Gasification furnace automatic optimization intelligent system

Abstract

The invention relates to the technical field of coal gasification, and designs an automatic optimization intelligent system for a gasification furnace in order to solve the problem of parameter control of an intermittent gasification furnace. The system integrates a plurality of modules of uplink and downlink temperature control, feeding time control and the like, and parameters of blowing time, steam quantity and the like are accurately adjusted by means of self-optimization, fuzzy control, predictive control algorithms and fusion of an in-furnace reaction model and an expert system. During actual operation, the adjustment amount of each parameter is calculated according to real-time data, and automatic intelligent control is achieved. The system is high in commissioning rate, effectively reduces labor intensity and management cost of operators, improves temperature control precision, stabilizes a material layer, reduces blow-over phenomena, guarantees stable, long-term and full-optimal operation of the gasification furnace, remarkably improves production efficiency and economic benefits, and promotes intelligent development of the coal gasification industry.

Description

Technical Field

[0001] The present invention relates to the field of gasifiers, and specifically to an automatic optimization intelligent system for gasifiers. Background Art

[0002] In the process of coal gasification production, the stable and efficient operation of intermittent gasifiers is of great importance. However, current intermittent gasifiers face many problems. On the one hand, there are numerous process parameters to be controlled, such as temperature, steam volume, feeding time, grate speed, carbon layer height, etc., and there are serious coupling relationships among these parameters. In different gasification stages, the controlled parameters are quite different, making it extremely difficult to determine the optimal working state of the gasifier, which is comprehensively restricted by various factors. On the other hand, the existing parameter operations mainly rely on manual adjustment by operators. Due to the uneven technical levels, large differences in operating habits, and the difficulty in unifying the grasp of the "degree" of adjustment among operators, the final adjustment effects are significantly different, and it is impossible to ensure the stable and efficient operation of the gasifier. This not only increases the labor intensity but also raises the management cost, restricting the improvement of production efficiency and product quality. Therefore, it is extremely urgent to develop an automatic optimization intelligent system for gasifiers that can achieve unified and intelligent operation. Summary of the Invention

[0003] The purpose of the present invention is to provide an automatic optimization intelligent system for gasifiers to solve the problems raised in the above background art.

[0004] To achieve the above purpose, the present invention provides the following technical solution: An automatic optimization intelligent system for gasifiers includes an upper and lower temperature control module, a feeding time control module, a grate temperature control module, a carbon layer height control module, an intelligent steam flow rate adjustment module, a fault diagnosis and early warning module, and an expert system optimization module. The upper and lower temperature control module calculates the given quantities of blowing, upblowing, and downblowing times for the next cycle based on the set values and actual values of the upper and lower temperatures, then calibrates the temperature response lag of the gasification layer based on historical operation big data analysis, accurately calculates the blowing time for the next cycle, and then calculates the ratio of the upblowing and downblowing times for the next cycle using the GPC algorithm according to the upper and lower temperature differences, and finally determines the upblowing and downblowing times in combination with the calculation formula for the sum of the upblowing and downblowing times;

[0005] The feeding time control module calculates the given quantity of feeding time within the selected gear time range according to the set value and actual value of the upper temperature;

[0006] The grate temperature control module calculates the given value of the grate rotation speed based on the set value and actual value of the grate temperature, as well as the trend of the downward temperature. When both the downward temperature and the grate temperature are decreasing, and the actual grate temperature is lower than the set temperature, the grate rotation speed is increased; conversely, the grate rotation speed is decreased. When the downward temperature remains unchanged and the grate temperature changes, the grate rotation speed remains unchanged.

[0007] The carbon layer height control module predicts the change trend of the carbon layer height in advance by establishing a carbon layer height prediction model.

[0008] The intelligent steam flow regulation module uses a flow sensor to continuously monitor the steam flow rate, and combines the load change of the gasifier and the pressure parameters in the furnace to dynamically adjust the opening of the steam regulating valve through a fuzzy control algorithm to achieve precise control of the steam flow rate.

[0009] The fault diagnosis and early warning module judges whether the gasifier has a fault by analyzing the data collected by the sensors in real time and combining with a fault diagnosis model. Once a fault is detected, an alarm is immediately issued, and relevant operators are notified by text message and email. At the same time, analysis of the cause of the fault and suggestions for solutions are provided.

