A coal-fired boiler combustion efficiency analysis and prediction method and system

Abstract

This invention relates to the field of boiler operation monitoring technology, specifically to a method and system for analyzing and predicting the combustion efficiency of coal-fired boilers. This invention collects real-time data on the received lower heating value, dry ash-free volatile matter, and received ash content of the coal entering the furnace. Combined with multi-point furnace temperatures and the volume concentrations of carbon monoxide and oxygen, it calculates the calorific value deviation, volatile matter deviation, ash influence coefficient, temperature dispersion, and flue gas component ratios. Furthermore, it integrates these data to generate a comprehensive combustion stability index and classify the dominant factors. This achieves real-time, quantitative diagnosis of combustion stability, accurately identifying the root causes of changes in combustion state caused by factors such as coal quality fluctuations, overcoming the problems of existing technologies that rely on single parameters or offline models for incomplete judgments.

Description

Technical Field

[0001] This invention relates to the field of boiler operation monitoring technology, specifically to a method and system for analyzing and predicting the combustion efficiency of coal-fired boilers. Background Technology

[0002] Boiler combustion optimization is a core element in ensuring the efficient, clean, and safe operation of thermal power plants. In actual operation, power plants generally face the problem of frequent fluctuations in the quality of coal fed into the boiler. Key indicators such as the net calorific value, dry ash-free volatile matter, and ash content vary greatly from coal of different origins and batches. This can easily lead to combustion vibration, reduced efficiency, and excessive pollutant emissions, affecting the safe and stable operation of boiler equipment and increasing operation and maintenance costs.

[0003] Existing technologies, such as Chinese invention patent publication number CN120724881A, disclose a boiler combustion optimization method, equipment, and medium based on CFD numerical simulation. This method constructs a three-dimensional combustion model and couples CFD numerical simulation with a multi-objective optimization algorithm to globally optimize parameters such as burner tilt angle and fuel ratio, thereby improving combustion efficiency and reducing pollutant emissions.

[0004] However, the aforementioned existing technologies have the following problems: 1. The optimization process is mainly based on deep offline CFD simulation, and the obtained optimization parameters are steady-state setpoints for specific coal quality and operating conditions, which are difficult to dynamically respond to real-time fluctuations in coal quality and complex operating condition disturbances. When the coal quality changes, the fixed optimization parameters may no longer match the current combustion state, resulting in control lag. 2. The analysis and judgment rely heavily on local parameters at preset locations, which are difficult to comprehensively and accurately reflect the complex combustion state inside the furnace. Furthermore, they lack the ability to analyze and quantify the impact of coal quality fluctuations on combustion stability based on multi-source coal quality parameters in real time, and to quickly match dynamic compensation strategies. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art, and to solve the problems that it is difficult to dynamically respond to real-time fluctuations in coal quality and complex operating conditions that cause control lag, as well as the lack of ability to analyze and quantify the impact of coal quality fluctuations on combustion stability based on real-time analysis of multi-source coal quality parameters and to quickly match dynamic compensation strategies.

[0006] The technical solution adopted by the present invention to solve its technical problem is: a method for analyzing and predicting the combustion efficiency of a coal-fired boiler, comprising the following steps: S1, the measured net calorific value, dry ash-free volatile matter, and net ash content of the coal fed into the boiler are respectively compared with the reference values ​​corresponding to the net calorific value, dry ash-free volatile matter, and net ash content of the coal designed for the current boiler, and then the difference is divided by the corresponding reference value to obtain the calorific value deviation, volatile matter deviation, and ash content influence coefficient.

[0007] S2. Divide the furnace into several combustion zones, calculate the absolute value of the temperature deviation between each temperature measuring point in the furnace and the average temperature of its respective combustion zone, and perform a weighted summation using the relative heat input level of each combustion zone as the weight to obtain the temperature dispersion; and obtain the volume concentrations of carbon monoxide and oxygen to calculate the flue gas composition ratio.

[0008] S3. Based on the calorific value deviation, volatile matter deviation, and ash content influence coefficient, combined with the temperature dispersion and flue gas composition ratio, a five-dimensional feature space is constructed and the Euclidean distance with the preset five-dimensional benchmark point is calculated. The Euclidean distance is converted into a numerical comprehensive combustion stability index according to the preset monotonically decreasing function. By comparing the period number of the inflection point of each feature in multiple consecutive sampling periods, the feature corresponding to the smallest period number is taken as the first dominant factor. Taking the inflection point of the first dominant factor as the benchmark, the data value change of the other features in a preset time window before and after the benchmark is calculated respectively. The feature corresponding to the data value change with the largest absolute value is taken as the second dominant factor.

[0009] S4. Input the historical sequence of the comprehensive combustion stability index and its adjacent time windows, as well as the classification of dominant factors, into the pre-constructed combustion state trend prediction model, and output the combustion stability prediction value at a future set time under at least one preset adjustment strategy.

