm is used to establish the input fuzzy set, and a membership degree mapping is performed to determine the fuzzy rule set for the non-uniformity of dust distribution.

[0077] If the membership degree of the fuzzy rule set is greater than the preset threshold, the pulse width and frequency are adjusted to generate an initial irregular waveform and obtain a waveform parameter set.

[0078] By analyzing the periodicity and offset of the particle motion trajectory collected in real time, the trajectory feature vector is obtained.

[0079] Based on the matching degree between the trajectory feature vector and the waveform parameter set, the particle swarm optimization algorithm is used to iteratively adjust the waveform parameters and determine the optimized waveform parameters.

[0080] If the matching degree between the optimized waveform parameters and the particle motion trajectory is lower than the preset threshold, the fuzzy rule set is updated through the feedback mechanism to obtain the updated waveform parameter set.

[0081] Based on the updated waveform parameter set, the final irregular waveform is generated, and a control signal matching the particle motion trajectory is output.

[0082] For example, in the operating environment of a high-voltage power supply system, the non-uniformity of dust distribution needs to be extracted through statistical analysis. Dust distribution data typically originates from particle concentration information collected in real time by sensors, reflecting the density variations of dust particles in space. This embodiment quantifies non-uniformity by calculating variance and skewness. For example, variance measures the degree of fluctuation in dust concentration across different areas, while skewness reflects whether the distribution is symmetrical. Assume the dust concentration data for a certain area is 10 mg / m³. 3 15mg / m 3 and 30mg / m 3 Variance calculations reveal significant concentration fluctuations, while skewness indicates a bias towards high-concentration regions, generating a feature vector containing both statistics. This feature vector provides a quantitative basis for subsequent control.

[0083] Specifically, when constructing a fuzzy control algorithm based on feature vectors, variance and skewness are used as input parameters to establish a fuzzy set. The fuzzy set maps non-uniform features to fuzzy states such as "low," "medium," and "high" through membership functions.

[0084] For example, variance values ​​in the range of 0-5 are considered "low," 5-20 are "medium," and values ​​greater than 20 are "high." The fuzzy rule set is defined as follows: if the variance is "high" and the skewness is "positively high," then the pulse width needs to be significantly adjusted. Assuming a scenario with a variance of 25 and a skewness of 0.8, after membership calculation, a rule is triggered, deciding to reduce the pulse width from 50μs to 30μs and increase the frequency from 100Hz to 150Hz, generating an initial irregular waveform parameter set. This waveform adjustment can better adapt to the dynamic changes in dust distribution.

[0085] In one embodiment, real-time acquisition of particle trajectories is achieved through laser scattering or image processing techniques. The periodicity and offset of the trajectory reflect the motion law of the particles under the influence of an electric field.

[0086] For example, the acquired trajectory display period is 0.2s, and the offset is 2mm, generating a trajectory feature vector. By matching this vector with a waveform parameter set, it is determined whether the waveform effectively drives particle motion. Assuming a matching degree of 85%, which is lower than the preset threshold of 90%, a particle swarm optimization algorithm is initiated, iteratively adjusting the pulse width to 35μs and the frequency to 140Hz. This optimization process gradually approximates the optimal waveform parameters by simulating the particle motion trajectory.

[0087] For example, in this embodiment, when updating the fuzzy rule set using a feedback mechanism, the membership function boundary is adjusted based on the deviation between trajectory features and waveform parameters. Assuming the initial "high" variance threshold is 20, and the matching degree is found to be low, the threshold is adjusted to 18, generating an updated waveform parameter set. Finally, the irregular waveform is output through a control signal, such as a pulse width of 32μs and a frequency of 145Hz, which can more accurately match the particle motion trajectory. The advantage of this method is that it dynamically adapts to changes in dust distribution and improves the matching accuracy of the control signal.

[0088] Understandably, the generated control signals need to be coordinated with the operating parameters of the high-voltage power supply system.

[0089] For example, the adjusted waveform must ensure a stable power supply output voltage to avoid system overload due to pulse variations. Real-time monitoring of operational data verifies whether the waveform parameters meet safe ranges. This method, through multi-level analysis and optimization, ensures the accuracy and stability of dust distribution control.

[0090] Furthermore, the process of obtaining the scenario switching instruction includes:

[0091] Temperature, pressure, and dust concentration are acquired through a sensor network and stored as a raw dataset;

[0092] Data preprocessing techniques are used to clean the original dataset, remove outliers, and generate a standardized dataset.

[0093] Based on the standardized dataset, waveform parameter features are extracted to construct a working condition scenario feature set;

[0094] The current working condition scenario type is determined by analyzing the feature set of the working condition scenario through a pre-established working condition scenario model.