[0010] The expert system optimization module contains the operation knowledge and experience of the gasifier, and is dynamically updated by combining real-time production data and industry achievements to evaluate and optimize the control strategy of the system in real time.

[0011] Preferably, in the upper and lower temperature control modules, the response lag of the gasification layer temperature is calibrated based on big data analysis of historical operation data, and the blowing time of the next cycle is accurately calculated. Subsequently, according to the temperature difference between the upper and lower temperatures, the ratio of the upper blowing time to the lower blowing time of the next cycle is calculated using the GPC algorithm. Then, combined with the calculation formula for the sum of the upper blowing time and the lower blowing time, the specific implementation steps for determining the upper blowing time and the lower blowing time are as follows:

[0012] Step 1: Collect and organize big data of historical operation

[0013] Data acquisition: Use sensors installed at different positions of the gasifier to continuously collect various parameter data during the operation of the gasifier. The above parameter data includes but is not limited to the upper temperature, lower temperature, blowing time, upper blowing steam volume, lower blowing steam volume, feeding time, grate rotation speed, and carbon layer height. The acquisition frequency is set according to the actual situation, and is collected once per minute or every few seconds to ensure the timeliness and accuracy of the data.

[0014] Data preprocessing: Clean and preprocess the collected raw data to remove noise, outliers, and missing values.

[0015] Step 2: Analyze and calibrate the response lag of the gasification layer temperature

[0016] Establish a time series model: Arrange the historical operation data in chronological order, construct time series of the upper temperature, lower temperature, and blowing time parameters, and model the lag of the gasification layer temperature response through the analysis of these time series using the autoregressive integrated moving average model;

[0017] Determine the lag time and response coefficient: Through the fitting and analysis of the time series model, determine the lag time when the gasification layer temperature begins to change significantly after the blowing time changes; at the same time, calculate the response coefficient of the gasification layer temperature under different blowing time changes, that is, the amplitude by which the gasification layer temperature rises or falls correspondingly when the blowing time increases or decreases by one unit;

[0018] Step 3: Calculate the blowing time for the next cycle

[0019] Set the target temperature: According to the production process requirements and actual operation conditions of the gasifier, set the target values of the upper temperature and the lower temperature;

[0020] Calculate the temperature deviation: Compare the actual values of the upper temperature and the lower temperature collected in the current cycle with the target values, and calculate the temperature deviation;

[0021] Use the GPC algorithm to calculate the blowing time: Take the temperature deviation, the lag time and response coefficient of the gasification layer temperature response, and the historical operation data as inputs, and use the generalized predictive control model and algorithm to calculate the given value of the blowing time for the next cycle;

[0022] Step 4: Calculate the ratio of the up-blowing time to the down-blowing time for the next cycle

[0023] Analyze the temperature difference between the upper and lower temperatures: Calculate the difference between the upper temperature and the lower temperature in the current cycle, that is, temperature difference = upper temperature - lower temperature; according to the magnitude and sign of the temperature difference, judge the combustion state and heat distribution in the gasifier;

[0024] Use the GPC algorithm to calculate the time ratio: Take the temperature difference between the upper and lower temperatures, the historical operation data, and the blowing time calculated previously as inputs, and use the GPC algorithm again to calculate the ratio of the up-blowing time to the down-blowing time for the next cycle. The GPC algorithm will optimize the allocation of the up-blowing time and the down-blowing time according to the change trend and target of the temperature difference to achieve the purpose of making the upper and lower temperatures more balanced;

[0025] Step 5: Determine the up-blowing and down-blowing times for the next cycle

[0026] Calculate the sum of the up-blowing and down-blowing times: According to the process requirements and operating cycle of the gasifier, determine a fixed time period. At the same time, subtract the fixed two-up time and purging time, and then subtract the blowing time of the next cycle calculated previously, to obtain the sum of the up-blowing and down-blowing times. Among them, the two-up time is fixed at 10s, and the purging time is fixed at 1s or 2s;

[0027] Determine the up-blowing and down-blowing times according to the time ratio: Distribute the sum of the up-blowing and down-blowing times according to the previously calculated ratio of the up-blowing and down-blowing times, to determine the up-blowing time and down-blowing time of the next cycle;

[0028] Step 6, Implement control and monitor feedback

[0029] Execute the control instruction: Convert the calculated blowing time, up-blowing time, and down-blowing time of the next cycle into control instructions, and send them to the actuator of the gasifier;

[0030] Monitor feedback in real time: During the operation of the gasifier, continuously monitor in real time the changes in parameters such as the up-temperatures and down-temperatures, and compare and analyze the actual operation data with the target values.

