Corrosion risk early warning method based on chlorine element monitoring and blending scheme optimization method

Abstract

The invention provides a corrosion risk early warning method based on chlorine element monitoring and a blending scheme optimization method. The corrosion risk early warning method comprises the following specific steps: monitoring the concentration of Cl element in flue gas in a horizontal flue at the tail of a boiler in real time by using a multispectral fusion probe; the equivalent chlorine deposition amount of the boiler flue heating surface is obtained through the Cl element concentration, the flue gas flow speed and the boiler flue heating surface; obtaining a corrosion risk early warning index according to the equivalent chlorine deposition amount, the sulfur content of the coal as fired and the metal wall surface temperature of the heating surface; obtaining a historical coal blending scheme and Cl release characteristics, obtaining a corresponding relation between the corrosion risk early warning index and the actual corrosion condition of the boiler, and dividing a corrosion threshold according to the corrosion condition; and comparing the corrosion risk early warning index with a corrosion threshold value to realize real-time as-fired coal corrosion risk early warning. The coal blending strategy is optimized through the corrosion risk early warning index, so that the problems of poor timeliness, inaccurate prediction and the like caused by off-line sampling and analysis in the prior art are effectively solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of monitoring the chlorine content of incoming coal, and in particular to a corrosion risk early warning method and a blending scheme optimization method based on chlorine element monitoring. Background Art

[0002] Currently, coal-fired power plants rely primarily on traditional offline sampling and analysis techniques to monitor chlorine in coal. Workers must manually collect coal samples from coal seams or conveyor belts, then send them to the laboratory for a complex series of pretreatment procedures, including crushing and ashing. Finally, chemical analysis methods, such as X-ray fluorescence, are used to indirectly deduce the chlorine content and classify the coal. This entire process is time-consuming, and monitoring results typically lag four to six hours. In actual power plant operation, the chlorine content of incoming coal fluctuates continuously with changes in coal source, quality, and the coal blending ratio. This long lag results in monitoring results significantly deviating from the real-time operating conditions of the incoming coal. In coal-fired power plants, this makes it impossible to dynamically optimize coal blending based on real-time chlorine content. When monitoring results indicate a low chlorine content, the actual chlorine content of the incoming coal may have already increased. Continuing with the previous blending plan can lead to a series of problems, such as increased corrosion of low-temperature heating surfaces, impacting the safe and stable operation of the power plant, and increasing equipment maintenance costs and the risk of downtime.

[0003] The migration and transformation behavior of chlorine in coal during combustion is extremely complex, resulting in a significant nonlinear correlation between the coal's chlorine content and the actual chlorine concentration in the flue gas. Data obtained from coal sampling monitoring cannot directly reflect the corrosion characteristics of the incoming coal. During combustion, chlorine reacts chemically with other elements in the coal to form various chlorides. The distribution and concentration of these chlorides in the flue gas are affected by factors such as combustion temperature, atmosphere, and residence time. For example, under different combustion conditions, even with the same coal chlorine content, the chlorine concentration in the flue gas can vary significantly. This leads to significant limitations in analysis and prediction based on offline coal quality data. Power plant operators cannot accurately predict the chlorine concentration in the flue gas based on the coal's chlorine content, making it difficult to assess the risk of chlorine corrosion on low-temperature heating surfaces. In actual operation, even if the coal's chlorine content appears low, the chlorine concentration in the flue gas can be very high due to the unique combustion conditions, leading to severe corrosion of the low-temperature heating surfaces.

[0004] In summary, existing technologies lack the ability to dynamically respond to the chlorine content of incoming coal, and it is difficult to accurately predict the risk of chlorine corrosion on low-temperature heating surfaces using offline coal quality data. This severely limits the accuracy and timeliness of corrosion prevention and control. Accuracy and timeliness are crucial in corrosion prevention and control. Without timely and accurate information on the chlorine content of incoming coal and the risk of corrosion on low-temperature heating surfaces, effective prevention and control measures cannot be implemented, potentially leading to the continued exacerbation of corrosion problems, impacting the service life of equipment and the economic benefits of the power plant. Summary of the Invention

[0005] In order to solve the problems existing in the prior art, the present invention provides a corrosion risk early warning method and a blending scheme optimization method based on chlorine element monitoring. By real-time monitoring of the chlorine content in the flue gas, combined with an intelligent algorithm, the corrosion risk of the coal entering the furnace is dynamically evaluated and the coal blending strategy is optimized, thereby effectively solving the problems of poor timeliness and inaccurate predictions caused by offline sampling and analysis in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a corrosion risk early warning method based on chlorine element monitoring, the specific steps are as follows:

[0007] The multi-spectral fusion probe is used to monitor the Cl element concentration in the flue gas in the horizontal flue at the tail of the boiler in real time;

[0008] The equivalent chlorine deposition on the boiler flue heating surface is obtained using Cl element concentration, flue gas flow rate and boiler flue heating surface;

[0009] The corrosion risk warning index is obtained based on the equivalent chlorine deposition amount, the sulfur content of the coal entering the furnace, and the metal wall temperature of the heating surface;

[0010] Obtain historical coal blending plans and Cl release characteristics, obtain the corresponding relationship between the corrosion risk warning index and the actual corrosion situation of the boiler, and divide the corrosion threshold according to the corrosion situation;

[0011] Comparing the corrosion risk warning index with the corrosion threshold value can realize real-time corrosion risk warning of coal quality entering the furnace.

