A method and system for predicting the combustion efficiency of a water jacket heating furnace

Abstract

This invention discloses a method and system for predicting the combustion efficiency of a water-jacketed furnace, specifically relating to the field of combustion efficiency prediction technology. It involves acquiring thermal image sequences of the furnace wall surface, extracting pixel temperature gradients and texture change features to identify abnormal heat transfer areas; performing wavelet decomposition on the temperature gradient features of these abnormal areas to calculate the total energy of high-frequency signals; extracting the patterns of scale formation and shedding based on the correlation with temperature fluctuation amplitude; using a dynamic Bayesian network to assess the risk of heat transfer efficiency fluctuations caused by scale changes and outputting the probability of heat transfer fluctuation risk; constructing a coupling model between the furnace combustion process and the probability of heat transfer fluctuation risk, and correcting the heat loss factor; and outputting a real-time predicted value of combustion efficiency based on the corrected heat loss factor, effectively improving the energy efficiency management level of the water-jacketed furnace.

Description

Technical Field

[0001] This invention relates to the field of combustion efficiency prediction technology, and more specifically, to a method and system for predicting the combustion efficiency of a water-jacketed furnace. Background Technology

[0002] The current evaluation of combustion efficiency of water-jacketed furnaces mainly relies on steady-state heat transfer models or empirical heat loss correction coefficients, which are difficult to accurately reflect the unsteady heat transfer fluctuations caused by the dynamic changes in scale on the furnace wall during actual operation.

[0003] Under conditions of frequent water quality changes and load fluctuations, the rate of scale formation and shedding on the furnace wall surface accelerates, causing drastic fluctuations in the instantaneous heat transfer performance of the furnace wall. This results in distortion of the input to the combustion efficiency calculation model, leading to deviations in the predicted combustion efficiency from the actual operating conditions. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a method and system for predicting the combustion efficiency of a water-jacketed heating furnace to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for predicting the combustion efficiency of a water-jacketed furnace includes the following steps:

[0007] S1: Collect thermal image sequences of the furnace wall surface of the water jacket heating furnace, extract pixel temperature gradient and texture change features within the thermal radiation area of ​​the furnace wall, and identify abnormal heat transfer areas of the furnace wall.

[0008] S2: Perform wavelet decomposition on the temperature gradient characteristics of the abnormal heat transfer area of ​​the furnace wall to obtain high-frequency signal components that reflect the process of scale adhesion and detachment, and calculate the energy of the high-frequency signal.

[0009] S3: By analyzing the correlation between high-frequency signal energy and the temperature fluctuation amplitude of the furnace wall, the patterns of scale formation and detachment on the furnace wall are extracted;

[0010] S4: Based on the laws of scale formation and shedding, a dynamic Bayesian network is used to assess the risk of furnace wall heat transfer efficiency fluctuation caused by scale changes, and output the probability of heat transfer fluctuation risk.

[0011] S5: Construct a coupled model between the furnace combustion process and the probability of heat transfer fluctuation risk, and correct the heat loss factor in the calculation of combustion efficiency;

[0012] S6: Based on the corrected heat loss factor, update the combustion efficiency prediction model parameters in real time and output the real-time predicted value of the combustion efficiency of the water jacket heater.

[0013] In a preferred embodiment, S1 specifically refers to:

[0014] Acquire a sequence of thermal images of the furnace wall surface of the water jacket heating furnace, delineate the boundary of the furnace wall thermal radiation area, and determine the temperature value of each pixel in the furnace wall thermal radiation area;

[0015] The temperature difference between any two adjacent pixels within the thermal radiation area of ​​the furnace wall is calculated using the spatial gradient calculation method, thus obtaining the temperature gradient value at each pixel location.

[0016] Energy, contrast and correlation indices were extracted from each frame of thermal image within the thermal radiation area of ​​the furnace wall using the gray-level co-occurrence matrix method to determine texture change characteristics.

[0017] A threshold is set based on the temperature gradient value and texture change characteristics to identify areas of abnormal heat transfer in the furnace wall.

[0018] In a preferred embodiment, S2 specifically refers to:

[0019] The temperature gradient values ​​of each pixel in the abnormal heat transfer area of ​​the furnace wall are decomposed based on the multi-scale wavelet transform method to obtain the high-frequency wavelet coefficients and low-frequency wavelet coefficients corresponding to the temperature gradient values.

[0020] The high-frequency wavelet coefficients are mapped to each pixel in the abnormal heat transfer area of ​​the furnace wall according to their spatial location to obtain the high-frequency signal component distribution of each pixel.

[0021] The high-frequency signal component distribution of each pixel in the abnormal heat transfer area of ​​the furnace wall is squared and calculated. Then, the high-frequency signal component values ​​of each pixel after squared calculation are accumulated point by point to determine the total energy value of the high-frequency signal component in the abnormal heat transfer area of ​​the furnace wall.

[0022] In a preferred embodiment, S3 specifically refers to:

[0023] A scatter plot is constructed with the total energy value of high-frequency signal components in the abnormal heat transfer area of ​​the furnace wall as the horizontal axis and the temperature fluctuation amplitude value of each pixel in the thermal radiation area of ​​the furnace wall as the vertical axis.

[0024] Regression analysis was used to fit the scatter plot to obtain a functional expression for the relationship between the total energy of the high-frequency signal components and the temperature fluctuation amplitude.

[0025] Based on the trend of the total energy of the high-frequency signal components changing over time and the functional expression, the periodic pattern and variation characteristics of scale formation and shedding in the furnace wall thermal radiation area are determined.

[0026] In a preferred embodiment, S4 specifically refers to:

[0027] Based on the periodic patterns and characteristics of scale formation and shedding within the furnace wall thermal radiation area, a dynamic Bayesian network containing hidden state nodes and observation nodes is constructed.

[0028] The periods of scale adhesion and scale shedding are used as hidden state nodes, and the total energy value of high-frequency signal components and the temperature fluctuation amplitude value are used as observation nodes. The probability value of furnace wall heat transfer efficiency fluctuation is determined by calculating the state transition probability and observation probability in the dynamic Bayesian network.

[0029] In a preferred embodiment, S5 specifically refers to:

[0030] Based on the calculation formula for heat loss during furnace combustion, a functional relationship is established between the furnace combustion process and the probability value of furnace wall heat transfer efficiency fluctuation.

[0031] The heat loss factor in the calculation formula for heat loss during furnace combustion is adjusted based on the probability value of the fluctuation of furnace wall heat transfer efficiency, and the correction value of the heat loss factor is determined.

[0032] The corrected heat loss factor is determined based on the corrected value of the heat loss factor.