[0031] Preferably, the specific working steps of the feed time control module are as follows:

[0032] Step A, Data collection and preparation

[0033] Obtain temperature data: Through the temperature sensors installed at specific positions of the gasifier, collect the actual values of the up-temperatures in real time;

[0034] Determine the set value: According to factors such as the production process requirements of the gasifier, the target gas quality, and the safe operating range of the equipment, preset a suitable up-temperature set value in advance;

[0035] Determine the gear time range: According to the operation experience and historical data of the gasifier, divide the feed time into several gears, and determine the corresponding time range for each gear;

[0036] Step B, Calculate the temperature deviation

[0037] Calculate the difference: Compare the actual value of the up-temperature collected with the set value, and calculate the difference between the two, that is, temperature deviation = actual value of up-temperature - up-temperature set value; Subsequently, according to the positive or negative of the calculated temperature deviation, judge whether the up-temperature is higher or lower than the set value;

[0038] Step C, Select the fuzzy control algorithm

[0039] By establishing a fuzzy rule base, input variables such as temperature deviation and deviation change rate are fuzzified, then inference is carried out according to the fuzzy rules, and finally the inference result is defuzzified to obtain the adjustment amount of the feeding time.

[0040] Step D: Calculate the adjustment amount of the feeding time

[0041] Combine with the current feeding time: Combine the calculated adjustment amount of the feeding time with the current feeding time to obtain a new given value of the feeding time.

[0042] Limit within the gear time range: Compare the new given value of the feeding time with the selected gear time range to ensure it is within this range. If it exceeds the gear time range, it needs to be adjusted to the nearest appropriate gear.

[0043] Step E: Implement control and feedback adjustment

[0044] Execute the adjustment of the feeding time: Convert the determined given value of the feeding time into a control signal and send it to the feeding equipment; during the feeding process, continuously and real - time monitor the change of the up - line temperature, compare the new actual temperature value with the set value, and calculate the temperature deviation again.

[0045] Preferably, the long - short - term memory network algorithm is introduced into the carbon layer height control module to establish a carbon layer height prediction model. The specific steps for predicting the change trend of the carbon layer height in advance are as follows:

[0046] Step Ⅰ: Data collection and pre - processing

[0047] Data collection: Collect historical data related to the carbon layer height. These data should include multiple features related to the change of the carbon layer height. The specific features include feeding time, blowing time, up - blowing / down - blowing steam volume, up - line temperature, down - line temperature, grate speed, and record the corresponding carbon layer height values at the same time.

[0048] Data pre - processing: Clean the collected data to remove noise, outliers, and missing values; then normalize all feature data and carbon layer height data and scale them to a fixed range; subsequently, divide the processed data into a training set, a validation set, and a test set according to a certain ratio, and the specific ratio is 7:2:1.

[0049] Step Ⅱ: Construct the LSTM model

[0050] Import necessary libraries: In Python, use the deep learning framework TensorFlow to construct the LSTM model.

[0051] Step Ⅲ: Model training

[0052] Convert the training set data into a format suitable for input to the LSTM model, i.e., a three-dimensional tensor, and then use the sliding window method to generate input sequences and corresponding target values; use the prepared training data to train the LSTM model, set appropriate training epochs and batch sizes. During the training process, the model will continuously adjust the parameters to minimize the loss function. At the same time, use the validation set data to validate the model and monitor the loss value of the validation set to prevent overfitting of the model.

[0053] Step IV. Model evaluation

[0054] After training is completed, use the test set data to evaluate the model and calculate the loss value and other evaluation metrics of the model on the test set.

[0055] Step V. Predict the change trend of the carbon layer height

[0056] Input real-time data: In practical applications, process the real-time collected feature data according to the same preprocessing method and then input it into the trained LSTM model.

[0057] Make predictions: The model will output the predicted carbon layer height value. According to the prediction results of multiple consecutive time steps, analyze the change trend of the carbon layer height.

[0058] Preferably, when the load of the gasifier increases, the intelligent steam flow regulation module automatically increases the steam flow; when the pressure in the furnace fluctuates abnormally, it timely adjusts the steam flow to maintain the stable progress of the reaction in the furnace.

