Heat exchange element deposition area positioning method and device based on temperature characteristic difference

Abstract

The application provides a heat exchange element deposition area positioning method and device based on temperature characteristic difference, which comprises the following steps: obtaining the working condition parameters of an air preheater, inputting a heat exchange efficiency prediction model, obtaining a gas flow side outlet pressure prediction value, and calculating an efficiency risk trend value; obtaining a gas flow side wall temperature measured matrix, inputting the efficiency risk trend value into a wall temperature prediction model, obtaining a flue gas side wall temperature prediction matrix, and constructing a wall temperature deviation matrix; monitoring the continuous abnormal time of the wall temperature deviation matrix, constructing a time loss matrix; fusing the wall temperature deviation matrix and the time loss matrix to obtain a deposition characteristic comprehensive index matrix; inputting the deposition characteristic comprehensive index matrix into a deposition area classification model, and outputting the deposition state label of each spatial partition to realize positioning. The efficiency risk trend value is used for early warning of deposition risk, the wall temperature deviation matrix is used for accurate perception of wall temperature abnormities, the time loss matrix is used for filtering noise, the classification model is used for accurate positioning of the deposition area, and predictive maintenance of the heat exchange element is realized.

Description

Technical Field

[0001] This invention relates to the field of air preheater condition monitoring and fault diagnosis technology, and in particular to a method and apparatus for locating the deposition zone of heat exchange elements based on temperature characteristic differences. Background Technology

[0002] Air preheaters are crucial heat exchange devices in boiler units of thermal power plants. They recover waste heat from boiler flue gas to heat the air needed for combustion, playing a key role in improving boiler thermal efficiency and reducing flue gas heat loss. In coal-fired power plants, air preheaters typically employ a rotary structure, with numerous metal heat storage elements inside. As the rotor slowly rotates, these elements alternately pass between the flue gas side and the air side, facilitating the transfer of heat from the flue gas to the air.

[0003] In existing technologies, the monitoring of ash accumulation and blockage in the heat exchange elements of air preheaters mainly employs the inlet-outlet temperature difference method and the pressure difference method. The inlet-outlet temperature difference method determines whether the heat exchange performance has declined by comparing the deviation of the air preheater outlet air temperature from the design value; the pressure difference method determines whether the flow channel is blocked by monitoring the changes in the inlet-outlet pressure difference on the flue gas side or the air side. Both of these methods are based on the overall inlet and outlet parameters of the equipment for judgment.

[0004] However, the existing technology has the following defects: (1) When the outlet air temperature drops or the flue gas temperature rises, the heat exchange elements have already become obviously blocked. This method cannot provide early warning. In severe cases, it will lead to a reduction in the cross-sectional area of ​​the flue gas flow channel, a sharp increase in the pressure difference on both sides, and even cause the unit to trip and shut down; (2) It is impossible to locate the specific location and range of the deposition zone. During maintenance, all heat exchange elements need to be fully inspected. The maintenance time is long and the maintenance cost is high, which seriously affects the power generation efficiency. Summary of the Invention

[0005] In view of the above-mentioned deficiencies of the prior art, the present invention provides a method and apparatus for locating the deposition zone of heat exchange elements based on temperature characteristic differences, so as to solve the technical problem that the deposition zone of heat exchange elements in air preheaters cannot be predicted in advance and accurately located.

[0006] To achieve the above and other related objectives, this invention provides a method for locating deposition zones in heat exchange elements based on temperature characteristic differences. The method includes: acquiring the operating parameters of an air preheater, inputting them into a heat exchange efficiency prediction model to obtain a predicted value of the airflow-side outlet pressure under healthy conditions, calculating the deviation between the predicted and measured values ​​of the airflow-side outlet pressure to obtain an efficiency risk trend value; acquiring the measured matrix of the airflow-side wall temperature for each spatial zone of the air preheater, combining it with the efficiency risk trend value and inputting it into a wall temperature prediction model to obtain a predicted matrix of the flue gas-side wall temperature for each spatial zone of the air preheater, constructing a wall temperature deviation matrix based on the deviation between the predicted and measured wall temperature matrices; monitoring the duration of abnormal conditions exceeding a preset threshold for each spatial zone in the wall temperature deviation matrix to construct a time loss matrix; fusing the wall temperature deviation matrix and the time loss matrix to obtain a comprehensive deposition characteristic index matrix; inputting the comprehensive deposition characteristic index matrix into a deposition zone classification model to output deposition status labels for each spatial zone, thereby locating the spatial position of the deposition zone on the end face of the air preheater.

