Industrial boiler combustion oscillation modal suppression system and method thereof

Abstract

The invention relates to the technical field of energy conservation and environmental protection, in particular to an industrial boiler combustion oscillation modal suppression system and a method thereof, and constructs a system which is highly intelligent, high in adaptability and remarkable in suppression effect by fusing innovative technologies such as multi-dimensional perception, intelligent identification, active suppression and adaptive optimization. The data acquisition module accurately captures hearth temperature, flame video and sound pressure signals; the signal processing module deeply analyzes and extracts oscillation modes of a temperature field and a flame area, and analyzes a sound pressure signal frequency spectrum through short-time Fourier transform; the mode identification module accurately identifies the combustion oscillation mode and determines the characteristic frequency and amplitude of the combustion oscillation mode; the suppression control module generates an infrasonic wave intervention signal to form an anti-phase sound field so as to effectively suppress oscillation; the fuel adjusting module adjusts the opening time sequence of the gas nozzle according to the instruction, and fuel supply is optimized. According to the system, the operation stability and heat efficiency of the boiler are remarkably improved, NOx emission is greatly reduced, and powerful technical support is provided for clean and efficient operation of the industrial boiler.

Description

Technical Field

[0001] The present invention relates to the technical field of energy conservation and environmental protection, in particular to an industrial boiler combustion oscillation mode suppression system and method. Background Art

[0002] As a key device for energy conversion, industrial boilers play a crucial role in industrial production. However, during the actual operation process, the combustion oscillation problem has always been one of the main obstacles that plague the stable operation of boilers. Combustion oscillation not only reduces the thermal efficiency of the boiler, but may also cause equipment damage and even lead to safety accidents. Therefore, effectively suppressing combustion oscillation is of great significance for improving the operation efficiency of boilers, extending the equipment life, and ensuring safe production.

[0003] In recent years, with the development of sensing technology, signal processing, and control theory, certain progress has been made in combustion oscillation suppression technology. Traditional methods mainly rely on PID control to suppress oscillation by adjusting the coal feeding amount and air volume. However, this method is insensitive to the complex combustion dynamic characteristics and difficult to cope with rapidly changing working conditions. In addition, the tuning of PID parameters often relies on experience and is difficult to meet the requirements of different working conditions.

[0004] Subsequently, the method based on fuzzy control was introduced into the field of combustion oscillation suppression. This method constructs fuzzy rules using expert experience and improves the adaptability of the system to a certain extent. However, the effect of fuzzy control highly depends on the quality and completeness of expert knowledge, and its performance still has limitations in the face of complex and changeable combustion processes.

[0005] Recently, some researchers have tried to apply machine learning technology to combustion oscillation suppression. Although these methods show promising effects in some aspects, there are still the following problems: First, most methods only rely on a single data source (such as pressure or temperature) and are difficult to comprehensively capture the complex dynamic characteristics of the combustion process; second, the existing identification algorithms are not precise enough in classifying combustion oscillation modes and are difficult to provide a reliable basis for subsequent suppression control; third, most of the existing suppression methods adopt passive control strategies and are difficult to quickly respond to and effectively suppress sudden strong oscillations; finally, the existing systems generally lack the adaptive ability and are difficult to cope with complex working conditions such as load changes and fuel property changes. Summary of the Invention

[0006] The present invention aims to solve the above technical problems and provides an intelligent system and method that can accurately identify, quickly respond to, and effectively suppress combustion oscillation.

[0007] The present invention proposes an industrial boiler combustion oscillation mode suppression system, including:

[0008] A data acquisition module, for:

[0009] Collect furnace temperature field data;

[0010] Obtain flame video data;

[0011] Receive sound pressure signals;

[0012] A signal processing module, electrically connected to the data acquisition module, for:

[0013] Analyze the furnace temperature field data and extract the temperature field oscillation mode;

[0014] Based on the flame video data, calculate the flame area oscillation mode;

[0015] Process the sound pressure signal using short-time Fourier transform to obtain spectral features;

[0016] A mode recognition module, data-connected to the signal processing module, for:

[0017] Based on the temperature field oscillation mode, flame area oscillation mode, and sound pressure signal spectral features, identify the combustion oscillation mode;

[0018] Determine the characteristic frequency and amplitude of the combustion oscillation mode;

[0019] A suppression control module, communicatively connected to the mode recognition module, for:

[0020] Generate an infrasound interference signal according to the characteristic frequency and amplitude of the combustion oscillation mode;

[0021] Control the infrasound emitter to output the infrasound interference signal to form an anti-phase sound field in the combustion chamber;

[0022] A fuel regulation module, signal-connected to the suppression control module, for:

[0023] Receive the regulation instruction sent by the suppression control module;

[0024] Based on the regulation instruction, adjust the opening timing of the gas nozzles of the fuel supply system.

[0025] Preferably, the data acquisition module includes:

[0026] A temperature acquisition unit, including multiple rows of thermocouples, arranged along the furnace length direction, 20 in each layer, a total of 160 thermocouples, with a sampling frequency of 1 Hz;

[0027] An image acquisition unit, including 60 industrial wide-angle cameras, installed at the furnace outlet position, and using the frame difference method to extract combustion frequency information;

[0028] The sound pressure acquisition unit includes several groups of sound / vibration composite sensors arranged inside the furnace, and the sampling frequencies of the sound pressure and vibration signals are 2 MHz.

[0029] Preferably, the signal processing module includes:

[0030] A temperature processing unit for performing spatial interpolation and time series analysis on the furnace temperature field data to generate temperature field oscillation characteristics;

[0031] An image processing unit for calculating the flame area using an improved adaptive threshold segmentation algorithm and performing time series analysis;

[0032] A sound pressure processing unit for performing short-time Fourier transform and blind source separation on the sound pressure signal to extract the characteristic frequencies of the combustion oscillation source signal.

