A kind of laser radar-based boiler feeding automatic control method and system

Abstract

The application discloses a kind of based on laser radar's boiler feed automatic control method and system, the method includes: by the point cloud data information and the oxygen concentration measured value of boiler furnace outlet obtained in real time, the optimal feed amount is obtained by fuzzy-predictive control algorithm calculation, and the optimal control frequency is calculated by the material thickness of conveying belt calculated in real time based on point cloud data information, to realize optimal control to the operating frequency of conveying belt.The present application provides a kind of high reliability, high stability of furnace outlet oxygen concentration automatic control system realized by fuzzy-predictive control method for the current unstable situation of furnace outlet oxygen concentration automatic control system, improve the stability and safety of boiler operation process, with simple equipment, installation is convenient, high economy, high stability, can be updated in real time and so on, provide support for the development of power plant boiler operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of solid fuel feeding technology, in particular to a laser radar-based boiler feeding automatic control method and system for controlling the feeding amount by automatically controlling the running frequency of the conveying belt to make the oxygen content at the furnace outlet tend to be constant. BACKGROUND

[0002] Direct combustion of solid fuel (solid waste, biomass, etc.) for power generation has the advantages of large scale, high efficiency and outstanding environmental benefits, but the power generation technology is not mature yet, and the boiler operation optimization, especially the feeding control system, needs to be improved. The uniformity of the particle size of solid fuel is often poor, which is greatly affected by the source, season and transportation procedure. When the material falls from the plowshare to the conveying belt, the stacking thickness is random and fluctuates greatly, with a maximum thickness difference of 50 cm. Due to the uneven thickness of the material on the conveying belt, it is difficult to accurately measure the thickness of the material on the conveying belt and the feeding amount per unit time.

[0003] Currently, the setting of system parameters and the control of equipment in the solid fuel power plant are generally adjusted by manual experience. Since the thickness of the material on the conveying belt can only be estimated by the naked eye by the technical personnel through monitoring equipment, and the running speed of the conveying belt is controlled manually to adjust the feeding amount of the material to ensure that the oxygen concentration at the furnace outlet is stable within the safe operation range, the oxygen concentration at the furnace outlet and the temperature in the furnace are prone to fluctuate. Frequent changes in the temperature in the furnace will reduce the thermal efficiency of the boiler, reduce the service life of the boiler, and seriously affect the heat transfer efficiency and the safe operation of the incineration system. SUMMARY

[0004] The present application provides a laser radar-based boiler feeding automatic control method and system, which aims to solve the problem that the feeding amount and the oxygen concentration at the furnace outlet cannot be guaranteed stable in the prior art, while reducing the labor intensity of workers. The system is stable and reliable, economical, simple in device, and accurate in measurement results.

[0005] The present application adopts the following technical solutions:

[0006] A laser radar-based boiler feeding automatic control method, which has the following specific steps:

[0007] 1) Obtain the historical operation data of the distributed computer control system (DCS) of the boiler, and establish a DCS operation database; the operation data includes the material conveying frequency of the conveying belt, the boiler feed water flow, the primary and secondary air flow in the furnace, and the oxygen content at the furnace outlet;

[0008] 2) Obtain the real-time feeding flow on the conveying belt based on the material conveying frequency of the conveying belt and the material thickness obtained by the laser radar, and establish a boiler feeding system operation database in combination with the DCS operation database in step 1).

[0009] 3) Take the real-time feeding flow, the boiler feed water flow, and the primary and secondary air flow in the furnace as input variables, and take the oxygen content at the furnace outlet as an output variable in the boiler feeding system operation database, and establish a transfer function mathematical model between the input and output variables by using the particle swarm algorithm; convert the transfer function mathematical model into a low-order finite step response (FSR) model, and use the FSR model as a prediction model; predict the future output of the boiler distributed computer control system based on the prediction model, and obtain the predicted value of the oxygen content at the furnace outlet; calculate the optimal feeding flow that makes the boiler distributed computer control system achieve the optimal economic performance by using an interior point method optimizer (IPOPT), and use the optimal feeding flow as the predicted feeding flow in this case;

[0010] 4) Take the deviation and the deviation change rate of the predicted value and the actual value obtained by the prediction model in step 3) as input variables, and take the real-time feeding flow as an output variable, and establish a fuzzy rule table between the input and output variables; according to the deviation and the deviation change rate of the predicted value and the actual value collected at present, perform fuzzy reasoning under the fuzzy rule table after fuzzy processing, and obtain the fuzzy feeding flow in this case after defuzzy processing;

[0011] 5) Use the fuzzy-predictive control algorithm, correct the predicted feeding flow by using the fuzzy feeding flow according to a set proportional factor, and obtain the corrected optimal feeding flow;

[0012] 6) Combine the size of the conveyor belt, the current material thickness obtained by the laser radar, and the optimal feeding flow obtained in step 5), and calculate the optimal control frequency in this case.