[0059] Preferably, the expert system optimization module evaluates and optimizes the control strategy of the system according to different coal characteristics, adjusts the blowing intensity and the ratio of the up-blow / down-blow steam volume. When encountering complex working conditions or abnormal situations, the expert system optimization module provides reasonable control suggestions according to the knowledge and inference rules in the knowledge base to assist the self-optimization algorithm in making decisions, enabling the system to adapt to different operating conditions and improving the reliability and stability of the system.

[0060] Compared with the prior art, the beneficial effects of the present invention are:

[0061] Precise control of reaction conditions: Each module in the system works in coordination to precisely control key parameters such as the temperature, steam flow rate, feeding time, and carbon layer height in the gasifier. For example, the upper and lower temperature control modules utilize the Generalized Predictive Control model and algorithm (GPC) to calibrate the temperature response time lag characteristics of the gasification layer based on the analysis of large amounts of historical operation data, and accurately calculate the blowing, upward blowing, and downward blowing times to ensure that the gasification reaction proceeds under the most suitable temperature conditions. The intelligent steam flow regulation module dynamically adjusts the opening degree of the steam regulating valve through a fuzzy control algorithm, combining parameters such as the load change of the gasifier and the pressure inside the furnace, to achieve precise control of the steam flow rate, enabling the fuel and steam to react fully and improving the efficiency of the gasification reaction;

[0062] Optimization of fuel utilization: The feeding time control module accurately calculates the given value of the feeding time within the selected gear time range according to the set value and actual value of the upward temperature, avoiding excessive or insufficient fuel input. The carbon layer height control module predicts the change trend of the carbon layer height in advance by fine-tuning the set value of the upward temperature and combining the carbon layer height prediction model established by machine learning algorithms, and reasonably adjusts the feeding time to keep the carbon layer height in the best state, ensuring the full gasification of the fuel, reducing energy waste, and improving energy utilization efficiency.

[0063] Real-time monitoring and fault warning: The fault diagnosis and warning module can timely detect abnormal situations during the operation of the gasifier, such as abnormal temperature, excessive pressure, feeding blockage, etc., by analyzing the data collected by each sensor in real time and combining the fault diagnosis model. Once a fault is detected, an alarm is immediately issued, and relevant operators are notified by means of text messages, emails, etc. At the same time, the cause analysis of the fault and suggestions for solutions are provided, enabling the operators to take timely measures to avoid the expansion of the fault and ensuring the stable operation of the gasifier;

[0064] Adaptive control and optimization: The expert system optimization module contains the knowledge and experience of the gasifier operation, and is dynamically updated by combining real-time production data and the latest research results in the industry. This module conducts real-time evaluation and optimization of the control strategy of the system. Description of the Drawings

[0065] Figure 1 It is a schematic diagram of the system structure of the present invention;

[0066] Figure 2 It is a schematic diagram of the working process structure of the upper and lower temperature control module of the present invention;

[0067] Figure 3 It is a schematic diagram of the working process structure of the feeding time control module of the present invention. Detailed Embodiment

[0068] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0069] Please refer to Figures 1-3 , the present invention provides a technical solution: an automatic optimization intelligent system for a gasifier, including upper and lower temperature control modules, feed time control module, grate temperature control module, carbon layer height control module, intelligent steam flow regulation module, fault diagnosis and early warning module, and expert system optimization module. The upper and lower temperature control modules calculate the blowing, upblowing, and downblowing time given values for the next cycle according to the set values and actual values of the upper and lower temperatures, and then calibrate the temperature response lag of the gasification layer based on the analysis of historical operation big data, accurately calculate the blowing time for the next cycle, and then calculate the ratio of the upblowing time to the downblowing time for the next cycle using the GPC algorithm according to the temperature difference between the upper and lower temperatures, and then determine the upblowing and downblowing times in combination with the calculation formula for the sum of the upblowing and downblowing times;

[0070] Among them, the specific implementation steps for the upper and lower temperature control modules to calibrate the temperature response lag of the gasification layer based on the analysis of historical operation big data, accurately calculate the blowing time for the next cycle, then calculate the ratio of the upblowing time to the downblowing time for the next cycle using the GPC algorithm according to the temperature difference between the upper and lower temperatures, and then determine the upblowing and downblowing times in combination with the calculation formula for the sum of the upblowing and downblowing times are as follows:

[0071] Step 1: Collect and organize historical operation big data

[0072] Data acquisition: Use sensors installed at different positions of the gasifier to continuously collect various parameter data during the operation of the gasifier. The above parameter data includes but is not limited to upper temperature, lower temperature, blowing time, upblowing steam volume, downblowing steam volume, feed time, grate speed, and carbon layer height. The acquisition frequency is set according to the actual situation, and is collected once per minute or every few seconds to ensure the timeliness and accuracy of the data;

[0073] Data preprocessing: Clean and preprocess the collected raw data to remove noise, outliers, and missing values;

[0074] Step 2: Analyze and calibrate the temperature response lag of the gasification layer

[0075] Build a time series model: Arrange the historical operation data in chronological order, construct time series of the upward temperature, downward temperature, and blowing time parameters, and model the lag of the gasification layer temperature response through the analysis of these time series using the autoregressive integrated moving average model;

[0076] Determine the lag time and response coefficient: Through the fitting and analysis of the time series model, determine the lag time when the gasification layer temperature starts to change significantly after the blowing time changes; at the same time, calculate the response coefficient of the gasification layer temperature under different blowing time changes, that is, the amplitude of the corresponding increase or decrease in the gasification layer temperature when the blowing time increases or decreases by one unit. For example, after a large amount of data analysis, it is found that when the blowing time increases by 1 second, after a lag of 5 minutes, the gasification layer temperature rises by an average of 2 °C. Then, 5 minutes can be determined as the lag time, and 2 °C / s can be determined as the response coefficient;

[0077] Step 3: Calculate the blowing time for the next cycle

[0078] Set the target temperature: According to the production process requirements and actual operation conditions of the gasifier, set the target values of the upward temperature and downward temperature;

[0079] Calculate the temperature deviation: Compare the actual values of the upward temperature and downward temperature collected in the current cycle with the target values to calculate the temperature deviation. If the target value of the upward temperature is 800 °C and the actual value is 780 °C, then the upward temperature deviation is -20 °C; if the target value of the downward temperature is 750 °C and the actual value is 760 °C, then the downward temperature deviation is +10 °C;

[0080] Calculate the blowing time using the GPC algorithm: Use the temperature deviation, the lag time and response coefficient of the gasification layer temperature response, and the historical operation data as inputs, and calculate the given value of the blowing time for the next cycle using the generalized predictive control model and algorithm;

[0081] Step 4: Calculate the ratio of the upblowing time to the downblowing time for the next cycle

[0082] Analyze the temperature difference between the upward and downward temperatures: Calculate the difference between the upward temperature and the downward temperature in the current cycle, that is, the temperature difference = upward temperature - downward temperature; according to the magnitude and sign of the temperature difference, judge the combustion state and heat distribution in the gasifier;

[0083] Calculate the ratio of the time for upward blowing to that for downward blowing using the GPC algorithm: Take the temperature difference between the upward and downward directions, historical operation data, and the blowing time obtained from the previous calculation as inputs, and use the GPC algorithm again to calculate the ratio of the upward blowing time to the downward blowing time for the next cycle. The GPC algorithm will optimize the allocation of the upward blowing time and the downward blowing time according to the change trend and target of the temperature difference to achieve the purpose of making the temperatures in the upward and downward directions more balanced. For example, according to the calculation by the GPC algorithm, in order to raise the upward temperature to the target value, the blowing time for the next cycle needs to be increased by 3 seconds.

[0084] Step 5. Determine the upward blowing time and the downward blowing time for the next cycle

[0085] Calculate the sum of the upward blowing time and the downward blowing time: According to the process requirements and operation cycle of the gasifier, determine a fixed time period. At the same time, subtract the fixed double-up time and the purging time, and then subtract the blowing time for the next cycle obtained from the previous calculation to obtain the sum of the upward blowing time and the downward blowing time. The double-up time is fixed at 10 s, and the purging time is fixed at 1 s or 2 s. For example, if the blowing time is 20 seconds, the double-up time is 10 seconds, and the purging time is 1 second, then the sum of the upward blowing time and the downward blowing time = 120 - 20 - 10 - 1 = 89 seconds.

