Intelligent electrostatic dust collection system suitable for multiple working conditions

Abstract

The invention discloses an intelligent electrostatic dust collection system suitable for multiple working conditions. Belongs to the technical field of environmental protection and energy comprehensive utilization. The sensor module monitors environmental parameters in real time, and a concentration sensor detects the dust concentration of a smoke inlet; the temperature sensor detects the temperature of flue gas at the flue gas absorption port; the flow velocity sensor detects the flow velocity of the smoke; the intelligent control system accurately analyzes various data of the sensor and automatically adjusts the pipeline state, the electric field size and the rapping frequency, so that the electrostatic dust collector is kept in the optimal operation state; the reaction module adopts a high-voltage power supply technology for providing stable high-voltage direct current, accurately receiving an instruction of a control system and quickly adjusting effective output voltage and current, so that an efficient dust removal effect can be kept under various complex working conditions; and meanwhile, the electrode plate is made of a special electrode material and an electrode structure, so that the uniformity and strength of an electric field are improved, and the charging effect and collection efficiency of dust are enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of environmental protection and comprehensive utilization of energy, and involves intelligent control, intelligent cleaning of smoke in factories, and a factory treatment system in cooperation with ESP technology; specifically, it involves an intelligent electrostatic precipitator system applicable to multiple working conditions. Background Art

[0002] With the acceleration of the industrialization process, the problem of air pollution has become increasingly serious, especially the particulate matter in industrial emissions poses a huge threat to the environment and human health. In order to effectively control and reduce these pollutants, the adaptive electrostatic precipitation technology (ESP) has emerged as the times require and has become one of the indispensable air purification technologies in the industrial field.

[0003] The electrostatic precipitation technology is a technology that uses an electrostatic field to capture and separate particulate matter in the air. This technology makes dust particles charged and then uses the electric field force to adsorb the charged particles onto the dust collecting plate, thereby achieving the purpose of purifying flue gas; the ESP technology is widely used in industries such as coal-fired boilers, cement kilns, and petrochemicals due to its high particulate matter removal ability. Although the traditional ESP technology has achieved remarkable results in removing particulate matter, there are still some limitations. For example, when the environmental temperature and humidity change, or after the ESP absorbs ash, its dust removal efficiency will change, and the efficiency change can even reach 3 times. At the same time, the generation amount of ozone is also unstable. In addition, with the increase in the use time, a large amount of particulate matter will accumulate on the dust collecting plate of the electrostatic precipitator, which not only increases the corona voltage and affects the dust removal efficiency, but also requires regular manual cleaning, increasing the maintenance cost and operation risk.

[0004] In order to overcome the limitations of the traditional ESP technology, the adaptive ESP technology has emerged as the times require. This technology automatically adjusts the operating parameters through an intelligent control system to respond to the changes in working conditions and achieve the optimal dust removal effect. For example, some adaptive ESP systems can control the dust removal effect and the generation amount of ozone by changing the voltage output, thereby effectively controlling the dust removal effect and the generation amount of ozone.

[0005] Although the adaptive ESP technology has made certain progress, it still faces technical challenges such as improving the dust removal efficiency, reducing energy consumption, and reducing the maintenance cost. Future research and development will focus on improving the intelligent control level of ESP, optimizing the electrode structure design, developing new and efficient electrode materials, and improving the automation and informatization level of the system.

[0006] In summary, as an efficient air purification technology, the adaptive electrostatic precipitation technology has broad application prospects in the industrial field. With the continuous progress and innovation of technology, the adaptive ESP technology will play an increasingly important role in environmental protection and air quality improvement. Summary of the Invention

[0007] In view of the above problems, the object of the present invention is to provide an intelligent electrostatic dust removal system applicable to multiple working conditions, aiming to break through the limitations of traditional electrostatic dust removal technology, improve the adaptability to changes in dust properties, temperature and humidity, effectively improve the stability and efficiency of dust removal, and save energy and manpower losses for factories and extend the service life of equipment.

[0008] The technical solution of the present invention is as follows: An intelligent electrostatic dust removal system applicable to multiple working conditions described in the present invention includes an air inlet pipe, a sensor module, a reaction module and a central control system which are connected to each other;

[0009] The air inlet pipe includes a relatively wide air pipe opening and a pretreatment system, which is used to directly intercept larger particulate matters in the air, such as dust, hair, fibers, etc., and at the same time plays a role in balancing the air flow and optimizing the dust distribution, making the dust more uniform and the air flow more stable, and extending the service life of the equipment and reducing the maintenance frequency of the equipment;

[0010] The sensor module includes multiple temperature sensors, flow velocity sensors and concentration sensors. The above sensors are installed in the ventilation pipeline to detect the actual temperature, flow velocity and concentration of the flue gas in the ventilation pipeline and transmit them to the central control module;

[0011] The central control system is centered on a multivariable control algorithm, and is combined with a fuzzy logic algorithm, a fault diagnosis and fault tolerance algorithm and an optimization scheduling algorithm;

[0012] The multivariable control algorithm: can simultaneously consider the influence of changes in multiple variables on the electrostatic dust removal effect, and realize the comprehensive regulation of multiple key parameters such as electric field strength, rapping frequency and fan speed; can real-time monitor the generation amount, properties of dust and changes in environmental conditions, and timely adjust the control strategy according to the changes, so that the electrostatic precipitator always maintains the best operating state and ensures efficient and stable dust removal effect;

[0013] The fuzzy logic algorithm: is used to process uncertainty and ambiguity, and enhance the adaptability and flexibility of the system;

[0014] The fault diagnosis and fault tolerance algorithm: is used to improve the reliability of the system and reduce the maintenance cost and risk;

[0015] The optimization scheduling algorithm: is used to reasonably utilize resources and improve the comprehensive performance of the system.

[0016] Furthermore, the reaction module includes a voltage regulating device, an intelligent pipeline and a rapping device;

[0017] The voltage regulating device and the intelligent pipeline include an electrostatic generating unit and a dust collecting unit;

[0018] The electrostatic generating unit is used to generate an electrostatic field, and includes a high-voltage power supply, an electrode assembly, and an insulating support component;

[0019] Among them, the high-voltage power supply outputs adjustable DC high voltage, with a range between 8 kV and 12 kV;

[0020] The electrode assembly is arranged in the dust removal channel and fixed on the outer housing through the insulating support component. It includes a discharge electrode and a dust collecting electrode that cooperate with each other. The discharge electrode is made of copper and its shape is one or a combination of needle-shaped, wire-shaped, or barbed-shaped. The tip curvature radius of the needle-shaped discharge electrode is 100 and 500 microns. The wire diameter of the wire-shaped discharge electrode is between 15 microns and a cross-sectional area of 22 square millimeters. The barb length of the barbed-shaped discharge electrode is between 20 and 50 millimeters. The dust collecting electrode is made of TMDs, graphene, black phosphorus, and MXene, and its shape is one or a combination of plate-shaped, cylindrical, or mesh-shaped. The thickness of the plate-shaped dust collecting electrode is between 0.1 and 1 millimeter. The wall thickness of the cylindrical dust collecting electrode is between 5 and 25 microns. The wire diameter of the mesh-shaped dust collecting electrode is between 0.2 millimeter and 0.25 millimeter, and the mesh size is between 0.31 and 0.79 millimeter;

[0021] The insulating support component is made of a high-strength porcelain support insulator;

[0022] The dust collecting unit is used to collect the dust adsorbed under the action of the electrostatic field and is connected to the dust removal channel. The dust collecting unit includes collecting components such as a dust collecting plate, a dust collecting cylinder, or a dust collecting bag. The surface of the dust collecting plate has a copper-graphene composite coating, which is used to enhance the dust adsorption ability and facilitate the sliding of dust. A spiral deflector is arranged inside the dust collecting cylinder to guide the dust to gather at the bottom. The dust collecting bag is made of a new type of metal decorative material and has air permeability, abrasion resistance, and anti-static performance.