[0012] Furthermore, the multi-spectral fusion probe is obtained by encapsulating a spectrometer, a high-energy pulse laser and a high-temperature resistant optical path; a hole is opened in the horizontal flue at the rear of the boiler, the multi-spectral fusion probe is hung outside the flue, and the optical path is arranged so that the laser focus is aimed at the flue gas in the flue.

[0013] Furthermore, the pulse energy of the high-energy pulse laser is greater than 200mJ and the frequency is 50Hz; the spectrometer adopts a high-resolution spectrometer, the wavelength range of the spectrometer is 300nm to 1100nm, including the characteristic emission line of the Cl element, and the resolution of the spectrometer is less than 0.15nm; the variable focal length lens group is a high-temperature resistant reflective lens group.

[0014] Furthermore, in the step of obtaining the equivalent chlorine deposition amount of the boiler flue heating surface using the Cl element concentration, the flue gas flow rate, and the boiler flue heating surface:

[0015] Filter the obtained Cl element concentration data, eliminate noise, and remove obvious erroneous data;

[0016] The processed Cl element concentration data is combined with the flue gas flow rate to establish a calculation model for the equivalent chlorine deposition on the boiler flue heating surface, as follows:

[0017] C d =C Cl vδ(T w )η w

[0018] In the formula, C d is the equivalent chlorine deposition amount, C Cl is the Cl element concentration data after treatment, v is the flue gas flow rate, δ(T w ) is the adsorption coefficient of the heated surface wall, which is a function of temperature, η w It is the split-screen correction factor, which is related to the contact area between the heated surface and the flue gas.

[0019] Furthermore, in the step of obtaining the corrosion risk warning index based on the equivalent chlorine deposition amount, the sulfur content of the coal entering the furnace, and the metal wall temperature of the heating surface, the specific calculation formula is as follows:

[0020] γ=a1C d +a2S+a3T m / 1000

[0021] In the formula, C d is the equivalent chlorine deposition; S is the sulfur content in coal; T m is the metal wall temperature of the heated surface; a1, a2, and a3 are weight coefficients, which are determined by regression of test data and historical corrosion cases.

[0022] Furthermore, the historical coal blending scheme and Cl release characteristics are obtained to obtain the corresponding relationship between the corrosion risk warning index and the actual corrosion situation of the boiler. The corrosion threshold is divided according to the corrosion situation, and the corrosion risk warning index is compared with the corrosion threshold to realize the corrosion risk warning step of the coal quality entering the furnace:

[0023] The corrosion thresholds include a medium corrosion level threshold and a severe corrosion level threshold. When the corrosion risk warning index exceeds the medium corrosion threshold, a corrosion risk warning of the coal quality entering the furnace is issued.

[0024] The present invention also provides a method for optimizing the coal blending scheme based on chlorine element monitoring, the specific steps of which are as follows:

[0025] Establish a database including Cl release characteristics of different coal qualities, boiler corrosion conditions, and coal quality industrial analysis data, and use the above-mentioned corrosion risk warning method based on chlorine element monitoring to obtain the corrosion risk warning index and corrosion level threshold when burning different coal qualities;

[0026] Establishing constraints for a neural network algorithm, using coal quality industrial analysis data and the total coal demand corresponding to the target load as input; using the blending ratio of each coal type as output, training the neural network algorithm through the database to obtain a calculation model for dynamically optimizing the coal blending scheme for the furnace;

[0027] The coal quality industrial analysis data is input into the calculation model of the dynamic optimization of the coal blending scheme for the furnace to obtain the blending ratio of each coal type.

[0028] Furthermore, the constraints of the neural network algorithm include:

[0029] Corrosion risk constraint: Corrosion risk warning index ≤ medium corrosion threshold;

[0030] Load constraint: The calorific value of the mixed coal after blending must meet Q 混合 ≥Q 负荷要求 ;

[0031] Environmental constraints: The sulfur and chlorine content of the blended coal must meet emission limits;

[0032] Blending ratio constraint: The sum of the blending ratios of each coal type is equal to 1.