[0095] If the current operating condition scenario type matches the preset scenario library, a corresponding scenario switching instruction will be generated;

[0096] If there is no match, record it as a new scenario type;

[0097] Adjust the sensor network's acquisition frequency according to the scenario switching command to obtain updated environmental parameters;

[0098] The waveform parameters are re-optimized by updating the environmental parameters to generate a new set of operating condition features.

[0099] Specifically, environmental parameters such as temperature, pressure, and dust concentration are acquired through sensor networks and stored as raw datasets.

[0100] The sensor network consists of multiple nodes distributed at key locations within the high-voltage power supply equipment, collecting real-time dynamic data on temperature, pressure, and dust concentration. Temperature sensors record ambient temperatures ranging from 20-50℃, pressure sensors monitor airflow pressure from 0.1-0.5 MPa, and dust concentration sensors detect particulate matter concentrations from 0-500 mg / m³. 3 These data are aggregated to a central database via a wireless communication module, forming a raw dataset containing timestamps and sensor numbers for subsequent processing. Data preprocessing techniques are then used to clean the raw dataset, removing outliers and generating a standardized dataset.

[0101] For example, outliers can be caused by sensor malfunctions or transient interference, such as a sudden temperature jump to 100°C or a negative dust concentration. These outliers are removed using median filtering, and the data is normalized to map temperature, pressure, and dust concentration to a standardized range of 0-1.

[0102] Preferably, for situations with large fluctuations in dust concentration, a sliding window averaging method is used to smooth the data, ensuring that the dataset reflects the characteristics of real operating conditions. Waveform parameter features are extracted from the standardized dataset to construct a feature set for operating conditions.

[0103] For example, when extracting waveform parameter features, we analyze the trend of dust concentration over time and calculate features such as peak frequency and amplitude change rate. If the dust concentration increases from 50 mg / m³ in 10 minutes... 3 Rapidly rises to 300 mg / m 3 This indicates the possibility of sudden dust emissions. By combining temperature and pressure variation trends, a feature set of operating conditions, including peak values, periods, and offsets, is constructed to comprehensively characterize the current environmental state. The feature set is then analyzed using a pre-established operating condition scenario model to determine the type of the current operating condition scenario.

[0104] For example, the operating condition scenario model is trained based on historical data and includes various scenario templates such as normal operation, high dust emission, and low airflow pressure. If the feature set shows high dust concentration and low pressure, the model can determine it as an "abnormal equipment dust emission" scenario.

[0105] Preferably, a clustering algorithm is used to match the feature set with the template to determine the scenario type. If the scenario type matches a preset scenario library, a corresponding scenario switching instruction is generated; otherwise, it is recorded as a new scenario type.

[0106] For example, when a "high dust emission" scenario is matched, instructions are generated to adjust the operating parameters of the spray system or pulse cleaning equipment. If the feature set shows abnormally high temperature accompanied by low dust concentration, and no preset scenario is matched, it is recorded as a new scenario and stored in the scenario library for subsequent model updates. The sensor network's acquisition frequency is adjusted according to the scenario switching instructions to obtain updated environmental parameters.

[0107] For example, in scenarios with high dust emissions, the sensor's data collection frequency is increased from once per minute to once per second to capture rapidly changing dust concentrations. The updated data reflects the adjusted operating conditions, such as a dust concentration decreasing to 100 mg / m³. 3 This indicates that the control measures are effective. The waveform parameters are then re-optimized using the updated environmental parameters to generate a new set of operating condition features.

[0108] For example, by recalculating waveform parameters based on updated data and adjusting the frequency and intensity of pulse cleaning, a new feature set is formed, ensuring that the system continuously adapts to dynamic operating conditions. This dynamic adjustment mechanism effectively improves the real-time performance and accuracy of dust control.

[0109] Furthermore, the process of obtaining the mode adjustment output includes:

[0110] Based on the scenario switching command, the corresponding scenario parameters are extracted from the preset command library to determine the current running scenario;

[0111] Based on the current operating scenario, the corresponding rule subset is matched from the fuzzy rule set to obtain the appropriate control rules;

[0112] The adapted control rules are quantized using membership functions to generate fuzzy control parameters.

[0113] Obtain the load change rate and, in conjunction with fuzzy control parameters, calculate the dynamic adjustment strategy;

[0114] If the dynamic adjustment strategy meets the preset threshold, the high-voltage power supply control mode is switched and a mode switching command is generated.

[0115] According to the mode switching command, adjust the output parameters of the high-voltage power supply to obtain the mode-regulated output;

[0116] By analyzing the output stability, the stability of the mode adjustment output is verified, and the final control signal is generated.