[0023] Furthermore, the rapping device includes a rapping device body, a driving mechanism, and a control module;

[0024] The rapping device body is used to rap and clean the dust collecting unit of the electrostatic dust removal device. The rapping device body is in contact with the dust collecting unit, and its rapping contact surface is made of titanium. This material has high hardness, high wear resistance, and good elasticity to ensure effective transmission of the rapping force during the rapping process and is not easily damaged;

[0025] The driving mechanism is connected to the main body of the rapping device and is used to provide rapping power for the main body of the rapping device. The driving mechanism is one of the electric, pneumatic or hydraulic driving methods. If it is an electric drive, its motor adopts a YZU series three-phase AC asynchronous motor, and the rated power of the motor is between 0.25 and 23.5 kW, the speed range is between 0 and 180 r / min, and the torque is between 0.84 and 23.5 N*m; if it is a pneumatic drive, its air source pressure is between 0.2 and 0.7 MPa, the piston diameter of the cylinder is between 8 mm and 320 mm, and the stroke is between 10 and 1500 mm; if it is a hydraulic drive, its hydraulic source pressure is between 10 and 31.5 MPa, the piston diameter of the hydraulic cylinder is between 8 and 320 mm, and the stroke is between 320 and 2000 mm;

[0026] The control module is electrically connected or signal-connected to the driving mechanism (determined according to the driving type) and is used to control the operating parameters of the driving mechanism, including the rapping frequency, rapping intensity and rapping duration; the control module includes a microprocessor and a storage unit storing a rapping control program. The microprocessor adopts a GEA electrostatic precipitator, and the rapping control program adjusts the rapping parameters according to the operating conditions of the electrostatic dust removal device and the dust accumulation situation of the dust collection unit.

[0027] Further, the multivariable control algorithm further includes a prediction model and an optimal control model;

[0028] Among them, the prediction model includes a state space model and a prediction equation.

[0029] Further, the operating steps of the state space model and the prediction equation are as follows:

[0030] (1). Establish a state space model:

[0031] Step 1: Select the dust concentration C, temperature T, humidity RH and some variables related to the equipment operating state (electric field strength E f , rapping frequency f) as state variables, denoted as x = [C, T, RH, E f , f] T ;

[0032] Step 2: The input variables of the system include the waste gas flow rate Q, the waste gas inlet temperature T i n, the dust generation rate P, denoted as u = [Q, T i n, P] T ;

[0033] The output variables of the system include the dust emission concentration C out , the dust removal efficiency η, denoted as y = [C o ut, η] T ;

[0034] Step 3: The state-space model of the system is expressed as: In the formula, is the derivative of the state variable, representing the rate of change of the state variable with time; A is the state matrix, describing the mutual relationship between the internal state variables of the system and its own dynamic characteristics; B is the input matrix, indicating the influence of the input variable on the state variable; w is the disturbance term of the system, including some uncertain factors and noises not considered in the model, etc.; y = Cx + Du + v; in the formula, C is the output matrix, describing the relationship between the state variable and the output variable; D is the direct transmission matrix, representing the direct influence of the input variable on the output variable; v is the measurement noise term.

[0035] For a simple case, assuming that the change in dust concentration is mainly affected by the waste gas flow rate and the electric field intensity, the change in temperature is mainly affected by the inlet temperature of the waste gas and the electric field energy loss, and the change in humidity is mainly affected by the moisture content in the waste gas and the temperature, the state matrix A, the input matrix B, the output matrix C, and the direct transmission matrix D are set as:

[0036]

[0037]

[0038]

[0039]

[0040] In the formula, S is the waste gas flow rate, η is the efficiency,

[0041] qv is the volume flow rate, V is the volume, Q out is the output heat, Q loss is the heat loss,

[0042] C p is the specific heat capacity at constant pressure, E v is the evaporation energy, C n is a constant, R is the gas constant, RH is the relative humidity, T in is the inlet temperature, α is the thermal diffusivity;

[0043] (2) Establish the prediction equation:

[0044] Step 1: In model predictive control, it is necessary to predict the output of the system within a certain period in the future based on the current state and input; assume the prediction time domain is N p , and the control time domain is N c (usually N c ≤N p );

[0045] Step 2: For a discrete-time system, assume the sampling time is Ts , at the k-th moment, the state of the system is updated through a discretized state space model:

[0046] x(k + 1) = Ax(k) + Bu(k);

[0047] Step 3: From this, the output of the next N p steps can be predicted:

[0048]

[0049] where, represents the predicted output value of the i-th step at the k-th moment.

[0050] Furthermore, the specific steps of the optimal control model are as follows:

[0051] Step 1: Construct the objective function:

[0052] (1): The objective function usually aims to minimize a performance index. For example, the deviation of the dust emission concentration and the energy consumption can be considered simultaneously. Let the set value of the dust emission concentration be C s et, and the actual predicted dust emission concentration sequence is:

[0053]

[0054] Let the electric field strength be E f , and the fan speed be n (unit: rpm). The energy consumption function per unit time is E(E f , n);

[0055] (2): Then the objective function is expressed as:

[0056]

[0057]

[0058] where, ω1 and ω2 are weight coefficients, which are used to adjust the relative importance of the dust emission concentration deviation and the energy consumption in the objective function. The selection of the weight coefficients needs to be weighed according to the actual process requirements and economic factors;

[0059] Step 2: Set the constraint conditions:

[0060] (1), Electric field strength constraint: E f ,min ≤ E f ≤ E f,max ; where, E f,min and E f,max are the lower and upper limits of the electric field strength respectively; this is determined by the electrical characteristics and safety requirements of the equipment;

[0061] (2) Fan speed constraint: n min ≤n≤n max ; where n min and n max are the allowable ranges of the fan speed;

[0062] (3) Rapping frequency constraint: f m in≤f≤f max ; where f is the rapping frequency;

[0063] Process requirement constraint:

[0064] (4) Dust emission concentration constraint:

[0065] where C max is the maximum allowable dust emission concentration, which should meet the environmental protection standards and production requirements;

[0066] (5) Gas flow rate constraint: Q min ≤Q≤Q max ; where Q is the exhaust gas flow rate, which should be within the processing capacity of the equipment and also consider the stability of the production process.

[0067] Furthermore, the fuzzy logic algorithm performs fuzzy processing on the output quantity, and the specific steps are as follows:

[0068] Step 1: Fuzzification of input and output variables:

[0069] (1): For the dust concentration deviation e C (the difference between the actual dust concentration C and the set value C s et), define the fuzzy set as:

[0070] {NB (Negative Big), NS (Negative Small), Z (Zero), PS (Positive Small), PB (Positive Big)};

[0071] (2): Assume that the measurement range of the dust concentration is [C min , C max , then divide it into several intervals corresponding to different fuzzy sets:

[0072] NB: NS: Z: PS: PB:

[0073] Use the membership function to determine the degree to which each actual dust concentration deviation value belongs to each fuzzy set. Commonly used membership functions include triangular functions, trapezoidal functions, Gaussian functions, etc. For the triangular membership function, when e CWhen in the NB interval, its membership function is expressed as:

[0074]

[0075] (3): For the temperature deviation e T (the difference between the actual temperature T and the set temperature T set ), the fuzzy set is {NB (Negative Big), NS (Negative Small), Z (Zero), PS (Positive Small), PB (Positive Big)}. Interval division is carried out according to the reasonable range of temperature, and the corresponding membership function is determined;

[0076] (4): The fuzzy sets of the output variables (such as the electric field strength adjustment amount ΔE f and the rapping frequency adjustment amount Δf) are defined accordingly; that is, the fuzzy set of the electric field strength adjustment amount can be {NB (Negative Big, that is, significantly reduce the electric field strength), NS (Negative Small, that is, appropriately reduce the electric field strength), Z (Zero, that is, keep the electric field strength unchanged), PS (Positive Small, that is, appropriately increase the electric field strength), PB (Positive Big, that is, significantly increase the electric field strength)};

[0077] The fuzzy set of the rapping frequency adjustment amount can be {NB (significantly reduce the rapping frequency), NS (appropriately reduce the rapping frequency), Z (keep the rapping frequency unchanged), PS (appropriately increase the rapping frequency), PB (significantly increase the rapping frequency)};

[0078] Step Two: Establish a fuzzy rule base:

[0079] Rule 1: If the dust concentration deviation is NB and the temperature deviation is NB, then the electric field strength adjustment amount is PB and the rapping frequency adjustment amount is PB;

[0080] Rule 2: If the dust concentration deviation is NB and the temperature deviation is Z, then the electric field strength adjustment amount is PB and the rapping frequency adjustment amount is PS;

[0081] Rule 3: If the dust concentration deviation is NS and the temperature deviation is PS, then the electric field strength adjustment amount is PS and the rapping frequency adjustment amount is Z;

[0082] Rule 4: If the dust concentration deviation is Z and the temperature deviation is PS, then the electric field strength adjustment amount is Z and the rapping frequency adjustment amount is PS;