[0010] S5. Select the target strategy from the preset adjustment strategies based on the combustion stability prediction value, calculate and output the compensation values ​​of boiler coal feed, primary air volume and secondary damper opening corresponding to the target strategy to the boiler control terminal for execution.

[0011] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention collects the received lower heating value, dry ash-free volatile matter, and received ash content of the coal entering the furnace in real time, and combines the furnace multi-point temperature, carbon monoxide and oxygen volume concentration to calculate the calorific value deviation, volatile matter deviation, ash influence coefficient, temperature dispersion and flue gas composition ratio, and further integrates them to generate a comprehensive combustion stability index and determine the classification of dominant factors. This realizes real-time and quantitative diagnosis of combustion stability, and can accurately identify the root causes of changes in combustion state caused by factors such as coal quality fluctuations, overcoming the problem of the existing technology relying on a single parameter or offline model for incomplete judgment.

[0012] 2. This invention constructs and utilizes a combustion state trend prediction model, taking the historical sequence of the comprehensive combustion stability index and its dominant factors as input, to predict the combustion stability prediction value at a future set time under different preset adjustment strategies; based on the prediction results, the optimal adjustment strategy is selected in advance, realizing proactive prediction and shortening the control response time after changes in combustion state caused by coal quality fluctuations.

[0013] 3. After determining the classification of dominant factors, this invention can quickly match and output the corresponding compensation values ​​for boiler coal feed, primary air volume and secondary damper opening. While ensuring the targeted control, it improves the control execution efficiency under the condition of frequent fluctuations in coal quality, which helps to improve the stability of boiler operation. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the prediction method of the present invention.

[0016] Figure 2 This is a schematic diagram of the system module connections of the present invention.

[0017] Figure 3 This is a schematic diagram illustrating the process of determining the temperature dispersion in this invention.

[0018] Figure 4 This is a schematic diagram illustrating the process for determining the comprehensive combustion stability index in this invention.

[0019] Figure 5 This is a flowchart illustrating the process of determining the classification of dominant factors in this invention. Detailed Implementation

[0020] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the invention. Furthermore, it should be understood that, for ease of description, the dimensions of the various parts shown in the drawings are not drawn to actual scale.

[0021] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use. Techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and apparatus should be considered part of the specification.

[0022] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0024] The specific scheme of the intelligent monitoring method for distribution network operation status based on wireless communication network provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0025] Please see Figure 1 The flowchart of a method for analyzing and predicting the combustion efficiency of a coal-fired boiler provided by the present invention is shown, which specifically includes the following steps: Step S1, based on the measured received lower heating value, dry ash-free volatile matter and received ash content of the coal entering the furnace, calculate the calorific value deviation, volatile matter deviation and ash content influence coefficient.

[0026] In practice, the first step is to obtain the baseline values ​​for the received lower heating value, dry ash-free volatile matter, and received ash content of the coal used in the current boiler design. These baseline values ​​can be obtained by consulting the boiler design documents or fuel procurement contracts.

[0027] If no direct data is available, the daily net calorific value, dry ash-free volatile matter, and ash content of the coal can be selected during a continuous 30-day period when the boiler load is stable at 80%-100% of the rated operating conditions, and their arithmetic average can be calculated as the benchmark value.

[0028] Subsequently, the measured values ​​of the coal quality entering the furnace are compared with the corresponding benchmark values, and the differences are divided by the benchmark values ​​to obtain the calorific value deviation, volatile matter deviation, and ash content influence coefficient.

[0029] When the calorific value deviation is too low, the heat released per unit of fuel is reduced. If the air supply remains unchanged, it may lead to a drop in furnace temperature and delayed combustion, which in turn may cause unstable combustion.

[0030] A decrease in volatile matter deviation will prolong the ignition time of pulverized coal, causing the combustion center to shift upwards and affecting the uniformity of temperature distribution.

[0031] The ash content impact coefficient indicates that increased ash content will dilute the fuel, reduce its calorific value, and may exacerbate slagging and wear, thus affecting heat transfer and combustion efficiency.

[0032] Step S2: Obtain temperature data from multiple temperature measurement points inside the furnace and calculate the temperature dispersion; and obtain the volume concentrations of carbon monoxide and oxygen and calculate the flue gas composition ratio.

[0033] The temperature at multiple points in the furnace reflects the uniformity of the combustion flame distribution. Local high or low temperatures indicate air distribution imbalance, uneven coal powder distribution, or abnormal ignition.

[0034] Carbon monoxide is a typical product of incomplete combustion of fuel, and its concentration directly indicates combustion efficiency; oxygen volume concentration reflects the matching relationship between the amount of air fed into the furnace and the amount of air required for fuel combustion.

[0035] Please see Figure 3 Step S21: Calculate the temperature dispersion. This invention does not directly perform simple statistical calculations on all temperature measurement points, but first divides the furnace into several combustion zones according to the arrangement of the boiler burners, with each combustion zone corresponding to at least one coal feed port; temperature analysis based on the combustion zones can more accurately pinpoint the source of the problem.