[0117] For example, in a scenario where a sensor network collects environmental parameters, after receiving a scenario switching command, the system extracts the corresponding scenario parameters from a preset command library. Assuming the sensor network is deployed in an industrial plant, monitoring temperature, pressure, and dust concentration, the command library contains different operating conditions, such as high temperature and high pressure, and normal temperature and low pressure. When extracting scenario parameters, the system matches the parameter set according to the command number. For instance, the high temperature and high pressure scenario corresponds to a temperature of 80℃, a pressure of 2.5MPa, and a dust concentration of 100mg / m³. 3 This determines that the current operating scenario is a high-temperature, high-pressure condition. This method ensures that the parameters are accurately matched to the operating conditions.

[0118] In one possible implementation, a subset of rules is matched from a fuzzy rule set based on the current operating scenario. The fuzzy rule set contains multiple rules, such as "if the temperature and pressure are high, the control strategy is to enhance cooling." The system matches the rule subset using scenario parameters, such as a temperature of 80°C, to obtain appropriate control rules, such as prioritizing cooling and reducing pressure. This matching, based on fuzzy logic, can handle parameter uncertainties and ensure that the control rules adapt to the dynamic environment.

[0119] For example, membership functions can be used to quantize control rules and generate fuzzy control parameters. Membership functions map parameters such as temperature and pressure to fuzzy sets; for instance, 80℃ is mapped to a "high temperature" membership degree of 0.9 and a "medium temperature" membership degree of 0.1, generating fuzzy control parameters such as a cooling intensity coefficient of 0.8 and a pressure regulation coefficient of 0.6. This quantization method transforms fuzzy rules into operational parameters, facilitating subsequent calculations.

[0120] Specifically, when acquiring the load change rate, the system monitors load changes in real time through sensors. For example, if the pressure suddenly changes from 2.5 MPa to 2.7 MPa, the change rate is calculated as 0.2 MPa / min. Combined with fuzzy control parameters, a dynamic adjustment strategy is calculated.

[0121] For example, with a cooling intensity coefficient of 0.8 and a change rate of 0.2 MPa / min, it is determined that the cooling water flow rate should be increased by 20 L / min. This strategy can quickly respond to load fluctuations and maintain system stability.

[0122] In one embodiment, if the dynamic adjustment strategy meets a preset threshold, such as the cooling water flow rate increment being less than 30 L / min, the system switches the high-voltage power supply control mode and generates a mode switching command.

[0123] For example, switching to energy-saving mode reduces the power supply output voltage from 220V to 200V. This switching is explicitly executed via instructions, reducing manual intervention.

[0124] For example, after adjusting the high-voltage power supply output parameters, a mode-adjustable output is obtained. In energy-saving mode, the power supply output stabilizes at 200V, and the current is adjusted to 5A. The system verifies the stability of the adjusted output through output stability analysis; for example, if the detected voltage fluctuation is less than ±5V, the output stability is confirmed. This analysis ensures the reliability of the control signal.

[0125] In one possible implementation, after generating the final control signal, the system sends the signal to the actuator, such as a cooling pump or a pressure valve.

[0126] For example, a control signal might specify a 10Hz increase in the operating frequency of the cooling pump and a 5% increase in the opening of the pressure valve. This signal directly drives equipment adjustments to ensure stable operation.

[0127] It should be noted that the above embodiments form a complete closed-loop control system from scenario parameter extraction to final control signal generation. Each step is closely linked, and based on sensor network data, the control logic ensures adaptability to complex operating conditions. Examples from different aspects, such as parameter matching, fuzzy rule application, and stability verification, collectively support the argument for efficient system operation, demonstrating the comprehensiveness and reliability of the solution.

[0128] Furthermore, the process of obtaining the response optimization parameters includes:

[0129] Voltage deviation and current fluctuation indicators are obtained from the mode regulation output. The time series data is decomposed by signal processing methods to obtain the quantized deviation and fluctuation values.

[0130] Based on the quantified deviation and fluctuation values, a fuzzy relation model is constructed, with the deviation and fluctuation values ​​as input and the load change trend as output.

[0131] If the load change trend shows rapid fluctuations, then the frequency and amplitude of the fluctuations can be extracted through time series analysis to obtain the characteristic values ​​of the rapid fluctuations.

[0132] Based on the rapidly fluctuating characteristic values, a real-time control algorithm is used to calculate the millisecond-level response strategy and generate the initial response parameters;

[0133] The optimized response parameters are obtained by iteratively adjusting the initial response parameters using an optimization algorithm.

[0134] Control commands are extracted from the optimized response parameters, and the output is adjusted by the command distribution mechanism to obtain a stable output state.

[0135] For example, in high-voltage power supply control systems, handling voltage deviation and current fluctuations is a core aspect. Voltage deviation refers to the difference between the actual output voltage and the target voltage, while current fluctuations reflect current instability caused by load changes. Time-series data is decomposed using signal processing methods, and wavelet transform techniques are employed to separate high-frequency and low-frequency components.