[0083] Rule 5: If the dust concentration deviation is PS and the temperature deviation is PB, then the electric field strength adjustment amount is NS and the rapping frequency adjustment amount is PB;

[0084] Step Three: Fuzzy inference and defuzzification:

[0085] (1) Fuzzy inference:

[0086] The Mamdani inference method is used to determine the fuzzy quantity of the output. For the membership values of the given dust concentration deviation and temperature deviation, inference is carried out according to the fuzzy rules;

[0087] Suppose the membership of the current dust concentration deviation is μ NB (e C ) = 0.6,

[0088] and the membership of the temperature deviation is μ Z (e T ) = 0.8. For Rule 2 (If the dust concentration deviation is NB and the temperature deviation is Z, then the electric field strength adjustment amount is PB and the rapping frequency adjustment amount is PS), the triggering strength of this rule is min(0.6, 0.8) = 0.6 o ;

[0089] For the fuzzy set PB of the electric field strength adjustment amount, fuzzy inference calculation is carried out according to the triggering strength and the membership function to obtain the fuzzy contribution of this rule to the electric field strength adjustment amount; similarly, corresponding calculations are also carried out for the fuzzy set PS of the rapping frequency adjustment amount;

[0090] Similar inference calculations are carried out for all rules to obtain the fuzzy contributions of each rule to the electric field strength adjustment amount and the rapping frequency adjustment amount;

[0091] (2) Defuzzification:

[0092] The centroid method is used for defuzzification. For the fuzzy set of the electric field strength adjustment amount, its centroid is calculated as the final precise adjustment amount;

[0093] Suppose the fuzzy set of the electric field strength adjustment amount is {(ΔE f1 , μ1), (ΔE f2 , μ2), …, (ΔE fn , μ n )}, where ΔE fi is an element in the fuzzy set (i.e., different values of the electric field strength adjustment amount), and μ i is its corresponding membership;

[0094] Then the defuzzified electric field strength adjustment amount

[0095] Furthermore, the fault diagnosis and fault tolerance algorithm includes multiple detection models, and its specific features are as follows:

[0096] Step 1, Fault diagnosis part:

[0097] (1) Parameter monitoring model:

[0098] Suppose there are several key parameters to be monitored in the electrostatic precipitation system, such as current I, voltage U, temperature T, and dust concentration C. Under normal operating conditions, these parameters have an expected range;

[0099] Taking the current as an example, assume that the current range during normal operation is I min ≤I≤I max , then establish a simple judgment function f I (I):

[0100]

[0101] For voltage U, temperature T, and dust concentration C, judgment functions f U (U), f T (T), and f C (C) can be established respectively;

[0102] (2), Fault comprehensive judgment model:

[0103] Suppose the fault types are divided into short - circuit fault F1, open - circuit fault F2, temperature - anomaly fault F3, and dust - concentration - anomaly fault F4; a fault vector F=(F1,F2,F3,F4) can be defined, and the values of its elements are determined by the above - mentioned parameter judgment functions; for example, when f I (I)=1 and other parameters are normal, it may be suspected of short - circuit or open - circuit faults. At this time, the value of F1 or F2 can be set to 1 (indicating that this fault may occur); the specific judgment rules are determined according to the actual physical process and experience. For example:

[0104]

[0105]

[0106]

[0107]

[0108] Step 2, Fault - tolerance algorithm part:

[0109] (1), Fault compensation model:

[0110] When an abnormal current is detected (assuming it is a short - circuit fault, F1 = 1), in order to maintain the dust - removal effect, the voltage can be adjusted; assume that under normal conditions, there is a relationship between the dust - removal efficiency η, current I, and voltage U: η = g(I,U);

[0111] Under fault conditions, by adjusting the voltage U ′ try to keep the dust - removal efficiency;

[0112] Suppose a compensation function U has been obtained through experiments′ = h(I, U, F1), when F1 = 1, U ′ The value of is adjusted according to the deviation degree of the current I and the current voltage U, and the equation is In the formula, k is a compensation coefficient, and its appropriate value is determined through experiments or simulations, so that the voltage can be appropriately increased to maintain the dust removal efficiency when the current is abnormal;

[0113] (2) System reliability evaluation model:

[0114] Assume that the system reliability R is related to the fault situation; it can be simply defined as:

[0115]

[0116] In the formula, w i is the weight corresponding to each fault type, indicating the influence degree of different faults on the system reliability;

[0117] The short - circuit fault may have a greater impact on the system. w1 can be set to 0.4, the open - circuit fault w2 = 0.3, the temperature - anomaly fault w3 = 0.2, and the dust - concentration - anomaly fault w4 = 0.1 o .

[0118] Furthermore, the optimization scheduling algorithm is composed of an objective function, constraint conditions, and an optimization algorithm model. The specific steps are as follows:

[0119] Step 1: Construction of the objective function:

[0120] (1) Energy consumption minimization:

[0121] Assume that the electrostatic precipitator has n working stages, and the power of each stage i is P i , and the working time is t i ; The total energy consumption E is:

[0122]

[0123] The goal is to minimize E while meeting the dust removal requirements;

[0124] (2) Dust removal efficiency maximization:

[0125] Assume that the dust removal efficiency is η, which is a function of factors such as current I, voltage U, and wind speed v, η = f(I, U, v); during the operation of the equipment, make η as large as possible; define a comprehensive objective function J, combining energy consumption and dust removal efficiency:

[0126]

[0127] Where η is the dust removal efficiency, E is the energy consumption, and α is a weight coefficient (0 < α < 1) used to balance the importance of energy consumption and dust removal efficiency;

[0128] Step 2: Set the constraint conditions:

[0129] (1) Equipment parameter constraints:

[0130] Current constraint: I min ≤ I ≤ I max , voltage has U m in ≤ U ≤ U max , wind speed has v min ≤ v ≤ v max ;

[0131] (2) Dust removal requirement constraints:

[0132] Let the required minimum dust removal efficiency be η min , then η ≥ η min ;

[0133] (3) Time allocation constraints:

[0134] The total working time T is fixed, ∑ i = 1 n t i = T, and the time t of each stage i ≥ 0;

[0135] Step 3: Optimize the algorithm model:

[0136] (1) Dynamic programming model:

[0137] Let J k (x) represent the optimal objective function value in the state x (current equipment parameters) after k stages;

[0138] The state transition equation is:

[0139]

[0140] Where U is the decision set (operations to adjust current, voltage, wind speed, etc. at each stage), x ′ is the state of the previous stage, and r(x ′ , u) is the immediate benefit (reduction in energy consumption or improvement in dust removal efficiency) obtained by taking the decision u in the state x ′ ;

[0141] (2) Genetic algorithm model:

[0142] a. Chromosome encoding:

[0143] Assume that the scheduling scheme is determined by parameters such as the current value, voltage value of each working stage, and the working time of each stage; for example, each parameter is binary - encoded, and then these encodings are concatenated to form a chromosome. Suppose the electrostatic precipitator has 3 working stages, and for each stage, 3 parameters, namely current \(I\), voltage \(U\), and working time \(t\), are considered; if the value range of current \(I\) is \([0, 100]\text{A}\) and the accuracy requirement is \(1\text{A}\), it needs to be represented by 7 - bit binary numbers (because \(2 7 =128>100\)); the value range of voltage \(U\) is \([0, 1000]\text{V}\), the accuracy is \(1\text{V}\), and it needs 10 - bit binary numbers to represent; the value range of working time \(t\) is \([0, 60]\) minutes, the accuracy is \(1\) minute, and it needs 6 - bit binary numbers to represent;

[0144] Then the length of a chromosome is:

[0145] (7 + 10+6)×3=69 bits;

[0146] Each segment of the chromosome corresponds to the current, voltage, and working time of one stage respectively;

[0147] b. Population initialization:

[0148] Randomly generate a certain number (set as \(N\)) of chromosomes to form the initial population; for example, \(N = 100\), and the genes of each chromosome are randomly generated; the parameter values represented by these randomly generated chromosomes are within a reasonable range, that is, the current, voltage, and working time should meet the physical limitations of the equipment (current \(I m in≤I≤I max );

[0149] c. Fitness function calculation:

[0150] Calculate the fitness of each chromosome according to the previously defined objective function For each chromosome (representing a scheduling scheme), first decode the current \(I i \), voltage \(U i \), and working time 1, 2, 3 represent 3 working stages from the binary encoding of the chromosome; then calculate the total energy consumption \(E=\sum i =1 3 P i ×t i , where the power \(P i \) can be calculated according to the current \(I i \) and voltage \(U i \) through the power characteristic formula of the equipment \(P = UI\);