[0007] In one embodiment of the present invention, the heat exchange efficiency prediction model is constructed in the following manner: historical operating parameters of the air preheater under healthy operating conditions are obtained as training samples, the historical operating parameters including at least unit load, flue gas inlet temperature, flue gas inlet pressure, airflow inlet temperature, and airflow inlet pressure; the historical operating parameters are used as input features, and the airflow outlet pressure is used as the prediction target, and a linear regression algorithm is used for training to obtain a first regression coefficient matrix, the first regression coefficient matrix and its corresponding linear regression structure constitute the heat exchange efficiency prediction model.

[0008] In one embodiment of the present invention, the air preheater is a three-compartment rotary air preheater, the airflow side of which includes a primary air side and a secondary air side; the outlet pressure of the airflow side is the outlet pressure of the secondary air side.

[0009] In one embodiment of the present invention, the wall temperature prediction model is constructed in the following manner: historical airflow sidewall temperature measurement data and corresponding historical flue gas sidewall temperature measurement data of each spatial zone of the air preheater under healthy operating conditions are obtained; the historical airflow sidewall temperature measurement data and the efficiency risk trend value are used as input features, and the historical flue gas sidewall temperature measurement data are used as the prediction target, and a linear regression algorithm is used for training to obtain a second regression coefficient matrix. The second regression coefficient matrix and its corresponding linear regression structure constitute the wall temperature prediction model.

[0010] In one embodiment of the present invention, the time loss matrix is ​​constructed in the following manner: continuously collecting the wall temperature deviation values ​​of each spatial partition in the wall temperature deviation matrix at each sampling time; when the wall temperature deviation value of any spatial partition first exceeds a preset threshold, the time is recorded as the start time; from the start time, continuously monitoring the wall temperature deviation value of the spatial partition; if the wall temperature deviation value continues to exceed the preset threshold, the difference between the current time and the start time is used as the current time loss value of the spatial partition; if the wall temperature deviation value falls back to not exceed the preset threshold, the timing is stopped and the time loss value is fixed as the total duration of this abnormality; arranging the time loss values ​​of all spatial partitions according to their corresponding positions to construct the time loss matrix.

[0011] In one embodiment of the present invention, the depositional characteristic comprehensive index matrix is ​​obtained by fusing the wall temperature deviation matrix and the time loss matrix, including: multiplying the element values ​​corresponding to the same spatial partition position in the wall temperature deviation matrix and the time loss matrix element by element to obtain the depositional characteristic comprehensive index of the spatial partition; and arranging the depositional characteristic comprehensive indices of all spatial partitions according to their corresponding positions to obtain the depositional characteristic comprehensive index matrix.

[0012] In one embodiment of the present invention, the sedimentation zone classification model is constructed in the following manner: obtaining sedimentation feature comprehensive index matrix samples of each spatial partition under historical operating conditions, and labeling the deposition state of the heat exchange element surface corresponding to each spatial partition, which is divided into two categories: normal state and abnormal state; using the sedimentation feature comprehensive index matrix samples as input features, and the corresponding deposition state labeling as prediction target, the random forest algorithm is used for training to obtain the sedimentation zone classification model.

[0013] In one embodiment of the present invention, the spatial partition is a fixed grid area formed by dividing the air preheater rotor in the radial and circumferential directions, and each spatial partition is respectively provided with flue gas sidewall temperature measuring point and airflow sidewall temperature measuring point.

[0014] In one embodiment of the present invention, the method further includes: visually marking the deposition status labels of each spatial partition at the corresponding spatial positions on the end face of the air preheater, and distinguishing between abnormal deposition status and normal status with different colors.

[0015] To achieve the above and other related objectives, the present invention also provides a heat exchange element deposition zone positioning device based on temperature characteristic differences, comprising: a data acquisition module for acquiring the operating parameters of the air preheater, the measured matrix of the airflow sidewall temperature of each spatial zone, and the measured matrix of the flue gas sidewall temperature; an efficiency risk calculation module for inputting the operating parameters into a heat exchange efficiency prediction model to obtain the predicted value of the airflow side outlet pressure under healthy conditions, and calculating the deviation between the predicted value and the measured value of the airflow side outlet pressure to obtain an efficiency risk trend value; and a wall temperature deviation calculation module for inputting the measured matrix of the airflow side wall temperature and the efficiency risk trend value into the wall temperature prediction model. The system employs a model to obtain a flue gas sidewall temperature prediction matrix for each spatial partition, and constructs a wall temperature deviation matrix based on the deviation between the predicted and measured wall temperature matrices. A time loss calculation module monitors the duration of abnormal fluctuations exceeding a preset threshold in each spatial partition within the wall temperature deviation matrix, and constructs a time loss matrix. A comprehensive index calculation module fuses the wall temperature deviation matrix and the time loss matrix to obtain a comprehensive depositional characteristic index matrix. A depositional zone positioning module inputs the comprehensive depositional characteristic index matrix into the depositional zone classification model and outputs depositional state labels for each spatial partition to locate the spatial position of the depositional zone on the air preheater end face.