[0036] Taking a tangential combustion boiler as an example, the division process is as follows: First, all burners located in the same corner of the furnace and their associated coal feeders are treated as an independent combustion unit.

[0037] Then, taking the bisector of the angle of the furnace wall where the combustion unit is located as the reference, and moving towards the center of the furnace, with the center of the outermost burner nozzle of the combustion unit as the vertex, offset by 1 / 2 of the design flame diffusion angle to both sides to form a horizontal fan-shaped area.

[0038] In this invention, the flame diffusion angle explicitly given in the boiler design data can be directly used. If no explicit value is given, it can be calculated and rounded down using the following formula: .

[0039] Where θ is the flame diffusion angle (degrees); D is the characteristic dimension (meters) of the furnace cross-section on the burner arrangement plane, for example, for a rectangular cross-section, it can be the distance between opposite sides along the center line of the burner nozzle; r is the distance (meters) from the center of the burner nozzle to the center of the furnace corner; H is the vertical distance (meters) from the burner nozzle to the projection plane of the furnace center, usually taken as half of the furnace depth; β is the horizontal swing angle of the burner (degrees); D, r, H and β can all be obtained by consulting the boiler design drawings or burner installation drawings.

[0040] At the same time, it covers the elevation range of all burner nozzles in the combustion unit in vertical height, and extends upward and downward by 0.5 to 1.5 times the burner layer spacing; in order to ensure basic coverage of the flame core area of ​​the burner itself, and also to better cover the area of ​​thermal influence of the flame rise on the adjacent burner layer space.

[0041] Ultimately, the space defined by the horizontal sector and the vertical range is considered as a combustion zone.

[0042] Based on the division of combustion zones, for the same sampling period, the product of the coal feed rate at the corresponding coal feed port of each combustion zone and the measured lower heating value of the coal entering the furnace at that coal feed port is calculated as the heat input rate of that combustion zone.

[0043] Furthermore, based on the installation position of each temperature measuring point in the furnace, it is assigned to the corresponding combustion zone, and the arithmetic mean of the temperatures of all temperature measuring points in each combustion zone is calculated as the average temperature.

[0044] Then, each heat input rate is divided by the sum of all heat input rates to obtain the relative heat input level of each combustion zone.

[0045] Finally, the absolute value of the deviation between the temperature at each measuring point and the corresponding average temperature is calculated. Considering that the temperature fluctuation of the combustion zone with a large proportion of heat input has a greater impact on the overall combustion stability of the furnace, the relative heat input level of this combustion zone is used as the weight to weight and sum all the absolute values ​​of deviation to obtain the temperature dispersion, which is dimensionless.

[0046] Step S22: Calculate the flue gas component ratio. The specific process is as follows: First, obtain the baseline value of oxygen volume concentration corresponding to the current operating load of the boiler. Then, collect the volume concentrations of carbon monoxide and oxygen in the flue gas based on the same sampling period.

[0047] The oxygen volume concentration benchmark value is usually derived from the boiler's thermal calculations. Alternatively, in the absence of direct data, it can be obtained similarly using the method described above for determining the received lower heating value.

[0048] If the measured oxygen volume concentration is less than the baseline value, it indicates that the actual air supply may be insufficient, posing a risk of oxygen-deficient combustion. If carbon monoxide levels rise, it is highly likely that incomplete combustion is caused by oxygen deficiency.

[0049] At this point, the difference between the two values ​​is calculated. When this difference exceeds a preset tolerance threshold, it indicates that the degree of oxygen deficiency has exceeded the allowable fluctuation range, posing a risk of generating a large amount of incomplete combustion products. The measured carbon monoxide volume concentration is then divided by this difference to obtain the flue gas component ratio, which is dimensionless.

[0050] The ratio of flue gas components characterizes the severity of incomplete combustion products under oxygen-deficient conditions; the higher the value, the more carbon monoxide is produced per unit of oxygen deficiency, and the more obvious the incomplete combustion.

[0051] If the measured value of oxygen volume concentration is greater than or equal to the reference value of oxygen volume concentration, it indicates that the overall oxygen is sufficient. If carbon monoxide is still present, it is mainly attributed to factors such as fuel mixing, furnace temperature or residence time, rather than oxygen content. Therefore, the flue gas composition ratio is set to zero.

[0052] It should be added that the preset threshold can be set according to the accuracy of CEMS, and the value is usually in the range of 0.3%-0.5% of the oxygen volume concentration. In this invention, 0.4% of the oxygen volume concentration can be used as an example. If the boiler has high requirements for combustion stability, it can be reduced to 0.2% to improve sensitivity; if the requirements for control stability are even higher, it can be increased to 0.6% to enhance anti-interference ability.