[0013] In the above technical solution, further, in step 2), the real-time feeding flow on the conveyor belt is obtained based on the material conveying frequency of the conveyor belt and the material thickness obtained by the laser radar, and specifically:

[0014] (1) Collect the point cloud data information of the material on the conveyor belt in the scanning area of the laser radar, and perform filtering and noise reduction; the filtering and noise reduction includes deleting the point cloud of non-material, removing environmental noise and outlier point cloud;

[0015] (2) Calculate the average distance d avg from the material to the laser radar according to the spatial position information included in the real-time material point cloud data obtained in step (1), and calculate the real-time average thickness h of the scanning area by combining the distance d tcb from the laser radar to the conveyor belt; wherein:

[0016] n is the number of effective point clouds, and d i is the distance information of the i-th effective point cloud to the laser radar;

[0017] h = davg -d bcb ;

[0018] (3) According to the real-time material average thickness h in step (2), combined with the material conveying frequency f of the real-time conveying belt and the belt size, the real-time feeding flow Q on the conveying belt is calculated:

[0019] Q = f · (2πR) · W · h, R is the radius of the pulley, and W is the width of the conveying belt.

[0020] Further, in the step 3), the real-time feeding flow in the boiler feeding system operation database, the boiler feeding flow, and the primary and secondary air flow in the furnace are taken as input variables, and the oxygen content at the furnace outlet is taken as an output variable, and a transfer function mathematical model between the input and output variables is established by using a particle swarm algorithm, and the specific steps are as follows:

[0021] (1) The gain of the input variable to the output variable is calculated by using a direct current gain method

[0022]

[0023] Wherein, N is the total amount of data in the boiler feeding system operation database, y i , and u i are the i-th output variable and input variable, respectively.

[0024] (2) The transfer function structure is determined to match the dynamic characteristics of the controlled object (i.e., the oxygen content at the furnace outlet), and a first-order inertia pure lag model G(s) is selected, and the structure is as follows:

[0025]

[0026] Wherein, e is the constant Euler number; s is a complex variable; T is a to-be-determined time constant, and τ is a to-be-determined time lag coefficient, both of which are model parameters.

[0027] (3) The transfer function mathematical model parameters are determined: for the first-order inertia pure lag model in step (2), the objective function is:

[0028] Maxγ = Maxγ(τ, T),

[0029] Wherein, γ is a sample correlation coefficient; through a particle swarm algorithm, the model parameters T and τ that make the sample correlation coefficient γ maximum are obtained, and the transfer function mathematical model is obtained.

[0030] Further, in the step 3), an interior point method optimizer (IPOPT) is used to calculate the optimal feeding flow that makes the boiler distributed computer control system achieve the optimal economic performance, and the optimal feeding flow is taken as the predicted feeding amount this time, and the specific steps include the following steps:

[0031] (1) using a quadratic function as the performance index J of the boiler distributed computer control system, specifically represented as:

[0032]

[0033] wherein Q represents an output error weighted coefficient matrix, R represents a control increment variation weighted coefficient matrix, represents a future output quantity prediction matrix, represents a future input quantity matrix, is the transpose of , and Yr represents an output quantity set trajectory;

[0034] (2) solving the above quadratic problem by IPOPT to obtain the future input quantity matrix The first element in the matrix is selected as the predicted feeding quantity of this time.

[0035] Further, in the step 5), using a fuzzy-predictive control algorithm, the predicted feeding quantity is corrected by the fuzzy feeding quantity according to the set proportional factor to obtain the corrected optimal feeding quantity, specifically:

[0036] Q ov =k·Q fzy +(1-k)·Q p ;

[0037] k is the proportional factor, Q ov is the optimal feeding quantity, Q fzy is the fuzzy feeding quantity, and Q p is the predicted feeding quantity.