[0136] For example, suppose a high-voltage power supply system has a target voltage of 10kV, an actual measured value of 10.2kV, and a voltage deviation of 0.2kV; the current fluctuation changes from 5A to 5.5A within 1 second. Through wavelet transform, the low-frequency trend of the deviation and the high-frequency features of the current fluctuation are extracted, and after quantization, a deviation value of 0.2 and a fluctuation value of 0.5 are obtained. This quantization facilitates subsequent fuzzy modeling.

[0137] In one possible implementation, when constructing the fuzzy relation model, the deviation value and fluctuation value are used as inputs and mapped to a fuzzy set of load change trends.

[0138] For example, a deviation of 0.2 is categorized as "low deviation," and a fluctuation of 0.5 is categorized as "medium fluctuation." The fuzzy relation model generates load change trends using predefined rules, such as "if the deviation is low and the fluctuation is moderate, then the trend is stable." If the trend shows rapid fluctuations, such as a fluctuation amplitude exceeding 0.3A within 1 second, further analysis of the fluctuation frequency and amplitude is required. Time series analysis uses Fast Fourier Transform to extract feature values, such as a fluctuation frequency of 10Hz and an amplitude of 0.4A. These feature values ​​reflect the dynamic characteristics of the load, providing a basis for real-time control.

[0139] Specifically, the real-time control algorithm is based on the design of a proportional-integral control strategy using eigenvalues.

[0140] For example, for a fluctuation frequency of 10Hz, the algorithm sets a millisecond-level response time to generate initial response parameters, such as adjusting the voltage gain to 1.05 times. The optimization algorithm further iterates and adjusts, for example, by using a genetic algorithm to optimize the gain to 1.03 times, generating optimized response parameters. This optimization ensures more precise control commands. The command distribution mechanism converts the optimized parameters into specific control signals, such as adjusting the duty cycle of the pulse width modulation signal to 60%, thereby changing the power supply output and ultimately achieving a stable output state, such as a voltage stable at 10±0.05kV and current fluctuation controlled within 0.2A.

[0141] It should be noted that the above methods are closely integrated with the application of scenario switching and fuzzy rule sets in historical dialogues.

[0142] For example, rapidly fluctuating feature values ​​can be used as additional inputs and combined with fuzzy rule sets to dynamically adjust the control mode. This method improves the system's adaptability, especially in scenarios with frequent load changes, effectively reducing output instability and improving power supply efficiency.

[0143] Furthermore, the process of obtaining the final adjustment instructions includes:

[0144] Real-time monitoring data is obtained from equipment sensors, and dust removal efficiency and energy consumption parameters are extracted to obtain an operating status dataset.

[0145] The operation status dataset is analyzed using a pre-set linear regression model to calculate the correlation between dust removal efficiency and energy consumption parameters, and to determine the initial weighting coefficients.

[0146] If the initial weight coefficient deviates from the preset threshold by more than a certain range, the gradient descent algorithm is used to adjust the weight coefficient to obtain the optimized weight coefficient.

[0147] Based on the optimized weight coefficients and the running status dataset, a multi-objective optimization algorithm is integrated to calculate the comprehensive control strategy.

[0148] A set of adjustment instructions is generated by combining comprehensive control strategies and real-time monitoring data;

[0149] If the matching degree between the adjustment instruction set and the running status dataset is lower than the preset threshold, then real-time monitoring data will be reacquired and the running status dataset will be updated.

[0150] Based on the updated operational status dataset, the optimization weight coefficients and comprehensive control strategy are repeatedly calculated to obtain the final adjustment instructions.

[0151] For example, when acquiring real-time monitoring data from equipment sensors, key parameters such as dust concentration, airflow velocity, and equipment power consumption are collected through multiple sensors installed on the dust removal equipment. Assume an industrial dust removal device collects data once per second, obtaining a dust concentration of 50 mg / m³. 3 Real-time data points with an airflow velocity of 2 m / s and a power consumption of 500 W were collected. These data constitute the operational status dataset, providing a foundation for subsequent analysis.

[0152] It should be noted that sensors need to be calibrated regularly to ensure data accuracy and avoid distortion in subsequent analysis due to deviations.

[0153] In one possible implementation, when analyzing the operational status dataset using a pre-defined linear regression model, dust concentration is used as the independent variable, and dust removal efficiency and energy consumption are used as dependent variables. The correlation between these two variables and dust concentration is calculated. Assuming the analysis results show a positive correlation between dust removal efficiency and dust concentration (correlation coefficient of 0.85), while the correlation coefficient between energy consumption and dust concentration is 0.60, it indicates that dust removal efficiency is significantly affected by dust concentration. The initial weighting is set to 0.7 for dust removal efficiency and 0.3 for energy consumption. If the deviation of the weighting coefficients from the pre-defined threshold of 0.5 exceeds ±0.2, further adjustments are required.