[0033] Furthermore, after the coal is blended and put into use according to the obtained blending ratio of each coal type, the corrosion risk warning index during combustion of different coal qualities is obtained through the above-mentioned corrosion risk warning method based on chlorine element monitoring; if the actual corrosion risk warning index is greater than the medium corrosion threshold, the calculation model of the coal blending scheme entering the furnace is dynamically optimized to recalculate the blending ratio.

[0034] Furthermore, when training the neural network algorithm, the coal quality parameters, blending ratios and corresponding corrosion risk warning indexes in the historical coal blending scheme are used as labels; a supervised learning training model is adopted, and the objective function is to minimize the error between the predicted and actual values.

[0035] Compared with the prior art, the present invention has at least the following beneficial effects:

[0036] The present invention provides a corrosion risk early warning method based on chlorine element monitoring. It uses a multi-spectral fusion probe to monitor the Cl element concentration in the flue gas in the horizontal flue at the rear of the boiler in real time. Compared with traditional offline sampling and analysis methods, it can timely obtain the dynamic changes of the chlorine content in the flue gas, avoiding the timeliness problem caused by the time lag of sampling and analysis. At the same time, combined with the dynamic evaluation of the intelligent algorithm, it improves the accuracy of the corrosion risk judgment of the coal entering the furnace, making the early warning results more in line with the actual situation.

[0037] Based on the Cl element concentration, a more comprehensive corrosion risk assessment system is established by comprehensively considering multiple factors such as flue gas flow rate, boiler flue heating surface, sulfur content of the coal entering the furnace, and metal wall temperature of the heating surface. This multi-factor comprehensive assessment method can more accurately reflect the actual corrosion status of the boiler, provide a more reliable basis for corrosion risk warning, and effectively reduce the deviation caused by single-factor assessment.

[0038] By acquiring historical coal blending plans and Cl release characteristics, we determine the relationship between the corrosion risk warning index and the actual corrosion situation of the boiler, and use this information to define corrosion thresholds. Comparing the real-time calculated corrosion risk warning index with the threshold enables accurate early warning of the corrosion risk of the coal entering the furnace, issuing timely warning signals so that appropriate measures can be taken to avoid or mitigate losses caused by boiler corrosion.

[0039] This invention provides a method for optimizing coal blending schemes based on chlorine monitoring. This method establishes a database containing information such as the Cl release characteristics of different coal qualities during combustion, boiler corrosion conditions, industrial coal quality analysis data, corrosion risk warning indexes, and corrosion level thresholds. This provides rich data support for neural network algorithms. A dynamic optimization model for coal blending schemes, trained based on this data, enables scientific blending decisions in a data-driven manner, improving the rationality and effectiveness of blending schemes.

[0040] The neural network algorithm incorporates multiple constraints, including corrosion risk, load, environmental, and blending ratio. These constraints ensure that the blended coal quality not only meets boiler corrosion risk control requirements, but also meets load demands, environmental emission standards, and the rationality of the blending ratio. This achieves multi-objective optimization of incoming coal blending, ensuring safe, economical, and environmentally friendly boiler operation.

[0041] After coal is blended and put into operation based on the calculated blending ratios for each coal type, the corrosion risk warning method is used to obtain the actual corrosion risk warning index. If the actual index exceeds the moderate corrosion threshold, the dynamic optimization model for the coal blending scheme recalculates the blending ratio, enabling dynamic adjustment and continuous optimization of the blending scheme. This feedback mechanism enables timely response to changes in actual operation, continuously improving the scientific nature and adaptability of coal blending, reducing boiler corrosion risks, and improving boiler operating efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Schematic diagram of the measurement point locations of the online monitoring system for Cl elements in flue gas.

[0043] Figure 2 This is a structural diagram of the online monitoring system for Cl elements in flue gas. DETAILED DESCRIPTION

[0044] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0045] The present invention is based on the online monitoring of chlorine in flue gas and provides a corrosion risk early warning method and a blending scheme optimization method based on chlorine monitoring, which includes the following steps:

[0046] (1) An online monitoring system for Cl elements in flue gas based on laser absorption spectroscopy was established. In this system, a spectrometer, a high-energy pulsed laser, and a high-temperature resistant optical path were packaged to obtain a multi-spectral fusion probe. The high-energy pulsed laser was connected to a multi-laser power supply box, which provided power support for the laser to ensure that the laser could operate normally and generate high-energy pulsed lasers. The high-energy pulsed laser emitted by the high-energy pulsed laser was focused by a variable focal length lens group (high-temperature resistant optical path) to form a laser beam suitable for penetrating Cl elements in flue gas. The variable focal length lens group can adjust the focusing position and focal length of the laser to adapt to different measurement requirements and flue conditions. The focused laser beam passes through the flue gas, penetrating the Cl elements in the flue gas, and generating plasma. The variable focal length lens group realizes laser focusing and plasma spectrum collection through a coaxial optical path. The collected plasma spectrum is transmitted to the spectrometer via an optical fiber. The optical fiber transmits the collected plasma spectrum to the spectrometer, which analyzes the spectrum, identifies the characteristic emission lines of the Cl element (837.59nm, 912.11nm), and measures their intensity, thereby calculating the concentration of the Cl element in the flue gas. The spectrometer transmits the analyzed spectral data to a computer, which processes the spectral data through data processing and analysis software, calculates the concentration of the Cl element in the flue gas, and displays and stores it. The control panel is used to control and operate the entire system, including starting and stopping the laser, setting spectrometer parameters, and collecting and displaying data, etc., to achieve real-time measurement of the chlorine concentration in the flue gas. The specific steps are as follows:

[0047] S1: Using high-energy pulse laser with pulse energy greater than 200mJ and frequency of 50Hz, it can break down Cl elements in flue gas;

[0048] The spectrometer uses a high-resolution spectrometer with a wavelength range of 300 to 1100 nm, including the characteristic emission lines of Cl element (837.59 nm, 912.11 nm), and a spectrometer resolution of less than 0.15 nm.

[0049] The variable focal length lens group selects a high temperature resistant reflective lens group to achieve laser focusing and plasma spectrum collection through a coaxial optical path.

[0050] S2: Calibrated using a standard HCl gas sample with a lower limit of 1 ppm and an upper limit of 1000 ppm. The relative error of the measured concentration is within 3%.

[0051] S3: A hole is opened in the horizontal flue at the rear of the boiler, and the multi-spectral fusion probe is hung outside the flue. The optical path is arranged and finely adjusted so that the laser focus is accurately aligned with the flue gas in the flue. The time resolution is 1 time per second, and the average value of the results of every 10 monitoring times is used as the output result to eliminate the influence of flue gas unevenness on the results.

[0052] (2) Construct a model for the total equivalent chlorine content in flue gas and dynamically evaluate the corrosion risk of coal entering the furnace based on the equivalent chlorine content. The specific steps are as follows:

[0053] S1: Filter the real-time monitored Cl element concentration data in the flue gas, eliminate noise and remove obvious erroneous data, and combine it with boiler operating parameters to extract key characteristic parameters affecting the migration and transformation of the Cl element, such as coal sulfur content and flue gas temperature.

[0054] S2: Combine the Cl element concentration data in the treated flue gas with the flue gas flow rate to establish a calculation model for the equivalent chlorine deposition on the boiler flue heating surface:

[0055] C d =C Cl vδ(T w )η w (1)

[0056] In the formula, C d is the equivalent chlorine deposition amount, C Cl is the real-time monitoring concentration of Cl element in flue gas, v is the flue gas flow rate, δ(T w ) is the adsorption coefficient of the heated surface wall, which is a function of temperature, η w It is the split-screen correction factor, which is related to the contact area between the heated surface and the flue gas.

[0057] S3: Corrosion risk warning index γ is obtained based on the equivalent chlorine deposition amount and the sulfur content of the coal entering the furnace:

[0058] γ=a1C d +a2S+a3T m / 1000(2)

[0059] In the formula, S is the sulfur content in coal, in %; T m is the metal wall temperature of the heated surface, in °C. a1, a2, and a3 are weight coefficients, which are determined by regression of test data and historical corrosion cases.

[0060] (3) Establish a coal quality-flue gas Cl element characteristic database, develop an intelligent coal blending optimization algorithm based on the total Cl element content control model in flue gas, and realize closed-loop control of corrosion prevention and control. The specific steps are as follows:

[0061] S1: Collect combustion records of different coal qualities, continuously record their Cl release characteristics and boiler corrosion conditions, and combine them with industrial analysis data of coal quality (ash content, volatile matter, sulfur content, etc.) to establish a coal quality-chlorine release-corrosion risk warning index database;

[0062] Taking coal type as the core index, the chlorine release characteristics during combustion (such as Cl element concentration peak, release rate curve), equivalent chlorine deposition (calculated by formula (1)) and corrosion risk warning index (calculated by formula (2)) are correlated to form a structured database:

[0063]

[0064] S2: Analyze the correspondence between the corrosion risk warning index in historical coal blending plans and the actual boiler corrosion situation, and establish a quantitative corrosion grade standard;

[0065] Data cleaning and grouping: Eliminate abnormal data (such as extreme values ​​caused by monitoring failures or operational errors); group historical cases by corrosion severity (such as no corrosion, mild corrosion, moderate corrosion, severe corrosion).

[0066] Statistical analysis: Calculate the distribution characteristics (such as mean, standard deviation, and quantile) of the corrosion risk warning index γ for each set of data; use cluster analysis or decision tree algorithm to identify the threshold boundaries of different corrosion levels.