[0038] Further, in the step 6), the optimal control frequency of this time is calculated by combining the conveyor belt size, the current material thickness obtained by the laser radar and the optimal feeding quantity obtained in the step 5), specifically:

[0039] f ov =Q ov / (2πR·W·h), f ov is the optimal control frequency, Qov is the optimal feeding quantity, R is the pulley radius, and W is the conveyor belt width.

[0040] The application also provides a boiler feeding automatic control system based on a laser radar, comprising:

[0041] A conveyor is used to transport materials; a laser radar is arranged above the conveyor and is used to scan the thickness of the materials on the conveyor; an industrial computer includes a processor and a memory connected to the processor, the memory stores a laser radar-based automatic control program for boiler feeding, and the laser radar-based automatic control program for boiler feeding is executed by the processor to realize the steps of the laser radar-based automatic control method for boiler feeding; and a frequency converter is used to receive signals from the industrial computer and control the running frequency of the conveyor, so as to realize optimal control of the feeding.

[0042] Further, the laser radar-based automatic control system for boiler feeding further includes a support frame, which is used to fix the laser radar above the conveyor.

[0043] The present application has the following advantages:

[0044] The method for controlling the running speed of the conveyor according to the oxygen concentration at the furnace outlet and the thickness fluctuation of the materials on the conveyor to ensure the stability of the oxygen concentration at the furnace outlet has important practical significance. Compared with the prior art, the system can realize automatic control of the volume of the materials entering the furnace according to the oxygen concentration at the furnace outlet, and the system is simple and economical and can adapt to complex industrial site environments.

[0045] The present application solves the problem that the oxygen concentration at the furnace outlet and the feeding amount cannot be ensured to be stable during the operation of the boiler of the solid fuel power plant, real-time analysis is performed on the oxygen concentration data at the furnace outlet, the industrial computer is used to calculate the recommended feeding amount interval under the current working condition, the running speed of the conveyor is controlled in real time after comparison with the current feeding amount, the oxygen concentration at the furnace outlet tends to be constant, the feeding control system is greatly simplified, and the safety of the operation of the boiler is improved.

[0046] The present application provides a laser radar and fuzzy control-based feeding automatic control method and system, which obtains point cloud data information and boiler furnace outlet oxygen concentration measurement values in real time, calculates an optimal feeding amount through a fuzzy-predictive control algorithm, combines a material thickness of a conveying belt calculated in real time based on the point cloud data information, and calculates a current optimal control frequency, so as to realize optimal control of the conveying belt operation frequency. The fuzzy-predictive control algorithm aims to combine the advantages of the two control methods and enhance the robustness of the system as a whole. A single predictive controller needs to be maintained frequently, has a short shelf life and a complex model identification, and a single fuzzy controller cannot effectively process a large time delay of the in-furnace combustion process. The fuzzy-predictive control algorithm is used to additionally compensate for errors generated by the predictive controller, the predictive controller is used as the main part, the fuzzy controller is used as the auxiliary part, and the stability and reliability of the automatic controller are ensured in the case of a serious mismatch between the predictive model and the actual process. The method provides a high-reliability and high-stability boiler furnace outlet oxygen concentration automatic control system realized through a fuzzy-predictive control method, improves the stability and safety of the boiler operation process, has the advantages of simple equipment, convenient installation, high economy, high stability, real-time updating, and the like, and provides support for the development of the power plant boiler operation. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 A laser radar-based boiler feeding automatic control method flowchart;

[0048] Figure 2 A fuzzy-predictive control algorithm flowchart in an industrial computer;

[0049] Figure 3 Oxygen amount error, error change rate, and corresponding membership function diagrams of feeding flow in a fuzzy control algorithm;

[0050] Figure 4 A fuzzy control rule table established;

[0051] Figure 5 A laser radar-based boiler feeding automatic control system schematic diagram

[0052] In the figure, 1 is a laser radar, 2 is a support frame, 3 is a conveying belt, 4 is a frequency converter, and 5 is an industrial computer. DETAILED DESCRIPTION

[0053] As shown in Figure 1 , the present application provides a laser radar-based boiler feeding automatic control method, which includes the following steps:

[0054] Step 101: Obtain historical operation data of a boiler distributed computer control system (DCS), and establish a DCS operation database; the historical operation data includes material conveying frequency of a conveying belt, boiler feed water flow, primary and secondary air flow in a furnace, and oxygen content at a furnace outlet;