[0053] Step S3: Based on the deviation of calorific value, deviation of volatile matter, and ash influence coefficient, combined with temperature dispersion and flue gas composition ratio, analyze and obtain the comprehensive combustion stability index and its dominant factor classification.

[0054] Please see Figure 4 Step S31: Determine the comprehensive combustion stability index. The specific process is as follows: Construct a five-dimensional feature space with calorific value deviation, volatile matter deviation, ash content influence coefficient, temperature dispersion, and flue gas composition ratio as coordinates.

[0055] Then, from the boiler historical database, select time period data that simultaneously meet all of the following limiting conditions: the main steam flow of the boiler is stable within ±5% of the rated flow; the average coal consumption for power supply is higher than 95% of the design value; and the average nitrogen oxide emission concentration is lower than 80% of the environmental protection limit.

[0056] Rated flow rate refers to the maximum continuous evaporation capacity indicated on the boiler nameplate. According to boiler operation specifications, the main steam flow rate is usually required to be stable within ±5% of the rated flow rate. Within this range, the boiler can be considered to be in quasi-steady-state operation.

[0057] Design values ​​can be obtained from the boiler design specifications. To eliminate the requirement for inefficient but stable operation due to non-combustion core factors such as reduced auxiliary equipment efficiency and steam-water system leaks, the average coal consumption for power supply should not exceed 95% of the design value.

[0058] Environmental limits are determined based on the latest environmental standards for the location of the boiler. In order to eliminate the illusion of combustion deterioration caused by excessive staged combustion while meeting environmental requirements, the average emission concentration is required to be less than 80% of the environmental limit.

[0059] Next, the arithmetic mean of the calorific value deviation, volatile matter deviation, ash influence coefficient, temperature dispersion, and flue gas composition ratio for each time period is calculated, and the coordinate point formed by these five average values ​​is defined as the reference point in the five-dimensional feature space.

[0060] Furthermore, the current calorific value deviation, volatile matter deviation, ash content influence coefficient, temperature dispersion, and flue gas composition ratio are used as coordinate values ​​in the five-dimensional feature space to form real-time feature points.

[0061] Based on this, the Euclidean distance between the real-time feature point and the five-dimensional reference point is calculated; according to the preset monotonically decreasing function, the Euclidean distance is converted into a numerical comprehensive combustion stability index.

[0062] The preset monotonically decreasing function can be of the exponential decay type, as shown below: .

[0063] in, The comprehensive combustion stability index is dimensionless and ranges from 0 to 100; d is the Euclidean distance.

[0064] α is the attenuation coefficient; in this invention, it can be taken as 0.5–1.0, which can be determined by fitting historical data. For example, at least 500 quasi-steady-state samples are selected from a historical database; the Euclidean distance between each sample and the five-dimensional reference point is calculated, and it is assumed that the ideal comprehensive combustion stability index of each sample is 90; (d, 90) is used as the sample point, substituted into the exponential attenuation function, and the least squares method is used to fit and solve for the attenuation coefficient that minimizes the overall error. In this invention, it can be specifically taken as 0.7.

[0065] Please see Figure 5 Step S32: Determine the classification of dominant factors. The specific process is as follows: First, obtain the calorific value deviation, volatile matter deviation, ash influence coefficient, temperature dispersion, and flue gas component ratio for multiple consecutive sampling periods to form a data value sequence.

[0066] Considering the thermal inertia of boiler combustion, which typically lasts from tens of seconds to several minutes, the sampling period should not be too short; for example, one minute can be used as a sampling period. Then, 10-30 consecutive sampling periods covering 10-20 minutes can typically capture the time required for the furnace to reach a steady state after a coal feeding or air distribution adjustment. In this invention, 20 consecutive sampling periods can be used as an example.

[0067] Next, for each data value sequence, using a sliding time window of length L, starting from the (L+1)th sampling period, the difference between the data value of the current sampling period and the mean of the previous sliding time window is calculated.

[0068] The sliding window length L can be 5-8, approximately 1 / 2 to 2 / 3 of the length of a single data value sequence, which can smooth noise while preserving trend inflection points. In specific implementation, L can be 6.

[0069] If the absolute value of the difference exceeds the preset threshold for a preset number of consecutive times, it indicates that the change is a continuous trend rather than a random fluctuation, and the starting sampling period is determined to be the inflection point of the change. Otherwise, it is not determined to be an inflection point and is considered as normal fluctuation.

[0070] Based on statistical principles, the preset threshold can be set to twice the standard deviation of the historical fluctuation of the corresponding feature. The preset number of trials can be set to three to avoid misjudgment of single-point mutations.

[0071] If no inflection point of change of any feature is detected in multiple consecutive sampling periods, the dominant factor is classified as steady state without dominant factor. At this time, the system considers the combustion state to be stable and there are no prominent contradictory factors. Therefore, it is neither necessary nor appropriate to perform the following determination steps for the first and second dominant factors.