[0033] Preferably, the modal recognition module includes:

[0034] A feature fusion unit for performing multi-dimensional data fusion on the temperature field oscillation characteristics, flame area oscillation characteristics, and sound pressure signal characteristic frequencies;

[0035] A modal classification unit for classifying and identifying the combustion oscillation modes based on the fused multi-dimensional features using a machine learning algorithm;

[0036] A stability evaluation unit for constructing a combustion stability map in combination with a computational fluid dynamics model to evaluate the stability of the current combustion state.

[0037] Preferably, the suppression control module includes:

[0038] A phase compensation unit for calculating the infrasound trigger delay and amplitude according to the identified combustion oscillation mode characteristic frequencies;

[0039] A waveform generation unit for generating an infrasound intervention signal based on the trigger delay and amplitude;

[0040] An emission control unit for controlling the infrasound emitter array arranged at the furnace mouth to output the infrasound intervention signal.

[0041] Preferably, the infrasound emitter array includes:

[0042] Six groups of infrasound emitters evenly arranged along the circumferential direction of the furnace, with each group containing 4 infrasound emitters;

[0043] The relative angular spacing between adjacent infrasound emitters within each group is 15°;

[0044] The infrasound emitters are fixedly installed along the height direction of the furnace.

[0045] Preferably, the fuel regulation module includes:

[0046] A nozzle control unit for independently controlling the opening sequence of multiple gas nozzles;

[0047] A supply optimization unit for optimizing the fuel supply ratio according to the concentration of flue gas components collected in real time;

[0048] An adaptive regulation unit for dynamically adjusting the fuel supply strategy according to the change trend of the combustion oscillation mode.

[0049] Preferably, it further includes a safety monitoring module, which is communicatively connected to the mode recognition module and the suppression control module and is used for:

[0050] Real-time monitoring of the type and amplitude of the combustion oscillation mode;

[0051] Judging the degree of system abnormality according to a preset threshold;

[0052] When a serious abnormality is detected, triggering emergency measures and sending an alarm signal to the staff.

[0053] Preferably, it further includes a performance optimization module, which is data-connected to the suppression control module and the fuel regulation module and is used for:

[0054] Comprehensively analyzing the furnace temperature field data, combustion stability index and suppression effect;

[0055] Based on a multi-objective optimization algorithm, dynamically adjusting the suppression strategy and fuel supply parameters;

[0056] Through iterative optimization, continuously improving the suppression effect and combustion efficiency of the system.

[0057] A method for suppressing the combustion oscillation mode of an industrial boiler includes the following steps:

[0058] S1. Multi-dimensional data acquisition:

[0059] Collecting temperature field data by using multiple rows of thermocouples arranged in the furnace;

[0060] Obtaining flame video data through an industrial wide-angle camera installed at the furnace outlet;

[0061] Collecting the furnace pressure fluctuation signal by using an acoustic / vibration composite sensor;

[0062] S2. Signal processing and feature extraction:

[0063] Performing spatial interpolation and time series analysis on the temperature field data to extract the temperature field oscillation characteristics;

[0064] Calculating the flame area by using an improved adaptive threshold segmentation algorithm and performing time series analysis;

[0065] Process the sound pressure signal using the short-time Fourier transform and blind source separation algorithm to obtain spectral features;

[0066] S3. Combustion oscillation mode identification:

[0067] Perform multi-dimensional data fusion on the oscillation characteristics of the temperature field, the oscillation characteristics of the flame area, and the spectral characteristics of the sound pressure signal;

[0068] Use machine learning algorithms to classify the fused multi-dimensional features and identify the combustion oscillation modes;

[0069] Combine with the computational fluid dynamics model to construct a combustion stability map and evaluate the current combustion state;

[0070] S4. Active suppression control:

[0071] Calculate the infrasound trigger delay and amplitude according to the identified characteristic frequency of the combustion oscillation mode;

[0072] Generate an infrasound intervention signal and control the infrasound emitter array at the furnace mouth to output this signal;

[0073] Utilize the Helmholtz resonance principle to form an anti-phase sound field in the combustion chamber to suppress the oscillation modes in specific frequency bands;

[0074] S5. Adaptive fuel regulation:

[0075] Real-time monitor the flue gas component concentration and optimize the fuel supply ratio;

[0076] According to the change trend of the combustion oscillation mode, dynamically adjust the opening sequence of multiple gas nozzles;

[0077] Through iterative optimization, continuously improve the combustion efficiency and system stability;

[0078] S6. Safety monitoring and performance optimization:

[0079] Continuously monitor the type and amplitude of the combustion oscillation mode and evaluate the degree of system abnormality;

[0080] When a serious abnormality is detected, trigger emergency measures and alarm;

[0081] Comprehensively analyze the system operation data and use multi-objective optimization algorithms to dynamically adjust the suppression strategy and operation parameters.

[0082] The industrial boiler combustion oscillation mode suppression system and method of the present invention have the following beneficial effects:

[0083] First, from a macroscopic perspective, the present invention constructs a complete closed-loop control system, realizing the full-process intelligent management from data acquisition, signal processing, modal identification to suppression control and fuel regulation. This system-level integration greatly improves the overall effect of combustion oscillation suppression, providing all-round guarantee for the stable operation of industrial boilers.