[0033] In a preferred embodiment, S6 specifically refers to:

[0034] The correction amount of the combustion efficiency prediction model parameters to be corrected in the combustion efficiency prediction model is calculated based on the corrected heat loss factor.

[0035] The correction amount of the combustion efficiency prediction model parameters is added to the initial value of the combustion efficiency prediction model parameters to obtain the real-time updated combustion efficiency prediction model parameters.

[0036] Based on the real-time updated combustion efficiency prediction model parameters, the temperature fluctuation amplitude of each pixel in the thermal radiation area of ​​the water jacket furnace wall, the total energy value of the high-frequency signal components, and the probability value of the furnace wall heat transfer efficiency fluctuation are input into the combustion efficiency prediction model, and the real-time predicted value of the combustion efficiency of the water jacket furnace is output.

[0037] On the other hand, the present invention provides a combustion efficiency prediction system for a water-jacketed heating furnace, comprising:

[0038] Feature extraction module: Collects thermal image sequences of the furnace wall surface of the water jacket heating furnace, extracts pixel temperature gradient and texture change features in the thermal radiation area of ​​the furnace wall, and identifies abnormal heat transfer areas of the furnace wall.

[0039] Wavelet decomposition module: Performs wavelet decomposition on the temperature gradient characteristics of the abnormal heat transfer area of ​​the furnace wall to obtain high-frequency signal components that reflect the process of scale adhesion and detachment, and calculates the energy of the high-frequency signal.

[0040] Pattern recognition module: By analyzing the correlation between high-frequency signal energy and the temperature fluctuation amplitude of the furnace wall, the pattern of scale formation and detachment on the furnace wall is extracted;

[0041] Risk assessment module: Based on the patterns of scale formation and shedding, a dynamic Bayesian network is used to assess the risk of furnace wall heat transfer efficiency fluctuations caused by scale changes, and outputs the probability of heat transfer fluctuation risk.

[0042] Factor correction module: Constructs a coupled model between the furnace combustion process and the probability of heat transfer fluctuation risk, and corrects the heat loss factor in the combustion efficiency calculation;

[0043] Efficiency prediction module: Based on the corrected heat loss factor, the combustion efficiency prediction model parameters are updated in real time, and the real-time predicted value of the combustion efficiency of the water jacket furnace is output.

[0044] The technical effects and advantages of the combustion efficiency prediction method and system for a water-jacketed heating furnace of the present invention are as follows:

[0045] By acquiring thermal image sequences and extracting image features, accurate identification of abnormal heat transfer areas on the furnace wall surface is achieved, improving the accuracy of identifying areas affected by scale. Wavelet decomposition is introduced to extract high-frequency signal components and calculate high-frequency signal energy, effectively characterizing the dynamic features of scale adhesion and detachment. By analyzing the correlation between high-frequency signal energy and temperature fluctuation amplitude, the periodic variation pattern of scale is accurately extracted. A dynamic Bayesian network enables probabilistic assessment of the risk of furnace wall heat transfer efficiency fluctuations. A coupling model between the furnace combustion process and heat transfer risk is constructed, and the heat loss factor is corrected, improving the dynamic response capability and stability of combustion efficiency calculation. By updating the combustion efficiency prediction model parameters in real time, real-time predicted values ​​of combustion efficiency are output, enhancing the accurate perception and dynamic control capability of the water-jacketed heating furnace's operating energy efficiency and improving the accuracy of thermal efficiency assessment under complex operating conditions. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of a method for predicting the combustion efficiency of a water-jacketed heating furnace according to the present invention;

[0047] Figure 2 This is a schematic diagram of the combustion efficiency prediction system for a water-jacketed heating furnace according to the present invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0049] Example 1

[0050] Figure 1 This invention provides a method for predicting the combustion efficiency of a water-jacketed heater, which includes the following steps:

[0051] S1: Collect thermal image sequences of the furnace wall surface of the water jacket heating furnace, extract pixel temperature gradient and texture change features within the thermal radiation area of ​​the furnace wall, and identify abnormal heat transfer areas of the furnace wall.

[0052] S2: Perform wavelet decomposition on the temperature gradient characteristics of the abnormal heat transfer area of ​​the furnace wall to obtain high-frequency signal components that reflect the process of scale adhesion and detachment, and calculate the energy of the high-frequency signal.

[0053] S3: By analyzing the correlation between high-frequency signal energy and the temperature fluctuation amplitude of the furnace wall, the patterns of scale formation and detachment on the furnace wall are extracted;

[0054] S4: Based on the laws of scale formation and shedding, a dynamic Bayesian network is used to assess the risk of furnace wall heat transfer efficiency fluctuation caused by scale changes, and output the probability of heat transfer fluctuation risk.

[0055] S5: Construct a coupled model between the furnace combustion process and the probability of heat transfer fluctuation risk, and correct the heat loss factor in the calculation of combustion efficiency;

[0056] S6: Based on the corrected heat loss factor, update the combustion efficiency prediction model parameters in real time and output the real-time predicted value of the combustion efficiency of the water jacket heater.

[0057] S1: Acquire a sequence of thermal images of the furnace wall surface of the water-jacketed heating furnace, extract pixel temperature gradients and texture change features within the thermal radiation area of ​​the furnace wall, and identify abnormal heat transfer areas of the furnace wall, including:

[0058] The thermal image sequence of the furnace wall surface of the water jacket heating furnace is acquired, the boundary position of the furnace wall thermal radiation area is delineated, and the temperature value of each pixel in the furnace wall thermal radiation area is determined.

[0059] Infrared thermal imaging devices are used to continuously capture images of the furnace wall surface of a water-jacketed heater, forming a time-series thermal image sequence. This sequence needs to cover the entire furnace wall to ensure comprehensive recording of its thermal radiation characteristics under different operating conditions. For example, thermal images of the furnace wall surface are captured continuously at predetermined time intervals, obtaining image data corresponding to multiple moments. The continuously captured image data constitutes the thermal image sequence of the furnace wall surface.

[0060] The boundary of the furnace wall thermal radiation region is delineated in the thermal image sequence. The specific method for delineating the boundary of the furnace wall thermal radiation region is as follows: First, select the area outside the furnace wall in the thermal image that is not affected by furnace wall heating as the background region. Calculate the average temperature value of all pixels within the background region to obtain the average temperature of the background region. Compare the temperature value of each pixel within the furnace wall region with the average temperature of the background region, and calculate the ratio of each pixel's temperature value to the average temperature of the background region. Mark all pixels whose ratio is greater than a specific percentage of the average temperature of the background region. The outermost line connecting all marked pixels is used as the boundary of the furnace wall thermal radiation region, where the specific percentage can be set between 10% and 20% of the average temperature of the background region.