[0072] Otherwise, compare the periodicity of the inflection points corresponding to the deviations in calorific value, volatile matter, ash content, temperature dispersion, and flue gas composition ratio, and take the characteristic corresponding to the smallest periodicity as the primary dominant factor.

[0073] The period number refers to the sequential number counted from the starting point of the data sequence being analyzed. For example, if an inflection point is detected in the 15th sampling period, its period number is 15.

[0074] When multiple characteristic inflection points share the same minimum cycle number, they can be selected according to a preset priority order. For example, following the analysis logic of first coal quality characteristics and then combustion state, the characteristic listed first can be taken as the primary dominant factor, in the order of calorific value deviation, volatile matter deviation, ash content influence coefficient, temperature dispersion, and flue gas composition ratio.

[0075] It should be noted that those skilled in the art will understand that the above priority order is merely an example, and other reasonable priority orders may be adopted in other implementations.

[0076] After identifying the primary dominant factor, using the inflection point of change of the primary dominant factor as a benchmark, calculate the changes in data values ​​of the remaining features within a preset time window before and after the benchmark. The changes in data values ​​can be the difference between the mean after the preset time window and the mean before the preset time window.

[0077] The length of the preset time window should be able to reasonably assess the change in parameters before and after the inflection point. In this invention, it can be the same as the length of the sliding window.

[0078] Finally, the feature corresponding to the largest absolute value change was selected as the second dominant factor.

[0079] Step S4: Input the historical sequence of the comprehensive combustion stability index and its adjacent time windows, as well as the classification of dominant factors, into the pre-constructed combustion state trend prediction model, and output the combustion stability prediction value at a future set time under at least one preset adjustment strategy.

[0080] The construction process of the combustion state trend prediction model is as follows: collect multiple combustion state adjustment operation records from the boiler's historical operation records; the records include the historical sequence of the comprehensive combustion stability index, the classification of dominant factors, the preset adjustment strategy adopted, and the actual combustion stability observation values ​​at future set times, which constitute a dataset.

[0081] It is important to consider that a certain amount of time is required from the issuance of control commands to the entry of coal into the furnace, the release of heat from the coal, the transfer of heat, and finally its reflection in the overall temperature field and flue gas composition data of the furnace. For medium and large boilers, this time is usually more than 10 minutes.

[0082] Therefore, in this invention, the future setting time is set to 15-30 minutes, specifically 20 minutes, which can give the control system enough time to stabilize its actions and state, while also ensuring the timeliness of the prediction.

[0083] The dataset is divided into training and test sets according to a set ratio. To optimize model performance and ensure its generalization ability, stratified sampling techniques can be used to maintain the distribution of various dominant factor classifications in the training and test sets consistent with the original dataset. Cross-validation can be used to select the specific ratio that minimizes the mean squared error of the test set as the final set ratio. In this invention, an 8:2 ratio can be used.

[0084] Using the historical sequence of the comprehensive combustion stability index and the classification of dominant factors in the training set as input features, and the actual combustion stability observation value at a future set time as the target variable, the time-series prediction model is trained. The model parameters are obtained by fitting and solving the model using the gradient descent method, and an initial combustion state trend prediction model is constructed.

[0085] Temporal prediction models include, but are not limited to, LSTM (Long Short-Term Memory) networks, GRU (Gated Recurrent Unit) systems, or Transformer models. This invention employs an LSTM network.

[0086] The combustion stability observations of the test set are predicted by an initial combustion state trend prediction model. The initial combustion state trend prediction model is then adjusted and optimized using mean square error and coefficient of determination to output the final combustion state trend prediction model.

[0087] The coefficient of determination represents the explanatory power of the historical sequence of the comprehensive combustion stability index and the classification of dominant factors predicted by the model for the predicted value of combustion stability at a future set time. The closer the value is to 1, the stronger the correlation between the input features and the target variable, and the better the model fit.

[0088] Mean square error (MSE) represents the overall deviation between the model's predicted combustion stability values ​​and the actual values; the smaller the value, the higher the model's prediction accuracy.

[0089] Adjusting and optimizing the combustion state trend prediction model can be achieved by adjusting hyperparameters such as the number of layers, number of neurons, and dropout rate of the LSTM network, or by using a more advanced optimizer, such as AdamW, to obtain a higher coefficient of determination and a lower mean square error on the test set, thereby outputting the final combustion state trend prediction model.

[0090] Step S5: Select the target strategy from the preset adjustment strategies based on the combustion stability prediction value, calculate and output the compensation values ​​of boiler coal feed, primary air volume and secondary damper opening corresponding to the target strategy to the boiler control terminal for execution.

[0091] The specific process is as follows: compare the combustion stability prediction values ​​corresponding to each preset adjustment strategy at the same future set time. Since the goal of this invention is to maximize combustion stability, the preset adjustment strategy corresponding to the highest value is selected as the target strategy.