[0084] Secondly, the present invention adopts multi-dimensional data fusion technology, comprehensively utilizing multi-source information such as temperature field, flame image and sound pressure signal, greatly enriching the sensing ability of the system. This multi-dimensional sensing not only improves the accuracy of oscillation identification, but also lays a solid data foundation for subsequent precise control. Especially under complex working conditions, the collaborative analysis of multi-dimensional data can capture subtle changes that are difficult to detect by a single data source, greatly enhancing the robustness of the system.

[0085] In terms of signal processing, the present invention adopts a series of advanced algorithms, such as improved adaptive threshold segmentation, short-time Fourier transform and blind source separation. The ingenious combination of these algorithms not only improves the accuracy of feature extraction, but also significantly enhances the real-time performance of the system. For example, the improved adaptive threshold segmentation algorithm can better adapt to the dynamic changes of flame brightness, providing more reliable input for subsequent flame area calculation.

[0086] The modal identification module is a major highlight of the present invention. By integrating multiple machine learning algorithms, such as random forest and support vector machine, the system can more accurately identify and classify different types of combustion oscillation modes. This accurate modal identification provides a precise decision-making basis for subsequent suppression control, enabling the system to adopt the most suitable suppression strategy for different types of oscillations.

[0087] In terms of suppression control, the present invention adopts an active suppression method based on the Helmholtz resonance principle. By precisely controlling the trigger delay and amplitude of infrasound, the system can form an effective anti-phase sound field in the combustion chamber, thereby quickly suppressing oscillations in a specific frequency band. This active suppression method has the advantages of faster response and better effect compared with traditional passive control.

[0088] The innovative design of the fuel regulation module further improves the adaptability of the system. By real-time optimizing the fuel supply ratio and nozzle opening timing, the system can actively adjust the combustion characteristics, reducing the generation of oscillations from the source. This feedforward + feedback control strategy greatly enhances the prevention ability and adjustment flexibility of the system.

[0089] It is worth mentioning that the present invention also incorporates advanced safety monitoring and performance optimization functions. The multi-level alarm mechanism and the emergency response system based on risk assessment greatly improve the safety of boiler operation. At the same time, the continuous optimization strategy based on deep reinforcement learning enables the system to continuously learn and improve, adapt to changing working conditions, and achieve long-term performance improvement.

[0090] From a microscopic perspective, the present invention has achieved breakthroughs at multiple key technical points. For example, in the data acquisition link, the use of a high-density thermocouple array and a high-speed CMOS camera has greatly improved the spatial and temporal resolution of the data. At the algorithm level, the improved STFT algorithm and the data fusion method based on the DS evidence theory have significantly enhanced the quality of feature extraction and information integration. The cumulative effect of these microscopic innovations is ultimately reflected in a qualitative leap in the overall performance of the system.

[0091] In summary, through the organic combination of innovative technologies such as multi-dimensional perception, intelligent recognition, active suppression, and adaptive optimization, the present invention constructs an industrial boiler combustion oscillation mode suppression system with high intelligence, strong self-adaptability, and remarkable suppression effect. This system can not only effectively improve the operation stability and thermal efficiency of the boiler, but also significantly reduce NOx emissions, providing strong technical support for the clean and efficient operation of industrial boilers. The successful application of the present invention will greatly promote the development of industrial boilers towards intelligence, high efficiency, and cleanliness, and has important practical significance for improving energy utilization efficiency and reducing environmental pollution. Brief Description of the Drawings

[0092] Figure 1 is the high-level logic diagram of the overall system of the present invention;

[0093] Figure 2 is the logic diagram of the data acquisition module of the present invention;

[0094] Figure 3 is the logic diagram of the signal processing module of the present invention;

[0095] Figure 4 is the logic diagram of the mode recognition module of the present invention;

[0096] Figure 5 is the logic diagram of the suppression control module of the present invention;

[0097] Figure 6 is the logic diagram of the fuel regulation module of the present invention;

[0098] Figure 7 is the logic diagram of the safety monitoring module of the present invention;

[0099] Figure 8 is the logic diagram of the performance optimization module of the present invention. Detailed Embodiments

[0100] Please refer to the attached Figure 1-8 , the present invention provides an industrial boiler combustion oscillation mode suppression system and its method. The system includes a data acquisition module 1, a signal processing module 2, a mode recognition module 3, a suppression control module 4, and a fuel regulation module 5.

[0101] The data acquisition module 1 is used to collect furnace temperature field data, obtain flame video data, and receive sound pressure signals. Preferably, the data acquisition module 1 includes a temperature acquisition unit 11, an image acquisition unit 12, and a sound pressure acquisition unit 13. The temperature acquisition unit 11 includes multiple rows of thermocouples arranged along the length direction of the furnace, with 20 thermocouples in each layer and a total of 160 thermocouples, and the sampling frequency is 1 Hz. This arrangement can comprehensively cover the furnace temperature distribution and provide a rich data basis for subsequent analysis. The image acquisition unit 12 includes 60 industrial wide-angle cameras installed at the furnace outlet. The frame difference method is used to extract combustion frequency information. Industrial wide-angle cameras usually adopt air-cooling or water-cooling technology and can work for a long time in a high-temperature environment of up to 2000 °C. The use of a large number of cameras can capture subtle changes in the flame morphology and improve the analysis accuracy. The sound pressure acquisition unit 13 includes several groups of sound / vibration composite sensors arranged inside the furnace, and the sampling frequencies of the sound pressure and vibration signals are 2 MHz. The high sampling frequency ensures the accurate capture of high-frequency oscillations.

