Real-time fuel quantity estimation and optimization control method for fluidized bed boiler

Abstract

The present application provides a kind of fluidized bed boiler real-time fuel quantity estimation and optimization control method, utilize dynamic parameter simulation fuel combustion process after entering furnace, combine heat release coefficient determined by boiler load, air-coal ratio and other factors to establish bed temperature prediction model, and based on historical operation production data of boiler, utilize heuristic optimization algorithm to solve model, so that optimization result can maximum approach real fuel combustion and heat release process;The quantitative estimation method of real-time fuel quantity in furnace can be obtained by transforming the combustion process equation obtained by optimization, and then the potential stored heat in the boiler is estimated and introduced into the control system as the main basis for accurate control of boiler load, solving the problem of insufficient control accuracy in fluidized bed boiler production control, thereby improving the overall safety and economic benefit of boiler operation, the present application can better handle the process with delay and inertia characteristics, improve the response speed and adjustment accuracy of boiler operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of general control or regulation system, in particular to a real-time fuel quantity estimation and optimization control method for fluidized bed boiler. BACKGROUND

[0002] Circulating fluidized bed technology is a new type of high-efficiency clean combustion technology, which is one of the main methods to solve the problem of coal pollution. Due to its unique advantages of direct desulfurization in the furnace and the use of low-quality coal, it is the focus of the development of clean coal combustion technology at home and abroad. However, due to its unique combustion mechanism and the characteristics of multi-coupling, nonlinearity and large time delay of the combustion system, the stability of the combustion system is more difficult to control than that of the pulverized coal boiler. The difficulty in controlling the circulating fluidized bed is an important factor affecting the rapid development of the circulating fluidized bed boiler. How to solve the stability and robustness of the combustion control of the circulating fluidized bed, especially the large-capacity circulating fluidized bed boiler, is an urgent problem that must be solved in the commercialization of large-scale boilers.

[0003] At present, the optimization control of the circulating fluidized bed boiler is mainly through analyzing the operation mechanism of the circulating fluidized bed and establishing a mechanism model, or analyzing the relationship between various variable parameters under different conditions based on the measured data of field tests, establishing a state quantity model required by the control system, or researching and using conventional instrument monitoring or advanced intelligent instrument and technology for monitoring.

[0004] The dynamic characteristics of the heat release of the fluidized bed boiler depend on the real-time fuel quantity in the furnace, which is a process state variable that cannot be measured by online instruments. The dynamic accumulation mechanism of the real-time fuel quantity in the furnace represents the essential process of the combustion of the circulating fluidized bed boiler, and is the fundamental reason for the large delay and inertia. Therefore, in order to solve the problem of large delay and inertia, the unknown to-be-burned fuel quantity must be estimated and introduced into the control system.

[0005] There is an urgent need for a method for introducing the to-be-burned fuel quantity into the control system. SUMMARY

[0006] The present application is to solve the problem that the to-be-burned fuel quantity cannot be combined in the control process of the fluidized bed boiler, and provides a real-time fuel quantity estimation and optimization control method for fluidized bed boiler. The dynamic characteristics of the heat release of the fluidized bed boiler are determined by using historical operation data, a calculation method for the real-time fuel quantity in the furnace is established, the estimated real-time fuel quantity and the combustion rate are introduced into the control system, the optimization control of the combustion of the fluidized bed boiler is realized, and the combustion stability is improved.

[0007] The present application provides a real-time fuel quantity estimation and optimization control method for fluidized bed boiler, comprising the following steps:

[0008] S1, Δ tThe time unit is obtained, the boiler operation history data with time sequence characteristics is obtained, and the dc t and dT t , wherein dc t is the change amount of the coal entering the furnace at t, dT t is the change amount of the bed temperature at t.

[0009] S2, a boiler heat release dynamic model is established, and heat release process dynamic weights [w1, w2,..., w L ] and heat release parameters a, b, g, d are obtained after the fuel enters the furnace.

[0010] S3, the real-time fuel H(t) at t is obtained according to the heat release process dynamic weights [w1, w2,..., w L ].