[0055] Step 102: Obtain real-time feeding flow on the conveying belt based on the material conveying frequency of the conveying belt and the material thickness obtained by the laser radar, and combine the DCS operation database in step 101 to establish a boiler feeding system operation database;

[0056] Step 103: Take the real-time feeding flow in the boiler feeding system operation database, the boiler feed water flow, and the primary and secondary air flow in the furnace as input variables, and take the oxygen content at the furnace outlet as an output variable, and establish a transfer function mathematical model between the input variables and the output variable by a particle swarm algorithm; convert the transfer function mathematical model into a low-order finite step response (FSR) model, and take the FSR model as a prediction model; predict the future output of the control system based on the prediction model to obtain a predicted value of the oxygen content at the furnace outlet; calculate the optimal feeding flow that makes the boiler distributed computer control system achieve the optimal economic performance by an interior point method optimizer (IPOPT), and take the optimal feeding flow as the predicted feeding amount of this time;

[0057] Step 104: Take the deviation and the deviation change rate of the predicted value and the actual value obtained by the prediction model in step 103 as input variables, and take the real-time feeding flow as an output variable, and establish a fuzzy rule table between the input variables and the output variable; according to the deviation and the deviation change rate of the predicted value and the actual value currently collected, perform fuzzy reasoning under the fuzzy rule table after fuzzification, and obtain the fuzzy feeding amount of this time after defuzzification;

[0058] Step 105: Use a fuzzy-prediction control algorithm, correct the predicted feeding amount by the fuzzy feeding amount according to a set proportional factor, and obtain the corrected optimal feeding amount;

[0059] Step 106: Combine the size of the conveying belt, the current material thickness obtained by the laser radar, and the optimal feeding amount obtained in step 105, and calculate the optimal control frequency of this time.

[0060] In step 101, the historical data of the DCS is specifically the material conveying frequency of the conveying belt, the boiler feed water flow, the primary and secondary air flow in the furnace, and the oxygen content at the furnace outlet. All DCS operation parameters in a long time are selected, and through Pearson correlation analysis and research on the combustion mechanism in the furnace, the material conveying frequency, the boiler feed water flow, and the primary and secondary air flow in the furnace are finally selected as the disturbance variables, and the oxygen content at the furnace outlet is selected as the final control variable, so that the DCS operation database under various operating conditions is obtained. The DCS operation database visualizes the parameters manually debugged for many years, facilitates the analysis of the operation under various operating conditions of the boiler, and provides data support for the research of the automatic control system.

[0061] In step 102, through investigation, it is found that most of the existing factories adopt the method of machine vision to observe the material thickness; and in actual operation, it is found that the machine vision technology has quite high requirements for the precision of the industrial camera, and has poor stability, high maintenance cost and frequent maintenance; therefore, unlike the prior art, the laser radar system with higher precision and better stability is installed in the process of feeding, which can accurately obtain the distance of the material from the radar within a certain range, so as to obtain accurate material thickness information, and the average distance of the material to the laser radar is d avg , the distance of the laser radar to the conveying belt is d tcb , and the calculation formula of the average thickness h is:

[0062] n is the number of effective point clouds, d i is the distance information of the i th effective point cloud to the laser radar;

[0063] h=d avg -d tcb ;

[0064] In combination with the real-time material conveying frequency f of the conveying belt fed back by the DCS system and the belt size, the real-time feeding flow Q on the conveying belt is calculated, and the specific calculation formula is:

[0065] Q=f·(2πR)·W·h, R is the radius of the pulley, and W is the width of the conveying belt.

[0066] In step 103, a prediction model for predicting the oxygen content at the furnace outlet is first established, and the specific formula is:

[0067] The DCS data of the boiler and the laser radar data in a long time are collected, and after data preprocessing, the boiler feed water flow, the primary and secondary air flow in the furnace, the oxygen content at the furnace outlet, and the real-time feed flow and other parameters are selected to construct the boiler feed system operation database. Selecting multiple groups of data with obvious fluctuations, divide them into identification set and test set according to the ratio of 8:2, and use particle swarm algorithm to optimize the parameters. The specific model parameter identification process is:

[0068] Calculate the gain of the input variable to the output variable by the direct current gain method

[0069]

[0070] Wherein, N is the total amount of data in the boiler feed system operation database, y i , u i The i-th output variable and the input variable are respectively;