[0061] After delineating the boundary of the furnace wall's thermal radiation area, the temperature value corresponding to each pixel within that area is determined. Specifically, the infrared radiation intensity value corresponding to each pixel within the furnace wall's thermal radiation area is extracted from each frame of the thermal image. This extraction is performed by using image data processing software to read the intensity information of the infrared radiation data recorded by each pixel. The extracted infrared radiation intensity value for each pixel is then converted into its corresponding actual temperature value using the calibration relationship of the infrared thermal imaging device. This calibration relationship employs the standard curve calibrated by the infrared thermal imaging device; that is, the infrared radiation intensity value is corrected by specific calibration coefficients within the instrument, environmental temperature compensation coefficients, and radiation correction factors. This process is repeated to obtain the accurate actual temperature value for each pixel within the furnace wall's thermal radiation area.

[0062] The temperature difference between any two adjacent pixels within the thermal radiation area of ​​the furnace wall is calculated using the spatial gradient calculation method, thus obtaining the temperature gradient value at each pixel location.

[0063] Based on the temperature values ​​of each pixel within the furnace wall's thermal radiation region, a pixel at any location within this region is selected. The temperature values ​​of the pixels adjacent to this pixel in the four cardinal directions (up, down, left, and right) are then determined. The absolute value of the temperature difference between each pixel and its adjacent pixels is calculated. These absolute values ​​are then summed and divided by the number of adjacent pixels involved in the calculation to obtain the temperature gradient value for that pixel. For example, the absolute values ​​of the temperature differences between a pixel and its four adjacent pixels in the four cardinal directions are summed and then divided by the total number of adjacent pixels involved in the difference calculation to obtain the spatial temperature gradient value for that pixel location. Through this process, the temperature gradient values ​​for all pixel locations within the furnace wall's thermal radiation region are obtained.

[0064] Energy, contrast and correlation indices were extracted from each frame of thermal image within the thermal radiation area of ​​the furnace wall using the gray-level co-occurrence matrix method to determine texture change characteristics.

[0065] Each frame of the thermal image is converted into a grayscale image. Specifically, the temperature value of each pixel within the thermal radiation area of ​​the furnace wall is used as the grayscale intensity value. Based on the grayscale image, a grayscale co-occurrence matrix (GCMM) is constructed. The construction method is as follows: within the thermal radiation area of ​​the furnace wall, taking each pixel as a reference, the frequency of the simultaneous occurrence of the grayscale intensity value of each pixel with the grayscale intensity values ​​of its right-hand adjacent pixel, its lower-hand adjacent pixel, and its diagonally adjacent pixel to the lower right is counted. The statistical results are recorded in the GCMM. The rows and columns of the GCMM represent the grayscale intensity values ​​of two adjacent pixels, and each element represents the frequency of occurrence of the grayscale value combination.

[0066] Based on the gray-level co-occurrence matrix (GLCM), energy, contrast, and correlation are calculated respectively. The energy index is calculated by squaring all element values ​​in the GLCM and then summing the squared values. The contrast index is calculated by first calculating the difference in gray intensity values ​​at each element's position (i.e., the difference between the row and column numbers in the GLCM), squaring this difference, multiplying it by the element value, and summing the calculated values ​​for all elements in the GLCM. The correlation index is calculated by calculating the average gray intensity values ​​in both the row and column directions of the GLCM. For each element, the difference between the gray intensity value at that position and the average value is calculated. This difference is multiplied and then multiplied by the element value. The sum of these values ​​is then normalized to obtain the correlation index.

[0067] Based on the temperature gradient value and texture change characteristics, a judgment threshold is set to identify abnormal heat transfer areas in the furnace wall.

[0068] Calculate the average value of the temperature gradient values ​​of all pixel locations within the thermal radiation area of ​​the furnace wall, and set the threshold for judging the temperature gradient values ​​to 120% of the average value; calculate the average values ​​of the energy index, contrast index, and correlation index in the texture change features respectively, and use the arithmetic mean of the energy index, contrast index, and correlation index multiplied by 120% as the corresponding texture change feature judgment threshold.

[0069] Within the furnace wall's thermal radiation area, each pixel is analyzed. First, it's determined whether the temperature gradient value corresponding to the pixel location exceeds a threshold value. Then, the texture change feature indicators corresponding to the pixel location—energy, contrast, and correlation—are each determined to be greater than their respective texture change feature threshold values. If both conditions are met simultaneously—the temperature gradient value at the pixel location exceeds the threshold value, and any one of the three texture change feature indicators exceeds its corresponding threshold value—then the pixel location is marked as an abnormal heat transfer area within the furnace wall. This method completes the identification of abnormal heat transfer areas within the furnace wall.

[0070] S2: Wavelet decomposition is performed on the temperature gradient characteristics of the abnormal heat transfer area of ​​the furnace wall to obtain high-frequency signal components reflecting the scale adhesion and detachment process, and the energy of the high-frequency signal is calculated, including:

[0071] The temperature gradient values ​​of each pixel in the abnormal heat transfer area of ​​the furnace wall are decomposed based on the multi-scale wavelet transform method to obtain the high-frequency wavelet coefficients and low-frequency wavelet coefficients corresponding to the temperature gradient values.

[0072] Temperature gradient values ​​corresponding to each pixel location within the abnormal heat transfer area of ​​the furnace wall are selected to form a one-dimensional spatial distribution sequence of temperature gradient values, which is the temperature gradient value sequence. A multi-scale wavelet transform method is used to progressively decompose the temperature gradient value sequence at multiple levels. The wavelet transform method employs pre-determined wavelet basis functions, such as orthogonal wavelet basis functions, to decompose the temperature gradient value sequence layer by layer according to the spatial scale from large to small. Each level of decomposition obtains high-frequency and low-frequency wavelet coefficients through sequence convolution. The high-frequency wavelet coefficients are used to represent the temperature gradient value sequence. Rapidly changing local features, low-frequency wavelet coefficients represent the slow overall trend of temperature gradient numerical sequence changes; the wavelet transform calculation method is as follows: multiply the temperature gradient value at each pixel location with the wavelet basis function point by point, and then sum the results point by point to obtain the high-frequency wavelet coefficients; multiply the temperature gradient value with the scaling function point by point and then sum the results to obtain the low-frequency wavelet coefficients; repeat the above calculation to complete the decomposition of all spatial scale levels, and finally obtain all high-frequency and low-frequency wavelet coefficients of each pixel location in the abnormal heat transfer area of ​​the furnace wall at multiple scales.

