A fuzzy control method, computer equipment, and computer-readable storage medium for stable operation of an electrostatic precipitator system.

Abstract

This invention discloses a fuzzy control method for the stable operation of an electrostatic precipitator (ESP) system, relating to the field of fuzzy control technology. The method includes: generating a first factor based on multi-physics field coupling analysis, where the multi-physics field includes temperature gradient, airflow vorticity, and charge field non-uniformity; simultaneously acquiring ash layer dielectric parameters, calculating ash layer resistivity, and generating a second factor; acquiring flue gas concentration deviation and energy consumption indicators, dynamically reconstructing the fuzzy membership function using the first and second factors, and generating a control rule base including a dielectric compensation term; inputting the flue gas concentration deviation and energy consumption indicators into the control rule base, executing multi-objective fuzzy decision-making, and outputting collaborative control parameters; triggering iteration of the first factor based on changes in ash layer dielectric parameters, and adjusting the second factor inversely based on dust removal efficiency deviation to optimize rule weights. This invention improves the operational stability of the ESP system by combining multi-physics field coupling analysis with dynamic detection of ash layer dielectric parameters.

Description

Technical Field

[0001] This invention relates to the field of fuzzy control technology, and in particular to a fuzzy control method for the stable operation of an electrostatic precipitator system. Background Technology

[0002] In the field of industrial flue gas purification, the intelligent control technology of electrostatic precipitators (ESPs) has gradually evolved from single-variable regulation to multi-parameter collaborative decision-making in recent years. Traditional fuzzy control methods achieve basic regulation by constructing a static rule base of flue gas concentration deviation and energy consumption indicators. Some studies have attempted to optimize control parameters using temperature field zoning monitoring or electric field strength feedback, but their input variables are mostly limited to single physical field characteristics. In existing technologies, fuzzy PID algorithms based on dynamic prediction of corona current can mitigate the impact of operating condition fluctuations, but their modeling under multi-field coupling effects is still insufficient, especially lacking the ability to quantitatively analyze nonlinear interactive effects such as changes in dust resistivity caused by temperature gradients and the decay of charge efficiency caused by airflow vortices. In addition, online detection of ash layer dielectric parameters mostly relies on contact electrode measurements, and the data update frequency and spatial resolution are difficult to meet the requirements of real-time control.

[0003] The current field of electrostatic precipitator control faces three bottlenecks: First, the lack of a dynamic integration mechanism for physical field coupling parameters results in insufficient adaptability of control rules to complex operating conditions such as airflow disturbance and temperature stratification, which can easily lead to secondary dust generation or electric field flashover. Second, the correlation model between the dielectric properties of the ash layer and the resistivity has not yet been established, and traditional detection methods under low resistivity conditions are lagging and cannot suppress back corona phenomena in a timely manner. Third, the boundaries of fuzzy membership functions are fixed in the multi-objective optimization process, making it impossible to dynamically adjust the domain range according to the game relationship between energy efficiency constraints and dust removal efficiency, which can easily lead to mismatch in the weight allocation of the rule base. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a fuzzy control method for the stable operation of an electrostatic precipitator system, solving the problem of poor stability of traditional electrostatic precipitators under the coupling of multiple physical fields.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a fuzzy control method for stable operation of an electrostatic precipitator system, comprising generating a first factor based on multi-physics field coupling analysis, wherein the multi-physics field includes temperature field gradient, airflow vorticity and charge field non-uniformity.

[0008] Simultaneously collect the dielectric parameters of the ash layer, calculate the specific resistivity of the ash layer, and generate the second factor;

[0009] Collect flue gas concentration deviation and energy consumption index, and dynamically reconstruct the fuzzy membership function using the first and second factors to generate a control rule base containing dielectric compensation terms.

[0010] Input the flue gas concentration deviation and energy consumption index into the control rule base, execute multi-objective fuzzy decision-making, and output collaborative control parameters;

[0011] The first factor iteration is triggered by the change in the dielectric parameters of the ash layer, and the second factor is adjusted in reverse according to the deviation in dust removal efficiency to optimize the rule weights.

[0012] As a preferred embodiment of the fuzzy control method for stable operation of the electrostatic precipitator system described in this invention, the first factor is obtained by normalizing the temperature field gradient, airflow vorticity, and charge field non-uniformity and then calculating using weighted summation. The temperature field gradient is collected in real time by a distributed temperature sensor array, the airflow vorticity is determined by multi-point measurements of the eddy current sensor at the electric field cross section, and the charge field non-uniformity is calculated based on the spatial distribution data of the charge density probe.

[0013] As a preferred embodiment of the fuzzy control method for stable operation of the electrostatic precipitator system described in this invention, the second factor is dynamically generated based on the real and imaginary parts of the ash layer dielectric parameters. When the ash layer resistivity is lower than the preset back corona risk threshold, a compensation mechanism is activated, and the amplitude of the second factor is adjusted by an exponential function. The ash layer resistivity is obtained by inversion calculation of the dielectric loss parameters collected by the microwave resonant cavity sensor.