[0154] For example, when adjusting weight coefficients using the gradient descent algorithm, iterative optimization brings the weight coefficients closer to actual operational requirements. Assuming initial weight coefficients are 0.7 and 0.3, after 10 iterations, the optimized weight coefficients are adjusted to 0.65 and 0.35. This adjustment better balances the priorities of dust removal efficiency and energy consumption.

[0155] Preferably, the gradient descent algorithm uses a small learning rate, such as 0.01, to ensure stable convergence.

[0156] In one possible implementation, when integrating a multi-objective optimization algorithm to calculate the comprehensive control strategy, the objectives of maximizing dust removal efficiency and minimizing energy consumption are considered simultaneously. Assume the algorithm outputs a strategy where, when the dust concentration exceeds 40 mg / m³... 3 At the same time, the airflow velocity is increased to 2.5 m / s while the power consumption is reduced to 450 W. This strategy can effectively improve dust removal efficiency while controlling energy consumption.

[0157] It should be noted that multi-objective optimization requires balancing the priorities of different objectives to avoid imbalances in other parameters caused by optimizing a single objective.

[0158] For example, when generating the adjustment instruction set, specific instructions are output based on the optimized weighting coefficients and real-time data, such as "increase the fan speed to 1200 rpm and reduce the heating power to 400W". If the matching degree between the instruction set and the operating status dataset is less than 80%, data needs to be collected again. Assume the new data point shows that the dust concentration has increased to 60 mg / m³. 3 The running status dataset needs to be updated, and the weight coefficients and control strategies need to be recalculated.

[0159] Preferably, the matching degree check is implemented through a data comparison algorithm to ensure that the instructions are highly consistent with the actual state.

[0160] In one possible implementation, a real-time feedback mechanism is introduced when repeatedly calculating the optimized weighting coefficients and the integrated control strategy. It is assumed that the updated data indicates the dust concentration remains stable at 55 mg / m³. 3 The optimization weighting coefficients were adjusted to 0.68 and 0.32, and the final adjustment command was "maintain the fan speed at 1200 rpm and reduce power consumption to 420W". This method can dynamically adapt to changes in equipment operating status and ensure efficient system operation.

[0161] It should be noted that the real-time feedback mechanism significantly improves the adaptability of instructions and reduces instability factors during equipment operation.

[0162] For example, the above scheme forms a complete closed-loop control process from core data acquisition to control strategy generation. The combination of multi-point sensor data acquisition, linear regression analysis, gradient descent optimization, and multi-objective optimization algorithms ensures the accuracy and adaptability of the commands.

[0163] Preferably, an anomaly detection module is added to detect when the dust concentration suddenly increases to 100 mg / m³. 3 In such cases, an emergency adjustment command is triggered, further enhancing the system's robustness. This multi-faceted support approach effectively improves the operating efficiency of dust removal equipment, reduces energy consumption, and adapts to complex operating conditions.

[0164] Furthermore, the process of dynamically adjusting the high-voltage power supply control parameters to obtain the dynamic adjustment results includes:

[0165] Acquire initial state data for multi-objective equilibrium and extract the current equilibrium state from the high-voltage power supply operating parameters;

[0166] If the current equilibrium state does not reach the preset threshold, the weight coefficients are iteratively optimized using the gradient descent algorithm to obtain the adjusted set of coefficients.

[0167] Based on the adjusted set of coefficients, new adjustment instructions are generated, and the execution sequence of the instructions is determined.

[0168] The balance effect of the adjustment command is verified by simulation, and the state data after simulation is obtained;

[0169] If the simulated state data still does not reach equilibrium, the weight coefficients are further adjusted based on the particle swarm optimization algorithm to obtain a refined set of coefficients.

[0170] The final control output is generated based on the set of refining coefficients, and the operating parameters of the high-voltage power supply are determined.

[0171] Dynamic adjustment is performed based on the high-voltage power supply control parameters to obtain dynamic adjustment results.

[0172] For example, in the optimization of high-voltage power supply operating parameters, acquiring initial state data is the core foundation of the entire process. Suppose a dust removal device's high-voltage power supply is running; parameters such as voltage, current, and power factor need to be extracted from sensors to assess the current equilibrium state. An initial state dataset is constructed by recording a voltage of 5000V, a current of 200mA, and a power factor of 0.85 in real time. This data reflects the device's operating characteristics under specific conditions, providing a basis for subsequent optimization.

[0173] It should be noted that the determination of the balance state is usually based on a preset threshold, such as a power factor of 0.9 or higher. If the current power factor is lower than this value, it indicates that the system has not reached ideal balance and further optimization is needed.

[0174] In one possible implementation, if the initial state is not satisfactory, the weighting coefficients are iteratively optimized using a gradient descent algorithm. The weighting coefficients determine the importance of voltage and current in the balance regulation.