[0071] Determine the transfer function structure to match the dynamic characteristics of the controlled object (i.e. the oxygen content at the furnace outlet), select a first-order inertia pure lag model G(s), and its structure is as follows:

[0072]

[0073] Wherein, e is the constant Euler number; s is a complex variable; T is a to-be-determined time constant, τ is a to-be-determined time lag coefficient, both of which are model parameters;

[0074] Determine the transfer function mathematical model parameters: for the first-order inertia pure lag model, the objective function is:

[0075] Maxγ=Maxγ(τ,T),

[0076] Wherein, γ is the sample correlation coefficient; through the particle swarm algorithm, the model parameters T and τ that make the sample correlation coefficient γ maximum are obtained, and the transfer function mathematical model is obtained.

[0077] After obtaining the transfer function mathematical model, it is verified that the model effect is good, and then it is converted into a low-order finite step response (FSR) prediction model using the tool provided by MATLAB. The next step is to solve the system performance index, which is:

[0078] A quadratic function is used as the performance index J of the boiler distributed computer control system, which is specifically represented as:

[0079]

[0080] Wherein, Q represents the output error weighted coefficient matrix, R represents the control increment change weighted coefficient matrix, represents the future output prediction matrix, represents a future input matrix, is the transpose of , Yr represents an output quantity setting trajectory;

[0081] The above quadratic programming problem is solved by an interior point method optimization solver (IPOPT) to obtain a future input matrix The first element in the matrix is selected as the predicted feed quantity of this time.

[0082] Further, the parameter settings of the predictive controller (including the overall control loop of the predictive control algorithm, such as the establishment of a prediction model, the prediction of future outputs according to input variables, and the solving of performance indicators to obtain a predicted feed quantity) of the present application are shown in Table 1:

[0083] Table 1: Parameters of the predictive controller

[0084] Parameter Symbol Value Sampling period Ts 10s Prediction step P 120 Control step M 20 Error weighting coefficient matrix Q 30I Control delta change weighting coefficient matrix R I

[0085] In step 104, unlike previous studies, the error E and the error change rate EC of the predicted value and the actual value of the output in the predictive control process are not directly used as input variables, and the feed flow F is used as an output variable, which aims to compensate the predictive controller through the fuzzy controller (including the overall control loop of the fuzzy control algorithm, such as the establishment of a fuzzy rule table, the fuzzification of input variables, fuzzy reasoning, and the defuzzification of fuzzy feed quantity). Specifically, the oxygen content prediction value V P obtained in step 103 and the actual oxygen content value V O2 are used to obtain the oxygen content error E and the real-time error change rate EC:

[0086] E = V P -V O2 ;

[0087] EC = (E - E') / t, t is the control step of the system, and E' is the oxygen content error at the time before t;

[0088] The basic domain of the oxygen content error E is set to [-5, 5], the basic domain of the real-time error change rate EC is set to [-1, 1], and the basic domain of the output feed flow F is set to [-5, 5]. The basic domains of the oxygen content error E, the real-time error change rate EC, and the output feed flow F are quantified using 5, 10, and 1 as the scale factors, respectively, to obtain the quantized domain of the oxygen content error E as [-25, 25], the quantized domain of the real-time error change rate EC as [-10, 10], and the quantized domain of the output feed flow F as [-5, 5]. The specific quantization formula is:

[0089] y = k(x-(xH -x L ) / 2);

[0090] x is the original value in the fundamental universe of discourse;

[0091] y is the quantized value in the quantization domain;

[0092] k is the scaling factor used for quantification;

[0093] x H x L These are the upper and lower bounds of the fundamental universe of discourse, respectively;

[0094] like Figure 3 As shown, a trapezoidal membership function is used to fuzzify the quantization domain obtained in step (4); and a method is used as follows: Figure 4 The fuzzy rules and Mamdani inference method shown are used to perform fuzzy inference on the input oxygen content error E and the real-time error change rate EC to obtain the feed flow rate F' in the fuzzy set; the centroid method is used to defuzzify the feed flow rate F' in the fuzzy set to obtain the fuzzy feed rate.

