Horizontal flue soot blowing control method and system adapting to deep peak regulation mode of thermal power generating unit
By arranging a capacitive level meter in the horizontal flue of the thermal power unit and using the TimeMixer prediction model, real-time monitoring and prediction of the ash accumulation height of the boiler horizontal flue is achieved, which solves the problem of boiler heated area ash in the deep peak-shaving mode of the thermal power unit, and realizes timely, thorough and economical soot blowing control of the boiler operation.
Patent Information
- Application Number
- CN202510191136.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-20
- Publication Date
- 2025-04-25
AI Technical Summary
When the thermal power unit is operating in deep peak-shaving mode, the fly ash is raised due to insufficient combustion on the boiler, and the flue gas flow rate is affected, so it cannot be taken away all the fly ash, resulting in ash accumulation in areas such as horizontal flue, affecting the uniformity of the boiler flue gas flow, increasing power consumption and aggravating the wear of the flow tube bundle. The existing soot blowing methods are not timely, thoroughly, and intelligent, which affects the safety and economicality of boiler operation.
A distributed sensor fusion system is used to arrange a capacitive level meter in the horizontal flue to monitor the ash accumulation height in real time, and combined with the TimeMixer prediction model, soot blowing discrimination conditions are set according to the ash deposition change curve and the level meter signal to achieve automated soot blowing control.
By monitoring and predicting the ash accumulation height in real time, and performing soot blowing operations in a timely manner, ensuring the economical and safe operation of the boiler, reducing the risk of wear and increased power consumption on the heated surface.
Smart Images
Figure CN119879221A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of sootblowing of a boiler tail heating surface, and in particular to a horizontal flue sootblowing control method and system adapted to a deep peak regulation mode of a thermal power unit. Background Art
[0002] Under the background of "dual carbon", domestic thermal power units have been transformed into flexible ones one after another. However, at the beginning of the design, most thermal power units did not consider that the units would operate in a deep peak-shaving mode for a long time. Therefore, the thermal power units in this deep peak-shaving operation mode will inevitably affect the operation safety of the boiler side, the turbine side and the auxiliary equipment. Among them, during the low-load operation of the boiler, incomplete combustion leads to an increase in fly ash, and the flue gas flow rate is affected and cannot take away all the fly ash, resulting in ash accumulation in areas such as horizontal flues. Ash accumulation causes a significant decrease in the heat exchange capacity of the heating surface itself, resulting in an increase in the outlet flue gas temperature, an increase in resistance, an increase in the output of the induced draft fan, and an increase in power consumption. The local ash accumulation phenomenon will seriously affect the uniformity of the boiler flue gas flow, causing the local flue gas flow rate to be high and aggravating the wear of the flow tube bundle.
[0003] Sootblowing is an effective way to remove dust from the heating surface. Existing sootblowing methods are basically achieved through strong flushing of some other gases or liquids, including steam sootblower, high-energy gas pulse sootblower and sonic sootblower, as well as high-pressure water flushing after shutdown. Existing sootblowing control strategies are mostly manual sootblowing or timed and quantitative sootblowing. During the operation of the unit, if the sootblower is used too many times, that is, the sootblowing cycle is too short, it will cause unnecessary sootblowing fluid consumption and wear and corrosion of the heating surface boiler tubes; if the cycle is too long, the sootblowing effect is not good, and the loss caused by dust on the heating surface is too large. Considering the safety and economy of boiler operation, it is particularly important to select the best sootblowing time.
[0004] Existing soot blowing control strategies are mostly manual soot blowing, timed and quantitative soot blowing, soot blowing based on the degree of soot accumulation determined by soot thermal resistance, and soot blowing based on unit operation big data, which have the following disadvantages:
[0005] 1. Manual soot blowing, which requires the unit to be shut down for operation, has a long soot blowing downtime period, which seriously affects the normal operation of the unit. In addition, this method requires manpower to disassemble and assemble the equipment, which is not smart enough.
[0006] 2. Timed and quantitative sootblowing. This method cannot be adjusted according to the actual blockage of the equipment. If the sootblower is used too many times, that is, the sootblowing cycle is too short, it will cause unnecessary sootblowing fluid consumption and wear and corrosion of the heating surface furnace tubes; if the cycle is too long, the sootblowing effect is not good, the loss caused by soot contamination on the heating surface is too great, and the sootblowing is not thorough.