[0016] The beneficial effects of this invention are as follows: This invention proposes a method and device for locating deposition zones in heat exchange elements based on temperature characteristic differences. This method constructs a heat exchange efficiency prediction model, uses operating parameters to predict the outlet pressure on the airflow side under healthy conditions, and compares this prediction with measured values ​​to obtain an efficiency risk trend value, thus achieving early warning of deposition risks. By constructing a wall temperature prediction model, it combines the efficiency risk trend value with measured values ​​of the airflow side wall temperature to predict the flue gas side wall temperature and compares this prediction with the measured values, constructing a wall temperature deviation matrix to achieve accurate perception of wall temperature anomalies. By introducing a time loss matrix, it integrates the amplitude and duration of the wall temperature deviation, effectively filtering out instantaneous noise interference. Finally, through a deposition zone classification model, it outputs deposition status labels for each spatial partition, achieving accurate location of the deposition zone and providing a reliable basis for predictive maintenance of heat exchange elements. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. The accompanying drawings are incorporated in and constitute a part of this specification, illustrating embodiments consistent with this application, and are used together with the description to explain the principles of this application. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a method for locating the deposition zone of a heat exchange element based on temperature characteristic differences, as provided in an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of a heat exchange element deposition zone positioning device based on temperature characteristic differences, provided in an embodiment of the present invention.

[0020] Figure labeling: 201, Data acquisition module; 202, Efficiency risk calculation module; 203, Wall temperature deviation calculation module; 204, Time loss calculation module; 205, Comprehensive index calculation module; 206, Deposition zone location module. Detailed Implementation

[0021] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. It should be noted that, unless otherwise specified, the following embodiments and features can be combined with each other. In addition to the specific methods, equipment, and materials used in the embodiments, based on the knowledge of the prior art and the description of the present invention by those skilled in the art, any prior art methods, equipment, and materials similar to or equivalent to the methods, equipment, and materials in the embodiments of the present invention can be used to implement the present invention.

[0022] It should be understood that the terminology used in the embodiments of this invention is for describing specific particular implementations and not for limiting the scope of protection of this invention. Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art.

[0023] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In some embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0024] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functions, and operations that may be implemented in the methods and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0025] Please see Figure 1 , Figure 1 An embodiment of the present invention provides a method for locating the deposition area of ​​a heat exchange element based on temperature characteristic differences, comprising steps S101 to S105.

[0026] Step S101: Obtain the operating parameters of the air preheater, input them into the heat exchange efficiency prediction model, obtain the predicted value of the airflow side outlet pressure under healthy conditions, calculate the deviation between the predicted and measured values ​​of the airflow side outlet pressure, and obtain the efficiency risk trend value. This step utilizes real-time collected operating parameters (including unit load, flue gas side and airflow side inlet temperature and inlet pressure, etc.), inputs them into the pre-trained heat exchange efficiency prediction model, and outputs the predicted value that the airflow side outlet pressure should reach under the current healthy conditions. By comparing the deviation between this predicted value and the measured values ​​of the field sensors, the efficiency risk trend value is obtained. This value can reflect in advance the changes in flow resistance caused by ash accumulation and blockage of heat exchange elements, realizing a preliminary assessment of the equipment's health status and providing a trigger signal for subsequent refined positioning.

[0027] In a specific embodiment of the present invention, the heat exchange efficiency prediction model is constructed in the following manner: (1) historical operating parameters of the air preheater under healthy operating conditions are obtained as training samples. The historical operating parameters include at least unit load, flue gas inlet temperature, flue gas inlet pressure, airflow inlet temperature, and airflow inlet pressure; (2) the historical operating parameters are used as input features, and the airflow outlet pressure is used as the prediction target. The linear regression algorithm is used for training to obtain the first regression coefficient matrix. The first regression coefficient matrix and its corresponding linear regression structure constitute the heat exchange efficiency prediction model. The construction of the heat exchange efficiency prediction model is completed offline and does not require repeated training during equipment operation. Specifically, the healthy operating data of the air preheater at the initial stage of commissioning or after thorough cleaning are selected as training samples. The unit load, flue gas inlet temperature / pressure, and airflow inlet temperature / pressure at the corresponding time are used as input features, and the airflow outlet pressure at the same time is used as the prediction target. The multiple linear regression method is used for training to obtain the regression coefficients corresponding to each feature, forming the first regression coefficient matrix. This model is essentially the "pressure response feature" of the equipment under healthy conditions. After training, it can be used for real-time online prediction.