[0014] As a preferred embodiment of the fuzzy control method for stable operation of the electrostatic precipitator system described in this invention, the multi-physics coupling analysis refers to the use of three-dimensional synchronous acquisition of temperature gradient sensor, eddy current sensor and charge density probe, real-time output of the maximum value of temperature field gradient, the integral value of airflow vorticity and the relative value of electric field non-uniformity, and the linear superposition of the three after normalization to generate the first factor.

[0015] As a preferred embodiment of the fuzzy control method for stable operation of the electrostatic precipitator system described in this invention, the specific steps of dynamically reconstructing the fuzzy membership function are as follows:

[0016] The fuzzy universe of discourse range of the flue gas concentration deviation is dynamically adjusted according to the first factor, so that the fuzzy universe of discourse range expands and contracts linearly with the physical field coupling strength.

[0017] The membership function center point of the energy consumption index is shifted according to the second factor, and the shift amount is proportional to the square of the ash layer resistivity.

[0018] The resistivity of the ash layer is detected, and an exponential compensation term is added to the membership calculation to enhance the control sensitivity under low resistivity conditions.

[0019] As a preferred embodiment of the fuzzy control method for stable operation of the electrostatic precipitator system described in this invention, the specific steps of performing multi-objective fuzzy decision-making are as follows:

[0020] The control rule group is selected based on the threshold. When the first factor exceeds the first preset threshold, the pulse-dominated rule group is activated; when the second factor exceeds the second preset threshold, the dielectric compensation rule group is activated.

[0021] Calculation of back corona risk coefficient based on gray layer dielectric parameters;

[0022] The parameters for coordinated control are generated by taking into account the deviation of flue gas concentration, energy consumption indicators and the risk of back corona discharge.

[0023] As a preferred embodiment of the fuzzy control method for stable operation of the electrostatic precipitator system described in this invention, the triggering of the first factor iteration refers to adjusting the weight coefficient of the first factor using the gradient descent method to match the changing trend of the first factor and the dielectric parameter of the ash layer.

[0024] As a preferred embodiment of the fuzzy control method for stable operation of the electrostatic precipitator system described in this invention, the reverse adjustment of the second factor and the optimization of the rule weights refer to adjusting the calculation coefficients of the second factor based on the deviation between the actual dust removal efficiency and the target dust removal efficiency through an error feedback mechanism, regenerating the second factor and updating the weight allocation in the control rule base.

[0025] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the fuzzy control method for stable operation of the electrostatic precipitator system as described in the first aspect of the present invention.

[0026] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the fuzzy control method for stable operation of the electrostatic precipitator system as described in the first aspect of the present invention.

[0027] The beneficial effects of this invention are as follows: By combining multi-physics coupling analysis with dynamic detection of gray layer dielectric parameters, a two-factor collaborative optimization mechanism is generated, realizing adaptive adjustment for complex working conditions; a three-dimensional sensor array is used to collect multi-source data in real time, and dynamic control factors are generated through normalization processing and weighted fusion, improving the system's response accuracy to physical field coupling effects; a gray layer resistivity compensation mechanism and an anti-corona risk suppression strategy are used, and the membership function and control rule base are dynamically corrected using an exponential function, effectively reducing energy consumption and corona breakdown risk while ensuring dust removal efficiency; and iterative optimization of control parameters is achieved through gradient descent and error feedback mechanisms, enabling the system to maintain stable operation even when the gray layer dielectric properties change abruptly. Attached Figure Description

[0028] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a flowchart of the overall fuzzy control method for stable operation of the electrostatic precipitator system in Example 1.

[0030] Figure 2 This is a flowchart of the process for generating the first factor in the multiphysics coupling analysis of Example 1.

[0031] Figure 3 This is a flowchart of the gray layer dielectric parameter processing and generation of the second factor in Example 1.

[0032] Figure 4 This is a flowchart of the dynamic reconstruction of the control rule base and multi-objective decision-making in Example 1. Detailed Implementation

[0033] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0034] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0035] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0036] Example 1, referring to Figures 1-4 This embodiment provides a fuzzy control method for stable operation of an electrostatic precipitator system, including the following steps:

[0037] S1: The first factor is generated based on multiphysics coupling analysis, wherein the multiphysics includes temperature field gradient, airflow vorticity and charge field inhomogeneity.

[0038] Specifically, it includes the following steps:

[0039] S1.1: Collect multiphysics data and extract features.

[0040] S1.1.1: Deploy a distributed temperature sensor array to collect temperature field gradients.

[0041] Specifically, a distributed array of temperature sensors (thermocouples or fiber optic temperature sensors) is deployed within the electric field region to collect temperature data from each area in real time. The local temperature gradient is calculated using the temperature difference between adjacent sensors, and the maximum temperature gradient value across the entire field is taken as the characteristic quantity of the temperature field gradient.

[0042] S1.1.2: Measure the airflow vortex using an vortex sensor.

[0043] Specifically, eddy current sensors (ultrasonic eddy current sensors or hot-wire anemometers) are installed on the electric field cross section to measure the airflow velocity vector at each point. The vorticity (curvature of the velocity vector) at all measuring points on the electric field cross section is integrated to calculate the total vorticity value as a characteristic quantity of airflow vorticity.