[0151] The dust - removal efficiency \(\eta=f(I_1,U_1,t_1,I_2,U_2,t_2,I_3,U_3,t_3)\) is calculated according to the specific dust - removal efficiency function;

[0152] Finally, calculate the fitness value. The higher the fitness value, the better the scheduling scheme;

[0153] d. Selection operation:

[0154] Adopt the roulette wheel selection method; calculate the proportion of the fitness of each chromosome in the total fitness of the population as the probability of being selected;

[0155] For example, the fitness of chromosome j is F j , and the total fitness of the population is Then the probability that chromosome j is selected According to these probabilities, select N / 2 chromosomes to enter the next-generation population by randomly simulating the roulette wheel;

[0156] e. Crossover operation:

[0157] Pair the selected chromosomes and adopt single-point crossover. Randomly select a crossover point position k (1 < k < chromosome length); for a pair of chromosomes, exchange the genes after the crossover point position; for example, there are two chromosomes A and B, the crossover point is k = 30, the first 30 genes of chromosome A remain the same as the first 30 genes of chromosome B, and the last 39 genes of A are exchanged with the last 39 genes of B to obtain two new offspring chromosomes; the crossover probability is generally set to P c = 0.7, that is, there is a probability of P c to perform the crossover operation on a pair of chromosomes;

[0158] f. Mutation operation:

[0159] For each gene position of each chromosome, perform mutation with a certain mutation probability P m = 0.01; the mutation operation is to reverse the binary value of the gene position; for example, if a certain gene position was originally 0, it becomes 1 after mutation.

[0160] g. Termination condition judgment:

[0161] Repeat steps c to e until the termination condition is met; common termination conditions include: reaching the maximum number of iterations (for example, iterating 1000 times) or the optimal fitness value of the population has not changed significantly for several consecutive generations; when the termination condition is met, the scheduling scheme represented by the chromosome with the highest fitness in the population is the optimized result.

[0162] The beneficial effects of the present invention are as follows: By means of the sensor module, the central control system and the reaction module, the present invention improves the adaptability to the properties of dust, temperature and humidity changes; at the same time, a variety of intelligent control algorithms are adopted to precisely control the system, effectively improving the stability and efficiency of dust removal; and it saves energy and manpower losses for the factory and extends the service life of the equipment. Brief Description of the Drawings

[0163] Figure 1 is a schematic diagram of the system structure of the present invention;

[0164] Figure 2 is a schematic diagram of the operation logic of the system of the present invention;

[0165] Figure 3 is a flowchart of the fuzzy logic algorithm adopted by the present invention;

[0166] Figure 4 is a flowchart of the genetic model algorithm adopted by the present invention. Detailed Description of the Invention

[0167] The following further elaborates on the specific technical solutions of the present invention with reference to specific examples.

[0168] As Figure 1 shown, an intelligent electrostatic dust removal system applicable to multiple working conditions according to the present invention includes an air inlet pipe, a sensor module, a reaction module and a central control system which are connected to each other;

[0169] As Figure 2 shown, the air inlet pipe includes a relatively wide ventilation pipe opening and a pretreatment system, which is used to directly intercept larger particulate matters in the air, such as dust, hair, fibers, etc., and at the same time plays a role in balancing the air flow and optimizing the dust distribution, making the dust more uniform, the air flow more stable, extending the service life of the equipment and reducing the maintenance frequency of the equipment;

[0170] The sensor module includes multiple temperature sensors, flow rate sensors and concentration sensors. The above sensors are installed in the ventilation pipeline to detect the actual temperature, flow rate and concentration of the flue gas in the ventilation pipeline and transmit them to the central control module;

[0171] The central control system takes the multivariable control algorithm as the center, and adds the fuzzy logic algorithm, the fault diagnosis and fault tolerance algorithm and the optimal scheduling algorithm;

[0172] The multi-variable control algorithm: It can simultaneously consider the influence of changes in multiple variables on the electrostatic dust removal effect, and achieve the comprehensive regulation of multiple key parameters such as electric field strength, rapping frequency, and fan speed; it can monitor the generation amount, properties of dust, and changes in environmental conditions in real time, and adjust the control strategy in a timely manner according to the changes, so that the electrostatic precipitator always maintains the best operating state and ensures efficient and stable dust removal effect;

[0173] The fuzzy logic algorithm: It is used to handle uncertainties and ambiguities, and enhance the adaptability and flexibility of the system;

[0174] The fault diagnosis and fault tolerance algorithm: It is used to improve the system reliability and reduce the maintenance cost and risk;

[0175] The optimization scheduling algorithm: It is used to rationally utilize resources and improve the comprehensive performance of the system.

[0176] Further, the reaction module includes a voltage regulating device, an intelligent pipeline, and a rapping device;

[0177] The voltage regulating device and the intelligent pipeline include an electrostatic generating unit and a dust collecting unit;

[0178] The electrostatic generating unit is used to generate an electrostatic field, and includes a high-voltage power supply, an electrode assembly, and an insulating support component;

[0179] Among them, the high-voltage power supply outputs adjustable DC high voltage, with a range between 8 kV and 12 kV;

[0180] The electrode assembly is installed in the dust removal channel and fixed on the outer casing through the insulating support component. It includes a discharge electrode and a dust collecting electrode that cooperate with each other. The discharge electrode is made of copper, and its shape is one or a combination of needle-shaped, linear-shaped, or spike-shaped. The tip curvature radius of the needle-shaped discharge electrode is 100 and 500 microns, the wire diameter of the linear-shaped discharge electrode is between 15 microns and a cross-sectional area of 22 square millimeters, and the spike length of the spike-shaped discharge electrode is between 20 and 50 millimeters; the dust collecting electrode is made of TMDs, graphene, black phosphorus, and MXene, and its shape is one or a combination of plate-shaped, cylindrical-shaped, or mesh-shaped. The thickness of the plate-shaped dust collecting electrode is between 0.1 and 1 millimeter, the wall thickness of the cylindrical-shaped dust collecting electrode is between 5 and 25 microns, the wire diameter of the mesh-shaped dust collecting electrode is between 0.2 millimeter and 0.25 millimeter, and the mesh hole size is between 0.31 and 0.79 millimeter;

[0181] The insulating support component is made of high-strength porcelain support insulators;

[0182] The dust collection unit is used to collect the dust adsorbed under the action of the electrostatic field and is connected to the dust removal channel. The dust collection unit includes collection components such as dust collection plates, dust collection cylinders or dust collection bags. The surface of the dust collection plate has a copper-graphene composite coating, which is used to enhance the dust adsorption ability and facilitate the sliding of dust. The inside of the dust collection cylinder is provided with spiral guide vanes for guiding the dust to gather at the bottom. The dust collection bag is made of a new type of metal decorative material and has the properties of air permeability, wear resistance and anti-static.

[0183] Further, the rapping device includes a rapping device body, a driving mechanism and a control module;

[0184] The rapping device body is used to rap and clean the dust collection unit of the electrostatic dust removal device. The rapping device body is in contact with the dust collection unit, and its rapping contact surface is made of titanium, which has high hardness, high wear resistance and good elasticity to ensure effective transmission of the rapping force and not easy to be damaged during the rapping process;

[0185] The driving mechanism is connected to the rapping device body and is used to provide rapping power for the rapping device body. The driving mechanism is one of electric, pneumatic or hydraulic driving methods. If it is electric drive, its motor uses a YZU series three-phase AC asynchronous motor, and the rated power of the motor is between 0.25 - 23.5kW, the speed range is between 0 - 180r / min, and the torque is between 0.84 - 23.5N*m; if it is pneumatic drive, its air source pressure is between 0.2 - 0.7MPa, the piston diameter of the cylinder is between 8mm - 320mm, and the stroke is between 10 - 1500mm; if it is hydraulic drive, its hydraulic source pressure is between 10 - 31.5MPa, the piston diameter of the hydraulic cylinder is between 8 - 320mm, and the stroke is between 320 - 2000mm;

[0186] The control module is electrically connected or signal-connected to the driving mechanism (determined according to the driving type) and is used to control the operating parameters of the driving mechanism, including rapping frequency, rapping intensity and rapping duration; the control module includes a microprocessor and a storage unit storing a rapping control program. The microprocessor uses a GEA electrostatic precipitator, and the rapping control program adjusts the rapping parameters according to the operating conditions of the electrostatic dust removal device and the dust accumulation situation of the dust collection unit.