[0102] The signal processing module 2 is electrically connected to the data acquisition module 1 and is used to analyze the furnace temperature field data, extract the oscillation mode of the temperature field; calculate the oscillation mode of the flame area based on the flame video data; and process the sound pressure signal using the short-time Fourier transform to obtain the spectral characteristics. In an embodiment of the present invention, the signal processing module 2 includes a temperature processing unit 21, an image processing unit 22, and a sound pressure processing unit 23. The temperature processing unit 21 performs spatial interpolation and time series analysis on the furnace temperature field data to generate the oscillation characteristics of the temperature field. The image processing unit 22 calculates the flame area using an improved adaptive threshold segmentation algorithm and performs time series analysis. The sound pressure processing unit 23 performs short-time Fourier transform and blind source separation on the sound pressure signal to extract the characteristic frequencies of the combustion oscillation source signal.

[0103] The mode recognition module 3 is data-connected to the signal processing module 2 and is used to identify the combustion oscillation mode based on the oscillation mode of the temperature field, the oscillation mode of the flame area, and the spectral characteristics of the sound pressure signal; determine the characteristic frequencies and amplitudes of the combustion oscillation mode. Preferably, the mode recognition module 3 includes a feature fusion unit 31, a mode classification unit 32, and a stability evaluation unit 33. The feature fusion unit 31 performs multi-dimensional data fusion on the oscillation characteristics of the temperature field, the oscillation characteristics of the flame area, and the characteristic frequencies of the sound pressure signal. The mode classification unit 32 classifies and identifies the combustion oscillation mode using a machine learning algorithm based on the fused multi-dimensional features. The stability evaluation unit 33 combines the computational fluid dynamics model to construct a combustion stability map and evaluate the stability of the current combustion state.

[0104] In practical applications, information transmission is achieved between the various modules and units of the system of the present invention through data lines or wireless communication means. For example, the data acquisition module 1 can be connected to the signal processing module 2 via a high-speed Ethernet to ensure real-time transmission of a large amount of data. The signal processing module 2 and the modal identification module 3 can be connected using a PCI bus to improve the data processing speed.

[0105] The industrial boiler combustion oscillation mode suppression system of the present invention achieves precise identification of combustion oscillations through multi-dimensional data acquisition and advanced signal processing techniques. The high-density thermocouple array (160 pieces) and multi-angle camera arrangement (60 pieces) adopted by the system greatly improve the spatial resolution of data acquisition. At the same time, the high-sampling-frequency sound pressure sensor with a frequency of 2 MHz ensures accurate capture of high-frequency oscillations. These hardware configurations lay a solid data foundation for subsequent signal processing and modal identification.

[0106] In the signal processing stage, the present invention adopts a variety of advanced algorithms. For example, for temperature field data, the system uses Kriging interpolation method for spatial interpolation, which takes into account spatial autocorrelation and can more accurately reconstruct the temperature field distribution. Wavelet transform method is used for time series analysis, which can obtain time-domain and frequency-domain information simultaneously and helps to capture transient changes. For flame image processing, the improved adaptive threshold segmentation algorithm adopts a method based on local entropy, which can better adapt to the dynamic changes of flame brightness. In the processing of sound pressure signals, the window length of the short-time Fourier transform is dynamically adjusted according to the actual working conditions, usually between 10 - 50 ms, to balance time-frequency resolution.

[0107] The core of the modal identification module 3 lies in the application of its multi-dimensional feature fusion and machine learning algorithms. The feature fusion adopts a method based on Dempster-Shafer evidence theory, which can effectively handle the uncertainty of data from different sources. Modal classification uses an ensemble learning method, combining random forest and support vector machine, to improve the accuracy and robustness of classification. The stability evaluation unit 33 constructs a CFD model using the large eddy simulation (LES) method, which can better capture the transient characteristics of turbulent combustion compared with the traditional RANS method.

[0108] A significant advantage of the system of the present invention lies in its adaptability and real-time performance. For example, the threshold of modal identification is dynamically adjusted according to the boiler load. When operating at full load, the system may set the oscillation amplitude threshold to 5% of the average pressure, while at low load, this threshold may be reduced to 3% to improve the sensitivity of the system.

[0109] Generally speaking, the combustion oscillation mode suppression system of the present invention realizes the accurate identification and classification of combustion oscillations through multi-dimensional data fusion, advanced signal processing, and intelligent mode recognition. The adaptability and real-time performance of the system enable it to adapt to different working conditions, providing a reliable decision-making basis for subsequent oscillation suppression and effectively improving the operating stability and efficiency of industrial boilers.

[0110] The suppression control module 4 of the present invention is communicatively connected to the mode recognition module 3 and is configured to generate a infrasonic wave intervention signal according to the characteristic frequency and amplitude of the combustion oscillation mode; control the infrasonic wave transmitter to output the infrasonic wave intervention signal to form an anti-phase sound field in the combustion chamber. Preferably, the suppression control module 4 includes a phase compensation unit 41, a waveform generation unit 42, and a transmission control unit 43.

[0111] The phase compensation unit 41 calculates the infrasonic wave trigger delay and amplitude according to the identified characteristic frequency of the combustion oscillation mode. In one embodiment of the present invention, the calculation formula for the trigger delay τ is as follows:

[0112]

[0113] where φ is the desired phase difference (usually π to form anti-phase interference), and f is the identified oscillation characteristic frequency.

[0114] The waveform generation unit 42 generates an infrasonic wave intervention signal based on the trigger delay and amplitude. The present invention adopts an improved sine sweep signal as the intervention waveform, and its expression is:

[0115] s(t) = Asin(2πf(t)t),

[0116] where A is the amplitude, and f(t) is the instantaneous frequency varying with time, which is defined as:

[0117]

[0118] Here, f 0 and f 1 are the starting and ending frequencies respectively, and T is the sweep period. This waveform design can cover a certain frequency range and improve the adaptability of the system to small frequency changes.