[0028] In a specific embodiment of the present invention, the air preheater is a three-compartment rotary air preheater, whose airflow side includes a primary air side and a secondary air side; the outlet pressure of the airflow side is the outlet pressure of the secondary air side. In the three-compartment rotary air preheater, the airflow side includes two independent channels: the primary air side and the secondary air side. Compared with the primary air side, the operating conditions of the secondary air side are more stable and less susceptible to drastic disturbances such as the start-up and shutdown of the coal mill and the switching of coal types. Its outlet pressure signal has a higher signal-to-noise ratio and can more accurately reflect the changes in the internal flow resistance of the air preheater. Therefore, this embodiment selects the outlet pressure of the secondary air side as the basis for calculating the efficiency risk trend value to improve the stability and reliability of monitoring.

[0029] In a specific embodiment of the present invention, the heat exchange efficiency prediction model can be expressed by the following formula:

[0030] ,

[0031] In the formula, This refers to values ​​of characteristics such as unit load, flue gas inlet temperature, flue gas inlet pressure, airflow inlet temperature, and airflow inlet pressure. The subscripts 0 to n represent different features, and the superscript t indicates the value collected at time t. Because the dimensions and numerical ranges of the input features differ significantly (e.g., load in MW, temperature in °C, pressure in Pa), each feature needs to be normalized before training the model to map the values ​​to a uniform interval, avoiding distortion of the regression coefficients due to differences in dimensions. Specifically, the Z-score standardization method (i.e., subtracting the mean and dividing by the standard deviation) can be used to ensure that all features are on the same order of magnitude, guaranteeing that the regression coefficients accurately reflect the influence of each feature on the prediction target. In the model construction phase, the first regression coefficient matrix... It can be recorded as:

[0032] .

[0033] The deviation between the predicted and measured values ​​of the airflow side outlet pressure can be calculated using the following formula:

[0034] ,

[0035] Alternatively, calculate using the following formula:

[0036] ,

[0037] Both formulas are used to calculate efficiency risk trend values, but they have different focuses. The first formula calculates the instantaneous absolute error between the predicted and measured values ​​at a single moment, suitable for reflecting the instantaneous deviation at the current moment; the second formula calculates the root mean square error of the prediction error sequence over the past N moments, which can smooth out instantaneous fluctuations and more robustly reflect the recent overall trend. In actual deployment, the applicable formula can be selected based on the noise level of the on-site data.

[0038] Step S102: Obtain the measured airflow sidewall temperature matrix for each spatial zone of the air preheater. Combine this matrix with the efficiency risk trend value and input it into the wall temperature prediction model to obtain the flue gas sidewall temperature prediction matrix for each spatial zone of the air preheater. Based on the deviation between the predicted flue gas sidewall temperature matrix and the measured wall temperature matrix, construct a wall temperature deviation matrix. This step introduces the efficiency risk trend value as a "global early warning factor" into the wall temperature prediction process. Specifically, the air preheater end face is divided into several zones according to spatial location. Measured values ​​of the primary air sidewall temperature and secondary air sidewall temperature for each zone are collected and input into the wall temperature prediction model along with the efficiency risk trend value to obtain the predicted flue gas sidewall temperature value that each zone should have under healthy conditions. This predicted value is compared with the measured flue gas sidewall temperature value of the corresponding zone, the deviation is calculated, and a wall temperature deviation matrix is ​​constructed. This decomposes the global efficiency anomaly signal into each spatial zone, achieving spatial location of the anomaly.

[0039] In a specific embodiment of the present invention, the wall temperature prediction model is constructed in the following manner: (1) historical airflow sidewall temperature measurement data and corresponding historical flue gas sidewall temperature measurement data of each spatial zone under healthy operating conditions are obtained; (2) the historical airflow sidewall temperature measurement data and efficiency risk trend value are used as input features, the historical flue gas sidewall temperature measurement data are used as prediction targets, and a linear regression algorithm is used for training to obtain a second regression coefficient matrix. The second regression coefficient matrix and its corresponding linear regression structure constitute the wall temperature prediction model. The construction of the wall temperature prediction model also adopts an offline training method. Historical data of the air preheater under healthy operating conditions are selected as training samples. The primary air sidewall temperature measurement value, secondary air sidewall temperature measurement value and efficiency risk trend value at that moment of each spatial zone are used as input features. The flue gas sidewall temperature measurement value at the same moment of the corresponding zone is used as the prediction target. The multiple linear regression method is used for training to obtain a second regression coefficient matrix. This model describes the heat transfer mapping relationship between "airflow sidewall temperature → flue gas sidewall temperature" under healthy conditions. After training, it can be used to predict the theoretical value of flue gas sidewall temperature in each zone in real time.