[0073] The high-frequency wavelet coefficients are mapped to each pixel in the abnormal heat transfer area of ​​the furnace wall according to their spatial location to obtain the high-frequency signal component distribution of each pixel.

[0074] The high-frequency wavelet coefficients are mapped to each pixel in the abnormal heat transfer area of ​​the furnace wall according to their spatial location to obtain the high-frequency signal component distribution of each pixel. Specifically, the high-frequency wavelet coefficients are mapped one by one to the spatial location information of the abnormal heat transfer area of ​​the furnace wall. The mapping method is to determine the value of the corresponding high-frequency wavelet coefficient for each pixel location based on the spatial coordinates of each pixel location in the image of the abnormal heat transfer area of ​​the furnace wall. For example, each pixel location corresponds to a set of high-frequency wavelet coefficients in the multi-scale wavelet transform process. The amplitude of this set of high-frequency wavelet coefficients is used as the intensity of the high-frequency signal component of the pixel. The amplitude of the high-frequency wavelet coefficients of each pixel location is marked at the corresponding pixel location to complete the spatial location mapping of the high-frequency signal component distribution of all pixels in the abnormal heat transfer area of ​​the furnace wall, thus obtaining the high-frequency signal component distribution of each pixel.

[0075] The high-frequency signal component distribution of each pixel in the abnormal heat transfer area of ​​the furnace wall is squared and then the high-frequency signal component values ​​of each pixel after squared calculation are accumulated point by point to determine the total energy value of the high-frequency signal component in the abnormal heat transfer area of ​​the furnace wall.

[0076] For the distribution of high-frequency signal components at each pixel within the abnormal heat transfer area of ​​the furnace wall, the amplitude of the high-frequency wavelet coefficient corresponding to each pixel is squared, that is, the value of each high-frequency signal component is multiplied by the value of each high-frequency signal component to obtain the energy value of the high-frequency signal component at each pixel position. The energy values ​​of the high-frequency signal components obtained by squaring all pixel positions within the abnormal heat transfer area of ​​the furnace wall are accumulated point by point. The final value obtained after accumulation is the total energy value of the high-frequency signal components within the abnormal heat transfer area of ​​the furnace wall. For example, the amplitude of the high-frequency wavelet coefficient at each pixel position is first multiplied by itself to obtain the high-frequency signal energy value at each pixel position, and then the addition operation is performed one by one, gradually accumulating until all pixel positions within the abnormal heat transfer area of ​​the furnace wall have been calculated, and finally the total energy value of the high-frequency signal components of the entire abnormal heat transfer area of ​​the furnace wall is obtained.

[0077] S3: By analyzing the correlation between high-frequency signal energy and the temperature fluctuation amplitude of the furnace wall, the patterns of scale formation and detachment on the furnace wall are extracted, including:

[0078] A scatter plot is constructed with the total energy value of high-frequency signal components in the abnormal heat transfer area of ​​the furnace wall as the horizontal axis and the temperature fluctuation amplitude value of each pixel in the thermal radiation area of ​​the furnace wall as the vertical axis.

[0079] The total energy value of the high-frequency signal components within the abnormal heat transfer area of ​​the furnace wall is used to determine the range of values ​​on the horizontal axis of the scatter plot. For each pixel location in the furnace wall's thermal radiation area, the actual temperature value of each pixel location at multiple consecutive time points is determined. The temperature fluctuation amplitude of each pixel location is determined by calculating the difference between the maximum and minimum temperature values ​​at consecutive time points. The calculation method for the temperature fluctuation amplitude is as follows: the temperature values ​​at multiple consecutive time points of each pixel location are compared, and the temperatures with the largest and smallest values ​​are selected. The largest temperature is then subtracted from the smallest temperature, and the result is the temperature fluctuation amplitude of the pixel location. The above steps are repeated for all pixel locations within the furnace wall's thermal radiation area to obtain the complete temperature fluctuation amplitude values, which are used as the range of values ​​on the vertical axis of the scatter plot. Using the total energy value of the high-frequency signal components within the abnormal heat transfer area of ​​the furnace wall and the temperature fluctuation amplitude values ​​of each pixel location within the furnace wall's thermal radiation area as the horizontal and vertical axes of the scatter plot, a scatter plot relating the total energy of the high-frequency signal components to the temperature fluctuation amplitude is constructed.

[0080] Regression analysis was used to fit the scatter plot to obtain a functional expression for the relationship between the total energy of the high-frequency signal components and the temperature fluctuation amplitude.

[0081] The appropriate function form is determined based on the data change trend reflected in the scatter plot. For example, when the data points in the scatter plot show an approximately linear relationship, a function with a linear form is selected as the basic function type; when the data points show a non-linear trend, a polynomial function can be selected as the basic function type; if the data points show a significant rapid growth or rapid decline trend, an exponential function can be selected as the basic function type.

[0082] After determining the basic function type, the least squares method is used for function fitting calculation. Specifically: First, determine the number of parameters in the function and the initial value of each parameter; for each data point in the scatter plot, determine the values ​​of the data point in the horizontal and vertical axes, and substitute the horizontal axis value into the function expression to calculate the corresponding theoretical prediction value; subtract the theoretical prediction value from the actual value of the vertical axis of the data point to obtain the error value of the data point; then square the error value to obtain the squared error value corresponding to the data point; calculate the squared error value of all data points in the scatter plot one by one, and then sum all the squared error values ​​to obtain the sum of the squared error values; by gradually adjusting the values ​​of each parameter in the function, repeatedly calculate the sum of the squared error values ​​until the optimal values ​​of each parameter are obtained when the sum of the squared error values ​​is minimized.

[0083] After obtaining the optimal values ​​of each parameter, substitute them into the determined basic function type to obtain the function expression; the function expression describes the correspondence between the total energy value of the high-frequency signal components in the abnormal heat transfer area of ​​the furnace wall and the temperature fluctuation amplitude value of each pixel in the thermal radiation area of ​​the furnace wall.

[0084] For example, if a function with a linear form is chosen as the basic function type, the function's parameters include the proportionality coefficient and a constant term. The least squares fitting method is used to determine the values ​​of the proportionality coefficient and the constant term that minimize the sum of squared errors in the scatter plot data.

[0085] Based on the trend of the total energy of the high-frequency signal components changing with time and the functional expression, the periodic pattern and variation characteristics of scale formation and shedding in the furnace wall thermal radiation area are determined.