[0011] S4, the parameter k that minimizes the loss function J1 is solved by using the least square method according to the real-time fuel H(t), and H(t0) = k*steam0 is obtained, wherein H(t0) is the target real-time fuel amount, and steam0 is the target load.

[0012] S5, when the real-time fuel amount estimation at t' and the optimization control of the target load are needed, step S6 is entered.

[0013] S6, the heat release process dynamic weights [w1, w2,..., w L ] are obtained according to the boiler data from t'-L+1 to t' using the method of step S2, the real-time fuel amount H(t') at t' and the parameter k' are obtained according to the method of step S3, and the target real-time fuel amount H(t0') at t' is obtained using the method of step S4 according to the current boiler load reaching the target value steam0'.

[0014] The optimization control method of the target load at t' is that when the current boiler load reaches the target value steam0', the current coal entering the furnace should be adjusted to coal target .

[0015]

[0016] , wherein coal t’ is the coal entering the furnace at the current time t', and coaol target is the target coal entering the furnace.

[0017] The real-time fuel amount estimation at t' and the optimization control of the target load are completed.

[0018] The fluidized bed boiler real-time fuel quantity estimation and optimization control method, as a preferred mode, the step S5 further comprises: when the optimization control of the bed temperature at the t' moment is needed, entering the step S7.

[0019] S7, according to the boiler data at t'-L+1 to t' moment, the heat release parameters alpha', beta', gamma', delta' are obtained by using the method of step S2, and the bed temperature T at t'+1 moment is recursively predicted according to the change amount of the coal into the furnace, the heat release coefficient, the bed temperature T at t' moment t’+1 , according to the bed temperature T at t'+1 moment t’+1 And the target bed temperature T0', the bed temperature at t'+1 moment is optimized and controlled by controlling the air-coal ratio, the primary and secondary air volume ratio, and the slag discharge amount slag t’ The optimization control of the bed temperature at t'+1 moment is completed.

[0020] The fluidized bed boiler real-time fuel quantity estimation and optimization control method, as a preferred mode, in the step S1, the boiler operation history data with time sequence characteristics includes the boiler load steam, the coal supply amount coal, the total air volume slag, the primary air volume wind1, the secondary air volume wind2, the bed temperature T and the slag discharge amount slag;The source of the boiler operation history data with time sequence characteristics is the boiler operation history operation data or the assay report data;

[0021] When the boiler operation history data with time sequence characteristics is missing, the spline interpolation method is used for filling;

[0022] The Savitzky-Golay smoothing filter is used for denoising;

[0023] dc t =coal t -coal t-1

[0024] dT t =T t -T t-1

[0025] Wherein, coal t is the coal into the furnace at t moment, T t is the bed temperature at t moment.

[0026] The fluidized bed boiler real-time fuel quantity estimation and optimization control method, as a preferred mode, in the step S1, the spline interpolation method is: using the python mathematical calculation library scipy to realize the cubic spline interpolation of the time sequence data.

[0027] The fluidized bed boiler real-time fuel quantity estimation and optimization control method provided by the application, as a preferred mode, in step S2, the loss function J0 of the boiler heat release dynamic model is:

[0028]

[0029] Wherein: N is the total amount of boiler operation history data with time sequence characteristics;

[0030] L is a positive integer, dc i-j is the change amount of the coal quantity entering the furnace at i-j moment, dT i is the bed temperature change amount at i moment, epsilon is a random item subject to standard normal distribution, and e is a heat release coefficient, which is determined by the boiler load, the air-coal ratio, the primary and secondary air quantity ratio, and the slag discharge quantity.

[0031] The fluidized bed boiler real-time fuel quantity estimation and optimization control method provided by the application, as a preferred mode, L is a positive integer in the interval [10, 20];

[0032]

[0033] Wherein: steam t is the boiler load at t moment, wind t is the total air quantity entering the furnace at t moment, coal t is the coal quantity entering the furnace at t moment, wind1 t is the primary air quantity at t moment, wind2 t is the secondary air quantity at t moment, and slag t is the slag discharge quantity at t moment.