[0187] Further, the multivariable control algorithm further includes a prediction model and an optimal control model;

[0188] Among them, the prediction model includes a state space model and a prediction equation.

[0189] Further, the operation steps of the state space model and the prediction equation are as follows:

[0190] (1). Establish a state space model:

[0191] Step 1: Select the dust concentration C, temperature T, humidity RH, and some variables related to the equipment operating status (electric field strength E f , rapping frequency f) as state variables, denoted as x = [C, T, RH, E f , f] T ;

[0192] Step 2: The input variables of the system include the waste gas flow rate Q, the waste gas inlet temperature T i n, dust generation rate P, denoted as u = [Q, T i n, P] T ;

[0193] The output variables of the system include the dust emission concentration C out , dust removal efficiency η, denoted as y = [C o ut, η] T ;

[0194] Step 3: The state - space model of the system is expressed as: In the formula, is the derivative of the state variable, representing the rate of change of the state variable with time; A is the state matrix, describing the mutual relationship between the internal state variables of the system and its own dynamic change characteristics; B is the input matrix, indicating the influence of the input variable on the state variable; w is the disturbance term of the system, including some uncertain factors and noises not considered in the model, etc.; y = Cx + Du + v; in the formula, C is the output matrix, describing the relationship between the state variable and the output variable; D is the direct transfer matrix, representing the direct influence of the input variable on the output variable; v is the measurement noise term;

[0195] For a simple case, assuming that the change in dust concentration is mainly affected by the waste gas flow rate and the electric field strength, the change in temperature is mainly affected by the waste gas inlet temperature and the electric field energy loss, and the change in humidity is mainly affected by the moisture content in the waste gas and the temperature, the state matrix A, input matrix B, output matrix C, and direct transfer matrix D are set as:

[0196]

[0197]

[0198]

[0199]

[0200] In the formula, S is the waste gas flow rate, η is the efficiency,

[0201] qv is the volume flow rate, V is the volume, Q out is the output heat, Q loss is the heat loss,

[0202] C p is the specific heat capacity at constant pressure, E v is the evaporation energy, C n is a constant, R is the gas constant, RH is the relative humidity, T in is the inlet temperature, and α is the thermal diffusivity;

[0203] (2) Establish a prediction equation:

[0204] Step 1: In model predictive control, it is necessary to predict the output of the system for a period of time in the future based on the current state and input; assume the prediction horizon is N p , and the control horizon is N c (usually N c ≤N p );

[0205] Step 2: For a discrete-time system, assume the sampling time is T s , then at the k-th moment, the state of the system is updated through a discretized state-space model:

[0206] x(k + 1) = Ax(k) + Bu(k);

[0207] Step 3: From this, the output for the next N p steps can be predicted:

[0208]

[0209]

[0210] wherein, represents the predicted output value at the i-th step at the k-th moment.

[0211] Furthermore, the specific steps of the optimization control model are as follows:

[0212] Step 1: Construct an objective function:

[0213] (1): The objective function usually aims to minimize a performance index. For example, the deviation of the dust emission concentration and the energy consumption can be considered simultaneously; let the set value of the dust emission concentration be C s et, and the actual predicted dust emission concentration sequence is:

[0214]

[0215] Let the electric field strength be E f , the fan speed be n (unit: rpm), and the energy consumption function per unit time be E(E f , n);

[0216] (2): Then the objective function is expressed as:

[0217]

[0218] In the formula, ω1 and ω2 are weight coefficients; they are used to adjust the relative importance of the dust emission concentration deviation and energy consumption in the objective function. The selection of the weight coefficients needs to be weighed according to the actual process requirements and economic factors;

[0219] Step Two: Set the constraint conditions:

[0220] (1) Electric field strength constraint: E f ,min ≤ E f ≤ E f,max ; In the formula, E f,min and E f,max are the lower and upper limits of the electric field strength respectively; this is determined by the electrical characteristics and safety requirements of the equipment;

[0221] (2) Fan speed constraint: n min ≤ n ≤ n max ; In the formula, n min and n max are the allowable ranges of the fan speed;

[0222] (3) Rapping frequency constraint: f m in ≤ f ≤ f max ; In the formula, f is the rapping frequency;

[0223] Process requirement constraint:

[0224] (4) Dust emission concentration constraint:

[0225] In the formula, C max is the maximum allowable dust emission concentration, which should meet the environmental protection standards and production requirements;

[0226] (5) Gas flow rate constraint: Q min ≤ Q ≤ Q max ; In the formula, Q is the waste gas flow rate, which should be ensured to be within the processing capacity of the equipment, and at the same time, the stability of the production process should also be considered.

[0227] As Figure 3 shown, further, the fuzzy logic algorithm performs fuzzy processing on the output quantity, and its specific steps are as follows:

[0228] Step One: Fuzzification of input and output variables:

[0229] (1): For the dust concentration deviation e C (actual dust concentration C - set value C s(Difference of et), the fuzzy sets are defined as:

[0230] {NB (Negative Big), NS (Negative Small), Z (Zero), PS (Positive Small), PB (Positive Big)};

[0231] (2): Assume the measurement range of dust concentration is [C min , C max , then it is divided into several intervals corresponding to different fuzzy sets:

[0232] NB: NS: Z: PS: PB:

[0233] The membership function is used to determine the degree to which each actual dust concentration deviation value belongs to each fuzzy set. Commonly used membership functions include triangular function, trapezoidal function, Gaussian function, etc. For the triangular membership function, when e C is within the NB interval, its membership function is expressed as:

[0234]

[0235] (3): For the temperature deviation e T (the difference between the actual temperature T and the set temperature T set ), the fuzzy sets are {NB (Negative Big), NS (Negative Small), Z (Zero), PS (Positive Small), PB (Positive Big)}. Interval division is carried out according to the reasonable range of temperature, and the corresponding membership function is determined;

[0236] (4): The fuzzy sets of the output variables (such as the electric field strength adjustment amount ΔE f and the rapping frequency adjustment amount Δf) are defined accordingly; that is, the fuzzy set of the electric field strength adjustment amount can be {NB (Negative Big, that is, significantly reduce the electric field strength), NS (Negative Small, that is, appropriately reduce the electric field strength), Z (Zero, that is, keep the electric field strength unchanged), PS (Positive Small, that is, appropriately increase the electric field strength), PB (Positive Big, that is, significantly increase the electric field strength)};

[0237] The fuzzy set of the rapping frequency adjustment amount can be {NB (significantly reduce the rapping frequency), NS (appropriately reduce the rapping frequency), Z (keep the rapping frequency unchanged), PS (appropriately increase the rapping frequency), PB (significantly increase the rapping frequency)};

[0238] Step 2: Establish a fuzzy rule base:

[0239] Rule 1: If the dust concentration deviation is NB and the temperature deviation is NB, then the electric field strength adjustment amount is PB and the rapping frequency adjustment amount is PB;

[0240] Rule 2: If the dust concentration deviation is NB and the temperature deviation is Z, then the electric field strength adjustment amount is PB, and the rapping frequency adjustment amount is PS;

[0241] Rule 3: If the dust concentration deviation is NS and the temperature deviation is PS, then the electric field strength adjustment amount is PS, and the rapping frequency adjustment amount is Z;

[0242] Rule 4: If the dust concentration deviation is Z and the temperature deviation is PS, then the electric field strength adjustment amount is Z, and the rapping frequency adjustment amount is PS;

[0243] Rule 5: If the dust concentration deviation is PS and the temperature deviation is PB, then the electric field strength adjustment amount is NS, and the rapping frequency adjustment amount is PB;

[0244] Step 3: Fuzzy inference and defuzzification:

[0245] (1) Fuzzy inference:

[0246] The Mamdani inference method is used to determine the fuzzy quantity of the output. For the given membership values of the dust concentration deviation and the temperature deviation, inference is carried out according to the fuzzy rules;

[0247] Suppose the membership of the current dust concentration deviation is μ NB (e C ) = 0.6,

[0248] The membership of the temperature deviation is μ Z (e T ) = 0.8. For Rule 2 (if the dust concentration deviation is NB and the temperature deviation is Z, then the electric field strength adjustment amount is PB, and the rapping frequency adjustment amount is PS), the triggering strength of this rule is min(0.6, 0.8) = 0.6 o ;

[0249] For the fuzzy set PB of the electric field strength adjustment amount, fuzzy inference calculation is carried out according to the triggering strength and the membership function to obtain the fuzzy contribution of this rule to the electric field strength adjustment amount; similarly, corresponding calculations are also carried out for the fuzzy set PS of the rapping frequency adjustment amount;