[0007] 3. Using thermal resistance to reflect the degree of ash accumulation for soot blowing. This method introduces an ash accumulation factor and uses the ash accumulation thermal resistance of the heated surface to determine whether to blow soot. This calculation method involves many heat sources, has poor anti-interference ability, and is easily interfered by other thermal resistances, resulting in inaccurate judgment of the degree of ash accumulation. In addition, this method is usually applicable to boiler heating surfaces and low-low economizers, and is not universal.
[0008] 4. Soot blowing is performed based on the unit operation big data to reflect the degree of ash accumulation and blockage in the horizontal flue. This soot blowing method is purely mathematically driven and lacks integration with the physical model. It does not consider the physical mechanism behind the ash accumulation and blockage in the horizontal flue. Based on the mathematically driven model, the amount of data calculation is too large and the calculation time is too long. The prediction results under different working conditions have large errors, which affects the reliability.
[0009] In summary, a horizontal flue soot blowing control method and system that is adapted to the deep peak-shaving mode of thermal power units is needed to solve the problems of imperfect existing soot blowing technology, untimely, incomplete, unintelligent and unsafe soot blowing. Summary of the invention
[0010] The purpose of the present invention is to provide a horizontal flue soot blowing control method and system adapted to the deep peak regulation mode of a thermal power unit.
[0011] A horizontal flue soot blowing control method adapted to the deep peak regulation mode of a thermal power unit comprises the following steps:
[0012] A capacitive level meter is arranged in the horizontal flue using a distributed sensor fusion system, and the ash accumulation height is monitored in real time by calculating the capacitance change between the level meter probes.
[0013] Establish the TimeMixer prediction model, select the operating parameters, and perform ash deposition prediction;
[0014] The soot blowing discrimination conditions are set according to the horizontal flue ash deposition change curve or the ash accumulation height signal transmitted by the level meter, and the soot blowing control is performed according to the ash deposition prediction results of the TimeMixer prediction model.
[0015] Furthermore, the operating parameters include unit load, coal consumption, excess air coefficient, operating time, and fuel ash content.
[0016] Furthermore, the specific steps of ash deposition prediction are as follows:
[0017] Collect historical operation data of DCS system, clean and filter the data, process missing values and eliminate noise;
[0018] Calculate the mutual information between each input variable and the dust accumulation height, and select the highly correlated input features;
[0019] By downsampling the input time series, time series of various time resolutions are generated, and each series is mapped to a high-dimensional feature space through an embedding layer;
[0020] The PDM blocks are decomposed, seasonally mixed, trend mixed, and information interacted and outputted;
[0021] Aggregate the outputs from the PDM block, utilizing both fine-grained and coarse-grained information, to generate future predictions of ash deposition.
[0022] Furthermore, the soot blowing judgment condition set according to the horizontal flue ash deposition variation curve is to predict that the ash deposition height in the horizontal flue will be greater than the soot blowing critical height at the future time T=2h.
[0023] The soot blowing judgment condition set according to the soot accumulation height signal transmitted by the material level meter is that the measured soot accumulation height is greater than the soot blowing critical height.
[0024] A control system for a horizontal flue soot blowing control method adapted to a deep peak regulation mode of a thermal power unit comprises a ball check valve, a filter, a pneumatic butterfly valve, a pressure switch, a soot blowing nozzle, a material level meter and an external automatic control system, wherein the ball check valve, the filter, the pneumatic butterfly valve, the pressure switch and the soot blowing nozzle are connected in sequence, a material level meter is arranged below the soot blowing nozzle, and the pneumatic butterfly valve and the pressure switch are connected to the external automatic control system.
[0025] Furthermore, the sootblowing nozzles are distributed at equal distances at the end of the pipeline.
[0026] The beneficial effects of the present invention are:
[0027] 1. Combining physical mechanism with mathematical drive, a soot accumulation prediction model is constructed. Taking into account relevant factors such as unit operating conditions and flue gas parameters, the actual blockage of the horizontal flue is determined, the soot deposition trend is predicted, and the appropriate soot blowing time is selected, so that soot blowing can be carried out in time.