[0086] The total energy value of high-frequency signal components is continuously collected at multiple consecutive moments within the abnormal heat transfer area of ​​the furnace wall. The interval between the multiple consecutive collection moments can be set from several minutes to tens of minutes, for example, the interval is set to five minutes, and the total energy value of high-frequency signal components is continuously collected over multiple hours.

[0087] The total energy values ​​of the high-frequency signal components collected at multiple consecutive time points are plotted as a trend graph with time on the horizontal axis and total energy value on the vertical axis. The trend graph is analyzed to determine whether there is a significant periodic trend in the total energy of the high-frequency signal components. The criteria for judging the periodic trend are as follows:

[0088] If the total energy value of the high-frequency signal components shows a continuous upward trend followed by a continuous downward trend, and the upward and downward trends repeat continuously, it is determined to have periodic change characteristics. The duration of the periodic change characteristics is determined by measuring the interval from when the total energy value of the high-frequency signal components begins to rise significantly to when it rises significantly again, which is determined as the periodic duration of scale formation and shedding.

[0089] Substitute the total energy value of the high-frequency signal components at each moment into the functional expression between the total energy of the high-frequency signal components and the temperature fluctuation amplitude to calculate the temperature fluctuation amplitude value at the corresponding moment within the thermal radiation area of ​​the furnace wall.

[0090] The calculated temperature fluctuation amplitude was plotted as a trend over time and compared with the periodic trend of the total energy of the high-frequency signal components. The analysis determined whether the temperature fluctuation amplitude increased significantly when the total energy of the high-frequency signal components reached a high level, and whether the temperature fluctuation amplitude decreased significantly when the total energy of the high-frequency signal components dropped to a low level.

[0091] By comparing and analyzing the two trend charts over multiple consecutive periods, we determined the characteristics of increased high-frequency signal energy and increased temperature fluctuation amplitude during the scale adhesion process, as well as the characteristics of decreased high-frequency signal energy and decreased temperature fluctuation amplitude during the scale shedding process.

[0092] By repeatedly observing and analyzing, the cycle duration of scale formation and shedding on the furnace wall surface was determined, such as several hours or several days. The maximum and minimum values ​​of temperature fluctuations in each cycle were recorded in detail, thereby obtaining the regularity and variation characteristics of scale formation and shedding.

[0093] S4: Based on the patterns of scale formation and shedding, a dynamic Bayesian network is used to assess the risk of furnace wall heat transfer efficiency fluctuations caused by scale changes, and the probability of heat transfer fluctuation risk is output, including:

[0094] Based on the periodic patterns and characteristics of scale formation and shedding within the furnace wall thermal radiation area, a dynamic Bayesian network containing hidden state nodes and observation nodes is constructed.

[0095] The periods of scale adhesion and scale shedding are used as hidden state nodes, and the total energy value of high-frequency signal components and the temperature fluctuation amplitude value are used as observation nodes. The probability value of furnace wall heat transfer efficiency fluctuation is determined by calculating the state transition probability and observation probability in the dynamic Bayesian network.

[0096] Hidden state nodes are used to describe states that cannot be directly observed but can be inferred from indirect data, while observed nodes are used to describe actual values ​​that can be measured.

[0097] The construction process of hidden state nodes is as follows: Based on the periodic pattern of scale formation and shedding within the furnace wall's thermal radiation area, the scale adhesion stage and the scale shedding stage are each set as two different hidden state nodes. Specifically, based on the periodic duration of the scale adhesion and shedding stages, two distinct hidden state nodes are set in the dynamic Bayesian network to represent different states of furnace wall heat transfer; for example, the stage in which scale gradually accumulates on the furnace wall surface, leading to a gradual increase in temperature fluctuation, is set as the scale adhesion period; the stage in which scale gradually sheds from the furnace wall surface, leading to a gradual decrease in temperature fluctuation, is set as the scale shedding period. Using the scale adhesion and shedding stages as hidden state nodes represents the implicit states with different furnace wall heat transfer efficiencies, thereby determining the definition and value range of the hidden state nodes.

[0098] The construction process of the observation nodes is as follows: the total energy value of the high-frequency signal components and the temperature fluctuation amplitude value within the furnace wall thermal radiation area are selected as observation nodes. Specifically, the total energy value of the high-frequency signal components and the temperature fluctuation amplitude value at each time moment are respectively used as two independent observation nodes in the dynamic Bayesian network; each observation node is represented in the dynamic Bayesian network by the actual measured values, such as the total energy value of the high-frequency signal components obtained during the measurement process at each time moment and the corresponding temperature fluctuation amplitude value, which are recorded in the observation node.

[0099] The state transition probability is calculated based on a dynamic Bayesian network. The state transition probability represents the change between hidden state nodes over time, specifically the likelihood of transitioning between the scale adhesion phase and the scale detachment phase. The calculation method is as follows: Based on historical data of the scale adhesion and detachment phases on the furnace wall surface, the frequency of transitioning from the scale adhesion phase to the detachment phase and vice versa are statistically analyzed. The ratio of the frequency of a specific phase transition to the total frequency of that specific phase is the state transition probability. For example, the proportion of times scale transitions from the adhesion phase to the detachment phase within a statistical period relative to the total number of times the scale adheres is the probability of transitioning from the adhesion phase to the detachment phase. The probability of transitioning from the detachment phase to the adhesion phase is calculated using the same method. Historical data includes the start time, end time, and number of transitions between the scale adhesion and detachment phases.

[0100] The observation probability is calculated based on a dynamic Bayesian network. This probability describes the likelihood of a hidden-state node's corresponding value occurring in a specific state. The calculation method involves dividing historical data on the total energy of the high-frequency signal components and the temperature fluctuation amplitude into multiple numerical intervals. For example, the total energy of the high-frequency signal is divided into high, medium, and low intervals; the temperature fluctuation amplitude is divided into large, medium, and small intervals. This division is based on uniformly dividing the data according to the maximum and minimum values ​​of the total energy and temperature fluctuation amplitude in the historical data.

[0101] The total energy value of the high-frequency signal components during the scale adhesion period is compared with each of the determined ranges of total high-frequency signal energy to determine the number of times the total energy value of the high-frequency signal components appears in each range; the temperature fluctuation amplitude value is compared with each of the temperature fluctuation amplitude values ​​to determine the number of times the temperature fluctuation amplitude value appears in each range.

[0102] Determine the total number of historical records during the period of scale adhesion, which is the total number of historical records of the total energy value of all high-frequency signals or the temperature fluctuation amplitude value during the period of scale adhesion.

[0103] Divide the frequency of occurrence of each numerical interval by the total number of historical records to obtain the observation probability value of each numerical interval during the scale adhesion stage.