[0034] The fluidized bed boiler real-time fuel quantity estimation and optimization control method provided by the application, as a preferred mode, the loss function J0 of the boiler heat release dynamic model is optimized by using the artificial bee colony algorithm, so that the heat release process dynamic weight [w1, w2,... w L ] and the heat release parameters alpha, beta, gamma and delta after the fuel enters the furnace are obtained.

[0035] The fluidized bed boiler real-time fuel quantity estimation and optimization control method provided by the application, as a preferred mode, in step S3, the heat release process dynamic weight [w1, w2,... w L ] after the fuel enters the furnace is that when 1 unit of fuel enters the furnace, w1, w2,... w L will be combusted and heat will be released at the next L moments respectively.

[0036]

[0037] As a preferred mode, in step S4 of the method,

[0038]

[0039] k is updated in real time in units of 6 hours to 24 hours.

[0040] As a preferred mode, in step S7 of the method,

[0041]

[0042] wherein T t， is the bed temperature at t', dc t’-i+1 is the change in the amount of coal entering the furnace at t'-i+1, e t’-i+1 is the heat release coefficient at t'-i+1.

[0043] The method has the following advantages:

[0044] The method uses dynamic parameters to simulate the combustion process after the fuel enters the furnace, establishes a bed temperature prediction model in combination with the heat release coefficient determined by the boiler load, the air-coal ratio, and the like, and solves the model by using a heuristic optimization algorithm based on historical operation and production data of the boiler, so that the optimization result can most closely approximate the real fuel combustion and heat release process. The combustion process equation obtained by the optimization can be transformed to obtain a quantitative estimation method of the real-time fuel amount in the furnace, and then the potential stored heat in the boiler can be estimated and introduced into the control system as the main basis for accurate control of the boiler load, thereby solving the problem of insufficient control accuracy in the production control of the fluidized bed boiler, and improving the overall safety and economic benefits of the boiler operation.

[0045] The combustion process parameters in the method can be dynamically updated under different production states or fuel characteristics, so that the estimation of the real-time fuel amount has self-adaptive capability. The determined process parameters also enable the calculation method to have time series prediction capability. The adjustment of the operation parameters of the boiler based on the predicted heat release process can better handle processes with delay and inertia characteristics, and improve the response speed and adjustment accuracy of the boiler operation. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 is a flowchart of a real-time fuel amount estimation and optimization control method for a fluidized bed boiler;

[0047] Figure 2 is a bed temperature prediction effect diagram of a real-time fuel amount estimation and optimization control method for a fluidized bed boiler;

[0048] Figure 3This is a dynamic weight distribution diagram during the combustion process of a real-time fuel quantity estimation and optimization control method for a fluidized bed boiler.

[0049] Figure 4 This is a diagram showing the boiler load and fuel quantity distribution in the furnace, used for a real-time fuel quantity estimation and optimization control method for fluidized bed boilers. Detailed Implementation

[0050] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0051] Example 1

[0052] like Figure 1 As shown, a method for real-time fuel quantity estimation and optimized control of a fluidized bed boiler is characterized by the following steps:

[0053] S1. Using Δt as the time unit, obtain historical boiler operation data with time-series characteristics, and obtain dc after denoising. t and dT t DC t Let dT be the change in the amount of coal fed into the furnace at time t. t Let t be the change in bed temperature at time t;

[0054] Historical boiler operation data with time-series characteristics includes boiler load steam, coal feed rate, total air volume slag, primary air volume wind1, secondary air volume wind2, bed temperature T, and ash discharge rate slag; the source of historical boiler operation data with time-series characteristics is historical boiler operation data or laboratory report data.

[0055] When there are missing data in the boiler operation history data with time-series characteristics, spline interpolation is used to fill the gaps. The spline interpolation method is to use the Python mathematical calculation library scipy to perform cubic spline interpolation on the time-series data.