[0044] S1.1.3: Collect charge density data using a charge density probe and calculate the charge field inhomogeneity.

[0045] Specifically, a three-dimensionally distributed charge density probe (electrostatic probe array) is used to collect charge density data at various spatial points within the electric field in real time. The spatial non-uniformity of charge density distribution is calculated using the standard deviation formula, which is expressed as follows:

[0046]

[0047] Where U is the relative mean of the charge field inhomogeneity, and C i Let be the charge density of the i-th probe. Where N is the average charge density, N is the total number of probes, and i is the probe number index.

[0048] S1.2: Normalize the multiphysics field characteristic quantities.

[0049] S1.2.1: Normalize the temperature field gradient.

[0050] Specifically, the maximum temperature gradient is determined by the temperature resistance of the equipment materials, and the characteristic quantity of the temperature field gradient is normalized as follows:

[0051]

[0052] Among them, G norm G is the normalized temperature field gradient characteristic quantity. max Let G be the maximum temperature gradient, and G be the characteristic quantity of the temperature field gradient.

[0053] When G>G max At that time, forced G norm=1.

[0054] S1.2.2: Normalize the airflow vorticity.

[0055] Specifically, the maximum vorticity integral value is determined based on historical data, and the airflow vorticity characteristic is normalized, expressed as:

[0056]

[0057] Among them, V norm V is the normalized characteristic quantity of airflow vorticity. max denoted as the maximum vorticity integral value, and V as the characteristic quantity of airflow vorticity.

[0058] S1.2.3: Normalize the non-uniformity of the charge field.

[0059] Specifically, the charge distribution under various typical operating conditions (such as voltage variation, temperature variation, material parameter disturbance, geometric deviation, etc.) is analyzed through finite element simulation, the non-uniformity index is extracted, and the threshold U of charge field non-uniformity is set based on the statistical results. th , as the normalization benchmark. The normalization expression is:

[0060]

[0061] Among them, U norm U is the normalized relative mean of the charge field inhomogeneity. th Let be the threshold value for charge field non-uniformity, and U be the relative mean value for charge field non-uniformity.

[0062] S1.3: Calculate the first factor based on the normalized temperature field gradient, airflow vorticity, and charge field inhomogeneity.

[0063] Specifically, the first factor is calculated by weighted summation, and the expression is:

[0064] F1 = w G ·G norm +w V ·V norm +w U ·U norm ;

[0065] Where F1 is the first factor (F1∈[0,1]), representing the multiphysics coupling state. The larger the value, the more significant the impact of multiphysics coupling on overall stability. G As the temperature field weight (taken as 0.4), w V w is the weight of the airflow vorticity (taken as 0.3). U This is the charge field weight (taken as 0.3).

[0066] The sum of the temperature field weight, airflow vorticity weight, and charge field weight is 1. The temperature field weight is increased in high-temperature environments, and the specific value is determined through historical data optimization and actual working conditions.

[0067] S1.3: Perform real-time timestamp synchronization and add a fault tolerance mechanism.

[0068] It should be noted that timestamp alignment is used to ensure the synchronous acquisition of temperature field, airflow field, and charge field data.

[0069] If the sensor malfunctions, interpolation compensation is performed based on adjacent data to ensure the continuity of feature calculation.

[0070] Ideally, distributed sensors are used to collect temperature gradient, airflow vorticity, and charge field data in real time, and normalization processing is combined to ensure that each physical quantity has a unified evaluation standard, avoiding deviations caused by differences in dimensions. The calculation of the first factor incorporates multi-dimensional influencing factors, giving it adaptive weight adjustment capabilities under different operating conditions, thereby optimizing the response to complex environments. Synchronization mechanisms and fault-tolerant compensation strategies enhance the continuity and reliability of temperature gradient, airflow vorticity, and charge field data acquisition, ensuring stable operation even in the event of sensor malfunction or short-term data loss.

[0071] S2: Synchronously collect the dielectric parameters of the ash layer, calculate the specific resistance of the ash layer, and generate the second factor.

[0072] Specifically, it includes the following steps:

[0073] S2.1: Collect the dielectric parameters of the gray layer and perform preprocessing.

[0074] S2.1.1: Measure the dielectric parameters of the gray layer using a microwave resonant cavity sensor array.

[0075] It should be noted that the dielectric parameter of the ash layer refers to the complex dielectric constant exhibited by the dust deposited on the surface of the electrostatic precipitator electrode plate under the influence of electromagnetic waves, including both real and imaginary parts. Obtaining this parameter through a microwave resonant cavity sensor array can indirectly reflect the conductivity and resistivity characteristics of the ash layer, providing fundamental information for electrostatic precipitator control.

[0076] Specifically, a microwave resonant cavity sensor array is uniformly arranged on the surface of the electrostatic precipitator plate. Each sensor transmits microwave signals to penetrate the ash layer, receives reflected signals, and extracts the real part (characterizing polarization capability) and imaginary part (characterizing dielectric loss) of the dielectric parameters.

[0077] The data is time-stamped with temperature field, airflow field and charge field data to ensure that the dielectric parameters of the gray layer are collected synchronously.