[0119] The transmission control unit 43 controls the infrasonic wave transmitter array arranged at the furnace mouth position to output the infrasonic wave intervention signal. In a preferred embodiment of the present invention, the infrasonic wave transmitter array includes six groups of infrasonic wave transmitters evenly arranged along the circumferential direction of the furnace chamber, and each group contains 4 infrasonic wave transmitters. The interval between each group of infrasonic wave transmitters is 60°, and the relative angular interval between adjacent infrasonic wave transmitters within each group is 15°. This arrangement can form an all-round sound field coverage and improve the uniformity of the suppression effect.

[0120] The fuel regulation module 5 of the present invention is signal-connected to the suppression control module 4 and is used to receive the regulation instruction sent by the suppression control module 4; based on the regulation instruction, the opening timing of the gas nozzles of the fuel supply system is adjusted. Preferably, the fuel regulation module 5 includes a nozzle control unit 51, a supply optimization unit 52, and an adaptive regulation unit 53.

[0121] The nozzle control unit 51 is used to independently control the opening timing of multiple gas nozzles. In an embodiment of the present invention, the system uses pulse width modulation (PWM) technology to control the gas nozzles, and its duty cycle D is calculated by the following formula:

[0122]

[0123] where Q required is the required gas flow rate, and Q max is the maximum flow rate of the nozzle.

[0124] The supply optimization unit 52 optimizes the fuel supply ratio according to the concentration of flue gas components collected in real time. The present invention uses a fuzzy control algorithm to achieve fine adjustment of the fuel ratio. The input variables are O 2 concentration and CO concentration, and the output variable is the fuel adjustment amount. An example of the fuzzy rule is as follows:

[0125] If the O 2 concentration is high and the CO concentration is low, then reduce the air volume

[0126] If the O 2 concentration is low and the CO concentration is high, then increase the air volume

[0127] The adaptive regulation unit 53 dynamically adjusts the fuel supply strategy according to the change trend of the combustion oscillation mode. This unit uses a method based on model predictive control (MPC), and its objective function J is defined as:

[0128]

[0129] where y is the system output (such as the oscillation amplitude), r is the reference trajectory, u is the control input (fuel supply amount), N p and N c are the prediction and control time domains respectively, and Q and R are weight matrices.

[0130] The system of the present invention further includes a safety monitoring module 6, which is communicatively connected to the mode recognition module 3 and the suppression control module 4. This module is used to monitor the type and amplitude of the combustion oscillation mode in real time; judge the degree of system abnormality according to the preset threshold; when a serious abnormality is detected, trigger emergency measures and send an alarm signal to the staff.

[0131] In a preferred embodiment of the present invention, the safety monitoring module 6 adopts a multi-level alarm mechanism. For example, when the oscillation amplitude reaches 8% of the average pressure, the system enters the yellow warning state and increases the data sampling frequency; when the amplitude reaches 12%, it enters the orange warning state and activates the standby suppression measures; when the amplitude exceeds 15%, it enters the red warning state, and the system will automatically reduce the boiler load and notify the operator. This grading mechanism can maximize the normal operation of the system while ensuring safety.

[0132] The present invention further includes a performance optimization module 7, which is data-connected to the suppression control module 4 and the fuel regulation module 5. This module is used to comprehensively analyze the furnace temperature field data, combustion stability index, and suppression effect; based on the multi-objective optimization algorithm, dynamically adjust the suppression strategy and fuel supply parameters; and continuously improve the suppression effect and combustion efficiency of the system through iterative optimization.

[0133] During the performance optimization process, the present invention adopts an improved particle swarm optimization (PSO) algorithm. Its velocity update formula is:

[0134]

[0135] where w is the inertia weight, c 1 、c 2 、c 3 are the acceleration constants, r 1 、r 2 、r 3 are random numbers, and p i 、p g 、p l are the individual best, global best, and local best positions respectively. This improvement takes into account the local best information, can better balance global exploration and local development, and improve the optimization efficiency.

[0136] Through the collaborative work of the above modules, the combustion oscillation mode suppression system of the industrial boiler of the present invention can achieve accurate identification, effective suppression, and continuous optimization of combustion oscillation. The self-adaptive and intelligent features of the system enable it to adapt to complex and changeable working conditions, significantly improving the operation stability and efficiency of the industrial boiler. At the same time, the design of the multi-level safety mechanism ensures the reliability of the system under abnormal conditions, providing a strong guarantee for industrial production.

[0137] In a preferred embodiment of the present invention, the safety monitoring module 6 further includes a risk assessment unit 61 and an emergency response unit 62. The risk assessment unit 61 adopts a dynamic risk assessment model based on a Bayesian network to calculate the risk index of the system in real time. This model considers multiple factors, including but not limited to combustion oscillation amplitude, frequency, duration, and boiler load. The calculation formula of the risk index R is as follows:

[0138]

[0139] Among them, w i is the weight of each risk factor, and P(E i |O) is the conditional probability of event E i occurring under the given observation condition O. This dynamic risk assessment method can more accurately reflect the real-time state of the system and provide a reliable basis for emergency decision-making.

[0140] According to the risk assessment results, the emergency response unit 62 automatically triggers corresponding levels of emergency measures. For example, when the risk index is between 0.6 and 0.8, the system will automatically adjust the combustion parameters and increase the monitoring frequency; when the risk index exceeds 0.8, the system will start an emergency load reduction program and send a high-priority alarm to the central control room. This hierarchical response mechanism can minimize the impact on normal production while ensuring safety.