[0175] For example, the initial weights can be set to 60% for voltage and 40% for current. Through multiple iterations, the algorithm adjusts the weights to 55% for voltage and 45% for current to better balance energy consumption and efficiency. This adjustment is based on trend analysis of historical operating data, ensuring that the weight coefficients more closely reflect actual operating conditions. The adjusted set of coefficients provides a basis for generating new instructions.

[0176] Specifically, when generating a new adjustment command, the new voltage and current setpoints are calculated based on the optimized weighting coefficients.

[0177] For example, the voltage is adjusted to 5100V and the current to 190mA, and the instructions are arranged in a time sequence to ensure a smooth transition. Subsequently, the effectiveness of the instructions is verified through simulation. The simulation system predicts the adjusted operating state, such as a power factor improvement to 0.88, close to but not fully reaching the target threshold. This simulation verification effectively assesses the applicability of the instructions and reduces risks in actual operation.

[0178] For example, if the simulation results still do not reach equilibrium, the particle swarm optimization algorithm can be used to further refine the weight coefficients. The particle swarm optimization algorithm searches for a better combination of weights by simulating the cooperative behavior of a group.

[0179] For example, a refined set of coefficients can be obtained with a 50% weight for voltage and a 50% weight for current. This method is better suited for handling complex nonlinear relationships than gradient descent and can further improve the balancing effect. The final control output is generated based on the refined coefficients, for example, setting the voltage to 5080V and the current to 195mA.

[0180] In one embodiment, the final control output is applied to the high-voltage power supply via a real-time monitoring system. The monitoring system sends instructions to the power controller to adjust operating parameters in real time and collect updated status data.

[0181] For example, the adjusted power factor improved to 0.91, meeting the preset threshold. This closed-loop control method ensures stable operation of the equipment under dynamic conditions, while optimizing energy consumption and efficiency. For instance, the application of a real-time monitoring system not only executes instructions but also further optimizes the operating status through a feedback mechanism. The monitoring system collects high-voltage power supply operating data, such as voltage, current, and power factor, every 5 seconds, generating dynamic state curves.

[0182] For example, the acquired data shows that the voltage is stable at 5080V, the current fluctuates slightly but remains within the range of 195mA±5mA, and the power factor is maintained above 0.91. This high-frequency data acquisition can promptly detect operational deviations and provide a basis for subsequent adjustments.

[0183] It should be noted that the core advantage of a real-time monitoring system lies in its closed-loop feedback capability, which dynamically corrects instructions based on the actual operating status, ensuring the adaptability of the equipment under different load conditions.

[0184] In one possible implementation, if the updated operating status data indicates that the system still has a slight imbalance, such as the power factor accidentally dropping to 0.89, then further optimization is achieved through local fine-tuning.

[0185] For example, based on the deviation trend, the monitoring system temporarily increases the voltage to 5100V while keeping the current constant to quickly restore balance. This fine-tuning method avoids large-scale parameter adjustments and reduces interference with equipment operation.

[0186] Preferably, the fine-tuning process also incorporates historical data analysis, such as the operating curve over the past hour, to ensure the long-term stability of the adjustments.

[0187] Specifically, the implementation of control outputs requires consideration of equipment response time and external environmental factors.

[0188] For example, the response time of high-voltage power supplies is typically in the millisecond range, so it is necessary to ensure that the frequency of command issuance matches the equipment's response capability. If the external ambient temperature rises, it can cause current fluctuations. In this case, a monitoring system can provide early warnings and dynamically adjust the weighting coefficients to cope with environmental changes.

[0189] For example, the current weight can be slightly increased to 52% to enhance adaptability to temperature fluctuations. This dynamic adjustment mechanism significantly improves the robustness of the system.

[0190] For example, in practical applications, the monitoring system also records the status data after each adjustment, forming a historical database for subsequent optimization.

[0191] For example, the database shows that under specific load conditions, a combination of 5080V voltage and 195mA current can consistently maintain a power factor above 0.9. This data provides a reference for parameter setting under similar operating conditions in the future, reducing the time spent on repeated optimization.

[0192] It should be noted that this data-driven optimization approach not only improves operational efficiency but also provides support for long-term equipment maintenance.

[0193] In one embodiment, if further expansion of the scheme is required, a predictive model can be introduced to assist in optimization.

[0194] For example, by analyzing historical operating data, the system can predict load changes over the next hour and adjust weighting coefficients in advance. This predictive adjustment effectively addresses sudden changes in operating conditions, such as a sudden increase in load, by raising the voltage to 5120V in advance to prevent a drop in the power factor. This extended solution seamlessly integrates with the core optimization process, further enhancing the system's intelligence.

[0195] The above are merely preferred embodiments 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.