[0056] Denoising was performed using a Savitzky-Golay smoothing filter;

[0057] Savitzky-Golay smoothing filters are widely used for data stream smoothing and noise reduction. They are a filtering method based on local polynomial least squares fitting in the time domain. The core idea is to perform k-order polynomial fitting on the data points within a certain length window, and then discretize the fitting result to ensure that the shape and width of the signal remain unchanged while filtering out noise.

[0058] This method utilizes the Python mathematical computation library scipy to filter and denoise boiler combustion data.

[0059] dc t = coal t -coal t-1 ;

[0060] dT t = T t -T t-1 ;

[0061] wherein, coal t is the amount of coal entering the furnace at time t, T t is the bed temperature at time t;

[0062] S2, a boiler heat release dynamic model is established to obtain the heat release process dynamic weight [w1, w2,..., w L ] and heat release parameters a, b, g, d after the fuel enters the furnace;

[0063] The loss function J0 of the boiler heat release dynamic model is:

[0064]

[0065] wherein:

[0066] L is a positive integer in the interval [10, 20], dc i-j is the amount of coal entering the furnace at time i-j, dT i is the bed temperature change at time i, and e is a random term subject to standard normal distribution, and e is the heat release coefficient, which is determined by the boiler load, the air-coal ratio, the primary-secondary air ratio, and the slag discharge amount;

[0067]

[0068] wherein: steam t is the boiler load at time t, wind t is the total amount of air entering the furnace at time t, coal t is the amount of coal entering the furnace at time t, wind1 t is the primary air amount at time t, wind2 t is the secondary air amount at time t, and slag t is the slag discharge amount at time t;

[0069] The loss function J0 of the boiler heat release dynamic model is optimized by using the artificial bee colony algorithm to obtain the heat release process dynamic weight [w1, w2,..., w L ] and heat release parameters a, b, g, d after the fuel enters the furnace;

[0070] Artificial bee colony algorithm is an optimization method simulating the foraging behavior of honeybee colony, which is a specific application of swarm intelligence. The main feature is that it does not need to understand the special information of the problem, only needs to compare the advantages and disadvantages of the problem, and finally makes the global optimal value emerge in the group through the local optimization behavior of each artificial bee individual. It is suitable for continuous and discrete optimization problems, and has a faster convergence speed.

[0071] The method uses artificial bee colony algorithm to optimize the loss function in step 1) above, and the optimization algorithm parameters are set as follows:

[0072] Number of bees: [10, 50];

[0073] Maximum number of iterations: [50, 200];

[0074] Maximum proportion of foraging bees: [10%, 50%];

[0075] Maximum number of foraging sources: [5, 20];

[0076] S3, according to the dynamic weight [w1, w2,..., w L ] of the heat release process, the real-time fuel H(t) at time t is obtained;

[0077] The dynamic weight [w1, w2,..., w L ] of the heat release process after the fuel enters the furnace is that when 1 unit of fuel enters the furnace, it will be burned w1, w2,..., w L and release heat at the next L time, and the unburned part is stored as real-time fuel in the furnace;

[0078] The real-time fuel in the furnace at time t can be represented by the following formula:

[0079]

[0080] Where: coal t-i+1 is the amount of coal entering the furnace at time t-i+1;

[0081] S4, according to the real-time fuel H(t), taking the boiler load steam as the independent variable and the real-time fuel amount ht as the dependent variable, using the least squares method to solve the parameter k that minimizes the loss function J1, to obtain ht=k*steam, in the fluidized bed boiler control process, when the boiler load changes, using the function relationship between the real-time fuel amount and the boiler load described in the above steps, calculate the boiler real-time fuel amount corresponding to the target load:

[0082] H(t0)=k*steam0, where H(t0) is the target real-time fuel amount and steam0 is the target load;

[0083]

[0084] k is updated in real time in 6 hours;

[0085] As Figure 2 shown, the boiler load and the amount of fuel implemented in the furnace are linearly distributed.