[0104] During the scale shedding period, the same method as the scale adhesion period was used to calculate the observation probability, that is, to divide the numerical intervals, count the number of occurrences of each numerical interval, and then divide the number of occurrences of each numerical interval by the total number of historical records during the scale shedding period to obtain the observation probability values ​​of each numerical interval during the scale shedding period.

[0105] After obtaining the state transition probability and observation probability, the probability value of furnace wall heat transfer efficiency fluctuation is calculated using the probabilistic inference algorithm of dynamic Bayesian network. The calculation method is as follows: the total energy value of the high-frequency signal components measured at each moment in the observation node and the temperature fluctuation amplitude value are substituted into the dynamic Bayesian network, and then the Bayesian probability calculation formula is used based on the state transition probability and observation probability. The Bayesian probability calculation formula is: the posterior probability is equal to the likelihood probability multiplied by the prior probability and then divided by the marginal probability of the observed data. Specifically, the state transition probability is used as the prior probability and the observation probability is used as the likelihood probability to calculate the posterior probability value of the node in a specific hidden state (scale adhesion or detachment state).

[0106] Under scale buildup conditions, the temperature fluctuation amplitude and total energy of high-frequency signal components at each pixel location within the furnace wall's thermal radiation area from historical data are used as sample data. A criterion for determining fluctuations in furnace wall heat transfer efficiency is set, for example, a temperature fluctuation amplitude exceeding 120% of the historical average temperature fluctuation. All historical data under scale buildup conditions are analyzed to determine the number of data points that meet the fluctuation criteria. The number of data points that meet the fluctuation criteria is then divided by the total number of historical data points under scale buildup conditions to calculate the conditional probability of significant fluctuations in furnace wall heat transfer efficiency under scale buildup conditions.

[0107] Under the condition of scale detachment, the conditional probability value of significant fluctuation in furnace wall heat transfer efficiency is determined using the same standard and method. Specifically, the ratio of the number of historical data that meet the fluctuation standard under the condition of scale detachment to the total number of historical data under the condition of scale detachment is counted to obtain the conditional probability of heat transfer efficiency fluctuation under the condition of scale detachment.

[0108] The posterior probabilities of scale adhesion and scale detachment states are multiplied by the conditional probabilities of furnace wall heat transfer efficiency fluctuations under the corresponding states, and then summed to obtain the final probability value of furnace wall heat transfer efficiency fluctuations. The probability value of furnace wall heat transfer efficiency fluctuations represents the likelihood of changes in furnace wall heat transfer efficiency under the two implicit states of scale adhesion and scale detachment. A higher probability value indicates that the furnace wall heat transfer efficiency fluctuations are more significant at the current moment, while a lower probability value indicates that the heat transfer efficiency fluctuations are more stable at the current moment.

[0109] The marginal probability is calculated as follows: the product of the prior probability of the scale attachment state and the observed probability of the scale attachment state, plus the product of the prior probability of the scale detachment state and the observed probability of the scale detachment state.

[0110] S5: Construct a coupled model between the furnace combustion process and the probability of heat transfer fluctuation risk, and correct the heat loss factor in the combustion efficiency calculation, including:

[0111] Based on the calculation formula for heat loss during furnace combustion, a functional relationship is established between the furnace combustion process and the probability value of furnace wall heat transfer efficiency fluctuation.

[0112] The formula for calculating heat loss during furnace combustion includes a heat loss factor. The formula is: Total heat loss during furnace combustion equals the total heat generated by combustion multiplied by the heat loss factor. The total heat generated by furnace combustion is determined by the product of the actual amount of fuel burned and the fuel's calorific value. The heat loss factor is a coefficient characterizing the proportion of furnace heat lost to the external environment due to ineffective utilization; a larger heat loss factor indicates a higher proportion of heat loss during furnace combustion, and consequently, a lower combustion efficiency in the water-jacketed furnace.

[0113] The method for establishing the functional relationship is as follows: The probability value of the furnace wall heat transfer efficiency fluctuation is used as the input variable of the functional relationship, and the value of heat loss during furnace combustion is used as the output variable to determine the form of the functional relationship. The method for determining the functional relationship is as follows: A basic functional form is set, for example, a function form with linear characteristics is selected. Specifically, the probability value of the furnace wall heat transfer efficiency fluctuation and the corresponding heat loss during furnace combustion are obtained through historical data. These values ​​are substituted into the selected basic functional form, and the error between the probability value of the furnace wall heat transfer efficiency fluctuation and the heat loss during furnace combustion is calculated. The error is calculated by squaring the difference between the heat loss during furnace combustion and the heat loss during furnace combustion predicted by the functional relationship, and then summing the squared errors of all historical data points to obtain the sum of squared errors. By adjusting the parameters of the basic functional form, the sum of squared errors is minimized, thus obtaining the optimal parameters of the basic functional form. For example, if the basic function is chosen to be a linear function, then the slope and intercept of the linear function are the values ​​of the sum of squared errors calculated from historical data that reach their minimum.

[0114] The heat loss factor in the calculation formula for heat loss during furnace combustion is adjusted based on the probability value of the fluctuation of furnace wall heat transfer efficiency, and the correction value of the heat loss factor is determined.

[0115] The adjustment method for the heat loss factor is as follows: When the probability value of furnace wall heat transfer efficiency fluctuation increases, it indicates that the instability of furnace wall heat transfer efficiency increases, and the heat loss during furnace combustion also increases accordingly. In this case, the heat loss factor needs to be increased. When the probability value of furnace wall heat transfer efficiency fluctuation decreases, it indicates that furnace wall heat transfer efficiency tends to stabilize, and the heat loss during furnace combustion decreases. In this case, the heat loss factor needs to be decreased. The adjustment method for the heat loss factor is as follows: the absolute value of the difference between the probability value of furnace wall heat transfer efficiency fluctuation and the historical average probability value of furnace wall heat transfer efficiency fluctuation is used as the adjustment coefficient. The adjustment coefficient is multiplied by the historical average value of the heat loss factor to obtain the corrected value of the heat loss factor. The historical average value of the heat loss factor is the arithmetic mean of the heat loss factor within the period of historical data pairs. For example, if the probability value of furnace wall heat transfer efficiency fluctuation is greater than the historical average probability value by a certain percentage, such as exceeding the historical average probability value by 10%, the historical average value of the heat loss factor is increased by a specific percentage, such as 5%; conversely, it is decreased by a specific percentage. Through the above calculation method, the corrected value of the heat loss factor is obtained.

[0116] Based on the corrected value of the heat loss factor, the corrected heat loss factor is determined.