[0250] Similar inference calculations are carried out for all rules to obtain the fuzzy contributions of each rule to the electric field strength adjustment amount and the rapping frequency adjustment amount;

[0251] (2) Defuzzification:

[0252] The centroid method is used for defuzzification. For the fuzzy set of the electric field strength adjustment amount, its centroid is calculated as the final precise adjustment amount;

[0253] Suppose the fuzzy set of the electric field strength adjustment amount is {(ΔEf1 , μ1), (ΔE f2 , μ2), …, (ΔE fn , μ n ),}, where ΔE fi is an element in the fuzzy set (i.e., the values of different electric field strength adjustment amounts), and μ i is its corresponding membership degree;

[0254] Then, the electric field strength adjustment amount after defuzzification

[0255] Furthermore, the fault diagnosis and fault tolerance algorithm includes multiple detection models, and its specific features are as follows:

[0256] Step 1, Fault diagnosis part:

[0257] (1), Parameter monitoring model:

[0258] Suppose there are several key parameters to be monitored in the electrostatic precipitation system, such as current I, voltage U, temperature T, and dust concentration C. Under normal operating conditions, these parameters have an expected range;

[0259] Taking the current as an example, assume that the current range during normal operation is I min ≤ I ≤ I max , then a simple judgment function f I (I):

[0260]

[0261] For voltage U, temperature T, and dust concentration C, judgment functions f U (U), f T (T), and f C (C) can be established respectively;

[0262] (2), Fault comprehensive judgment model:

[0263] Suppose the fault types are divided into short - circuit fault F1, open - circuit fault F2, temperature anomaly fault F3, and dust concentration anomaly fault F4; a fault vector F = (F1, F2, F3, F4) can be defined, and the values of its elements are determined by the above - mentioned parameter judgment functions; for example, when f I (I) = 1 and other parameters are normal, it may be suspected of short - circuit or open - circuit faults. At this time, the value of F1 or F2 can be set to 1 (indicating that this fault may occur); the specific judgment rules are determined according to the actual physical process and experience, for example:

[0264]

[0265]

[0266]

[0267]

[0268] Step 2, Fault Tolerance Algorithm Part:

[0269] (1), Fault Compensation Model:

[0270] When abnormal current is detected (assuming a short - circuit fault, F1 = 1), in order to maintain the dust removal effect, the voltage can be adjusted; assume that under normal conditions, there is a relationship between the dust removal efficiency η, current I, and voltage U: η = g(I, U);

[0271] In the fault state, by adjusting the voltage U ′ Try to maintain the dust removal efficiency;

[0272] Assume that a compensation function U ′ = h(I, U, F1) has been obtained through experiments. When F1 = 1, the value of U ′ is adjusted according to the deviation degree of the current I and the current voltage U. Then the equation is In the formula, k is a compensation coefficient, and its appropriate value is determined through experiments or simulations, so that when the current is abnormal, the voltage can be appropriately increased to maintain the dust removal efficiency;

[0273] (2), System Reliability Evaluation Model:

[0274] Assume that the system reliability R is related to the fault situation; it can be simply defined as:

[0275]

[0276] In the formula, w i is the weight corresponding to each fault type, indicating the degree of influence of different faults on the system reliability;

[0277] A short - circuit fault may have a greater impact on the system. w1 can be set to 0.4, an open - circuit fault w2 = 0.3, a temperature - anomaly fault w3 = 0.2, and a dust - concentration - anomaly fault w4 = 0.1 o .

[0278] As Figure 4 shown, further, the optimization scheduling algorithm consists of an objective function, constraint conditions, and an optimization algorithm model. The specific steps are as follows:

[0279] Step 1, Construction of the Objective Function:

[0280] (1), Energy Consumption Minimization:

[0281] Suppose the electrostatic precipitator has n working stages, and the power of each stage i is P i , and the working time is t i ; The total energy consumption E is:

[0282]

[0283] The goal is to minimize E while meeting the dust removal requirements;

[0284] (2) Maximize the dust removal efficiency:

[0285] Suppose the dust removal efficiency is η, which is a function of factors such as current I, voltage U, and wind speed v, η = f(I, U, v); During the operation of the equipment, make η as large as possible; Define a comprehensive objective function J, combining energy consumption and dust removal efficiency:

[0286]

[0287] In the formula, η is the dust removal efficiency, E is the energy consumption, and α is a weight coefficient (0 < α < 1) used to balance the importance of energy consumption and dust removal efficiency;

[0288] Step 2. Set the constraint conditions:

[0289] (1) Equipment parameter constraints:

[0290] Current constraint: I min ≤ I ≤ I max , the voltage has U m in ≤ U ≤ U max , the wind speed has v min ≤ v ≤ v max ;

[0291] (2) Dust removal requirement constraints:

[0292] Suppose the required minimum dust removal efficiency is η min , then η ≥ η min ;

[0293] (3) Time allocation constraints:

[0294] The total working time T is fixed, ∑ i = 1 n t i = T, and the time t of each stage i ≥ 0;

[0295] Step 3. Optimize the algorithm model:

[0296] (1) Dynamic programming model:

[0297] Suppose J kLet \(f_k(x)\) denote the optimal objective function value at state \(x\) (the current parameters of the device) after \(k\) stages.

[0298] The state transition equation is:

[0299]

[0300] where \(U\) is the decision set (operations such as adjusting current, voltage, wind speed, etc. at each stage), \(x_{k - 1}\) is the state of the previous stage, and \(r(x_{k - 1},u)\) is the immediate reward (reduction in energy consumption or improvement in dust removal efficiency) obtained by taking decision \(u\) at state \(x_{k - 1}\). ′ is the state of the previous stage, and \(r(x ′ ,u)\) is the immediate reward (energy consumption reduction or dust removal efficiency improvement) obtained by taking decision \(u\) at state \(x ′ .

[0301] (2) Genetic algorithm model:

[0302] a. Chromosome encoding:

[0303] Suppose the scheduling scheme is determined by parameters such as the current value, voltage value, and working time of each working stage. For example, each parameter is encoded in binary, and then these encodings are concatenated to form a chromosome. Assume that the electrostatic precipitator has 3 working stages, and 3 parameters, namely current \(I\), voltage \(U\), and working time \(t\), are considered for each stage. If the value range of current \(I\) is \([0,100]\ A\) and the accuracy requirement is \(1\ A\), it needs to be represented by 7 - bit binary numbers (because \(2^7 = 128>100\)); the value range of voltage \(U\) is \([0,1000]\ V\) with an accuracy of \(1\ V\), which requires 10 - bit binary numbers; the value range of working time \(t\) is \([0,60]\) minutes with an accuracy of \(1\) minute, which requires 6 - bit binary numbers. 7 Then the length of a chromosome is:

[0304] (7 + 10 + 6)×3 = 69 bits;

[0305] Each segment of the chromosome corresponds to the current, voltage, and working time of one stage respectively.

[0306]

[0307] b. Population initialization:

[0308] Randomly generate a certain number (denoted as \(N\)) of chromosomes to form the initial population. For example, \(N = 100\), and the genes of each chromosome are randomly generated; the parameter values represented by these randomly generated chromosomes are within a reasonable range, that is, the current \(I\), voltage \(U\), and working time must satisfy the physical limitations of the device (\(I_{min}\leq I\leq I_{max}\)). m in≤I≤I max )

[0309] c. Fitness function calculation:

[0310] According to the objective function defined above Calculate the fitness of each chromosome. For each chromosome (representing a scheduling scheme), first decode the current I for each stage from the binary encoding of the chromosome i , voltage U i and working time 1, 2, 3 represent 3 working stages; then calculate the total energy consumption E = ∑ i = 1 3 P i ×t i , where the power P i can be calculated according to the current I i and voltage U i through the power characteristic formula of the device P = UI;

[0311] The dust removal efficiency η = f(I1, U1, t1, I2, U2, t2, I3, U3, t3), which is calculated according to the specific dust removal efficiency function;

[0312] Finally, calculate the fitness value. The higher the fitness value, the better the scheduling scheme;

[0313] d. Selection operation:

[0314] Adopt the roulette wheel selection method; calculate the proportion of the fitness of each chromosome in the total fitness of the population as the probability of being selected;

[0315] For example, the fitness of chromosome j is F j , and the total fitness of the population is Then the probability of chromosome j being selected According to these probabilities, select N / 2 chromosomes to enter the next generation population by randomly simulating the roulette wheel;

[0316] e. Crossover operation:

[0317] Pair the selected chromosomes and adopt single-point crossover. Randomly select a crossover point position k (1 < k < chromosome length); for a pair of chromosomes, exchange the genes after the crossover point position; for example, there are two chromosomes A and B, the crossover point is k = 30, the first 30 genes of chromosome A remain the same as the first 30 genes of chromosome B, and the last 39 genes of A are exchanged with the last 39 genes of B to obtain two new offspring chromosomes; the crossover probability is generally set to P c = 0.7, that is, there is a probability of P c to perform the crossover operation on a pair of chromosomes;

[0318] f. Mutation operation:

[0319] For each gene position of each chromosome, with a certain mutation probability P m=0.01 to mutate; the mutation operation is to invert the binary value of the gene bit; for example, a gene bit was originally 0, and after mutation it becomes 1.