[0078] S2.1.2: Preprocess the dielectric parameters of the ash layer.

[0079] Specifically, wavelet noise reduction is performed on the collected gray layer dielectric parameters to eliminate high-frequency electromagnetic interference (such as pulse noise generated by high-voltage electric fields).

[0080] Meanwhile, outliers are corrected. If the real part of the dielectric parameter of the gray layer collected by the sensor is less than 1 or the imaginary part is less than 0, interpolation replacement is performed based on the data of adjacent sensors to ensure spatial continuity.

[0081] S2.2: Based on dynamic calculation of gray layer resistivity.

[0082] S2.2.1: Gray layer resistivity is derived from dielectric loss inversion.

[0083] Specifically, the loss tangent is calculated based on the real and imaginary parts of the dielectric parameters of the gray layer, and the expression is:

[0084]

[0085] Where tanδ is the dielectric loss tangent, ∈ ′ ∈″ represents the real part of the dielectric parameter of the gray layer, and ∈″ represents the imaginary part of the dielectric parameter of the gray layer.

[0086] The expression for calculating the resistivity of the ash layer is:

[0087]

[0088] Among them, R dust Here, k is the resistivity of the ash layer, and k is a material characteristic constant (calibrated through ash sample testing; typical value k = 10). 12 Ω·cm·rad), exp is the natural exponential function, α is the temperature correction factor (taken as 0.02), T dust This refers to the temperature of the ash layer.

[0089] S2.2.2: The resistivity of the ash layer is calibrated based on the percentage of flue gas humidity and the thickness of the ash layer.

[0090] It should be noted that a humidity sensor is used to collect the percentage of flue gas humidity, and a laser thickness gauge is used to measure the ash layer thickness. The ash layer resistivity is calibrated based on the percentage of flue gas humidity and the ash layer thickness to solve the problem of inaccurate ash layer resistivity calculation caused by excessive flue gas humidity or excessive ash layer thickness.

[0091] Specifically, if the flue gas humidity percentage is >15%, the ash layer resistivity is corrected based on the flue gas humidity percentage, as expressed by the following expression:

[0092] R new =R dust ·(1-0.05·(H-15%));

[0093] Among them, R newThe resistivity of the ash layer is the humidity-corrected value, where H is the percentage of humidity in the flue gas. For every 1% increase in humidity, the resistivity decreases by 5%.

[0094] Water vapor in flue gas has extremely high electronegativity. When humidity exceeds 15%, it significantly affects the conductivity of the discharge space and the polarization degree of the ash layer, leading to an overestimation of the resistivity measured by dielectric loss. The correction factor, which reduces resistance by approximately 5% for every 1% increase in humidity, is a linear empirical parameter obtained by fitting laboratory flue gas test and simulation data. When humidity is ≤15%, the water vapor content has a weak effect on the polarization path, and the change in dielectric response is not significant; therefore, no correction is made to avoid overfitting and unnecessary calculations.

[0095] When the ash layer thickness is greater than 5mm, ash layer thickness compensation is performed, and the expression is:

[0096] R thick =R dust ·(1+0.1·(d dust -5));

[0097] Where, d dust R represents the thickness of the ash layer. thick This represents the resistivity of the ash layer after thickness compensation. A greater thickness results in a longer charge conduction path within the ash layer, leading to a non-linear increase in resistivity. This upward trend is particularly pronounced when the thickness exceeds 5 mm. A correction factor increasing the resistance by 10% for every 1 mm increase in thickness was obtained through calibration using on-site industrial ash thickness data.

[0098] S2.3: Calculate the second factor and make dynamic adjustments.

[0099] S2.3.1: Initial calculation of the second factor.

[0100] Specifically, the second factor is generated by combining the real and imaginary parts of the dielectric parameters of the gray layer, and its expression is as follows:

[0101] F2=β·∈′+γ·∈″;

[0102] Where F2 is the second factor, β is the real part weight (taken as 0.6), and γ is the imaginary part weight (taken as 0.4), satisfying β+γ=1.

[0103] S2.3.2: Perform low specific resistance compensation.

[0104] Specifically, when R dust <R th At that time, the compensation mechanism is activated to adjust the second factor, as expressed in the following expression:

[0105]

[0106] Among them, F 2,compThe adjusted second factor, λ is the compensation strength coefficient (taken as 2), R th To preset the back corona risk threshold, a value of 10 is used. 10 Ω·cm (set based on industry experience, when the specific resistance is less than 10 Ω·cm) 9 ~10 10 Back corona is more likely to occur at Ω·cm.

[0107] S2.3.3: Adjust the weights based on the rate of change of the resistivity of the ash layer.

[0108] Specifically, the real and imaginary weights are dynamically updated based on the rate of change of the gray layer resistivity, as expressed by the following expression:

[0109]

[0110] γ = 1 - β;

[0111] Where, ΔR dust Δt represents the rate of change of the resistivity of the ash layer, and Δt represents the rate of change over time.

[0112] When the resistivity of the gray layer decreases rapidly, the weight of the imaginary part is increased to respond more quickly to sudden changes in dielectric loss.