[0092] If the combustion stability prediction value corresponding to all preset adjustment strategies is lower than the preset stability threshold, or if no matching preset adjustment strategy is found according to the current dominant factor classification, an instruction to maintain the current operating parameters unchanged will be output, and a warning that manual intervention is required will be issued.

[0093] The stability threshold can be determined as follows: select time period data that meet all the above-mentioned limiting conditions from the boiler historical operation database, calculate the comprehensive combustion stability index of each reference time period, and take its lower quartile as the stability threshold.

[0094] After determining the target strategy, based on the pre-stored compensation scheme in the target strategy and combined with the current boiler load, total air volume and main steam pressure operating parameters, calculate the boiler coal feed compensation value, primary air volume compensation value and secondary damper opening compensation value.

[0095] In the preset adjustment strategy, each compensation scheme corresponds to a set of quantitative compensation coefficients for boiler coal feed rate, primary air volume, and secondary damper opening, such as percentage changes. The percentage change is the percentage adjustment relative to the current operating parameters.

[0096] The quantitative compensation coefficient can be obtained by conducting multi-condition tests and verifications on the boiler digital twin through a simulation platform and then solidifying it; the specific process is existing technology and will not be described in detail in this invention.

[0097] The following lists several compensation schemes, namely, the combination of each typical primary dominant factor and the corresponding quantitative compensation coefficient, as shown in Table 1.

[0098] Table 1. Typical combinations of dominant factors and corresponding quantitative compensation coefficients.

[0099]

[0100] Taking the deviation of calorific value (negative value) as an example: increase the current total coal feed by 2.0% to compensate for the heat input; reduce the primary air volume by 1.0% to prevent the combustion temperature from dropping further due to an excessively high air-coal ratio; at the same time, increase the opening of the secondary air damper by 1.5% overall or specifically in the low-temperature region to help supplement oxygen and improve the uneven temperature field that may be caused by combustion delay.

[0101] Finally, the boiler coal feed compensation values, primary air volume compensation values, and secondary damper opening compensation values ​​are converted into control commands that the DCS control system can receive, such as changes in the coal feeder speed setpoint, primary air fan damper opening commands, or secondary damper actuator position commands. These control commands are then output to the boiler control terminal for execution.

[0102] After the compensation value is applied, the following data can be collected again: the lower heating value on the received basis, the volatile matter on the dry ash-free basis, and the ash content on the received basis of the coal entering the furnace, as well as the temperature, carbon monoxide and oxygen volume concentrations at multiple temperature measurement points in the furnace. The comprehensive combustion stability index and the classification of dominant factors are updated and used for online learning and parameter updating of the combustion state trend prediction model.

[0103] Please see Figure 2 As shown, a coal-fired boiler combustion efficiency analysis and prediction system includes: a quantitative calculation module, which is used to calculate the difference between the measured net calorific value, dry ash-free volatile matter, and net ash content of the coal entering the boiler and the corresponding benchmark values ​​of net calorific value, dry ash-free volatile matter, and net ash content of the coal used in the current boiler design, and then divide the difference by the corresponding benchmark value to obtain the calorific value deviation, volatile matter deviation, and ash content influence coefficient.

[0104] The feature extraction module is used to divide the furnace into several combustion zones, calculate the absolute value of the temperature deviation between each temperature measuring point in the furnace and the average temperature of its respective combustion zone, and perform weighted summation using the relative heat input level of each combustion zone as the weight to obtain the temperature dispersion; and obtain the volume concentrations of carbon monoxide and oxygen to calculate the flue gas composition ratio.

[0105] The fusion diagnostic module is used to construct a five-dimensional feature space based on calorific value deviation, volatile matter deviation, and ash content influence coefficient, and calculate the Euclidean distance with the preset five-dimensional benchmark point. The Euclidean distance is converted into a numerical comprehensive combustion stability index according to a preset monotonically decreasing function. By comparing the period number of the inflection point of each feature in multiple consecutive sampling periods, the feature corresponding to the smallest period number is taken as the first dominant factor. Taking the inflection point of the first dominant factor as the benchmark, the data value change of the other features in a preset time window before and after the benchmark is calculated. The feature corresponding to the data value change with the largest absolute value is taken as the second dominant factor.

[0106] The simulation prediction module is used to input the comprehensive combustion stability index and its historical sequence, as well as the classification of dominant factors, into a pre-constructed combustion state trend prediction model, and output the combustion stability prediction value at a future set time under at least one preset adjustment strategy.

[0107] The decision and output module is used to select the target strategy based on the combustion stability prediction value, calculate and output the compensation values ​​of boiler coal feed, primary air volume and secondary damper opening corresponding to the target strategy to the boiler control terminal for execution.

[0108] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product.

[0109] 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 implementations should not be considered beyond the scope of this invention.

[0110] In addition, the functional modules in the various embodiments of the present invention 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.