[0028] 2. A level meter is arranged in the horizontal flue to monitor the ash accumulation height in the horizontal flue in real time and use this as one of the conditions for judging soot blowing, which complements the mathematical simulation. When the ash accumulation height exceeds the set threshold, soot blowing is carried out in time to ensure the economy of the unit operation and soot blowing process.
[0029] 3. Use a large number of measured operating parameters to train and test the dust accumulation prediction model to ensure the accuracy of the prediction results. A large amount of data combined with mathematical drive prediction results have smaller errors and higher reliability.
[0030] 4. The interval between two adjacent sootblowing operations does not exceed the threshold time. Considering the complexity of unit operation under the background of deep peak regulation, a sootblowing strategy combining intelligent sootblowing and timed sootblowing can be adopted. On the basis of intelligent sootblowing, if there is no sootblowing within the specified time, sootblowing will be forced to start once, which improves the safety of unit operation and sootblowing process. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] Figure 1 It is the overall logic diagram of the horizontal flue soot blowing control method of the present invention adapted to the deep peak regulation mode of the thermal power unit;
[0032] Figure 2 It is the layout diagram of the horizontal flue material level meter;
[0033] Figure 3 This is the basic structure diagram of the TimeMixer model;
[0034] Figure 4 It is the logic diagram of sootblowing rule construction and verification;
[0035] Figure 5 It is a diagram of a horizontal flue soot blowing control system adapted to the deep peak regulation mode of thermal power units. DETAILED DESCRIPTION
[0036] The present invention proposes a horizontal flue soot blowing control method and system adapted to the deep peak regulation mode of a thermal power unit. The present invention is further described below in conjunction with the accompanying drawings and specific embodiments.
[0037] The present invention uses the real-time monitoring value of the dust accumulation height and the TimeMixer model to predict the dust accumulation blockage of the flue, judge the actual blockage of the horizontal flue and select the appropriate soot blowing time, so as to automatically control the soot blowing system to achieve the purpose of timely and thorough soot blowing, which is safer, energy-saving and environmentally friendly. The overall logic is as follows Figure 1 shown.
[0038] 1. Real-time monitoring of dust accumulation height
[0039] Capacitive level meters are arranged in the horizontal flue using a distributed sensor fusion system. Distributed capacitive level meters have low communication bandwidth requirements, fast calculation speed, good reliability and continuity, and are suitable for high-temperature operating environments in boilers. The ash accumulation height is monitored in real time by calculating the capacitance changes between the level meter probes. Figure 2 This is the layout diagram of the horizontal flue level meter.
[0040] 2. TimeMixer prediction model
[0041] Figure 3 It is the basic structure diagram of the TimeMixer model, including the following aspects:
[0042] 1. Operation parameter selection
[0043] The degree of ash accumulation and blockage of the horizontal flue on site can be directly reflected by the ash deposition at the inlet and outlet of the horizontal flue. The ash accumulation height is directly measured by the level meter arranged in the horizontal flue. The horizontal flue ash deposition is mainly affected by the unit load, coal consumption, excess air coefficient, operating time, and fuel ash content. The basis for parameter selection is as follows:
[0044] During the low-load operation of the boiler, the flue gas velocity is affected and cannot carry away all the fly ash. Some fly ash accumulates in the convection superheater, low-temperature superheater and horizontal flue and slowly forms ash. Due to the difference in the composition of the coal used and the designed coal, the fly ash content in the flue gas increases after the coal is burned. Some fly ash cannot be carried away by the flue gas and forms physical deposition on the convection heating surface. In addition, the amount of ash generated in the furnace increases with the increase in coal consumption and the passage of time.
[0045] 2. Ash deposition prediction
[0046] Ash deposition prediction is divided into five steps, namely data collection and preprocessing, initial feature selection, multi-scale decomposition, PDM block processing and future prediction (FMM).
[0047] Step 1: Data collection and preprocessing
[0048] Collect historical operation data of the DCS system, clean and filter the data, process missing values and eliminate noise.
[0049] Step 2: Initial feature selection
[0050] The mutual information (MI) between each input variable and the dust accumulation height is calculated, and the highly correlated input features are selected.