[0086] Determine the upward blowing time and the downward blowing time according to the ratio of the time: Allocate the sum of the upward blowing time and the downward blowing time according to the ratio of the upward blowing time to the downward blowing time obtained from the previous calculation to determine the upward blowing time and the downward blowing time for the next cycle. For example, if the ratio of the upward blowing time to the downward blowing time is 1:1.2 and the sum of the upward blowing time and the downward blowing time is 89 seconds, then the upward blowing time = 89×1 / (1 + 1.2) ≈ 40.5 seconds, and the downward blowing time = 89×1.2 / (1 + 1.2) ≈ 48.5 seconds.

[0087] Step 6. Implement control and monitor feedback

[0088] Execute the control instruction: Convert the calculated blowing time, upward blowing time, and downward blowing time for the next cycle into control instructions and send them to the actuator of the gasifier.

[0089] Monitor the feedback in real time: During the operation of the gasifier, continuously monitor the changes in parameters such as the upward temperature and the downward temperature in real time, and compare and analyze the actual operation data with the target values.

[0090] The grate temperature control module calculates the given value of the grate speed according to the set value and the actual value of the grate temperature, as well as the trend of the downward temperature. When both the downward temperature and the grate temperature are decreasing and the actual grate temperature is lower than the set temperature, increase the grate speed; otherwise, decrease the grate speed. When the downward temperature remains unchanged and the grate temperature changes, the grate speed remains unchanged.

[0091] The feed time control module calculates the given feed time within the selected gear time range according to the set value and actual value of the upstream temperature.

[0092] The specific working steps of the feed time control module are as follows:

[0093] Step A: Data acquisition and preparation

[0094] Temperature data acquisition: The actual value of the upstream temperature is collected in real time through a temperature sensor installed at a specific position in the gasifier.

[0095] Set value determination: A suitable set value of the upstream temperature is preset according to factors such as the production process requirements of the gasifier, the target gas quality, and the safe operating range of the equipment.

[0096] Gear time range determination: According to the operating experience and historical data of the gasifier, the feed time is divided into several gears, and the corresponding time range for each gear is determined.

[0097] Step B: Calculate the temperature deviation

[0098] Calculate the difference: Compare the actual value of the upstream temperature collected with the set value, and calculate the difference between the two, that is, temperature deviation = actual value of upstream temperature - set value of upstream temperature; then, according to the positive or negative of the calculated temperature deviation, judge whether the upstream temperature is higher or lower than the set value.

[0099] Step C: Select the fuzzy control algorithm

[0100] By establishing a fuzzy rule base, the input variables such as temperature deviation and deviation change rate are fuzzified, then reasoning is carried out according to the fuzzy rules, and finally the reasoning result is defuzzified to obtain the adjustment amount of the feed time.

[0101] Step D: Calculate the adjustment amount of the feed time

[0102] Combine with the current feed time: Combine the calculated adjustment amount of the feed time with the current feed time to obtain a new given feed time.

[0103] Limit within the gear time range: Compare the new given feed time with the selected gear time range to ensure that it is within this range. If it exceeds the gear time range, it needs to be adjusted to the nearest appropriate gear.

[0104] Step E: Implement control and feedback adjustment

[0105] Execute the feed time adjustment: Convert the determined feed time given value into a control signal and send it to the feeding equipment; During the feeding process, continuously monitor the change of the upward temperature in real time, compare the new actual temperature value with the set value, and calculate the temperature deviation again.

[0106] The carbon layer height control module establishes a carbon layer height prediction model by introducing the long short-term memory network algorithm to predict the change trend of the carbon layer height in advance;

[0107] The specific steps to establish a carbon layer height prediction model and predict the change trend of the carbon layer height in advance are as follows:

[0108] Step I. Data collection and preprocessing

[0109] Data collection: Collect historical data related to the carbon layer height. These data should include multiple features related to the change of the carbon layer height. The specific features include feed time, blowing time, up / down blowing steam volume, upward temperature, downward temperature, grate speed, and record the corresponding carbon layer height value at the same time;

[0110] Data preprocessing: Clean the collected data to remove noise, outliers, and missing values; Then normalize all feature data and carbon layer height data and scale them to a fixed range; Subsequently, divide the processed data into a training set, a validation set, and a test set according to a certain ratio, and the specific ratio is 7:2:1;

[0111] Step II. Build an LSTM model

[0112] Import necessary libraries: In Python, use the deep learning framework TensorFlow to build an LSTM model;