[0111] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0112] Finally, 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 coal-fired boiler, characterized by, include: S1. The measured net calorific value, dry ash-free volatile matter, and net ash content of the coal fed into the furnace are compared with the reference values ​​corresponding to the net calorific value, dry ash-free volatile matter, and net ash content of the coal for the current boiler design. The difference is then divided by the corresponding reference value to obtain the calorific value deviation, volatile matter deviation, and ash content influence coefficient. S2. Divide the furnace into several combustion zones, calculate the absolute value of the temperature deviation between each temperature measuring point in the furnace and the average temperature of the combustion zone, and perform weighted summation with the relative heat input level of each combustion zone as the weight to obtain the temperature dispersion; and obtain the volume concentrations of carbon monoxide and oxygen, and calculate the flue gas composition ratio. S3. Based on the calorific value deviation, volatile matter deviation, and ash content influence coefficient, combined with the temperature dispersion and flue gas composition ratio, a five-dimensional feature space is constructed and the Euclidean distance with the preset five-dimensional benchmark point is calculated. The Euclidean distance is converted into a numerical comprehensive combustion stability index according to the preset monotonically decreasing function. By comparing the period number of the inflection point of each feature in multiple consecutive sampling periods, the feature corresponding to the smallest period number is taken as the first dominant factor. Taking the inflection point of the first dominant factor as the benchmark, the data value change of the other features in a preset time window before and after the benchmark is calculated respectively. The feature corresponding to the data value change with the largest absolute value is taken as the second dominant factor. S4. Input the historical sequence of the comprehensive combustion stability index and its adjacent time windows, as well as the classification of dominant factors, into the pre-constructed combustion state trend prediction model, and output the combustion stability prediction value at a future set time under at least one preset adjustment strategy. S5. Select the target strategy from the preset adjustment strategies based on the combustion stability prediction value, calculate and output the compensation values ​​of boiler coal feed, primary air volume and secondary damper opening corresponding to the target strategy to the boiler control terminal for execution.

2. The method for analyzing and predicting the combustion efficiency of a coal-fired boiler according to claim 1, characterized in that, The specific process for determining the temperature dispersion is as follows: The furnace is divided into several combustion zones according to the arrangement of the boiler burners, and each combustion zone corresponds to at least one coal feed port. For the same sampling period, the product of the coal feed rate at the corresponding coal feed port in each combustion zone and the measured lower heating value of the coal entering the furnace at that coal feed port is calculated as the heat input rate; Based on the installation position of each temperature measuring point in the furnace, it is assigned to the corresponding combustion zone, and the arithmetic mean of the temperature of all temperature measuring points in each combustion zone is calculated as the average temperature. Divide each heat input rate by the sum of all heat input rates to obtain the relative heat input level of each combustion zone; Calculate the absolute value of the deviation between the temperature at each temperature measurement point and the corresponding average temperature, and then perform a weighted summation using the relative heat input level of the combustion zone as the weight to obtain the temperature dispersion.

3. The method for analyzing and predicting the combustion efficiency of a coal-fired boiler according to claim 2, characterized in that, The process of dividing the furnace into several combustion zones is as follows: Based on the installation position of the boiler burners in the circumferential and vertical directions of the furnace, each coal feed port or each group of coal feed ports on the same side, at the same angle, and on the same layer is designated as an independent combustion unit. Taking each combustion unit as the center, the dominant combustion zone is determined based on the flame diffusion range and flue gas flow path.

4. The method for analyzing and predicting the combustion efficiency of a coal-fired boiler according to claim 1, characterized in that, The calculation process for the ratio of the flue gas components is as follows: Obtain the baseline value of oxygen volume concentration corresponding to the current operating load of the boiler; collect the volume concentrations of carbon monoxide and oxygen in the flue gas based on the same sampling period; If the measured oxygen volume concentration is less than the reference value, the difference between the two is calculated. When the difference is greater than the preset tolerance threshold, the measured carbon monoxide volume concentration is divided by the difference to obtain the flue gas composition ratio. Otherwise, the flue gas composition ratio is set to zero.

5. The method for analyzing and predicting the combustion efficiency of a coal-fired boiler according to claim 1, characterized in that, The specific process for determining the comprehensive combustion stability index is as follows: A five-dimensional feature space is constructed with calorific value deviation, volatile matter deviation, ash content influence coefficient, temperature dispersion, and flue gas composition ratio as coordinates; Based on historical boiler operation data, a five-dimensional reference point is pre-defined in the five-dimensional feature space; The current calorific value deviation, volatile matter deviation, ash content influence coefficient, temperature dispersion, and flue gas composition ratio are used as coordinate values ​​to form real-time feature points; Calculate the Euclidean distance between real-time feature points and five-dimensional reference points; convert the Euclidean distance into a numerical comprehensive combustion stability index based on a preset monotonically decreasing function.