[0040] In a specific embodiment of the present invention, a typical air preheater includes a three-tier structure, each tier containing seven heat exchange element layers, for a total of Tier 3 heat exchange element configuration. For an air preheater with a three-tier structure and seven heat exchange element layers per layer, there are two approaches: First, train a set of regression coefficients for each layer separately and calculate the wall temperature deviation matrix for that layer individually; second, merge all layers into a single matrix (e.g., 21 rows × 10 columns), train a unified set of regression coefficients, and calculate the wall temperature deviation at all locations at once. The former offers higher accuracy, while the latter is simpler to implement and can be chosen based on the actual arrangement of measurement points.

[0041] In a specific embodiment of the present invention, the wall temperature prediction model can be expressed as:

[0042] ,

[0043] in, This is the flue gas sidewall temperature prediction matrix for each spatial zone of the air preheater. and These are the measured wall temperature matrices for the primary and secondary air sides of each spatial zone of the air preheater; their dimensions are... , This is the second regression coefficient matrix.

[0044] Understandably, the calculation process of the above formula is as follows: the measured matrix of primary wind sidewall temperature for each spatial zone is compared with the first regression coefficient. Multiplication, measured matrix of secondary wind sidewall temperature and second regression coefficient Multiplication, efficiency risk trend value and third regression coefficient Multiplying these three values ​​and adding them together yields the flue gas sidewall temperature prediction matrix for that spatial partition. If... , and If the dimensions are inconsistent (e.g., the airflow side is 7 rows × 5 columns and the flue gas side is 7 rows × 10 columns), then spatial interpolation or partition merging methods are needed to align the matrices to the same dimensions before performing calculations.

[0045] In a specific embodiment of the present invention, in the step of constructing the wall temperature deviation matrix based on the deviation between the flue gas sidewall temperature prediction matrix and the measured wall temperature matrix, the calculation can also be performed using two methods based on the efficiency risk trend value, which will not be elaborated here. The final wall temperature deviation matrix can be denoted as: .

[0046] Step S103: Monitor the duration of continuous abnormality exceeding a preset threshold in each spatial partition of the wall temperature deviation matrix, and construct a time loss matrix. This step introduces time dimension information to distinguish between "instantaneous random fluctuations" and "continuous anomalies." By continuously monitoring the deviation values ​​of each spatial partition in the wall temperature deviation matrix and recording the duration for which the deviation value of each partition exceeds the preset threshold, a time loss matrix is ​​constructed. Only persistent wall temperature deviations are identified as genuine deposition anomalies, thereby effectively filtering out false alarms caused by sensor noise and transient fluctuations in operating conditions, and significantly improving the reliability of deposition zone identification.

[0047] In a specific embodiment of the present invention, the time loss matrix is ​​constructed in the following manner: (1) continuously collecting the wall temperature deviation values ​​of each spatial partition in the wall temperature deviation matrix at each sampling time; (2) when the wall temperature deviation value of any spatial partition exceeds a preset threshold for the first time, recording that time as the start time; (3) continuously monitoring the wall temperature deviation value of the spatial partition from the start time. If the wall temperature deviation value continues to exceed the preset threshold, the difference between the current time and the start time is used as the current time loss value of the spatial partition. If the wall temperature deviation value falls back to not exceed the preset threshold, the timing is stopped and the time loss value is fixed as the total duration of this abnormality; (4) arranging the time loss values ​​of all spatial partitions according to their corresponding positions to construct the time loss matrix. The time loss matrix can be denoted as: Its size is the same as that of the wall temperature deviation matrix.

[0048] The preset threshold can be determined based on the statistical distribution of wall temperature deviations in each spatial zone of the air preheater under healthy operating conditions. For example, the threshold can be the average deviation under healthy conditions plus 2 to 3 times the standard deviation, or it can be set directly as a fixed value (e.g., 5℃) based on field experience. When the wall temperature deviation of a certain zone first exceeds the threshold, the start time is recorded. If the deviation continues to exceed the threshold, the time loss value accumulates. If the deviation falls back below the threshold, the timing stops and the accumulated time is fixed. In this way, the time loss matrix retains historical anomaly information and reflects the current anomaly duration in real time.

[0049] Step S104: Combine the wall temperature deviation matrix and the time loss matrix to obtain the comprehensive deposition characteristic index matrix.

[0050] In a specific embodiment of the present invention, step S104 specifically includes: multiplying the element values ​​corresponding to the same spatial partition position in the wall temperature deviation matrix and the time loss matrix element by element to obtain the comprehensive depositional characteristic index of the spatial partition; arranging the comprehensive depositional characteristic indices of all spatial partitions according to their corresponding positions to obtain the comprehensive depositional characteristic index matrix. This step can be expressed by the formula:

[0051] .

[0052] This step involves element-wise multiplying the wall temperature deviation matrix, reflecting the "abnormal amplitude," with the time loss matrix, reflecting the "abnormal persistence," to obtain a comprehensive depositional characteristic index for each spatial partition. This index considers both the severity and duration of the wall temperature deviation—only areas with large deviation amplitudes and long durations will obtain high index values, thus being identified as high-risk depositional zones. Compared to traditional methods that rely solely on deviation values ​​at a single moment, this fusion approach significantly improves anti-interference capabilities and identification accuracy.