[0067] Threshold determination logic:

[0068] Moderate corrosion threshold: defined as the highest allowable safety value (i.e., the upper limit constraint of the neural network algorithm in step S3), above which the boiler corrosion risk increases significantly;

[0069] Severe corrosion threshold: defined as the critical value at which the coal blending plan must be adjusted immediately.

[0070] S3: In order to meet the boiler load requirements and environmental emission limits (such as SO2, NO x Under the premise of emission), by blending different coal qualities, the corrosion risk warning index γ is controlled below the medium corrosion threshold, while optimizing the combustion efficiency.

[0071] A neural network algorithm is used to calculate a coal blending scheme that meets load requirements based on a database of coal quality, chlorine release, and corrosion risk warning index, with moderate corrosion tendency as the upper limit of the corrosion level. After the blending scheme is put into use, the coal blending scheme can be dynamically optimized according to the corrosion level to ensure that the corrosion risk warning index is controlled within a safe range while ensuring that the load capacity and environmental emissions do not exceed the limit. The specific calculation logic framework is as follows:

[0072] 1) The input of the neural network algorithm includes industrial analysis data of coal quality and the total coal demand corresponding to the target load;

[0073] 2) Constraints of the neural network algorithm include:

[0074] Corrosion risk constraint: Corrosion risk warning index ≤ medium corrosion threshold;

[0075] Load constraint: The calorific value of the mixed coal after blending must meet Q 混合 ≥Q 负荷要求 ;

[0076] Environmental constraints: The sulfur and chlorine content of the blended coal must meet emission limits;

[0077] Blending ratio constraint: The sum of the blending ratios of each coal type is equal to 1.

[0078] 3) The neural network algorithm includes input layer, hidden layer, and hidden layer, among which:

[0079] The input layer inputs industrial analysis data of coal quality (such as sulfur content, chlorine content, ash content, etc., with the dimension of n coal types × number of features);

[0080] The hidden layer uses a multi-layer perceptron (MLP) to capture the complex mapping relationship between coal quality, chlorine release, and corrosion risk through nonlinear activation functions (such as ReLU).

[0081] The output layer outputs the blending ratio of each coal type.

[0082] 4) When training the neural network algorithm, the coal quality parameters, blending ratios, and corresponding corrosion risk warning indexes in the historical coal blending scheme are used as labels. A supervised learning training model is used, and the objective function is to minimize the error between the predicted and actual values ​​(such as the mean square error (MSE)).

[0083] 5) Optimize algorithm process

[0084] Initialization: Randomly generate a set of blending ratios that meet the load requirements.

[0085] Forward calculation:

[0086] Calculate the sulfur content and chlorine content of the mixed coal according to the blending ratio;

[0087] The concentration of Cl element in the flue gas is predicted by the laser absorption spectroscopy monitoring model and substituted into formula (1) to calculate the equivalent chlorine deposition amount, which is then substituted into formula (2) to calculate the corrosion risk warning index.

[0088] If the corrosion risk warning index exceeds the medium corrosion threshold, adjust the blending ratio (such as reducing the proportion of high-chlorine and high-sulfur coal); if the load or environmental protection constraints are not met, readjust the ratio.

[0089] Iterative optimization:

[0090] The blending ratio output by the neural network is used as the initial solution, and combined with genetic algorithm or particle swarm optimization (PSO) for global optimization, minimizing the corrosion risk while meeting multiple objective constraints.

[0091] 6) Closed-loop control logic

[0092] Real-time feedback: After the coal blending scheme is put into use, the concentration of Cl in the flue gas is monitored in real time through the laser absorption spectroscopy system to calculate the actual corrosion risk warning index;

[0093] Dynamic adjustment: If the actual corrosion risk warning index approaches the medium corrosion threshold, the algorithm is triggered to recalculate the blending ratio to reduce the proportion of high-risk coal types; if the corrosion level increases (for example, from mild to medium), it is forced to switch to an alternative coal blending plan with low corrosion risk.

[0094] Example 1:

[0095] The present invention provides a corrosion risk early warning method and a blending scheme optimization method based on chlorine element monitoring, comprising the following steps:

[0096] Step 1: Establish an online monitoring system for chlorine in flue gas based on laser absorption spectroscopy to achieve real-time measurement of flue gas chlorine concentration.