[0113] Step III. Model training

[0114] Convert the training set data into a format suitable for input to the LSTM model, that is, a three-dimensional tensor, and then use the sliding window method to generate input sequences and corresponding target values; Use the prepared training data to train the LSTM model, set appropriate training epochs and batch sizes. During the training process, the model will continuously adjust the parameters to minimize the loss function. At the same time, use the validation set data to validate the model and monitor the loss value of the validation set to prevent the model from overfitting;

[0115] Step IV. Model evaluation

[0116] After training is completed, use the test set data to evaluate the model and calculate the loss value and other evaluation metrics of the model on the test set;

[0117] Step V. Predict the change trend of the carbon layer height

[0118] Input real-time data: In practical applications, the feature data collected in real time is processed according to the same preprocessing method and then input into the trained LSTM model;

[0119] Make predictions: The model will output the predicted values of the carbon layer height. According to the prediction results of multiple consecutive time steps, analyze the change trend of the carbon layer height.

[0120] Intelligent steam flow regulation module: Use a flow sensor to monitor the steam flow in real time. Combine the load change of the gasifier and the pressure parameters in the furnace. Dynamically adjust the opening of the steam control valve through a fuzzy control algorithm to achieve precise control of the steam flow; among them, when the load of the gasifier increases, the intelligent steam flow regulation module automatically increases the steam flow; when the pressure in the furnace fluctuates abnormally, adjust the steam flow in time to maintain the stable operation of the reaction in the furnace.

[0121] Fault diagnosis and early warning module: Through real-time analysis of the data collected by the sensor and combined with the fault diagnosis model, judge whether the gasifier has a fault; once a fault is detected, immediately issue an alarm and notify the relevant operators by text message and email. At the same time, provide an analysis of the cause of the fault and suggestions for solutions;

[0122] Expert system optimization module: Contains common knowledge and experience of gasifier operation, and is dynamically updated in combination with real-time production data and the latest research results in the industry; conduct real-time evaluation and optimization of the control strategy of the system. When encountering complex working conditions or abnormal situations, provide reasonable control suggestions according to the knowledge and reasoning rules in the knowledge base to assist the self-optimization algorithm in making decisions. Among them, the expert system optimization module evaluates and optimizes the control strategy of the system according to different coal characteristics, and adjusts the blowing intensity and the ratio of the upblow / downblow steam volume.

[0123] The present invention focuses on the field of coal gasification technology. The core is to construct an automatic optimization intelligent system for gasifiers, aiming to solve the control problems of traditional intermittent gasifiers.

[0124] The process parameters of existing intermittent gasifiers are complex to control, with serious coupling between parameters. The controlled parameters are different in different gas production stages, relying on manual adjustment, which is greatly affected by the technology and habits of operators, and it is difficult to ensure the stable and efficient operation of the gasifier. The automatic optimization intelligent system of the present invention covers basic modules such as upper and lower row temperature control, feed time control, grate temperature control, and carbon layer height control, and also adds modules such as intelligent steam flow regulation, fault diagnosis and early warning, and expert system optimization. Each module operates in coordination, and through advanced algorithms such as self-optimization, fuzzy control, and predictive control, combined with the gasification reaction model in the furnace and the expert system, accurately regulate the key parameters of the gasifier.

[0125] In actual operation, the system calculates the adjustment amounts of various parameters based on the real-time collected data. For example, the upper and lower temperature control modules use the generalized predictive control model and algorithm (GPC), and calculate the blowing, upward blowing, and downward blowing times in combination with the lag of the temperature response in the gasification layer; the feeding time control module calculates the feeding time based on the set value and the actual value of the upper row temperature, etc. The system realizes automatic intelligent control, has a high fully automatic operation rate, and greatly reduces the labor intensity of operators and management costs. The temperature control accuracy is significantly improved, and the temperature fluctuations of the upper row, lower row, and grate are small; the material layer is stable and the control accuracy is high; the phenomenon of blowing over is basically eliminated, effectively improving the stability, efficiency, and safety of the gasifier operation, and creating good economic benefits.