[0053] Step S105: Input the comprehensive sedimentation characteristic index matrix into the sedimentation zone classification model and output the sedimentation state labels for each spatial partition (e.g., a state label of 1 indicates abnormal wall temperature, and a label of 0 indicates normal wall temperature) to locate the spatial position of the sedimentation zone on the air preheater end face. This step uses the trained sedimentation zone classification model to map the aforementioned calculated comprehensive sedimentation characteristic index matrix into sedimentation state labels for each spatial partition. By arranging these labels according to spatial position, the distribution of sedimentation zones can be visually presented on the air preheater end face. Maintenance personnel can then accurately locate the heat exchange element areas requiring maintenance, avoiding a comprehensive inspection of the entire air preheater and significantly shortening maintenance time.

[0054] In a specific embodiment of the present invention, the sedimentation zone classification model is constructed in the following way: (1) Obtain the sedimentation feature comprehensive index matrix samples of each spatial partition under historical operating conditions, and label the deposition state of the heat exchange element surface corresponding to each spatial partition, which is divided into normal state and abnormal state; (2) Use the sedimentation feature comprehensive index matrix samples as input features, and the corresponding deposition state label as prediction target, and use the random forest algorithm for training to obtain the sedimentation zone classification model. The construction of the sedimentation zone classification model requires the use of historical maintenance data as the labeling basis. Specifically, collect the sedimentation feature comprehensive index matrix samples at multiple times during the historical operation of the air preheater, and combine them with the maintenance records at the corresponding times to manually label whether there is deposition on the surface of the heat exchange element of each spatial partition (normal label is 0, and abnormal deposition label is 1), forming a labeled training dataset. Use the sedimentation feature comprehensive index of each spatial partition as input features, and the corresponding deposition state label as prediction target, and use the random forest algorithm for training to obtain the classification model. As an ensemble learning algorithm, random forest can effectively handle nonlinear classification problems and has good generalization performance. After training, it can be used for real-time online classification.

[0055] In a specific embodiment of the present invention, the spatial partitioning is a fixed grid area formed by dividing the air preheater rotor radially and circumferentially. Each spatial partition is respectively equipped with flue gas sidewall temperature measuring points and airflow sidewall temperature measuring points. The spatial partitioning is a fixed spatial position mark on the end face of the air preheater and does not change with rotor rotation. Specifically, the rotor is divided into several layers (e.g., 7 layers) in the radial direction according to different radii, and into several columns (e.g., 10 columns) in the circumferential direction according to different angles, forming a fixed grid area. At a fixed position corresponding to each spatial partition, wall temperature thermocouples are installed in the flue gas side chamber and the airflow side chamber, respectively, to measure the flue gas sidewall temperature and the airflow sidewall temperature of the heat storage element at that spatial position. This fixed coordinate system setting enables the algorithm to accurately locate temperature changes to specific spatial areas, providing a physical basis for subsequent deposition zone location.

[0056] In a specific embodiment of the present invention, the positioning method further includes: visually marking the deposition status labels of each spatial partition at the corresponding spatial location on the air preheater end face, and distinguishing between abnormal and normal deposition states using different colors. The deposition status labels of each spatial partition output by the model are binary data of 0 / 1. To facilitate quick identification by maintenance personnel, the label matrix can be overlaid on the schematic diagram of the air preheater end face in the form of a heat map. Specifically, the air preheater end face diagram is divided into grids according to the actual location of the spatial partitions, and partitions with a deposition status label of 1 are marked in red (representing abnormal deposition), while partitions with a label of 0 are marked in green (representing normal). Maintenance personnel can clearly grasp the distribution range and severity of the deposition area through this visualization interface, thereby arranging targeted cleaning and maintenance work.

[0057] It should be noted that the steps of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they contain the same logical relationship, they are all within the protection scope of this invention. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, without changing the core design of the algorithm and process, are also within the protection scope of this invention.