[0117] The corrected heat loss factor is obtained by adding the corrected value of the heat loss factor to the historical average value of the heat loss factor.

[0118] S6: Based on the corrected heat loss factor, update the combustion efficiency prediction model parameters in real time and output the real-time predicted value of the combustion efficiency of the water-jacketed heater, including:

[0119] The correction amount of the combustion efficiency prediction model parameters to be corrected in the combustion efficiency prediction model is calculated based on the corrected heat loss factor.

[0120] Based on the corrected heat loss factor, the correction amount of the parameters to be corrected in the combustion efficiency prediction model is calculated: The combustion efficiency prediction model contains multiple parameters, such as fuel combustion completeness parameters, furnace heat transfer efficiency parameters, flue gas heat loss parameters, and incomplete combustion heat loss parameters, etc. Each parameter affects the accuracy of the output of the combustion efficiency prediction model; The combustion efficiency prediction model predicts the combustion efficiency of the water jacketed furnace through the functional relationship between the parameters and the input values.

[0121] The method for calculating the correction amount of the combustion efficiency prediction model parameters is as follows: Based on the changes in the heat loss factor during furnace combustion, the correction amount of the flue gas heat loss parameter in the combustion efficiency prediction model is determined. The correction amount is calculated as follows: Calculate the difference between the corrected heat loss factor and the historical average heat loss factor; divide the difference by the historical average heat loss factor to obtain the correction coefficient; multiply the correction coefficient by the initial value of the flue gas heat loss parameter in the combustion efficiency prediction model parameters to calculate the correction amount of the flue gas heat loss parameter. For example, if the calculated corrected heat loss factor increases by 5% compared to the historical average heat loss factor, the flue gas heat loss parameter in the combustion efficiency prediction model will also increase by 5% accordingly; if the corrected heat loss factor is lower than the historical average heat loss factor, the value of the flue gas heat loss parameter will decrease accordingly. Through the above calculation method, the correction amount of each parameter in the combustion efficiency prediction model is obtained, ensuring that the combustion efficiency prediction model can reflect the actual changes in furnace heat loss.

[0122] The correction amount of the combustion efficiency prediction model parameters is added to the initial value of the combustion efficiency prediction model parameters to obtain the real-time updated combustion efficiency prediction model parameters.

[0123] The initial values ​​of the combustion efficiency prediction model parameters are the baseline values ​​of parameters determined based on the actual furnace combustion conditions within the historical measurement period, such as the historical average values ​​of fuel combustion completeness parameters, furnace heat transfer efficiency parameters, flue gas heat loss parameters, and incomplete combustion heat loss parameters. The correction values ​​of the combustion efficiency prediction model parameters are added to the initial values ​​of the combustion efficiency prediction model parameters to obtain the real-time updated combustion efficiency prediction model parameters. This ensures that the combustion efficiency prediction model parameters can be dynamically updated and adjusted in a timely manner to follow the actual furnace combustion conditions.

[0124] Based on the real-time updated combustion efficiency prediction model parameters, the temperature fluctuation amplitude of each pixel in the thermal radiation area of ​​the water jacket furnace wall, the total energy value of the high-frequency signal components, and the probability value of the furnace wall heat transfer efficiency fluctuation are input into the combustion efficiency prediction model, and the real-time predicted value of the combustion efficiency of the water jacket furnace is output.

[0125] The temperature fluctuation amplitude, the total energy of the high-frequency signal components, and the probability of furnace wall heat transfer efficiency fluctuation are used as input variables for the combustion efficiency prediction model. These variables are substituted into the combustion efficiency prediction model, and the real-time predicted value of the combustion efficiency of the water jacket furnace is calculated based on the real-time updated parameters of the combustion efficiency prediction model and the functional relationship between the parameters and the input values.

[0126] Specifically, if a linear function is adopted as the functional relationship of the combustion efficiency prediction model, then: the product of the fuel combustion completeness parameter and the furnace wall temperature fluctuation amplitude is added to the product of the furnace heat transfer efficiency parameter and the probability value of the furnace wall heat transfer efficiency fluctuation, then the product of the flue gas heat loss parameter and the total energy value of the high-frequency signal component is subtracted, and then the product of the incomplete combustion heat loss parameter and the probability value of the furnace wall heat transfer efficiency fluctuation is subtracted to obtain the real-time predicted value of the combustion efficiency of the water jacket heater.

[0127] Example 2

[0128] The difference between Embodiment 2 and Embodiment 1 is that this embodiment introduces a combustion efficiency prediction system for a water-jacketed heating furnace.

[0129] Figure 2 A schematic diagram of a combustion efficiency prediction system for a water-jacketed furnace according to the present invention is provided. The system includes:

[0130] Feature extraction module: Collects thermal image sequences of the furnace wall surface of the water jacket heating furnace, extracts pixel temperature gradient and texture change features in the thermal radiation area of ​​the furnace wall, and identifies abnormal heat transfer areas of the furnace wall.

[0131] Wavelet decomposition module: Performs wavelet decomposition on the temperature gradient characteristics of the abnormal heat transfer area of ​​the furnace wall to obtain high-frequency signal components that reflect the process of scale adhesion and detachment, and calculates the energy of the high-frequency signal.

[0132] Pattern recognition module: By analyzing the correlation between high-frequency signal energy and the temperature fluctuation amplitude of the furnace wall, the pattern of scale formation and detachment on the furnace wall is extracted;

[0133] Risk assessment module: Based on the patterns of scale formation and shedding, a dynamic Bayesian network is used to assess the risk of furnace wall heat transfer efficiency fluctuations caused by scale changes, and outputs the probability of heat transfer fluctuation risk.

[0134] Factor correction module: Constructs a coupled model between the furnace combustion process and the probability of heat transfer fluctuation risk, and corrects the heat loss factor in the combustion efficiency calculation;

[0135] Efficiency prediction module: Based on the corrected heat loss factor, the combustion efficiency prediction model parameters are updated in real time, and the real-time predicted value of the combustion efficiency of the water jacket furnace is output.