[0141] The performance optimization module 7 of the present invention further includes a data mining unit 71 and a policy generation unit 72. The data mining unit 71 uses deep learning algorithms to extract valuable patterns and rules from a large amount of historical operation data. The present invention adopts a time series prediction model based on the long short-term memory network (LSTM), and its structure is as follows:

[0142] f t =σ(W f ·[h t-1 ,x t +b f ),

[0143] i t =σ(W i ·[h t-1 ,x t +b i ),

[0144]

[0145] o t =σ(W o ·[h t-1 ,x t +b o ),

[0146] h t =o t *tanh(C t ),

[0147] Among them, f t 、i t 、o t are the forget gate, input gate and output gate respectively, and C tis the unit state, h t is the hidden state, and W and b are the weight and bias parameters. This model can effectively capture the long-term dependence of combustion oscillations and improve the prediction accuracy.

[0148] Based on the results of data mining and combined with the reinforcement learning algorithm, the policy generation unit 72 automatically generates an optimization policy. The method adopted in the present invention is based on the deep Q-network (DQN), and its Q-value update formula is:

[0149]

[0150] where s t and a t are the current state and action respectively, r t is the immediate reward, γ is the discount factor, and α is the learning rate. This method can learn the optimal control policy in a complex state space and continuously improve the overall performance of the system.

[0151] The method for suppressing the combustion oscillation mode of the industrial boiler of the present invention includes the following steps:

[0152] S1. Multidimensional data acquisition: Using multiple rows of thermocouples arranged in the furnace to collect temperature field data; obtaining flame video data through an industrial wide-angle camera installed at the furnace outlet; using a sound / vibration composite sensor to collect the furnace pressure fluctuation signal.

[0153] In this step, the distributed optical fiber sensing technology is adopted for the acquisition of temperature field data. Compared with traditional thermocouples, it has a higher spatial resolution (up to 0.1 m) and a faster response speed (<1 s). The high-speed CMOS camera is used for flame image acquisition, and the frame rate can reach 1000 fps, which can capture the rapidly changing flame morphology. The piezoelectric sensor array is used for the acquisition of sound pressure signals, and the sensitivity reaches 1 mV / Pa, which can accurately capture tiny pressure fluctuations.

[0154] S2. Signal processing and feature extraction: Conducting spatial interpolation and time series analysis on the temperature field data to extract the temperature field oscillation characteristics; calculating the flame area using an improved adaptive threshold segmentation algorithm and conducting time series analysis; processing the sound pressure signal using the short-time Fourier transform and blind source separation algorithm to obtain the spectral characteristics.

[0155] In the processing of temperature field data, the interpolation method based on the radial basis function (RBF) is adopted in the present invention, and its interpolation function is:

[0156]

[0157] where P = (x, y, z) is the point to be interpolated, P iis a known temperature point, φ is a radial basis function, and usually a Gaussian function is selected. This method can better handle irregularly distributed temperature data.

[0158] For flame image processing, an adaptive threshold segmentation algorithm based on local entropy is adopted, and its threshold calculation formula is:

[0159]

[0160] where p 1 (t) and p 2 (t) are the probability distributions of the foreground and background respectively. This method can better adapt to the dynamic changes of flame brightness.

[0161] In the processing of sound pressure signals, the present invention adopts an improved short-time Fourier transform (STFT) algorithm, and its expression is:

[0162]

[0163] where w(t) is a window function, and the present invention adopts a variable-length Kaiser window, which can better balance the time-frequency resolution.

[0164] S3. Combustion oscillation mode identification: Multidimensional data fusion of the temperature field oscillation characteristics, flame area oscillation characteristics, and sound pressure signal spectrum characteristics; Using machine learning algorithms to classify the fused multidimensional features to identify the combustion oscillation mode; Combining with the computational fluid dynamics model to construct a combustion stability map to evaluate the current combustion state.

[0165] In the data fusion stage, the present invention adopts a method based on Dempster-Shafer evidence theory. For features A and B, its fusion rule is:

[0166]

[0167] where m(C) is the basic probability assignment after fusion. This method can effectively handle the uncertainty of data from different sources.

[0168] The mode identification adopts an ensemble learning method, combining random forest and support vector machine. The decision function of the random forest is:

[0169]

[0170] where h k is the prediction result of the k-th decision tree. The decision function of the support vector machine is:

[0171]

[0172] where K(x i, x) is the kernel function, and the present invention adopts the RBF kernel. This integration method can improve the accuracy and robustness of classification.

[0173] Through the above steps, the method for suppressing combustion oscillation modes of industrial boilers in the present invention can achieve accurate identification and classification of combustion oscillations. The innovation of this method is mainly reflected in the comprehensive utilization of multi-dimensional data, the application of advanced signal processing algorithms, and the construction of intelligent recognition models. This method not only improves the accuracy of oscillation recognition, but also provides a reliable decision-making basis for subsequent suppression control, which is of great significance for improving the operation stability and efficiency of industrial boilers.

[0174] To verify the effectiveness of the system and method for suppressing combustion oscillation modes of industrial boilers in the present invention, the present invention selects a certain 300MW circulating fluidized bed boiler as the research object and conducts a comprehensive simulation test. This boiler adopts low-temperature and low-pressure circulating fluidized bed combustion technology, with pulverized coal as the main fuel and a designed evaporation capacity of 1000t / h.

[0175] The simulation conditions are set as follows:

[0176] Boiler load: 70%, 85%, 100%;

[0177] Primary air velocity: 4.5m / s, 5.0m / s, 5.5m / s;

[0178] Secondary air volume ratio: 30%, 35%, 40%;

[0179] Bed temperature: 850°C, 880°C, 910°C;

[0180] The present invention selects three working conditions for comparative testing:

[0181] Example 1: Adopt the system and method for suppressing combustion oscillation modes of industrial boilers in the present invention.