[0058] Please see Figure 2 , Figure 2 An embodiment of the present invention provides a heat exchange element deposition zone positioning device based on temperature characteristic differences, comprising a data acquisition module 201, an efficiency risk calculation module 202, a wall temperature deviation calculation module 203, a time loss calculation module 204, a comprehensive index calculation module 205, and a deposition zone positioning module 206. The data acquisition module 201 is used to acquire the operating parameters of the air preheater, the measured matrix of the airflow side wall temperature of each spatial zone, and the measured matrix of the flue gas side wall temperature. The efficiency risk calculation module 202 is used to input the operating parameters into the heat exchange efficiency prediction model to obtain the predicted value of the airflow side outlet pressure under healthy conditions, and calculate the deviation between the predicted value and the measured value of the airflow side outlet pressure to obtain the efficiency risk trend value. The wall temperature deviation calculation module 203 is used to input the measured matrix of the airflow side wall temperature and the efficiency risk trend value into the wall temperature prediction model to obtain the predicted matrix of the flue gas side wall temperature of each spatial zone, and... Based on the deviation between the flue gas sidewall temperature prediction matrix and the measured wall temperature matrix, a wall temperature deviation matrix is ​​constructed; the time loss calculation module 204 is used to monitor the continuous abnormal time exceeding the preset threshold in each spatial partition of the wall temperature deviation matrix and construct a time loss matrix; the comprehensive index calculation module 205 is used to fuse the wall temperature deviation matrix and the time loss matrix to obtain the comprehensive index matrix of deposition characteristics; the deposition area positioning module 206 is used to input the comprehensive index matrix of deposition characteristics into the deposition area classification model and output the deposition status label of each spatial partition to locate the spatial position of the deposition area on the end face of the air preheater.

[0059] It should be noted that the positioning device in this embodiment corresponds to the positioning method described above, and the functional modules in the positioning device may correspond to the corresponding steps in the positioning method. The positioning device in this embodiment can be implemented in conjunction with the positioning method; that is, without conflict, the relevant technical details mentioned in the positioning method of the above embodiments can also be applied to the positioning device in this embodiment.

[0060] In summary, the present invention has the following advantages over the prior art: (1) By predicting the outlet pressure on the airflow side in real time and comparing it with the measured value through the heat exchange efficiency prediction model, the efficiency risk signal can be captured several minutes in advance before the heat exchange element is severely blocked, thus realizing early warning of deposition risk and effectively avoiding unplanned unit shutdowns caused by severe blockage; (2) By decomposing the overall efficiency risk signal into each spatial partition through the wall temperature prediction model and constructing the wall temperature deviation matrix, the precise spatial positioning of the abnormal wall temperature location is realized, providing maintenance personnel with a clear target area for maintenance; (3) By introducing the time loss matrix, the amplitude and duration of the wall temperature deviation are integrated into a comprehensive deposition characteristic index, which effectively suppresses the instantaneous false alarms caused by sensor noise and operating condition fluctuations, and greatly improves the reliability of identification and anti-interference ability; (4) By outputting the deposition status label of each spatial partition through the deposition zone classification model and visualizing it, maintenance personnel can directly locate the specific spatial partition for targeted cleaning without having to conduct a comprehensive inspection of all heat exchange elements, which significantly shortens the maintenance time, reduces maintenance costs, and improves the availability of power generation equipment.

[0061] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for locating the deposition zone of a heat exchange element based on temperature characteristic differences, characterized in that, include: The operating parameters of the air preheater are obtained, and the heat exchange efficiency prediction model is input to obtain the predicted value of the airflow side outlet pressure under healthy conditions. The deviation between the predicted value and the measured value of the airflow side outlet pressure is calculated to obtain the efficiency risk trend value. Obtain the measured airflow sidewall temperature matrix of each spatial zone of the air preheater, and input the efficiency risk trend value into the wall temperature prediction model to obtain the flue gas sidewall temperature prediction matrix of each spatial zone of the air preheater. Based on the deviation between the flue gas sidewall temperature prediction matrix and the measured wall temperature matrix, construct the wall temperature deviation matrix. Monitor the duration of abnormal fluctuations exceeding a preset threshold in each spatial partition of the wall temperature deviation matrix, and construct a time loss matrix. By fusing the wall temperature deviation matrix and the time loss matrix, a comprehensive deposition characteristic index matrix is ​​obtained; The sedimentation feature comprehensive index matrix is ​​input into the sedimentation zone classification model, and the sedimentation state labels of each spatial partition are output to locate the spatial position of the sedimentation zone on the air preheater end face.

2. The method for locating the deposition zone of a heat exchange element based on temperature characteristic differences according to claim 1, characterized in that, The heat exchange efficiency prediction model is constructed in the following way: Historical operating parameters of the air preheater under healthy operating conditions are obtained as training samples. The historical operating parameters include at least the unit load, flue gas inlet temperature, flue gas inlet pressure, airflow inlet temperature, and airflow inlet pressure. Using the historical operating parameters as input features and the airflow side outlet pressure as the prediction target, a linear regression algorithm is used for training to obtain a first regression coefficient matrix. The first regression coefficient matrix and its corresponding linear regression structure constitute the heat exchange efficiency prediction model.

3. The method for locating the deposition zone of a heat exchange element based on temperature characteristic differences according to claim 2, characterized in that, The air preheater is a three-compartment rotary air preheater, and its airflow side includes a primary air side and a secondary air side; the outlet pressure of the airflow side is the outlet pressure of the secondary air side.