[0113] Ideally, by acquiring and analyzing the dielectric parameters of the ash layer in real time, accurate calculation and dynamic adjustment of the ash layer resistivity are achieved. A microwave resonant cavity sensor array is used to ensure spatial coverage of the acquisition and improve measurement reliability. Wavelet denoising and outlier correction are applied to the ash layer dielectric parameters to eliminate high-frequency electromagnetic interference. The method of inverting ash layer resistivity based on dielectric loss can reflect the conductivity characteristics of the ash layer, and the accuracy of the ash layer resistivity calculation is ensured through temperature, humidity, and thickness compensation mechanisms. The calculation of the second factor combines the real and imaginary parts of the ash layer dielectric properties, and the weights are adjusted by the dynamic rate of change of resistivity, improving the method's adaptability to low resistivity risks. This effectively enhances the real-time perception capability of the electrostatic precipitator of the ash layer accumulation state, helps to provide early warning of back corona risks, optimizes the operation strategy of dust removal methods, and improves emission control accuracy and equipment operation stability.

[0114] S3: Collect flue gas concentration deviation and energy consumption indicators, and dynamically reconstruct the fuzzy membership function using the first and second factors to generate a control rule base containing dielectric compensation terms.

[0115] Specifically, it includes the following steps:

[0116] S3.1: Collect flue gas concentration deviation and energy consumption indicators through a laser dust sensor.

[0117] Specifically, a laser dust sensor measures the concentration of particulate matter in the flue gas at the flue outlet and compares it with a preset environmental limit to obtain the flue gas concentration deviation. The preset environmental limit is the upper limit of the target emission concentration, used to assess whether the flue gas purification effect meets the emission standards. It is set according to internal control requirements, with common settings ranging from 10 to 30 mg / m³. 3 .

[0118] Real-time current and voltage signals from the high-voltage power supply are collected to calculate the instantaneous input power. The average power value within a certain time window is then calculated to obtain the average corona power. This value is compared with the rated power to obtain a normalized energy consumption index, which is used to characterize the energy efficiency per unit dust removal effect and serves as one of the input parameters for fuzzy control.

[0119] S3.2: Reconstruct the fuzzy membership function by adjusting the universe of discourse, shifting the center point, and adding a dielectric compensation term.

[0120] S3.2.1: Adjustment of the fuzzy domain of flue gas concentration deviation.

[0121] It should be noted that the domain of discourse refers to the domain range of the input variable (i.e., flue gas concentration deviation) of the fuzzy membership function, which is used to determine the mapping boundary of the fuzzy quantity.

[0122] Specifically, when the first factor is greater than 0.6 (high multiphysics coupling strength), the domain range is linearly expanded from the default [-50%, +50%] to [-60%, +60%].

[0123] When the first factor is less than 0.3, the universe of discourse shrinks to [-40%, +40%] to prevent over-control in a low-coupling state.

[0124] S3.2.2: Energy consumption index belongs to the center point offset.

[0125] Specifically, the membership function center point of the energy consumption index is adjusted from the default 0.5, and the expression is:

[0126]

[0127] Among them, P center The energy consumption index belongs to the center point, and η is the offset coefficient (taken as 0.01).

[0128] When R dust <R th When the specific resistance is low, the energy consumption index shifts to the left, prioritizing energy reduction. When R... dust ≥R th The energy consumption index shifts to the right of the center point, allowing for increased energy consumption to maintain dust removal efficiency.

[0129] S3.2.3: Add dielectric compensation term.

[0130] Specifically, an exponential compensation term is added to the membership degree calculation, with the following expression:

[0131]

[0132] Among them, K comp This is an exponential compensation term.

[0133] When R dust Less than R th At that time, K comp →0, amplify the control quantity to suppress back corona.

[0134] When R dust ≥R th At that time, K comp ≈1, the compensation item is invalid.

[0135] S3.3: Reconstruct the control rule base based on the first and second factors.

[0136] S3.3.1: Assigning rule weights.

[0137] Specifically, if the first factor is greater than 0.7, the pulse control rule group (weight w) is activated. pulse =0.7), prioritize adjusting the corona pulse frequency.

[0138] If the first factor is less than or equal to 0.7, activate the basic control rule group (weight w). base =0.5).

[0139] When the second factor is greater than 0.8 (gray layer dielectric anomaly), increase the weight of the dielectric compensation term to w. comp =0.6, forcibly reducing the intensity of the vibration.

[0140] Furthermore, the pulse control rule group rapidly suppresses electric field fluctuations through high-frequency, short-duration corona pulses, preventing a decrease in dust removal efficiency or the risk of back corona caused by multi-physical field coupling, including:

[0141] When the flue gas concentration deviation is PB, the pulse frequency is increased to enhance dust removal efficiency;

[0142] When the charge field non-uniformity is high and the resistivity of the gray layer is low, the pulse width is shortened to reduce the risk of back corona.

[0143] The basic control rules group is responsible for maintaining a balance between dust removal efficiency and energy consumption under normal operating conditions, avoiding over-control, including:

[0144] When the flue gas concentration deviation is PM and the energy consumption index is M (medium energy consumption, energy consumption index 0.3), <P norm When the value is less than 0.7, the corona voltage is slightly increased to optimize efficiency;

[0145] When the resistivity of the dust layer is normal and the risk of back corona is low, extend the rapping cycle to reduce secondary dust generation.