[0097] A high-energy pulsed laser with a pulse energy of 250 mJ and a frequency of 50 Hz is used to penetrate the chlorine in the flue gas. A high-resolution spectrometer with a wavelength range of 300 to 1100 nm, encompassing the characteristic chlorine emission lines at 837.59 nm and 912.11 nm, and a resolution of 0.1 nm, is used. A high-temperature-resistant reflective lens set is used to achieve laser focusing and plasma spectrum collection via a coaxial optical path. The spectrometer, high-energy pulsed laser, and high-temperature-resistant optical path are encapsulated. Calibration is performed in the laboratory using a standard hydrogen chloride gas sample with a lower limit of 5 ppm and an upper limit of 800 ppm, achieving a relative error of less than 2% in measured concentration. A hole is opened in the horizontal flue, and the device is mounted externally. The optical path is arranged and adjusted to ensure the laser focus is accurately aligned with the flue gas. The time resolution is 1 time every 2 seconds, and the average of 20 monitoring results is used as the output to eliminate the influence of flue gas unevenness.

[0098] Step 2: Construct a model for the total equivalent chlorine content in flue gas and dynamically evaluate the corrosion risk of coal entering the furnace based on the equivalent chlorine content.

[0099] The real-time monitoring data of flue gas chlorine element is filtered to eliminate noise and obvious erroneous data. The key characteristic parameters affecting the migration and transformation of chlorine element are extracted by combining the boiler operating parameters such as coal sulfur content and flue gas temperature.

[0100] Using the gas phase chlorine results obtained by monitoring, the equivalent chlorine deposition amount C on the heated surface is calculated according to formula (1): d ; Calculate the corrosion risk warning index γ according to formula (2);

[0101] Step 3: Establish a coal quality-flue gas chlorine element characteristic database and develop an intelligent coal blending optimization algorithm based on the flue gas total chlorine element control model.

[0102] Combustion records of different coal qualities were collected, and their chlorine release characteristics and boiler corrosion conditions were continuously recorded. Combined with industrial coal analysis data such as ash, volatile matter, and sulfur content, a multidimensional database was established for model training and optimization. A neural network algorithm was used to analyze historical coal blending plans and chlorine release characteristics, and the corrosion risk warning index thresholds were determined to be 0.6 for moderate corrosion and 0.8 for severe corrosion.

[0103] Step 4: Based on the results of the intelligent coal blending optimization algorithm, optimize the coal blending plan to achieve closed-loop control of corrosion prevention and control.

[0104] Combined with the equivalent chlorine content, the coal blending plan entering the furnace is dynamically optimized to ensure that the corrosion risk warning index is controlled within a safe range below 0.6 while ensuring that the boiler's load capacity and environmental emissions do not exceed the limit.

[0105] Example 2:

[0106] The present invention provides a corrosion risk early warning method and a blending scheme optimization method based on chlorine element monitoring, comprising the following steps:

[0107] Step 1: Establish an online monitoring system for chlorine in flue gas based on laser absorption spectroscopy to achieve real-time measurement of flue gas chlorine concentration. A high-energy pulsed laser with a pulse energy of 300mJ and a frequency of 60Hz is used to penetrate the chlorine in the flue gas. A high-resolution spectrometer with a wavelength range of 320-1080nm, including the characteristic chlorine emission lines of 837.59nm and 912.11nm, and a resolution of 0.08nm is configured. A high-temperature-resistant reflective lens set is selected, and a coaxial optical path is used to achieve laser focusing and plasma spectrum collection. The spectrometer, high-energy pulsed laser, and high-temperature-resistant optical path are packaged together. Calibration is performed in the laboratory using standard hydrogen chloride gas samples, with a lower limit of 3ppm and an upper limit of 900ppm. The relative error in the measured concentration is within 1.5%. A hole is opened in the horizontal flue, the device is hung outside the flue, the optical path is arranged and adjusted to ensure that the laser focus is accurately aligned with the flue gas. The time resolution is 1 time per 1 second, and the average value of the results of 15 monitoring times is used as the output result to eliminate the influence of uneven flue gas on the results.

[0108] Step 2: Construct a model for the total equivalent chlorine content in flue gas and dynamically evaluate the corrosion risk of coal entering the furnace based on the equivalent chlorine content.

[0109] The real-time monitoring data of flue gas chlorine element is filtered to eliminate noise and obvious erroneous data. The key characteristic parameters affecting the migration and transformation of chlorine element are extracted by combining the boiler operating parameters such as coal sulfur content and flue gas temperature.

[0110] Using the gas phase chlorine results obtained by monitoring, the equivalent chlorine deposition amount C on the heated surface is calculated according to formula (1): d ; Calculate the corrosion risk warning index γ according to formula (2);

[0111] Step 3: Establish a coal quality-flue gas chlorine element characteristic database and develop an intelligent coal blending optimization algorithm based on the flue gas total chlorine element control model.

[0112] Collect combustion records of different coal qualities, continuously record their chlorine release characteristics and boiler corrosion conditions, and combine them with industrial analysis data of coal quality such as ash content, volatile matter, sulfur content, etc. to establish a multi-dimensional database for model training and optimization.