Claims

1. A method for dynamic adjustment of high-voltage power supply in an electrostatic precipitator based on fuzzy control, characterized in that, include: The real-time operating parameters of the high-voltage power supply of the electrostatic precipitator are obtained. Based on the real-time operating parameters, impedance change and dust concentration are extracted as input variables. The tripping risk level is determined through a fuzzy inference mechanism to obtain the risk prediction result. Based on the risk prediction results, the adjustment requirements for the output characteristics of the high-voltage power supply are determined. If the risk level is higher than the preset threshold, the output voltage amplitude is reduced to obtain a preliminary adjustment scheme. The non-uniformity of dust distribution is obtained from the preliminary adjustment scheme. The pulse width and frequency are adjusted by using a fuzzy control algorithm to generate an irregular waveform output that matches the particle motion trajectory, thereby obtaining optimized waveform parameters. Environmental parameter data is collected through a sensor network. A working condition scenario model is constructed by combining the environmental parameter data with optimized waveform parameters. A scenario recognition algorithm is used to determine the current scenario type and obtain a scenario switching command. Based on the scenario switching instruction, the corresponding fuzzy rule set and membership function are selected to drive the high-voltage power supply switching control mode and incorporate the load change rate as an additional input to obtain the mode adjustment output. Voltage deviation and current fluctuation indicators are extracted from the mode regulation output. A fuzzy relation model is used to predict the load change trend. If the load change trend shows rapid fluctuation, a millisecond-level response strategy is calculated to obtain response optimization parameters. By integrating dust removal efficiency and energy consumption targets through the aforementioned response optimization parameters, assigning dynamic weight coefficients and calculating a comprehensive control strategy, the final adjustment command is obtained. The multi-objective balance state is verified according to the final adjustment command. If the balance state is not met, the weight coefficients are iteratively adjusted to obtain the high-voltage power supply control parameters. Dynamic adjustment is then performed based on the high-voltage power supply control parameters to obtain the dynamic adjustment result.

2. The method according to claim 1, characterized in that, The process of determining the tripping risk level and obtaining risk prediction results through fuzzy reasoning mechanisms includes: Based on the voltage, current and power data obtained from the high voltage power supply of the electrostatic precipitator, the impedance change is calculated and the impedance change value is obtained. Based on the impedance change value and the dust concentration data collected by the dust concentration sensor, an input variable set is constructed to obtain standardized input variables; A fuzzy inference mechanism is used to fuzzify the standardized input variables, generate fuzzy membership functions, and obtain a fuzzy input set. By using a pre-established fuzzy rule base, reasoning is performed on the fuzzified input set to determine the tripping risk level and obtain a risk level assessment value. If the risk level assessment value exceeds the preset threshold, the risk level is classified using the support vector machine algorithm to obtain the classified risk prediction result. Based on the risk prediction results after classification, time series analysis is used to predict the future tripping risk trend and obtain the risk trend prediction value. By comparing the predicted risk trends with historical data and adjusting the parameters of the fuzzy rule base, an optimized fuzzy inference model is obtained.

3. The method according to claim 1, characterized in that, The process of obtaining a preliminary adjustment plan includes: Based on the risk prediction results, the current output voltage amplitude and operating status parameters are extracted from the high-voltage power supply system operating data to obtain risk level data. Using threshold comparison logic, if the risk level data is higher than a preset threshold, a voltage regulation mechanism is triggered to determine the output voltage amplitude that needs to be reduced. Based on the determined output voltage amplitude that needs to be reduced, the target output characteristics of the high-voltage power supply system are calculated, and a preliminary adjustment scheme is generated. By simulating the operating environment, the impact of the preliminary adjustment scheme on the high-voltage power supply system is verified, and the adjusted power supply operating status is obtained. If the adjusted power supply operating state meets the preset safety range, the preliminary adjustment scheme will be converted into an executable control command to generate the final adjustment scheme. Based on the final adjustment scheme, update the control parameters of the high-voltage power supply system, obtain the updated operating data, and determine the system stability. By analyzing the degree of matching between the updated operational data and the risk prediction results, the threshold comparison logic is optimized, resulting in an improved adjustment mechanism.

4. The method according to claim 1, characterized in that, The process of obtaining optimized waveform parameters includes: Non-uniformity characteristics are obtained from dust distribution data, and statistical analysis methods are used to calculate the variance and skewness of the distribution to obtain the feature vector. Based on the feature vector, an input fuzzy set is established using a fuzzy control algorithm. Membership degree mapping is performed to address the non-uniformity of dust distribution, and a fuzzy rule set is determined. If the membership degree of the fuzzy rule set is greater than a preset threshold, the pulse width and frequency are adjusted to generate an initial irregular waveform and obtain a waveform parameter set. By analyzing the periodicity and offset of the particle motion trajectory collected in real time, the trajectory feature vector is obtained. Based on the matching degree between the trajectory feature vector and the waveform parameter set, the waveform parameters are iteratively adjusted using the particle swarm optimization algorithm to determine the optimized waveform parameters; If the matching degree between the optimized waveform parameters and the particle motion trajectory is lower than a preset threshold, the fuzzy rule set is updated through a feedback mechanism to obtain the updated waveform parameter set. Based on the updated waveform parameter set, the final irregular waveform is generated, and a control signal matching the particle motion trajectory is output.