4. The method for locating the deposition zone of a heat exchange element based on temperature characteristic differences according to claim 1, characterized in that, The wall temperature prediction model is constructed in the following way: Acquire historical measured airflow sidewall temperature data and corresponding historical measured flue gas sidewall temperature data for each spatial zone of the air preheater under healthy operating conditions; Using the historical measured sidewall temperature data and the efficiency risk trend value as input features, and the historical measured flue gas sidewall temperature data as the prediction target, a linear regression algorithm is used for training to obtain a second regression coefficient matrix. The second regression coefficient matrix and its corresponding linear regression structure constitute the wall temperature prediction model.

5. The method for locating the deposition zone of a heat exchange element based on temperature characteristic differences according to claim 1, characterized in that, The time loss matrix is ​​constructed in the following way: The wall temperature deviation values ​​of each spatial partition in the wall temperature deviation matrix are continuously collected at each sampling time. When the wall temperature deviation of any spatial partition first exceeds the preset threshold, that moment is recorded as the start time; Starting from the initial time, the wall temperature deviation value of the space partition is continuously monitored. If the wall temperature deviation value continues to exceed the preset threshold, the difference between the current time and the initial time is used as the current time loss value of the space partition. If the wall temperature deviation value falls back to not exceed the preset threshold, the timing is stopped and the time loss value is fixed as the total duration of this abnormality. Arrange the time loss values ​​of all spatial partitions according to their corresponding positions to construct the time loss matrix.

6. The method for locating the deposition zone of a heat exchange element based on temperature characteristic differences according to claim 1, characterized in that, By integrating the wall temperature deviation matrix and the time loss matrix, a comprehensive deposition characteristic index matrix is ​​obtained, including: The element-wise multiplication of the wall temperature deviation matrix and the element values ​​corresponding to the same spatial partition position in the time loss matrix yields the comprehensive depositional characteristic index of the spatial partition; the comprehensive depositional characteristic indices of all spatial partitions are arranged according to their corresponding positions to obtain the comprehensive depositional characteristic index matrix.

7. The method for locating the deposition zone of a heat exchange element based on temperature characteristic differences according to claim 1, characterized in that, The sedimentary zone classification model is constructed in the following way: The comprehensive index matrix of deposition characteristics of each spatial partition under historical operating conditions was obtained, and the deposition state of the heat exchange element surface corresponding to each spatial partition was marked, which was divided into two categories: normal state and abnormal state. Using the sedimentary feature comprehensive index matrix samples as input features and the corresponding sedimentary state labels as prediction targets, the random forest algorithm is used for training to obtain the sedimentary area classification model.

8. The method for locating the deposition zone of a heat exchange element based on temperature characteristic differences according to claim 1, characterized in that, The spatial partition is a fixed grid area formed by dividing the rotor of the air preheater in the radial and circumferential directions. Each spatial partition is respectively provided with flue gas sidewall temperature measuring point and airflow sidewall temperature measuring point.

9. The method for locating the deposition zone of a heat exchange element based on temperature characteristic differences according to claim 1, characterized in that, The method further includes: The deposition status labels of each spatial partition are marked in a visual manner at the corresponding spatial locations on the end face of the air preheater, and different colors are used to distinguish between abnormal deposition status and normal status.

10. A device for locating the deposition zone of a heat exchange element based on temperature characteristic differences, characterized in that, include: The data acquisition module is used to acquire the operating parameters of the air preheater, the measured matrix of the airflow sidewall temperature of each spatial zone, and the measured matrix of the flue gas sidewall temperature. The efficiency risk calculation module is used to input the operating parameters into the heat exchange efficiency prediction model, obtain the predicted value of the airflow side outlet pressure under healthy conditions, and calculate the deviation between the predicted value and the measured value of the airflow side outlet pressure to obtain the efficiency risk trend value. The wall temperature deviation calculation module is used to input the measured wall temperature matrix of the airflow side wall and the efficiency risk trend value into the wall temperature prediction model to obtain the flue gas side wall temperature prediction matrix for each spatial zone, and to construct the wall temperature deviation matrix based on the deviation between the flue gas side wall temperature prediction matrix and the measured wall temperature matrix. The time loss calculation module is used to monitor the continuous abnormal time when each spatial partition in the wall temperature deviation matrix exceeds a preset threshold, and to construct the time loss matrix. The comprehensive index calculation module is used to fuse the wall temperature deviation matrix and the time loss matrix to obtain the comprehensive index matrix of deposition characteristics. The sedimentation zone positioning module is used to input the sedimentation feature comprehensive index matrix into the sedimentation zone classification model and output the sedimentation state labels of each spatial partition to locate the spatial position of the sedimentation zone on the air preheater end face.

Images