[0146] S3.3.2: Set anti-corona risk suppression rules.

[0147] Combining the resistivity of the ash layer and the non-uniformity of the charge field, the back corona risk coefficient is calculated, and the expression is:

[0148]

[0149] Among them, K risk This represents the risk factor for anti-corona discharge.

[0150] When K risk When the value is greater than 0.5, the risk of back corona is considered low, and the membership output of the rapping intensity is suppressed to 50% of the original value to prevent secondary dust generation.

[0151] Ideally, by collecting flue gas concentration deviation and energy consumption indicators, and combining the first and second factors to dynamically reconstruct the fuzzy membership function, a more precise electrostatic precipitator (ESP) control strategy is achieved. Employing fuzzy domain adjustment based on coupling strength improves the method's adaptability to different operating conditions; dynamic offsetting of the energy consumption membership center point balances dust removal efficiency and energy cost; and the addition of a dielectric compensation term strengthens control under low resistivity conditions, effectively suppressing the risk of back corona. A hierarchical control rule base is constructed to optimize the pulse control mechanism in high-intensity coupling scenarios, while maintaining a balanced adjustment of energy consumption and dust removal efficiency under normal operating conditions. By calculating the back corona risk coefficient and dynamically suppressing the rapping intensity, the risk of secondary dust generation is reduced, improving the method's stability and operating efficiency. While ensuring dust removal effectiveness, energy consumption is reduced, enhancing the intelligence level of the ESP.

[0152] S4: Input the flue gas concentration deviation and energy consumption index into the control rule base, execute multi-objective fuzzy decision-making, and output collaborative control parameters.

[0153] Specifically, it includes the following steps:

[0154] S4.1: Fuzzyize the input variables.

[0155] S4.1.1: Fuzzification of flue gas concentration deviation (E).

[0156] Specifically, based on the reconstructed universe of discourse, it is divided into 5 fuzzy sets:

[0157] NB (Negative Large): E≤-40%;

[0158] NM (Negative Medium): -40% <E≤-20%;

[0159] ZE (zero): -20% <E<+20%;

[0160] PM (center): +20% ≤ E < +40%;

[0161] PB (Chia Tai): E≥+40%;

[0162] The triangular membership function is used to map the flue gas concentration deviation to the membership degree of a fuzzy set. For example, the membership degree expression for the center (PM) is:

[0163]

[0164] Where, μ PM (E) represents the membership degree of the flue gas concentration deviation PM, max is the maximum value function, and E is the flue gas concentration deviation.

[0165] S4.1.2: Energy consumption index (P) norm Blurring.

[0166] Specifically, based on energy consumption indicators, three fuzzy sets are defined:

[0167] L (Low Energy Consumption): P norm ≤0.3;

[0168] M (medium energy consumption): 0.3 <P norm <0.7;

[0169] H (High Energy Consumption): P norm ≥0.7;

[0170] Based on the membership center of the energy consumption index, a trapezoidal membership function is used to map the energy consumption index to the membership degree of a fuzzy set. For example, the membership degree expression for medium energy consumption (M) is:

[0171]

[0172] Where, μ M (P norm ) represents the membership degree of energy consumption index M, min is the minimum value function, and P norm This is an energy consumption indicator.

[0173] Furthermore, for the flue gas concentration deviation variable, a triangular membership function is used to enhance the sensitivity to the deviation trend and achieve a continuous and smooth control response; while for the energy consumption index variable, a trapezoidal membership function is used with a plateau interval to enhance overall robustness, avoid frequent control adjustments caused by energy consumption fluctuations, and thus ensure the stability and practical adaptability of the fuzzy control rules.

[0174] S4.2: Generate coordinated control parameters.

[0175] It should be noted that the final control quantity is calculated using the center-of-gravity method, taking into account flue gas concentration deviation, energy consumption indicators, and back corona risk, to output coordinated control parameters. These coordinated control parameters include corona voltage, pulse frequency, and rapping cycle.

[0176] Ideally, by fuzzifying the flue gas concentration deviation and energy consumption indicators, and combining this with the reconstructed control rule base, more precise collaborative control of the electrostatic precipitator (ESP) is achieved. Optimized fuzzy set partitioning and adaptive membership functions improve adaptability to different operating conditions. Through multi-objective fuzzy decision-making, considering dust removal efficiency, energy consumption optimization, and back corona risk, the control strategy maintains high efficiency and robustness under dynamically changing conditions. Finally, the center-of-gravity method is used to calculate collaborative control parameters, enabling adaptive adjustment of corona voltage, pulse frequency, and rapping cycle. This allows the ESP to minimize energy consumption and reduce back corona phenomenon while ensuring low emissions, thereby improving overall operating efficiency and stability.

[0177] S5: Trigger the first factor iteration based on the change in the dielectric parameter of the ash layer, and adjust the second factor in reverse based on the dust removal efficiency deviation to optimize the rule weights.