[0113] By analyzing historical coal blending plans and chlorine release characteristics, the moderate corrosion threshold of the corrosion risk warning index was determined to be 0.65, and the severe corrosion threshold was 0.85.

[0114] Step 4: Based on the results of the intelligent coal blending optimization algorithm, optimize the coal blending plan to achieve closed-loop control of corrosion prevention and control.

[0115] Combined with the equivalent chlorine content, the coal blending plan for the furnace is dynamically optimized to ensure that the corrosion risk warning index is controlled within a safe range below 0.65 while ensuring that the boiler's load capacity and environmental emissions do not exceed the limit.

[0116] Example 3

[0117] A corrosion risk early warning method based on chlorine element monitoring and the measurement point location in the blending scheme optimization method are as follows: Figure 1 As shown, a multi-spectral fusion probe was installed in the horizontal flue at the rear of a 660MW counter-hedge boiler. The probe integrates a high-energy pulsed laser, a high-resolution spectrometer, and a high-temperature variable-focus lens. The high-energy pulsed laser has a pulse energy of 220mJ and a frequency of 50Hz. The high-resolution spectrometer has a wavelength range of 300-1100nm and an average spectral resolution of 0.12nm. The high-temperature variable-focus lens has a maximum temperature resistance of 600°C. After coaxial alignment of the optical path, the laser focus is aligned with the central airflow of the flue.

[0118] The quantitative relationship between the Cl element characteristic spectrum intensity and Cl element concentration was determined by laboratory calibration. The standard concentration gradient of 1ug / m 3 , 5ug / m 3 、10ug / m 3 、50ug / m 3 、100ug / m 3 、1000ug / m 3 The HCl gas was used, and considering the self-absorption effect, the nonlinear response of the characteristic peaks of the spectrometer at 837.59nm and 912.11nm was confirmed. Through calibration tests on the same sample, the relative error was less than 5%.

[0119] The Cl element concentration in the flue gas is collected once every 1 second on site, and the average value of every 10 measurements is output to eliminate the influence of flue gas turbulence fluctuations.

[0120] The average concentration of Cl in the flue gas at the high-temperature reheater is monitored in real time and is 8.3 ug / m 3 , synchronously obtain boiler parameters: average sulfur content of coal entering the furnace is 0.5%, flue gas flow rate is 12m / s, heating surface temperature is 270℃, and the adsorption coefficient δ(T w )=0.15. Since the contact area between the high temperature reheater and the flue gas is large, the split screen correction factor η is taken. w =1.2. According to formula (1), the equivalent chlorine deposition amount Cd = 0.179 mg / m 3 , substituted into formula (2), selecting a1, a2 and a3 as 0.6, 0.3 and 0.1 respectively, and calculating the corrosion risk warning index γ = 0.284.

[0121] During the daily operation of the boiler, the coal quality of the incoming coal, including its sulfur content, is recorded. The above steps are continuously performed to continuously calculate the combustion of different coal qualities. A database of coal quality-chlorine release-corrosion risk warning index is established. The corresponding relationship between the corrosion risk warning index and the actual corrosion situation of the boiler is obtained, and the thresholds of moderate and severe corrosion of the corrosion risk warning index are determined:

[0122] When γ<0.35, it is in a slight corrosion tendency;

[0123] When 0.35<γ<0.5, it is in a medium corrosion tendency;

[0124] When γ>0.5, it is in a tendency of severe corrosion.

[0125] Under the premise of ensuring the load capacity and NOx and SOx emissions do not exceed the limit, the corrosion risk warning index is controlled within 0.5, and the dynamic optimization of the coal blending scheme calculation model is used to obtain the blending ratio of high-chlorine coal to be 25%.

Claims

1. A corrosion risk early warning method based on chlorine element monitoring, characterized in that: The specific steps are as follows: The multi-spectral fusion probe is used to monitor the Cl element concentration in the flue gas in the horizontal flue at the tail of the boiler in real time; The equivalent chlorine deposition on the boiler flue heating surface is obtained using Cl element concentration, flue gas flow rate and boiler flue heating surface; The corrosion risk warning index is obtained based on the equivalent chlorine deposition amount, the sulfur content of the coal entering the furnace, and the metal wall temperature of the heating surface; Obtain historical coal blending plans and Cl release characteristics, obtain the corresponding relationship between the corrosion risk warning index and the actual corrosion situation of the boiler, and divide the corrosion threshold according to the corrosion situation; Comparing the corrosion risk warning index with the corrosion threshold value can realize real-time corrosion risk warning of coal quality entering the furnace.

2. A corrosion risk early warning method based on chlorine element monitoring according to claim 1, characterized in that: The multi-spectrum fusion probe is obtained by encapsulating a spectrometer, a high-energy pulse laser and a high-temperature resistant optical path; a hole is opened in the horizontal flue at the rear of the boiler, the multi-spectrum fusion probe is hung outside the flue, and the optical path is arranged so that the laser focus is aimed at the flue gas in the flue.