[0182] Comparative Example 1: Adopt the traditional PID control method to suppress combustion oscillation by adjusting the coal feeding amount and air volume.

[0183] Comparative Example 2: Adopt a combustion oscillation suppression method based on fuzzy control and construct fuzzy rules using expert experience.

[0184] The test indexes and their detection methods are as follows:

[0185] 1. Combustion oscillation amplitude: Measure the furnace pressure fluctuation using a pressure sensor and obtain the amplitude of the main oscillation frequency through Fourier analysis.

[0186] 2. Oscillation recognition accuracy rate: Manually set different types of combustion oscillations and count the proportion correctly recognized by the system.

[0187] 3. Suppression response time: The time interval from detecting significant oscillation to effective suppression.

[0188] 4. Boiler thermal efficiency: Measured according to GB / T 10184 - 2015 "Regulations for Thermal Performance Tests of Industrial Boilers".

[0189] 5. NOx emission concentration: Continuously monitored using an on - line flue gas analyzer.

[0190] 6. System stability: Operate continuously for 100 hours and record the number of times the system experiences abnormalities or requires manual intervention.

[0191] The test results are shown in the following table:

[0192]

[0193] Analysis and discussion are as follows:

[0194] 1. Combustion oscillation amplitude: The method of the present invention (Example 1) significantly reduces the amplitude of combustion oscillation, reducing it by 58.3% and 37.5% compared to Comparative Example 1 and Comparative Example 2 respectively. This is mainly due to the multi - dimensional data fusion and precise modal recognition techniques adopted by the present invention, which can capture the characteristics of combustion oscillation more accurately, thus achieving more effective suppression.

[0195] 2. Oscillation recognition accuracy: The method of the present invention reaches a high recognition accuracy of 95.3%, far exceeding 82.7% of Comparative Example 2. This is attributed to the advanced signal processing algorithms and machine learning models adopted by the present invention, which can better handle the complex combustion dynamic characteristics.

[0196] 3. Suppression response time: The method of the present invention has the fastest response speed and can effectively suppress oscillation in only 2.5 ± 0.5 seconds. This benefits from the real - time data processing ability and model - based predictive control strategy of the present invention, which can make a quick response and take targeted suppression measures.

[0197] 4. Boiler thermal efficiency: The method of the present invention slightly improves the boiler thermal efficiency while ensuring the oscillation suppression effect. This is because the system of the present invention can control the combustion process more precisely, reducing unnecessary energy losses.

[0198] 5. NOx emission concentration: The method of the present invention also performs well in NOx emission reduction, reducing it by 14.3% and 8.7% compared to Comparative Example 1 and Comparative Example 2 respectively. This is because the system of the present invention can better optimize combustion parameters to achieve low - NOx combustion.

[0199] 6. System stability: During the 100-hour continuous operation test, the method of the present invention only required manual intervention once, which is much lower than the other two methods. This highlights the high automation and reliability of the system of the present invention.

[0200] Generally speaking, the industrial boiler combustion oscillation mode suppression system and method of the present invention show obvious advantages in all key indicators. It can not only more effectively suppress combustion oscillation, improve the operation stability of the boiler, but also contribute to improving energy efficiency and reducing emissions. These excellent performances are mainly due to the innovative technologies adopted in the present invention, such as multi-dimensional data fusion, advanced signal processing, intelligent mode recognition, and adaptive control.

[0201] It is worth noting that the method of the present invention shows good adaptability under different working conditions. For example, during the process of the boiler load changing from 70% to 100%, the increase amplitude of the combustion oscillation amplitude does not exceed 0.2 kPa, while the traditional method often shows obvious performance degradation when the load changes. This strong adaptability makes the system of the present invention particularly suitable for industrial boilers with frequent load changes.

[0202] In addition, the high stability and reliability shown by the method of the present invention during long-term operation greatly reduce the need for manual intervention, which is beneficial to reducing operation and maintenance costs and improving the overall economy of the boiler. This has important practical application value for modern industrial boilers pursuing efficient, stable, and environmentally friendly operation.

[0203] In summary, the industrial boiler combustion oscillation mode suppression system and method of the present invention show significant technical advantages and economic benefits in practical applications, providing strong support for the intelligent, efficient, and clean operation of industrial boilers.

[0204] It should be noted that the above are only the preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. Industrial boiler combustion oscillation mode suppression system, characterized by ,include: Data acquisition module for: Collect furnace temperature field data; Get flame video data; receiving a sound pressure signal; The signal processing module is electrically connected to the data acquisition module and is used to: Analyzing the furnace temperature field data and extracting the temperature field oscillation mode; Based on the flame video data, calculating the flame area oscillation mode; Processing the sound pressure signal by short-time Fourier transform to obtain frequency spectrum characteristics; The modal identification module is data-connected to the signal processing module and is used to: Based on the temperature field oscillation mode, the flame area oscillation mode and the sound pressure signal spectrum characteristics, the combustion oscillation mode is identified; determining a characteristic frequency and amplitude of the combustion oscillation mode; The suppression control module is communicatively connected with the modality identification module and is used to: generating an infrasonic intervention signal according to the characteristic frequency and amplitude of the combustion oscillation mode; Controlling the infrasound transmitter to output the infrasound wave intervention signal to form an anti-phase sound field in the combustion chamber; A fuel regulating module, signal-connected to the suppression control module, is used to: receiving an adjustment instruction sent by the inhibition control module; Based on the adjustment instruction, the opening timing of the gas nozzle of the fuel supply system is adjusted.