[0086] S5, when real-time fuel amount estimation and target load optimization control at time t' is needed, go to step S6; when bed temperature optimization control at time t'+1 is needed, go to step S7;

[0087] S6, according to the boiler data at time t'-L+1 to t', use the method of step S2 to obtain the heat release process dynamic weight [w1, w2,... w L ], as Figure 3 shown, according to the method of step S3, obtain real-time fuel H(t') at time t', parameter k', and according to the current boiler load reaching the target value steam0', use the method of step S4 to obtain the target real-time fuel amount H(t0') at time t';

[0088] The optimization control method of target load at time t' is: under the condition that the combustion state is unchanged, when the current boiler load reaches the target value steam0', the current amount of coal entering the furnace should be adjusted to coal target ;

[0089]

[0090] Wherein, coal t’ is the amount of coal entering the furnace at the current time t', coal target is the target amount of coal entering the furnace;

[0091] Real-time fuel amount estimation and target load optimization control at time t' are completed;

[0092] S7, according to the boiler data at time t'-L+1 to t', use the method of step S2 to obtain heat release parameters α', β', γ', g', and then according to the amount of coal entering the furnace, the heat release coefficient, and the bed temperature at time t', predict and recursively the bed temperature T t’+1 at time t'+1, according to the bed temperature T t’+ at time t'+1 and the target bed temperature T0', through the control of the air-coal ratio, the primary and secondary air volume ratio, and the amount of slag slag t′ , perform the optimization control of the bed temperature at time t'+1,

[0093]

[0094] Wherein, T t′ is the bed temperature at time t', dc t′-i+1is the change of the coal quantity at the time t'-i+1, e t’-i+1 is the heat release coefficient at the time t'-i+1;

[0095] In the control process of the fluidized bed boiler, under the condition that the coal quantity is constant, the bed temperature can be controlled by controlling the ratio of the air and coal the ratio of the primary air and the secondary air the slag quantity sl t .

[0096] The optimized control of the bed temperature at the time t'+1 is completed.

[0097] The application of the method to the used data and the difference between the actual value and the predicted value of the bed temperature are shown in Tables 1-2 and Figure 4 Fig. 1, and the prediction precision of the method is better.

[0098] Table 1: Boiler operation data and actual value of the bed temperature

[0099]

[0100] Table 2: Comparison between the actual value and the predicted value of the bed temperature

[0101]

[0102] The above description is only the preferred specific implementation of the present application, but the protection scope of the present application is not limited to this, any skilled person in the art can make equivalent replacement or change according to the technical solution and the inventive concept of the present application within the technical range disclosed by the present application, which should be covered in the protection scope of the present application.

Claims

1. A method for real-time fuel mass estimation and optimal control of a fluidized bed boiler, characterized by: The method comprises the following steps: S1, obtain the boiler operation history data with timing characteristics in Δt time unit, and obtain dc after denoising t and dT t , wherein dc t is the change amount of the coal quantity entering the furnace at t, and dT t is the change amount of the bed temperature at t; The boiler operation history data with time sequence characteristics comprises a boiler load steam, a coal supply amount coal, a total air amount slag, a primary air amount wind1, a secondary air amount wind2, a bed temperature T, and a slag discharge amount slag; the boiler operation history data with time sequence characteristics is obtained from boiler operation history operation data or assay report data; S2, a boiler heat release dynamic model is established to obtain heat release process dynamic weights [w1, w2, …, w L ] and heat release parameters a, b, g, d after the fuel enters the furnace. The loss function J0 of the boiler heat release dynamic model is: wherein: N is the total amount of boiler operation history data with timing characteristics; L is a positive integer, dc i-j is the change of the amount of coal into the furnace at i-j moment, dT i is the change of the bed temperature at i moment, ε is a random item obeying standard normal distribution, e is the heat release coefficient, which is determined by the boiler load, the air-coal ratio, the primary and secondary air volume ratio, and the amount of slag discharge; wherein: e t is the heat release coefficient at time t, steam t is the boiler load at time t, wind t is the total air flow into the furnace at time t, coal t is the amount of coal into the furnace at time t, wind1 t is the primary air flow at time t, wind2 t is the secondary air flow at time t, slag t is the amount of slag at time t. The loss function J0 of the boiler heat release dynamic model is optimized by using an artificial bee colony algorithm to obtain dynamic weights [w1, w2, … w L ] and heat release parameters a, b, g, d of the heat release process after the fuel enters the furnace. S3, dynamically weight [w1, w2, …, w L ] to obtain the real-time fuel H(t) at time t; S4, according to the real-time fuel H(t), the parameter k is solved by using the least square method to minimize the loss function J1, and H(t0) = k*steam0 is obtained, wherein H(t0) is a target real-time fuel amount, and steam0 is a target load; S5, when real-time fuel amount estimation at t' moment and target load optimization control are needed, step S6 is entered; S6. Using the boiler data at time t' - L + 1 to t', the method of step S2 is used to obtain the heat release process dynamic weights [w1, w2,..., w L ], the method of step S3 is used to obtain the real-time fuel quantity H(t') at time t' and the parameter k', and according to the current boiler load reaching the target value steam0', the method of step S4 is used to obtain the target real-time fuel quantity H(t0') at time t'. The optimization control method of the target load at time t' is: when the current boiler load reaches the target value steam0', the current coal quantity entering the boiler should be adjusted to coal target ; wherein, coal t’ is the current amount of coal at time t', coal target is the target amount of coal; Real-time fuel amount estimation at t' moment and target load optimization control are completed.