[0095] In step 105, a fuzzy control algorithm is used to compensate for the errors generated by the predictive controller. The predictive controller is primary, and the fuzzy controller is secondary, ensuring that the automatic controller maintains a certain level of stability and reliability even when there is a significant mismatch between the predictive model and the actual process. Specifically, the predicted feed rate obtained in step 103 and the fuzzy feed rate obtained in step 104 are weighted and averaged according to a set scaling factor to obtain the optimal feed rate. The specific calculation formula is as follows:

[0096] Q ov =k·Q fzy +(1-k)·Q p ;

[0097] k is a scaling factor, Q ov For the optimal feed rate, Q fzy For fuzzy feed rate, Q p To predict the feed rate;

[0098] In step 106, the obtained optimal feed rate needs to be converted into an executable optimal operating frequency. The specific calculation method is as follows:

[0099] f ov =Q ov / (2πR·W·h), f ov The optimal control frequency is Qov, the optimal feed rate is R, the pulley radius is W, and the conveyor belt width is W.

[0100] This invention also provides an automatic boiler feeding control system based on lidar, such as...Figure 5 As shown, the system comprises: a conveyor belt 3 for conveying materials; a laser radar 1 arranged above the conveyor belt 3 for scanning the thickness of the materials on the conveyor belt 3; an industrial computer 5 comprising a processor and a memory connected to the processor, the memory storing a laser radar-based automatic control program for boiler feeding, the laser radar-based automatic control program for boiler feeding being executed by the processor to realize the steps of the laser radar-based automatic control method for boiler feeding as described above; and a frequency converter 4 for receiving signals from the industrial computer 5 to control the running frequency of the conveyor belt 3, thereby realizing optimal control of the feeding. The system further comprises a support frame 2 for fixing the laser radar 1 above the conveyor belt 3.

[0101] In summary, the laser radar and fuzzy control-based automatic control method and system for feeding of the present application, by using the laser radar point cloud data information and the oxygen content at the outlet of the boiler furnace obtained in real time, the optimal feeding amount is calculated by using the fuzzy-predictive control algorithm, and combined with the average thickness of the materials obtained by the laser radar, the running frequency of the conveyor belt 3 is optimized by using the industrial computer 5. The method provides a high-reliability and high-stability automatic control system for the oxygen concentration at the outlet of the furnace realized by the fuzzy-predictive control method, which improves the stability and safety of the boiler operation process, has the advantages of simple equipment, convenient installation, high economy, high stability, real-time updating, etc., and provides support for the development of power plant boiler operation.

Claims

1. An automatic control method for boiler feeding based on lidar, characterized in that, Includes the following steps: 1) Obtain historical operating data of the boiler distributed computer control system (DCS) and establish a DCS operating database; the historical operating data includes the material conveying frequency of the conveyor belt, boiler feedwater flow rate, primary and secondary air flow rates in the furnace, and oxygen content at the furnace outlet; 2) Based on the material conveying frequency of the conveyor belt and the material thickness obtained by the lidar, the real-time feeding flow rate on the conveyor belt is obtained. Combined with the DCS operation database mentioned in step 1), a boiler feeding system operation database is established. 3) Using the real-time feed flow rate, boiler feedwater flow rate, and primary and secondary air flow rates in the boiler feeding system operation database as input variables, and the oxygen content at the furnace outlet as the output variable, a mathematical model of the transfer function between the input and output variables is established using the particle swarm optimization algorithm. The transfer function mathematical model is then converted into a low-order finite step response model, which is used as the prediction model. Based on the prediction model, the future output of the boiler distributed computer control system is predicted to obtain the predicted value of the oxygen content at the furnace outlet. The optimal feed flow rate that enables the boiler distributed computer control system to achieve the best economic performance is calculated using the interior point optimization solver IPOPT, and this is used as the predicted feed flow rate for this operation. 4) Using the deviation and rate of change of the predicted value and the actual value obtained by the prediction model in step 3) as input variables and the real-time feed flow rate as output variables, a fuzzy rule table between the input and output quantities is established; based on the deviation and rate of change of the currently collected predicted value and the actual value, fuzzification is performed and fuzzy inference is performed under the fuzzy rule table, and the fuzzy feed quantity is obtained after defuzzification. 5) Using a fuzzy-predictive control algorithm, the predicted feed rate is corrected by the fuzzy feed rate based on the set proportional factor, so as to obtain the corrected optimal feed rate; 6) Combining the conveyor belt size, the current material thickness obtained by the lidar, and the optimal feed rate obtained in step 5), the optimal control frequency for this operation is calculated. In step 2), the real-time feed flow rate on the conveyor belt is obtained based on the material conveying frequency of the conveyor belt and the material thickness obtained by the lidar. This specifically includes the following steps: (1) Collect point cloud data of materials on the conveyor belt within the laser radar scanning area and perform filtering and noise reduction; the filtering and noise reduction includes deleting point clouds of non-materials, removing environmental noise and outlier point clouds; (2) Calculate the average distance from the material to the lidar based on the spatial location information included in the material point cloud data acquired in real time in step (1). The distance from laser radar to the conveyor belt Calculate the real-time average material thickness h in the scanned area: Where n is the number of valid point clouds, d i This represents the distance information from the i-th valid point cloud to the lidar. (3) Based on the real-time average material thickness h in step (2), combined with the real-time material conveying frequency f and belt size, the real-time feed flow rate Q on the conveyor belt is calculated: Where R is the radius of the pulley and W is the width of the conveyor belt.