2. The system according to claim 1, characterized in that , the data acquisition module includes: The temperature acquisition unit includes multiple rows of thermocouples arranged along the length of the furnace, with 20 thermocouples in each layer, a total of 160 thermocouples, and a sampling frequency of 1 Hz; The image acquisition unit, including 60 industrial wide-angle cameras, is installed at the furnace exit and uses the frame difference method to extract the combustion frequency information; The sound pressure collection unit includes a plurality of groups of sound / vibration composite sensors arranged inside the furnace, and the sampling frequency of the sound pressure and vibration signals is 2 MHz.

3. The system according to claim 1, characterized in that , the signal processing module comprises: A temperature processing unit, used for performing spatial interpolation and time series analysis on the furnace temperature field data to generate temperature field oscillation characteristics; An image processing unit, used to calculate the flame area using an improved adaptive threshold segmentation algorithm and perform time series analysis; The sound pressure processing unit is used to perform short-time Fourier transform and blind source separation on the sound pressure signal to extract the characteristic frequency of the combustion oscillation source signal.

4. The system according to claim 1, characterized in that , the modality recognition module includes: A feature fusion unit is used to fuse the temperature field oscillation feature, the flame area oscillation feature and the characteristic frequency of the sound pressure signal in multiple dimensions; A mode classification unit, used for classifying and identifying the combustion oscillation mode by using a machine learning algorithm based on the fused multi-dimensional features; The stability assessment unit is used to combine the computational fluid dynamics model, construct a combustion stability map, and assess the stability of the current combustion state.

5. The system according to claim 1, characterized in that , the inhibition control module includes: A phase compensation unit, used to calculate the infrasound triggering delay and amplitude according to the identified combustion oscillation mode characteristic frequency; A waveform generating unit, configured to generate an infrasound wave intervention signal based on the trigger delay and the amplitude; The emission control unit is used to control the infrasound emitter array arranged at the furnace mouth to output the infrasound wave intervention signal.

6. The system according to claim 5, characterized in that , the infrasound emitter array comprises: Six groups of infrasound emitters are evenly arranged along the circumference of the furnace, each group containing four infrasound emitters; Each group of infrasound emitters is spaced 60° apart, and the relative angular spacing between adjacent infrasound emitters within each group is 15°; The infrasound emitter is fixedly installed along the height direction of the furnace.

7. The system according to claim 1, characterized in that , the fuel adjustment module comprises: Nozzle control unit, used to independently control the opening sequence of multiple gas nozzles; Supply optimization unit, used to optimize the fuel supply ratio according to the real-time collected flue gas component concentration; The adaptive regulation unit is used to dynamically adjust the fuel supply strategy according to the changing trend of the combustion oscillation mode.

8. The system according to claim 1, characterized in that , further comprising a safety monitoring module, which is in communication connection with the modal identification module and the inhibition control module, and is used to: Real-time monitoring of the type and amplitude of combustion oscillation modes; Determine the degree of system abnormality based on preset thresholds; When serious anomalies are detected, emergency measures are triggered and an alarm signal is sent to the staff.

9. The system according to claim 1, characterized in that , further comprising a performance optimization module, data-connected to the suppression control module and the fuel adjustment module, for: Comprehensive analysis of furnace temperature field data, combustion stability index and suppression effect; Dynamically adjust the suppression strategy and fuel supply parameters based on a multi-objective optimization algorithm; Through iterative optimization, the suppression effect and combustion efficiency of the system are continuously improved.

10. A method for suppressing combustion oscillation modes of industrial boilers, characterized in that , including the following steps: S1. Multidimensional data collection: The temperature field data is collected using multiple rows of thermocouples arranged in the furnace; Obtain flame video data through an industrial wide-angle camera installed at the furnace exit; Use acoustic / vibration composite sensor to collect furnace pressure fluctuation signal; S2.Signal processing and feature extraction: Perform spatial interpolation and time series analysis on temperature field data to extract temperature field oscillation characteristics; The flame area is calculated using an improved adaptive threshold segmentation algorithm, and time series analysis is performed. Use short-time Fourier transform and blind source separation algorithm to process sound pressure signals and obtain spectrum characteristics; S3. Combustion Oscillation Mode Identification: Multi-dimensional data fusion of temperature field oscillation characteristics, flame area oscillation characteristics and sound pressure signal spectrum characteristics; The machine learning algorithm is used to classify the fused multi-dimensional features and identify the combustion oscillation mode; Combined with the computational fluid dynamics model, a combustion stability map is constructed to evaluate the current combustion state; S4. Active Inhibitory Control: According to the identified combustion oscillation mode characteristic frequency, the infrasound triggering delay and amplitude are calculated; Generate an infrasound intervention signal and control the infrasound transmitter array at the furnace mouth position to output the signal; By using the Helmholtz resonance principle, an anti-phase sound field is formed in the combustion chamber to suppress the oscillation mode in a specific frequency band; S5. Adaptive fuel regulation: Real-time monitoring of flue gas component concentration and optimization of fuel supply ratio; According to the changing trend of the combustion oscillation mode, the opening sequence of multiple gas nozzles is dynamically adjusted; Continuously improve combustion efficiency and system stability through iterative optimization; S6. Security monitoring and performance optimization: Continuously monitor the type and amplitude of combustion oscillation modes and evaluate the degree of system abnormality; When a serious anomaly is detected, emergency measures are triggered and an alarm is issued; Comprehensively analyze system operation data and use multi-objective optimization algorithms to dynamically adjust suppression strategies and operating parameters.

Images