[0051] Step 3: Multi-scale decomposition
[0052] The multi-scale decomposition process generates time series of various time resolutions by downsampling the input time series X, which is expressed as X = {X 0 ,X 1 ,...,X M}. Specifically:
[0053] X 0 : represents the original input time series, capturing the finest temporal details.
[0054] X M : Represents the coarsest time scale series, highlighting the macroscopic time trend.
[0055] Each X m(where m=0,...,M) is obtained by gradually applying average pooling, gradually reducing the temporal resolution. This process enables the model to separate fine-grained patterns from broader temporal trends and effectively capture multi-scale temporal dynamics.
[0056] After generating these multi-scale time series, each series X m Mapped to high-dimensional feature space through the embedding layer. The formula for this process is:
[0057]
[0058] here, Indicates m -t h The embedding layer transforms each time series into a deep feature space, providing a basis for subsequent decomposition and integration steps.
[0059] Step 4: PDM block processing
[0060] For each PDM block l, perform the following steps:
[0061] (1) Decomposition
[0062] Each time scale series X m Decomposed into seasonal components S m and trend component T m :
[0063] X m =S m +T m
[0064] (2) Seasonal Mix
[0065] Seasonal component S m Mixing is performed in a bottom-up manner, gradually transferring information from fine scale to coarse scale to generate mixed seasonal components
[0066]
[0067] (3) Trend Mix
[0068] Trend component T m Mixing is performed in a top-down manner, gradually transferring information from coarse scale to fine scale to generate mixed trend components
[0069]
[0070] (4) Information interaction and output
[0071] Seasonal component after mixing and trend component Further processing through the feedforward neural network generates the lth -th The output of the block X l :
[0072]
[0073] Step 5: Future Prediction (FMM)
[0074] The Future Prediction Module (FMM) integrates multi-scale features by aggregating the outputs from the aforementioned PDM blocks. This integration utilizes both fine-grained and coarse-grained information to generate accurate future predictions. The final prediction results are expressed as: y = DeepNN (Integrated Prediction Results).
[0075] Collect the furnace load, coal consumption, excess air coefficient, operating time, coal ash content and the corresponding measured ash height value of the material level meter. Use TimeMixer to establish a prediction model with 70% of the data as the training set, and use 30% of the data as the test set to test the rules of the prediction model and test the model accuracy. After the model prediction accuracy meets the requirements of soot blowing in industrial production, the prediction ash accumulation model is used to control soot blowing. Figure 4 It is the logic diagram of sootblowing rule construction and verification.
[0076] 3. Sootblowing control
[0077] Set the soot blowing control loop, visualize the prediction results of the TimeMixer model, and obtain the horizontal flue ash deposition change curve. Set the soot blowing critical height L lj =20%L, L is the height of the horizontal flue. The curve directly reflects the growth rate of ash deposition in the horizontal flue. When the ash deposition height H in the horizontal flue is predicted at the time T=2h in the future, lj Greater than the critical height of sootblowing L lj In addition, the ash accumulation height signal from the material level meter arranged in the horizontal flue is used for judgment. When the measured ash accumulation height H is greater than the critical height of ash blowing L lj , perform soot blowing action.
[0078] In addition, the soot blowing frequency and soot blowing interval can be limited by combining the timed soot blowing strategy. For example, the number of soot blowing operations for the horizontal flue in a single day does not exceed the first preset value; to prevent frequent soot blowing, it is advisable to set the soot blowing frequency to no more than 2 to 3 times per day.
[0079] The interval between two adjacent sootblowing operations does not exceed the threshold time. Considering the complexity of unit operation under the background of deep peak regulation, a sootblowing strategy combining intelligent sootblowing and timed sootblowing can be adopted. Since the unit has been in low-load operation mode for a long time, there is less soot accumulation in the horizontal flue, so that the pressure difference growth rate is always lower than the set critical value. In order to prevent the soot accumulation from hardening and scaling, on the basis of intelligent sootblowing, it is calculated as 8 hours. If there is no sootblowing within 8 hours, sootblowing will be forced to start once.
[0080] Through the above control ideas, the physical mechanism drive and mathematical model drive are combined in the sootblowing control method to grasp the best time for sootblowing and realize the timeliness, thoroughness, economy and safety of the unit sootblowing system.