[0178] Specifically, it includes the following steps:

[0179] S5.1: Perform the first factor iteration based on the dielectric parameters of the gray layer.

[0180] Specifically, when the rate of change of the resistivity of the ash layer is greater than 10... 9 When the abrupt change in the real and imaginary parts of the Ω·cm / min or gray layer dielectric parameter exceeds 20%, the weighting iteration of the first factor is triggered.

[0181] Specifically, a loss function is constructed based on the correlation between the real and imaginary parts of the gray layer dielectric parameters and the first factor, and its expression is as follows:

[0182] L F1 =α1·(∈′-∈′) ref ) 2 +α2·(∈″-∈″ ref ) 2 ;

[0183] Among them, L F1 Let be the first factor iterative loss function, α1 be the real part iterative weight (taken as 0.6), α2 be the imaginary part iterative parameter (taken as 0.4), ∈′ ref For the real part reference value, ∈″ ref This is a reference value for the imaginary part, obtained through statistical analysis of historical stable operating data.

[0184] Update w using gradient descent G ,w V ,w UFirst, the weights w of the three physical fields are adjusted according to the loss function. G ,w V ,w U Calculate the partial derivatives for each factor and use gradient descent to calculate the update. After each update, normalize the new weight values ​​so that the sum of the three is 1, thus ensuring the interpretability and stability of the weighted combination of the first factor.

[0185] The expression is:

[0186]

[0187] Where, η GD The learning rate is set to 0.01. Δw is the identifier for the partial derivative. i For w G ,w V ,w U The change value.

[0188] S5.2: Adjust the second factor in reverse based on the dust removal efficiency deviation.

[0189] Specifically, the expression for calculating the dust removal efficiency deviation is:

[0190]

[0191] Among them, E η For the deviation in dust removal efficiency, η target For the target dust removal efficiency (taken as 99.5%), C out C represents the concentration of the flue gas at the outlet. in This refers to the concentration of inlet flue gas.

[0192] By correcting the calculated coefficients β, γ, λ of the second factor through error feedback, the gray layer state characterization is optimized. The PID feedback expression is as follows:

[0193]

[0194] Among them, K p K is the PID proportional coefficient (taken as 0.5). q K is the integral coefficient of the PID controller (taken as 0.2). d For the PID differential coefficient (taken as 0.1), sign(E) η ) is the deviation sign function (positive deviation is +1, negative deviation is -1), Δβ is the change in β, Δγ is the change in γ, and Δλ is the change in λ.

[0195] Restricting β∈[0.3,0.7] and λ∈[1,3] prevents over-adjustment.

[0196] S5.3: Optimize rule weights based on the adjusted first and second factors.

[0197] Specifically, the weights of the adjustment rules are based on the updated first and second factors, expressed as follows:

[0198]

[0199] The constant 0.1 is used to avoid the denominator being zero, w pulse For the weights of the pulse control rule group, w comp For the dielectric compensation term weight, w base The weights of the basic control rule group.

[0200] When the weight change rate When using this method, a first-order low-pass filter is employed for smoothing.

[0201] If E η If the weight exceeds 5% for 5 minutes, the weight will be forcibly reset to the default value, and an alarm will be triggered.

[0202] Ideally, dynamic correction of the dielectric state of the ash layer and intelligent optimization of the control strategy are achieved through iterative first factor, reverse adjustment of the second factor, and rule weight optimization. Based on abrupt changes in the ash layer resistivity and dielectric parameters, the weight adjustment of the first factor is triggered, and the multi-physics coupling strength is optimized using gradient descent, making the method more sensitive to changes in the dielectric environment. According to the dust removal efficiency deviation, a PID feedback mechanism is used to dynamically correct the calculated coefficients of the second factor, ensuring the accuracy of the ash layer state characterization and improving the stability of dust removal efficiency. The weight allocation of the control rules is dynamically adjusted based on the first and second factors, ensuring that the control strategy can adaptively adjust under different operating conditions, avoiding over-control or response lag. Low-pass filtering smooths weight changes, and alarms and resets are triggered when abnormal deviations persist, enhancing overall robustness and reliability.

[0203] This embodiment also provides a computer device applicable to the fuzzy control method for stable operation of an electrostatic precipitator system, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the fuzzy control method for stable operation of the electrostatic precipitator system as proposed in the above embodiment.

[0204] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0205] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the fuzzy control method for achieving stable operation of the electrostatic precipitator system as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0206] In summary, this invention achieves adaptive adjustment under complex operating conditions by combining multi-physics coupling analysis with dynamic detection of gray layer dielectric parameters to generate a two-factor collaborative optimization mechanism; it improves the system's response accuracy to physical field coupling effects by using a three-dimensional sensor array to collect multi-source data in real time, and generating dynamic control factors through normalization processing and weighted fusion; it uses a gray layer resistivity compensation mechanism and an anti-corona risk suppression strategy, and utilizes an exponential function to dynamically correct the membership function and control rule base, effectively reducing energy consumption and corona breakdown risk while ensuring dust removal efficiency; and iterative optimization of control parameters is achieved through gradient descent and error feedback mechanisms, enabling the system to maintain stable operation even when gray layer dielectric properties change abruptly.