6. The method for analyzing and predicting the combustion efficiency of a coal-fired boiler according to claim 1, characterized in that, The specific process for determining the primary dominant factor is as follows: The calorific value deviation, volatile matter deviation, ash content influence coefficient, temperature dispersion, and flue gas composition ratio were obtained for multiple consecutive sampling periods to form a data value sequence. For each data value sequence, using a sliding time window of length L, starting from the (L+1)th sampling period, calculate the difference between the data value of the current sampling period and the mean of the previous sliding time window; If the absolute value of the difference exceeds the preset threshold for a preset number of consecutive times, the starting sampling period is determined to be the inflection point of change. If no inflection point of change of any feature is detected in multiple consecutive sampling periods, the dominant factor is classified as steady state with no dominant factor. Otherwise, compare the periodicity of the inflection points corresponding to the deviations in calorific value, volatile matter, ash content, temperature dispersion, and flue gas composition ratio, and take the characteristic corresponding to the smallest periodicity as the primary dominant factor.

7. The method for analyzing and predicting the combustion efficiency of a coal-fired boiler according to claim 1, characterized in that, The combustion state trend prediction model is constructed in the following ways: Collect multiple combustion state adjustment operation records from the boiler's historical operation records; the records include the historical sequence of the comprehensive combustion stability index, the classification of dominant factors, the preset adjustment strategies adopted, and the actual combustion stability observation values ​​at future set times, forming a dataset; The dataset is divided into training and testing sets according to a set ratio. The historical sequence of the comprehensive combustion stability index and the classification of dominant factors in the training set are used as input features, and the actual combustion stability observation values ​​at a future set time are used as target variables. The time series prediction model is trained by inputting these values, and the model parameters are obtained by fitting and solving the model using the gradient descent method. The initial combustion state trend prediction model is then constructed. The combustion stability observations of the test set are predicted by an initial combustion state trend prediction model. The initial combustion state trend prediction model is then adjusted and optimized using mean square error and coefficient of determination to output the final combustion state trend prediction model.

8. The method for analyzing and predicting the combustion efficiency of a coal-fired boiler according to claim 6, characterized in that, Step S5 specifically involves: By comparing the combustion stability prediction values ​​corresponding to each preset adjustment strategy at the same future set time, the preset adjustment strategy corresponding to the highest value is selected as the target strategy; each compensation scheme in the preset adjustment strategy corresponds to a set of quantitative compensation coefficients for boiler coal feed, primary air volume, and secondary damper opening; the quantitative compensation coefficients are percentage changes; Based on the percentage change corresponding to the pre-stored compensation scheme in the target strategy, and on the basis of the current boiler load, total air volume and main steam pressure operating parameters, the boiler coal feed compensation value, primary air volume compensation value and secondary damper opening compensation value are obtained by increasing or decreasing the percentage change. The boiler coal feed compensation value, primary air volume compensation value, and secondary damper opening compensation value are converted into control commands and output to the boiler control terminal for execution.

9. A system for analyzing and predicting the combustion efficiency of a coal-fired boiler, characterized in that, include: The quantitative calculation module is used to calculate the difference between the measured net calorific value, dry ash-free volatile matter, and net ash content of the coal fed into the boiler and the corresponding benchmark values ​​for the net calorific value, dry ash-free volatile matter, and net ash content of the coal designed for the current boiler. The difference is then divided by the corresponding benchmark value to obtain the calorific value deviation, volatile matter deviation, and ash content influence coefficient. The feature extraction module is used to divide the furnace into several combustion zones, calculate the absolute value of the temperature deviation between each temperature measuring point in the furnace and the average temperature of its respective combustion zone, and perform weighted summation using the relative heat input level of each combustion zone as the weight to obtain the temperature dispersion; and obtain the volume concentrations of carbon monoxide and oxygen to calculate the flue gas composition ratio. The fusion diagnostic module is used to construct a five-dimensional feature space based on calorific value deviation, volatile matter deviation, and ash content influence coefficient, and calculate the Euclidean distance with a preset five-dimensional benchmark point. The Euclidean distance is converted into a numerical comprehensive combustion stability index according to a preset monotonically decreasing function. By comparing the period number of the inflection point of each feature in multiple consecutive sampling periods, the feature corresponding to the smallest period number is taken as the first dominant factor. Taking the inflection point of the first dominant factor as the benchmark, the data value change of the other features in a preset time window before and after the benchmark is calculated. The feature corresponding to the data value change with the largest absolute value is taken as the second dominant factor. The simulation prediction module is used to input the comprehensive combustion stability index and its historical sequence, as well as the classification of dominant factors, into a pre-built combustion state trend prediction model, and output the combustion stability prediction value at a future set time under at least one preset adjustment strategy. The decision and output module is used to select the target strategy based on the combustion stability prediction value, calculate and output the compensation values ​​of boiler coal feed, primary air volume and secondary damper opening corresponding to the target strategy to the boiler control terminal for execution.

Images