2. The automatic control method for boiler feeding based on lidar according to claim 1, characterized in that, In step 3), the real-time feed flow rate, boiler feedwater flow rate, and primary and secondary air flow rates in the boiler feeding system operation database are used as input variables, and the oxygen content at the furnace outlet is used as the output variable. A mathematical model of the transfer function between the input and output variables is established using the particle swarm optimization algorithm. The specific steps are as follows: (1) Calculate the gain of the input variable on the output variable using the DC gain method. : ， Where N is the total amount of data in the boiler feeding system's operating database, and y i u i These are the i-th output variable and input variable in the database, respectively; (2) Determine the transfer function structure to match the dynamic characteristics of the controlled object. Select the first-order inertial pure time delay model G(s), whose structure is as follows: ， Where e is the constant Euler number; For complex variables; T is the undetermined time constant. These are the time delay coefficients to be determined, and both are undetermined model parameters. (3) Determine the mathematical model parameters of the transfer function: For the first-order inertial pure time delay model described in step (2), its objective function is: in, Let be the sample correlation coefficient; obtain the sample correlation coefficient using the particle swarm optimization algorithm. The largest model parameter T, Thus, the mathematical model of the transfer function is obtained.

3. The automatic control method for boiler feeding based on lidar according to claim 1, characterized in that, In step 3), the optimal feed flow rate that enables the boiler distributed computer control system to achieve the best economic performance is calculated using the interior point optimization solver IPOPT, and this is used as the predicted feed rate for this operation. The specific steps include: (1) A quadratic function is used as the performance index J of the boiler distributed computer control system, specifically expressed as: ； Where Q represents the output error weighting coefficient matrix, and R represents the control increment change weighting coefficient matrix. This represents a matrix predicting future output. The matrix representing future input quantities. for The transpose of Yr represents the output setting trajectory; (2) Solve the above quadratic function using IPOPT to obtain the future input matrix. The first element in the matrix is ​​selected as the predicted feed amount for this test.

4. The automatic control method for boiler feeding based on lidar according to claim 1, characterized in that, In step 5), a fuzzy predictive control algorithm is used to correct the predicted feed rate based on a set proportional factor and the fuzzy feed rate, thereby obtaining the corrected optimal feed rate. Specifically: ； k is a scaling factor, Q ov For the optimal feed rate, Q fzy For fuzzy feed rate, Q p To predict the feed rate.

5. The automatic control method for boiler feeding based on lidar according to claim 1, characterized in that, In step 6), the optimal control frequency is calculated by combining the conveyor belt size, the current material thickness obtained by the lidar, and the optimal feed rate obtained in step 5), specifically as follows: f ov For the optimal control frequency, Q ov For the optimal feed rate, R is the pulley radius and W is the conveyor belt width.

6. An automatic boiler feeding control system based on lidar, characterized in that, The system includes: A conveyor belt for transporting materials; a lidar sensor positioned above the conveyor belt for scanning the thickness of the material on the conveyor belt; an industrial control computer including a processor and a memory connected to the processor, the memory storing a lidar-based automatic boiler feeding control program, which, when executed by the processor, implements the steps of the lidar-based automatic boiler feeding control method as described in any one of claims 1-5; and a frequency converter for receiving signals from the industrial control computer and controlling the operating frequency of the conveyor belt to achieve optimal feeding control.

7. The automatic feeding control system based on lidar according to claim 6, characterized in that, The system also includes a support frame for fixing the lidar above the conveyor belt.

Images