[0207] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A fuzzy control method for stable operation of an electrostatic precipitator system, characterized in that: include, The first factor is generated based on multiphysics coupling analysis, wherein the multiphysics includes temperature field gradient, airflow vorticity and charge field inhomogeneity. Simultaneously collect the dielectric parameters of the ash layer, calculate the specific resistivity of the ash layer, and generate the second factor; Collect flue gas concentration deviation and energy consumption index, and dynamically reconstruct the fuzzy membership function using the first and second factors to generate a control rule base containing dielectric compensation terms. Input the flue gas concentration deviation and energy consumption index into the control rule base, execute multi-objective fuzzy decision-making, and output collaborative control parameters; The first factor iteration is triggered by the change in the dielectric parameter of the ash layer, and the second factor is adjusted in reverse according to the dust removal efficiency deviation to optimize the rule weights. The specific steps for dynamically reconstructing the fuzzy membership function are as follows: The fuzzy universe of discourse range of the flue gas concentration deviation is dynamically adjusted according to the first factor, so that the fuzzy universe of discourse range expands and contracts linearly with the physical field coupling strength. The membership function center point of the energy consumption index is shifted according to the second factor, and the shift amount is proportional to the square of the ash layer resistivity. The resistivity of the ash layer is detected, and an exponential compensation term is added to the membership calculation to enhance the control sensitivity under low resistivity conditions. The triggering of the first factor iteration refers to adjusting the weight coefficient of the first factor using the gradient descent method to match the changing trend of the first factor and the dielectric parameter of the gray layer. The loss function is constructed based on the correlation between the real and imaginary parts of the gray layer dielectric parameter and the first factor, and its expression is: L F1 ＝α1·(∈′-∈′ ref ) 2 +α2·(∈”-∈″ ref ) 2 ； Among them, L F1 Let α1 be the real part of the iterative loss function, α2 be the imaginary part of the iterative parameter, and α' be the real part of the iterative loss function. ref For the real part reference value, ∈″ ref ∈ is the imaginary reference value, ∈′ is the real part of the gray layer dielectric parameter, and ∈” is the imaginary part of the gray layer dielectric parameter; Update w using gradient descent G ,w V ,w U ; Based on the loss function, the weights w of the three physical fields G ,w V ,w U Find the partial derivatives of w respectively. G For the temperature field weights, w V For the weight of airflow vorticity, w U For charge field weights; The reverse adjustment of the second factor and optimization of the rule weight refers to adjusting the calculation coefficient of the second factor based on the deviation between the actual dust removal efficiency and the target dust removal efficiency through an error feedback mechanism, regenerating the second factor and updating the weight allocation in the control rule base. The expression for adjusting the rule weights based on the updated first and second factors is: The constant 0.1 is used to avoid the denominator being zero, w pulse For the weights of the pulse control rule group, w comp For the dielectric compensation term weight, w base The weights of the basic control rule group are: F1 is the first factor, and F2 is the second factor.

2. The fuzzy control method for stable operation of the electrostatic precipitator system as described in claim 1, characterized in that: The first factor is obtained by normalizing the temperature field gradient, airflow vorticity and charge field non-uniformity and then calculating using weighted summation. The temperature field gradient is collected in real time by a distributed temperature sensor array, the airflow vorticity is determined by multi-point measurements of the electric field cross section by eddy current sensors, and the charge field non-uniformity is calculated based on the spatial distribution data of the charge density probe.

3. The fuzzy control method for stable operation of the electrostatic precipitator system as described in claim 1, characterized in that: The second factor is dynamically generated based on the real and imaginary parts of the gray layer dielectric parameters. When the gray layer resistivity is lower than the preset back corona risk threshold, the compensation mechanism is activated, and the amplitude of the second factor is adjusted by an exponential function. The gray layer resistivity is obtained by inversion calculation of the dielectric loss parameters collected by the microwave resonant cavity sensor.

4. The fuzzy control method for stable operation of the electrostatic precipitator system as described in claim 2, characterized in that: The multiphysics coupling analysis refers to the use of three-dimensional synchronous acquisition of temperature gradient sensor, eddy current sensor and charge density probe, real-time output of maximum temperature field gradient, airflow vortex integral value and electric field inhomogeneity relative value, and linear superposition of the three after normalization to generate the first factor.

5. The fuzzy control method for stable operation of the electrostatic precipitator system as described in claim 1, characterized in that: The specific steps for performing multi-objective fuzzy decision-making are as follows: The control rule group is selected based on the threshold. When the first factor exceeds the first preset threshold, the pulse-dominated rule group is activated; when the second factor exceeds the second preset threshold, the dielectric compensation rule group is activated. Calculation of back corona risk coefficient based on gray layer dielectric parameters; The parameters for coordinated control are generated by taking into account the deviation of flue gas concentration, energy consumption indicators and the risk of back corona discharge.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the fuzzy control method for stable operation of the electrostatic precipitator system according to any one of claims 1 to 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the fuzzy control method for stable operation of the electrostatic precipitator system according to any one of claims 1 to 5.

Images