[0081] Figure 5 This is a diagram of a horizontal flue sootblowing control system that is suitable for the deep peak-shaving mode of thermal power units. The air intake duct is arranged on the external piping system of the boiler system. The inlet of the air intake duct is connected to a stable compressed air source through a high-temperature resistant sealing material, and the air source pressure is set to 0.5-0.8Mpa. From the starting end of the air intake duct, a ball check valve 1, a filter 2, a pneumatic butterfly valve 3 and a pressure switch 4 are connected in sequence to ensure that the airflow can only flow in one direction to prevent backflow from causing system failure. Sootblowing nozzles 5 are arranged at equal distances at the end of the pipeline, and a material level meter 6 is set below the sootblowing nozzles 5. The pneumatic butterfly valve 3 and the pressure switch 4 in the sootblowing system are connected to an external automatic control system to achieve safe and normal operation of the compressed air sootblowing system for the horizontal flue suitable for the deep peak-shaving mode of thermal power units.
[0082] The present invention can achieve timely soot blowing, ensure the economy of the unit operation and soot blowing process. In addition, the accuracy of the prediction result can be ensured, so that the prediction result error is small and the reliability is higher.
Claims
1. A horizontal flue soot blowing control method adapted to the deep peak regulation mode of a thermal power unit, characterized in that: The following steps are involved: A capacitive level meter is arranged in the horizontal flue using a distributed sensor fusion system, and the ash accumulation height is monitored in real time by calculating the capacitance change between the level meter probes. Establish the TimeMixer prediction model, select the operating parameters, and perform ash deposition prediction; The soot blowing discrimination conditions are set according to the horizontal flue ash deposition change curve or the ash accumulation height signal transmitted by the level meter, and the soot blowing control is performed according to the ash deposition prediction results of the TimeMixer prediction model.
2. The horizontal flue soot blowing control method adapted to the deep peak load regulation mode of a thermal power unit according to claim 1 is characterized in that: Operating parameters include unit load, coal consumption, excess air coefficient, operating time, and fuel ash content.
3. The horizontal flue soot blowing control method adapted to the deep peak load regulation mode of a thermal power unit according to claim 2 is characterized in that: The specific steps of ash deposition prediction are as follows: Collect historical operation data of DCS system, clean and filter the data, process missing values and eliminate noise; Calculate the mutual information between each input variable and the dust accumulation height, and select the highly correlated input features; By downsampling the input time series, time series of various time resolutions are generated, and each series is mapped to a high-dimensional feature space through an embedding layer; The PDM blocks are decomposed, seasonally mixed, trend mixed, and information interacted and outputted; Aggregate the outputs from the PDM block, utilizing both fine-grained and coarse-grained information, to generate future predictions of ash deposition.
4. The horizontal flue soot blowing control method adapted to the deep peak load regulation mode of a thermal power unit according to claim 1 is characterized in that: The soot blowing judgment condition set according to the horizontal flue ash deposition change curve is to predict that the ash deposition height in the horizontal flue will be greater than the soot blowing critical height at the future time T=2h.
5. The horizontal flue soot blowing control method adapted to the deep peak load regulation mode of a thermal power unit according to claim 1 is characterized in that: The soot blowing judgment condition set according to the soot accumulation height signal transmitted by the material level meter is that the measured soot accumulation height is greater than the soot blowing critical height.
6. A control system for the horizontal flue soot blowing control method adapted to the deep peak load regulation mode of a thermal power unit according to claim 1, characterized in that: The invention comprises a ball check valve (1), a filter (2), a pneumatic butterfly valve (3), a pressure switch (4), a soot blowing nozzle (5), a material level meter (6) and an external automatic control system, wherein the ball check valve (1), the filter (2), the pneumatic butterfly valve (3), the pressure switch (4) and the soot blowing nozzle (5) are connected in sequence, a material level meter (6) is arranged below the soot blowing nozzle (5), and the pneumatic butterfly valve (3) and the pressure switch (4) are connected to the external automatic control system.
7. The control system of the horizontal flue soot blowing control method adapted to the deep peak load regulation mode of thermal power units according to claim 6 is characterized in that: The soot blowing nozzles (5) are distributed at equal distances at the end of the pipeline.