5. The method according to claim 1, characterized in that, The process of receiving a scenario switching instruction includes: Temperature, pressure, and dust concentration are acquired through a sensor network and stored as a raw dataset; The original dataset is cleaned using data preprocessing techniques to remove outliers and generate a standardized dataset. Based on the standardized dataset, waveform parameter features are extracted to construct a working condition scenario feature set; The current working condition scenario type is determined by analyzing the feature set of the working condition scenario through a pre-established working condition scenario model. If the current operating condition scenario type matches the preset scenario library, a corresponding scenario switching instruction will be generated; If there is no match, record it as a new scenario type; Adjust the sensor network's acquisition frequency according to the scenario switching command to obtain updated environmental parameters; The waveform parameters are re-optimized by updating the environmental parameters to generate a new set of operating condition features.

6. The method according to claim 1, characterized in that, The process of obtaining the mode adjustment output includes: Based on the scenario switching instruction, the corresponding scenario parameters are extracted from the preset instruction library to determine the current running scenario; Based on the current operating scenario, the corresponding rule subset is matched from the fuzzy rule set to obtain the appropriate control rules; The adapted control rules are quantized using membership functions to generate fuzzy control parameters. Obtain the load change rate and, in conjunction with the fuzzy control parameters, calculate the dynamic adjustment strategy; If the dynamic adjustment strategy meets the preset threshold, the high-voltage power supply control mode is switched and a mode switching command is generated. According to the mode switching command, adjust the output parameters of the high-voltage power supply to obtain the mode regulation output; By analyzing the output stability, the stability of the mode adjustment output is verified, and the final control signal is generated.

7. The method according to claim 1, characterized in that, The process of obtaining response optimization parameters includes: Voltage deviation and current fluctuation indicators are obtained from the mode regulation output. The time series data is decomposed by signal processing methods to obtain the quantized deviation and fluctuation values. Based on the quantified deviation and fluctuation values, a fuzzy relation model is constructed, with the deviation and fluctuation values ​​as input and the load change trend as output. If the load change trend shows rapid fluctuations, then the frequency and amplitude of the fluctuations can be extracted through time series analysis to obtain the characteristic values ​​of the rapid fluctuations. Based on the rapidly fluctuating characteristic values, a real-time control algorithm is used to calculate the millisecond-level response strategy and generate the initial response parameters; The initial response parameters are iteratively adjusted using an optimization algorithm to obtain the optimized response parameters; Control commands are extracted from the optimized response parameters, and the output is adjusted by the command distribution mechanism to obtain a stable output state.

8. The method according to claim 1, characterized in that, The process of obtaining the final adjustment instructions includes: Real-time monitoring data is obtained from equipment sensors, and dust removal efficiency and energy consumption parameters are extracted to obtain an operating status dataset. The operating status dataset is analyzed using a pre-defined linear regression model to calculate the correlation between dust removal efficiency and energy consumption parameters, and to determine the initial weighting coefficients. If the initial weight coefficient deviates from the preset threshold by more than a certain range, the gradient descent algorithm is used to adjust the weight coefficient to obtain the optimized weight coefficient. Based on the optimized weight coefficients and the running status dataset, a multi-objective optimization algorithm is fused to calculate a comprehensive control strategy. A set of adjustment instructions is generated based on the comprehensive control strategy and real-time monitoring data. If the matching degree between the adjustment instruction set and the running status dataset is lower than a preset threshold, then real-time monitoring data is reacquired and the running status dataset is updated. Based on the updated operational status dataset, the optimization weight coefficients and comprehensive control strategy are repeatedly calculated to obtain the final adjustment instructions.

9. The method according to claim 1, characterized in that, The process of dynamically adjusting the high-voltage power supply based on the control parameters to obtain the dynamic adjustment result includes: Acquire initial state data for multi-objective equilibrium and extract the current equilibrium state from the high-voltage power supply operating parameters; If the current equilibrium state does not reach the preset threshold, the weight coefficients are iteratively optimized using the gradient descent algorithm to obtain the adjusted set of coefficients. Based on the adjusted set of coefficients, a new adjustment instruction is generated, and the execution sequence of the instruction is determined; The balance effect of the adjustment command is verified by simulation, and the state data after simulation is obtained; If the simulated state data still does not reach equilibrium, the weight coefficients are further adjusted based on the particle swarm optimization algorithm to obtain a refined set of coefficients. The final control output is generated based on the set of refining coefficients, and the operating parameters of the high-voltage power supply are determined. Dynamic adjustment is performed based on the high-voltage power supply control parameters to obtain dynamic adjustment results.

Images