2. A method for real-time fuel estimation and optimal control of a fluidized bed boiler according to claim 1, characterized in that: Step S5 further comprises: when optimization control of the bed temperature at t'+1 moment is needed, step S7 is entered; S7, according to the boiler data at t'-L+1 to t' moment, using the method of step S2 to obtain heat release parameters α', β', γ', δ', and then according to the change amount of the coal into the furnace, the heat release coefficient, and the bed temperature T at t' moment, the bed temperature T at t'+1 moment is predicted recursively t’+1 , according to the bed temperature T at t'+1 moment t’+1 and the target bed temperature T0', the bed temperature at t'+1 moment is optimized and controlled by controlling the air-coal ratio, the primary and secondary air volume ratio, and the slag amount slag t ', and the optimization control of the bed temperature at t'+1 moment is completed.

3. A method for real-time fuel estimation and optimal control of a fluidized bed boiler according to claim 1, characterized in that: In step S1, the boiler operation history data with time sequence characteristics is obtained from boiler operation history operation data or assay report data; When there is a missing value in the boiler operation history data with time sequence characteristics, a spline interpolation method is used for filling; A Savitzky-Golay smoothing filter is used for denoising; dc t = coal t -coal t-1 dT t = T t - T t-1 wherein coal t is the amount of coal entering the furnace at time t, T t is the bed temperature at time t.

4. A method for real-time fuel estimation and optimal control of a fluidized bed boiler according to claim 3, characterized in that: In step S1, the spline interpolation method is: a cubic spline interpolation is performed on time sequence data by using a python mathematical calculation library scipy.

5. A method for real-time fuel estimation and optimal control of a fluidized bed boiler according to claim 1, characterized in that: L is a positive integer in the interval [10, 20].

6. A method for real-time fuel estimation and optimal control of a fluidized bed boiler according to claim 1, characterized in that: In step S3, the dynamic weight [w1, w2, … wL] of the heat release process after the fuel entering the furnace is calculated. L When 1 unit of fuel enters the furnace, w1, w2, … wL will be burned respectively at the next L time points and release heat. L When 1 unit of fuel enters the furnace, w1, w2, … wL will be burned respectively at the next L time points and release heat.

7. A method for real-time fuel estimation and optimal control of a fluidized bed boiler according to claim 1, characterized in that: In step S4, k is updated in real time in units of 6 hours to 24 hours.

8. A method of real-time fuel estimation and optimal control of a fluidized bed boiler according to claim 2, characterized in that: In step S7, Wherein, T t’ is the bed temperature at t' moment, dc t’-i+1 is the change of coal quantity at t'-i+1 moment, e t’-i+1 is the heat release coefficient at t'-i+1 moment.

Images