[0320] g. Termination condition judgment:

[0321] Repeat steps c to e until the termination condition is met. Common termination conditions include: reaching the maximum number of iterations (for example, 1000 iterations) or the optimal fitness value of the population does not change significantly in several consecutive generations. When the termination condition is met, the scheduling plan represented by the chromosome with the highest fitness in the population is the optimized result.

Claims

1. An intelligent electrostatic precipitator system applicable to multiple working conditions, characterized in that It includes an interconnected intake pipe, a sensor module, a reaction module, and a central control system; The intake pipe includes a ventilation duct and a pretreatment system, which is used to directly intercept larger particulate matters in the air and play a role in balancing the air flow and optimizing the dust distribution; The sensor module includes multiple temperature sensors, flow rate sensors, and concentration sensors. These sensors are installed in the ventilation duct and are used to detect the actual temperature, flow rate, and concentration of the flue gas in the ventilation duct and transmit them to the central control module; The reaction module includes a voltage regulation device, an intelligent pipeline, and a rapping device; The central control system is centered on a multivariable control algorithm, supplemented by a fuzzy logic algorithm, a fault diagnosis and fault tolerance algorithm, and an optimization scheduling algorithm; The multivariable control algorithm realizes the comprehensive regulation of the electric field strength, rapping frequency, and fan speed; It monitors the generation amount, properties of dust, and changes in environmental conditions in real time, and adjusts the control strategy in a timely manner according to the changes; The fuzzy logic algorithm is used to handle uncertainties and fuzziness, enhancing the adaptability and flexibility of the system; The fault diagnosis and fault tolerance algorithm improves the system reliability and reduces the maintenance cost and risk; The optimization scheduling algorithm utilizes resources to improve the comprehensive performance of the system.

2. The intelligent electrostatic precipitator system applicable to multiple working conditions according to claim 1, wherein, The voltage regulation device and the intelligent pipeline include an electrostatic generation unit and a dust collection unit; The electrostatic generation unit is used to generate an electrostatic field and includes a high-voltage power supply, an electrode assembly, and an insulating support component; The dust collection unit is used to collect the dust adsorbed under the action of the electrostatic field and is connected to the dust removal channel. It includes a dust collection component such as a dust collection plate, a dust collection cylinder, or a dust collection bag; The surface of the dust collection plate has a copper-graphene composite coating, which is used to enhance the dust adsorption ability and facilitate the sliding of dust; The inside of the dust collection cylinder is provided with spiral guide vanes, which are used to guide the dust to gather at the bottom; The dust collection bag is made of a metal decorative material.

3. An intelligent electrostatic precipitator system applicable to multiple working conditions according to claim 1, characterized in that, The rapping device includes a rapping device body, a driving mechanism, and a control module; The rapping device body is used to rap and clean the dust collection unit of the electrostatic precipitator. The rapping device body is in contact with the dust collection unit, and its rapping contact surface is made of titanium; The driving mechanism is connected to the rapping device body and is used to provide rapping power for the rapping device body; The control module is electrically connected or signal-connected to the driving mechanism and is used to control the operating parameters of the driving mechanism, including the rapping frequency, rapping intensity, and rapping duration; the control module includes a microprocessor and a storage unit storing a rapping control program. The microprocessor uses a GEA electrostatic precipitator, and the rapping control program adjusts the rapping parameters according to the operating conditions of the electrostatic precipitator and the dust accumulation situation of the dust collection unit.

4. An intelligent electrostatic precipitation system applicable to multiple working conditions according to claim 1, characterized in that, The multivariable control algorithm also includes a prediction model and an optimal control model; Among them, the prediction model includes a state space model and a prediction equation.

5. An intelligent electrostatic precipitation system applicable to multiple working conditions according to claim 4, characterized in that, The operating steps of the state space model and the prediction equation are as follows: (1) The state space model: Step 1: Select the dust concentration C, temperature T, humidity RH, and electric field strength E f , and the rapping frequency f as state variables, denoted as x = [C, T, RH, E f , f] T ; Step 2: The input variables of the system include the waste gas flow rate Q, the waste gas inlet temperature T i n, and the dust generation rate P, denoted as u = [Q, T i n, P] T ; The output variables of the system include the dust emission concentration C out , and the dust removal efficiency η, denoted as y = [C o ut, η] T ; Step 3: The state space model of the system is expressed as: In the formula, is the derivative of the state variable, representing the rate of change of the state variable with respect to time; A is the state matrix, which describes the mutual relationship between the internal state variables of the system and its own dynamic change characteristics; B is the input matrix, which represents the influence of the input variables on the state variables; w is the disturbance term of the system; y = Cx + Du + v; where C is the output matrix, which describes the relationship between the state variables and the output variables; D is the direct transmission matrix, which represents the direct influence of the input variables on the output variables; v is the measurement noise term; Assume that the change in dust concentration is affected by the waste gas flow rate and the electric field strength, the change in temperature is affected by the waste gas inlet temperature and the electric field energy loss, the change in humidity is affected by the moisture content in the waste gas and the temperature, and the state matrix A, input matrix B, output matrix C, and direct transmission matrix D are set as: Wherein, S is the exhaust gas flow rate, η is the efficiency, qv is the volume flow rate, V is the volume, Q out is the output heat, Q loss is the heat loss C p is the specific heat capacity at constant pressure, E v is the evaporation energy, C n is a constant, R is the gas constant, RH is the relative humidity, T in is the inlet temperature, α is the thermal diffusivity; (2) Prediction equation: Step 1: In model predictive control, predict the output of the system for a period of time in the future based on the state and input at the current moment; assume the prediction horizon is N p , and the control horizon is N c ; Step 2: For a discrete-time system, assume the sampling time is T s , then at the k-th moment, the state of the system is updated through a discretized state-space model: x(k + 1) = Ax(k) + Bu(k); Step 3: Predict the future N p step output: ...... In the formula, represents the predicted output value at the \(i\)-th step at the \(k\)-th moment.

6. An intelligent electrostatic precipitator system applicable to multiple working conditions according to claim 4, characterized in that, The specific steps of the optimization control model are as follows: Step 1. Construct the objective function: (1): The objective function usually aims to minimize a performance metric while considering the deviation of dust emission concentration and energy consumption. Let the set value of dust emission concentration be C s et, and the actual predicted dust emission concentration sequence is as follows: Let the electric field strength be E f , and the fan speed be n. Then the energy consumption function per unit time is E(E f , n); (2) Then the objective function is expressed as: Where ω1 and ω2 are weight coefficients; Step 2. Set the constraint conditions: (1) Electric field strength constraint: E f ,min ≤ E f ≤ E f,max ; where E f,min and E f,max are the lower and upper limits of the electric field strength respectively; (2) Fan speed constraint: n min ≤ n ≤ n max ; where n min and n max are the allowable ranges of the fan speed; (3) Vibration frequency constraint: f m in≤f≤f max ; where f is the vibration frequency; Process requirement constraints: (4) Dust emission concentration constraint: In the formula, C max is the maximum allowable dust emission concentration; (5) Gas flow constraint: Q min ≤ Q ≤ Q max ; where Q is the exhaust gas flow rate.