[0136] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0137] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0138] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0139] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0140] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0141] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0142] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0143] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0144] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0145] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting the combustion efficiency of a water-jacketed heating furnace, characterized in that, Includes the following steps: S1: Collect thermal image sequences of the furnace wall surface of the water jacket heating furnace, extract pixel temperature gradient and texture change features within the thermal radiation area of ​​the furnace wall, and identify abnormal heat transfer areas of the furnace wall. S2: Perform wavelet decomposition on the temperature gradient characteristics of the abnormal heat transfer area of ​​the furnace wall to obtain high-frequency signal components that reflect the process of scale adhesion and detachment, and calculate the energy of the high-frequency signal. S3: By analyzing the correlation between high-frequency signal energy and the temperature fluctuation amplitude of the furnace wall, the patterns of scale formation and detachment on the furnace wall are extracted; S4: Based on the laws of scale formation and shedding, a dynamic Bayesian network is used to assess the risk of furnace wall heat transfer efficiency fluctuation caused by scale changes, and output the probability of heat transfer fluctuation risk. S5: Construct a coupled model between the furnace combustion process and the probability of heat transfer fluctuation risk, and correct the heat loss factor in the calculation of combustion efficiency; S6: Based on the corrected heat loss factor, update the combustion efficiency prediction model parameters in real time and output the real-time predicted value of the combustion efficiency of the water jacket heater.

2. The method for predicting the combustion efficiency of a water-jacketed furnace according to claim 1, characterized in that, S1, specifically: Acquire a sequence of thermal images of the furnace wall surface of the water jacket heating furnace, delineate the boundary of the furnace wall thermal radiation area, and determine the temperature value of each pixel in the furnace wall thermal radiation area; The temperature difference between any two adjacent pixels within the thermal radiation area of ​​the furnace wall is calculated using the spatial gradient calculation method, thus obtaining the temperature gradient value at each pixel location. Energy, contrast and correlation indices were extracted from each frame of thermal image within the thermal radiation area of ​​the furnace wall using the gray-level co-occurrence matrix method to determine texture change characteristics. A threshold is set based on the temperature gradient value and texture change characteristics to identify areas of abnormal heat transfer in the furnace wall.

3. The method for predicting the combustion efficiency of a water-jacketed furnace according to claim 2, characterized in that, S2, specifically: The temperature gradient values ​​of each pixel in the abnormal heat transfer area of ​​the furnace wall are decomposed based on the multi-scale wavelet transform method to obtain the high-frequency wavelet coefficients and low-frequency wavelet coefficients corresponding to the temperature gradient values. The high-frequency wavelet coefficients are mapped to each pixel in the abnormal heat transfer area of ​​the furnace wall according to their spatial location to obtain the high-frequency signal component distribution of each pixel. The high-frequency signal component distribution of each pixel in the abnormal heat transfer area of ​​the furnace wall is squared and calculated. Then, the high-frequency signal component values ​​of each pixel after squared calculation are accumulated point by point to determine the total energy value of the high-frequency signal component in the abnormal heat transfer area of ​​the furnace wall.

4. The method for predicting the combustion efficiency of a water-jacketed furnace according to claim 3, characterized in that, S3, specifically: A scatter plot is constructed with the total energy value of high-frequency signal components in the abnormal heat transfer area of ​​the furnace wall as the horizontal axis and the temperature fluctuation amplitude value of each pixel in the thermal radiation area of ​​the furnace wall as the vertical axis. Regression analysis was used to fit the scatter plot to obtain a functional expression for the relationship between the total energy of the high-frequency signal components and the temperature fluctuation amplitude. Based on the trend of the total energy of the high-frequency signal components changing over time and the functional expression, the periodic pattern and variation characteristics of scale formation and shedding in the furnace wall thermal radiation area are determined.

5. The method for predicting the combustion efficiency of a water-jacketed furnace according to claim 4, characterized in that, S4, specifically: Based on the periodic patterns and characteristics of scale formation and shedding within the furnace wall thermal radiation area, a dynamic Bayesian network containing hidden state nodes and observation nodes is constructed. The periods of scale adhesion and scale shedding are used as hidden state nodes, and the total energy value of high-frequency signal components and the temperature fluctuation amplitude value are used as observation nodes. The probability value of furnace wall heat transfer efficiency fluctuation is determined by calculating the state transition probability and observation probability in the dynamic Bayesian network.

6. The method for predicting the combustion efficiency of a water-jacketed furnace according to claim 5, characterized in that, S5, specifically: Based on the calculation formula for heat loss during furnace combustion, a functional relationship is established between the furnace combustion process and the probability value of furnace wall heat transfer efficiency fluctuation. The heat loss factor in the calculation formula for heat loss during furnace combustion is adjusted based on the probability value of the fluctuation of furnace wall heat transfer efficiency, and the correction value of the heat loss factor is determined. The corrected heat loss factor is determined based on the corrected value of the heat loss factor.

7. The method for predicting the combustion efficiency of a water-jacketed furnace according to claim 6, characterized in that, S6, specifically: The correction amount of the combustion efficiency prediction model parameters to be corrected in the combustion efficiency prediction model is calculated based on the corrected heat loss factor. The correction amount of the combustion efficiency prediction model parameters is added to the initial value of the combustion efficiency prediction model parameters to obtain the real-time updated combustion efficiency prediction model parameters. Based on the real-time updated combustion efficiency prediction model parameters, the temperature fluctuation amplitude of each pixel in the thermal radiation area of ​​the water jacket furnace wall, the total energy value of the high-frequency signal components, and the probability value of the furnace wall heat transfer efficiency fluctuation are input into the combustion efficiency prediction model, and the real-time predicted value of the combustion efficiency of the water jacket furnace is output.

8. A combustion efficiency prediction system for a water-jacketed furnace, used to implement the combustion efficiency prediction method for a water-jacketed furnace as described in any one of claims 1-7, characterized in that, include: Feature extraction module: Collects thermal image sequences of the furnace wall surface of the water jacket heating furnace, extracts pixel temperature gradient and texture change features in the thermal radiation area of ​​the furnace wall, and identifies abnormal heat transfer areas of the furnace wall. Wavelet decomposition module: Performs wavelet decomposition on the temperature gradient characteristics of the abnormal heat transfer area of ​​the furnace wall to obtain high-frequency signal components that reflect the process of scale adhesion and detachment, and calculates the energy of the high-frequency signal. Pattern recognition module: By analyzing the correlation between high-frequency signal energy and the temperature fluctuation amplitude of the furnace wall, the pattern of scale formation and detachment on the furnace wall is extracted; Risk assessment module: Based on the patterns of scale formation and shedding, a dynamic Bayesian network is used to assess the risk of furnace wall heat transfer efficiency fluctuations caused by scale changes, and outputs the probability of heat transfer fluctuation risk. Factor correction module: Constructs a coupled model between the furnace combustion process and the probability of heat transfer fluctuation risk, and corrects the heat loss factor in the combustion efficiency calculation; Efficiency prediction module: Based on the corrected heat loss factor, the combustion efficiency prediction model parameters are updated in real time, and the real-time predicted value of the combustion efficiency of the water jacket furnace is output.

Images