3. A corrosion risk early warning method based on chlorine element monitoring according to claim 2, characterized in that: The pulse energy of the high-energy pulse laser is greater than 200mJ and the frequency is 50Hz; the spectrometer adopts a high-resolution spectrometer, the wavelength range of the spectrometer is 300nm to 1100nm, including the characteristic emission line of the Cl element, and the resolution of the spectrometer is less than 0.15nm; the variable focal length lens group is a high-temperature resistant reflective lens group.

4. The corrosion risk early warning method based on chlorine element monitoring according to claim 1 is characterized in that: In the step of obtaining the equivalent chlorine deposition amount of the boiler flue heating surface using the Cl element concentration, flue gas flow rate and boiler flue heating surface: Filter the obtained Cl element concentration data, eliminate noise, and remove obvious erroneous data; The processed Cl element concentration data is combined with the flue gas flow rate to establish a calculation model for the equivalent chlorine deposition on the boiler flue heating surface, as follows: C d =C Cl vδ(T w )or w In the formula, C d is the equivalent chlorine deposition amount, C Cl is the Cl element concentration data after treatment, v is the flue gas flow rate, δ(T w ) is the adsorption coefficient of the heated surface wall, which is a function of temperature, η w It is the split-screen correction factor, which is related to the contact area between the heated surface and the flue gas.

5. The corrosion risk early warning method based on chlorine element monitoring according to claim 1 is characterized in that: In the step of obtaining the corrosion risk warning index based on the equivalent chlorine deposition amount, the sulfur content of the coal entering the furnace, and the metal wall temperature of the heating surface, the specific calculation formula is as follows: γ=a1C d +a2S+a3T m / 1000 In the formula, C d is the equivalent chlorine deposition; S is the sulfur content in coal; T m is the metal wall temperature of the heated surface; a1, a2, and a3 are weight coefficients, which are determined by regression of test data and historical corrosion cases.

6. A corrosion risk early warning method based on chlorine element monitoring according to claim 1, characterized in that: Obtain historical coal blending plans and Cl release characteristics, obtain the corresponding relationship between the corrosion risk warning index and the actual corrosion situation of the boiler, divide the corrosion threshold according to the corrosion situation, and compare the corrosion risk warning index with the corrosion threshold to implement the corrosion risk warning step of the coal quality entering the furnace: The corrosion thresholds include a medium corrosion level threshold and a severe corrosion level threshold. When the corrosion risk warning index exceeds the medium corrosion threshold, a corrosion risk warning of the coal quality entering the furnace is issued.

7. A method for optimizing coal blending scheme based on chlorine element monitoring, characterized in that: The specific steps are as follows: Establish a database including Cl release characteristics of different coal qualities, boiler corrosion conditions, coal quality industrial analysis data, and corrosion risk warning index and corrosion level threshold when different coal qualities are burned using the method described in any one of claims 1 to 6; Establishing constraints for a neural network algorithm, using coal quality industrial analysis data and the total coal demand corresponding to the target load as input; using the blending ratio of each coal type as output, training the neural network algorithm through the database to obtain a calculation model for dynamically optimizing the coal blending scheme for the furnace; The coal quality industrial analysis data is input into the calculation model of the dynamic optimization of the coal blending scheme for the furnace to obtain the blending ratio of each coal type.

8. The method for optimizing coal blending scheme based on chlorine element monitoring according to claim 7, characterized in that: The constraints of the neural network algorithm include: Corrosion risk constraint: Corrosion risk warning index ≤ medium corrosion threshold; Load constraint: The calorific value of the mixed coal after blending must meet Q 混合 ≥Q 负荷要求 ; Environmental constraints: The sulfur and chlorine content of the blended coal must meet emission limits; Blending ratio constraint: The sum of the blending ratios of each coal type is equal to 1.

9. The method for optimizing coal blending scheme based on chlorine element monitoring according to claim 7, characterized in that: After blending the coals according to the obtained blending ratios of the various coal types and putting them into use, the corrosion risk warning index during combustion of different coal qualities is obtained by the method according to any one of claims 1 to 6; If the actual corrosion risk warning index is greater than the medium corrosion threshold, the calculation model for the dynamic optimization of the coal blending scheme for the furnace will recalculate the blending ratio.

10. The method for optimizing coal blending scheme based on chlorine element monitoring according to claim 7, characterized in that: When training the neural network algorithm, the coal quality parameters, blending ratios and corresponding corrosion risk warning indexes in the historical coal blending plan are used as labels; a supervised learning training model is used, and the objective function is to minimize the error between the predicted and actual values.

Images