7. An intelligent electrostatic dust removal system applicable to multiple working conditions according to claim 1, characterized in that The fuzzy logic algorithm performs fuzzy processing on the output quantity, and its specific steps are as follows: Step 1. Fuzzification of input and output variables: (1): For the dust concentration deviation e C , the fuzzy sets are defined as: {NB (Negative Big), NS (Negative Small), Z (Zero), PS (Positive Small), PB (Positive Big)}; (2): Assume the measurement range of the dust concentration is [C min , C max , then it is divided into several intervals corresponding to different fuzzy sets: NB: NS: Z： PS: PB: The membership function is used to determine the degree to which each actual dust concentration deviation value belongs to each fuzzy set, and then a triangular membership function is established. When e C is within the NB interval, its membership function is expressed as: (3): For the temperature deviation e T , the fuzzy sets are {NB (Negative Big), NS (Negative Small), Z (Zero), PS (Positive Small), PB (Positive Big)}. Interval division is carried out according to the reasonable range of temperature, and the corresponding membership functions are determined; (4) The fuzzy sets of the output variables are defined accordingly; the fuzzy set of the electric field strength adjustment amount is {NB, NS, Z, PS, PB}; the fuzzy set of the rapping frequency adjustment amount is {NB, NS, Z, PS, PB}; Step 2. Establish a fuzzy rule base: Rule 1: If the dust concentration deviation is NB and the temperature deviation is NB, then the electric field strength adjustment amount is PB and the rapping frequency adjustment amount is PB; Rule 2: If the dust concentration deviation is NB and the temperature deviation is Z, then the electric field strength adjustment amount is PB and the rapping frequency adjustment amount is PS; Rule 3: If the dust concentration deviation is NS and the temperature deviation is PS, then the electric field strength adjustment amount is PS and the rapping frequency adjustment amount is Z; Rule 4: If the dust concentration deviation is Z and the temperature deviation is PS, then the electric field strength adjustment amount is Z and the rapping frequency adjustment amount is PS; Rule 5: If the dust concentration deviation is PS and the temperature deviation is PB, then the electric field strength adjustment amount is NS and the rapping frequency adjustment amount is PB; Step 3. Fuzzy inference and defuzzification: (1) Fuzzy inference: The Mamdani inference method is used to determine the fuzzy quantity of the output. For the given membership degree values of the dust concentration deviation and the temperature deviation, reasoning is carried out according to the fuzzy rules; Suppose the membership degree of the current dust concentration deviation is μ NB (e C ) = 0.6, The membership degree of the temperature deviation is μ Z (e T ) = 0.

8. For Rule 2, the triggering strength of this rule is min(0.6, 0.8) = 0.6 o ; For the fuzzy set PB of the electric field strength adjustment amount, fuzzy inference calculation is carried out according to the triggering strength and the membership function to obtain the fuzzy contribution of this rule to the electric field strength adjustment amount; Similarly, corresponding calculations are also carried out for the fuzzy set PS of the rapping frequency adjustment amount; Similar reasoning calculations are carried out for all rules to obtain the fuzzy contributions of each rule to the electric field strength adjustment amount and the rapping frequency adjustment amount; (2) Defuzzification: The centroid method is used for defuzzification. For the fuzzy set of the electric field strength adjustment amount, its centroid is calculated as the final precise adjustment amount. Assume that the fuzzy set of the electric field strength adjustment amount is {(ΔE f1 , μ1), (ΔE f2 , μ2), …, (ΔE fn , μ n )}, where ΔE fi is an element in the fuzzy set and μ i is its corresponding membership degree; Then the adjustment amount of the electric field strength after defuzzification 8. An intelligent electrostatic precipitator system applicable to multiple working conditions according to claim 1, characterized in that, The fault diagnosis and fault tolerance algorithm includes multiple detection models, and its specific characteristics are as follows: Step 1. Fault diagnosis part: (1), Parameter Monitoring Model: Suppose the parameters of current I, voltage U, temperature T, and dust concentration C in the electrostatic precipitation system need to be monitored; (2), Fault Comprehensive Judgment Model: Assume the fault types are short - circuit fault F1, open - circuit fault F2, temperature anomaly fault F3, and dust concentration anomaly fault F4; Define a fault vector F=(F1,F2,F3,F4), and the value of its elements is determined by the above - mentioned parameter judgment function; Then the established equation is: Step Two, Fault - Tolerant Algorithm Part: (1), Fault Compensation Model: When abnormal current is detected, to maintain the dust removal effect, the voltage is adjusted; Suppose there is a relationship between the dust removal efficiency η and current I and voltage U under normal conditions, η = g(I,U); In the fault state, the dust removal efficiency is maintained by adjusting the voltage U'; assume that a compensation function U' = h(I, U, F1) is obtained through experiments. When F1 = 1, the value of U' is adjusted according to the deviation degree of the current I and the current voltage U. In the formula, k is a compensation coefficient, and its appropriate value is determined through experiments or simulations. (2), System Reliability Evaluation Model: Suppose the system reliability R is related to the fault situation; Define as: where w i is the weight corresponding to each fault type, indicating the impact degree of different faults on the system reliability.

9. An intelligent electrostatic dust removal system applicable to multiple working conditions according to claim 1, characterized in that, The optimization scheduling algorithm consists of an objective function, constraint conditions, and an optimization algorithm model, and its specific steps are as follows: Step One, Construction of the Objective Function: (1), Energy Consumption Minimization: Suppose the electrostatic precipitator has n working stages, and the power of each stage i is P i , and the working time is t i ; The total energy consumption E is: The goal is to minimize E under the condition of meeting the dust removal requirements; (2), Dust Removal Efficiency Maximization: Suppose the dust removal efficiency is η, which is a function of factors such as current I, voltage U, and wind speed v, η = f(I,U,v); During the operation of the equipment, make η as large as possible; Construct the function J: In the formula, η is the dust removal efficiency, E is the energy consumption, and α is a weight coefficient used to balance the importance of energy consumption and dust removal efficiency; Step Two, Set Constraint Conditions: (1), Equipment Parameter Constraints: Current constraint: I min ≤I≤I max and the voltage has U m in≤U≤U max and the wind speed has v min ≤v≤v max ; (2), Dust Removal Requirement Constraints: Let the required minimum dust removal efficiency be η min , then η ≥ η min ; (3), Time Allocation Constraints: The total working time T is fixed, ∑ i = 1 n t i = T, and the time t for each stage i ≥ 0; Step Three, Optimization Algorithm Model: (1), Dynamic Programming Model: Let J k (x) denote the optimal objective function value in state x after k stages; The state - transfer equation is: In the formula, U is the decision - making set, x′ is the state of the previous stage, and r(x′,u) is the immediate benefit obtained by taking decision u in state x′; (2), Genetic Algorithm Model: a. Chromosome Encoding: Suppose the scheduling scheme is determined by the parameters of the current value, voltage value during the working stage, and the working time of each stage; Assume there are 3 working stages in the electrostatic precipitation equipment, and each stage considers 3 parameters: current I, voltage U, and working time t; If the value range of current I is [0,100] А and the accuracy requirement is 1 A, 7 - bit binary numbers are needed; The value range of voltage U is [0,1000] V and the accuracy is 1 V, which can be represented by 10 - bit binary numbers; The value range of working time t is [0,60] minutes and the accuracy is 1 minute, which requires 6 - bit binary numbers; Then the length of a chromosome is: (7 + 10 + 6)×3 = 69 bits; Each segment of the chromosome corresponds to the current, voltage, and working time of a stage respectively; b. Population Initialization: Randomly generate a certain number N of chromosomes to form the initial population; c. Fitness Function Calculation: According to the objective function defined above Calculate the fitness of each chromosome. For each chromosome, first decode the current I at each stage from the binary encoding of the chromosome i , voltage U i and working time Then calculate the total energy consumption E = ∑ i = 1 3 P i ×t i , where the power P i is calculated according to the current I i and voltage U i through the power characteristic formula of the device P = UI; The dust removal efficiency η = f(I1,U1,t1,I2,U2,t2,I3,U3,t3), which is calculated according to the specific dust removal efficiency function; Finally, calculate the fitness value. The higher the fitness value, the better the scheduling scheme; d. Selection Operation: Adopt the roulette - wheel selection method; Calculate the proportion of the fitness of each chromosome in the total fitness of the population as the probability of being selected; e. Crossover operation: Pair the selected chromosomes and perform single-point crossover. Randomly select a crossover point position k. For a pair of chromosomes, exchange the genes after the crossover point position; f. Mutation operation: For each gene locus of each chromosome, a mutation is carried out with a certain mutation probability P m = 0.01; the mutation operation is to invert the binary value of the gene locus; g. Termination condition judgment: Repeat steps c to e until the termination condition is met. Common termination conditions include: reaching the maximum number of iterations or the optimal fitness value of the population not changing significantly for several consecutive generations. When the termination condition is met, the scheduling scheme represented by the chromosome with the highest fitness in the population is the optimized result.

Images