[0126] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An automatic optimization intelligent system for a gasifier, characterized in that, It includes upper and lower temperature control modules, feeding time control module, grate temperature control module, carbon layer height control module, intelligent steam flow regulation module, fault diagnosis and early warning module, and expert system optimization module. The upper and lower temperature control modules calculate the given quantities of blowing, upblowing, and downblowing time for the next cycle based on the set values and actual values of the upper and lower temperatures. Then, based on the big data analysis of historical operation, the response lag of the gasification layer temperature is calibrated, and the blowing time for the next cycle is accurately calculated. Subsequently, according to the temperature difference between the upper and lower temperatures, the ratio of upblowing time to downblowing time for the next cycle is calculated using the GPC algorithm, and the upblowing and downblowing times are determined by combining the calculation formula for the sum of upblowing and downblowing times. The feeding time control module calculates the given quantity of feeding time within the selected gear time range based on the set value and actual value of the upper temperature. The grate temperature control module calculates the given quantity of grate speed based on the set value and actual value of the grate temperature, as well as the trend of the lower temperature. When both the lower temperature and the grate temperature are decreasing, and the actual grate temperature is lower than the set temperature, the grate speed is increased; otherwise, the grate speed is decreased. When the lower temperature remains unchanged and the grate temperature changes, the grate speed remains unchanged. The carbon layer height control module predicts the change trend of the carbon layer height in advance by establishing a carbon layer height prediction model. The intelligent steam flow regulation module uses a flow sensor to monitor the steam flow in real time, combines the load change of the gasifier and the furnace pressure parameters, and dynamically adjusts the opening of the steam control valve through a fuzzy control algorithm to achieve precise control of the steam flow. The fault diagnosis and early warning module analyzes the data collected by the sensors in real time, combines the fault diagnosis model, and determines whether the gasifier has a fault. Once a fault is detected, an alarm is immediately issued, and relevant operators are notified by text message and email. At the same time, the cause analysis of the fault and suggestions for solutions are provided. The expert system optimization module contains the operation knowledge and experience of the gasifier, is dynamically updated by combining real-time production data and industry achievements, and evaluates and optimizes the control strategy of the system in real time.

2. The automatic optimization intelligent system of a gasifier according to claim 1, characterized in that: The specific implementation steps for the upper and lower temperature control modules to calibrate the response lag of the gasification layer temperature based on the big data analysis of historical operation, accurately calculate the blowing time for the next cycle, then calculate the ratio of upblowing time to downblowing time for the next cycle according to the temperature difference between the upper and lower temperatures, and finally determine the upblowing and downblowing times by combining the calculation formula for the sum of upblowing and downblowing times are as follows: Step 1: Collect and organize the big data of historical operation Data acquisition: Use sensors installed at different positions of the gasifier to continuously collect various parameter data during the operation of the gasifier. The above parameter data includes but is not limited to upper temperature, lower temperature, blowing time, upblowing steam volume, downblowing steam volume, feeding time, grate speed, and carbon layer height. The acquisition frequency is set according to the actual situation, which can be once per minute or once every few seconds to ensure the timeliness and accuracy of the data. Data preprocessing: Clean and preprocess the collected raw data to remove noise, outliers, and missing values. Step 2: Analyze the lag of the calibrated gasification layer temperature response Establish a time series model: Arrange the historical operation data in chronological order, construct time series of the upward temperature, downward temperature, and blowing time parameters, and model the lag of the gasification layer temperature response through the analysis of these time series using the autoregressive integrated moving average model; Determine the lag time and response coefficient: Through the fitting and analysis of the time series model, determine the lag time when the gasification layer temperature starts to change significantly after the blowing time changes; at the same time, calculate the response coefficient of the gasification layer temperature under different blowing time changes, that is, the amplitude by which the gasification layer temperature rises or falls correspondingly when the blowing time increases or decreases by one unit; Step 3: Calculate the blowing time for the next cycle Set the target temperature: According to the production process requirements and actual operation conditions of the gasifier, set the target values of the upward temperature and downward temperature; Calculate the temperature deviation: Compare the actual values of the upward temperature and downward temperature collected in the current cycle with the target values to calculate the temperature deviation; Calculate the blowing time using the GPC algorithm: Use the temperature deviation, the lag time and response coefficient of the gasification layer temperature response, and the historical operation data as inputs, and use the generalized predictive control model and algorithm to calculate the given value of the blowing time for the next cycle; Step 4: Calculate the ratio of the upblow time to the downblow time for the next cycle Analyze the difference between the upward and downward temperatures: Calculate the difference between the upward temperature and the downward temperature in the current cycle, that is, temperature difference = upward temperature - downward temperature; according to the magnitude and sign of the temperature difference, judge the combustion state and heat distribution in the gasifier; Calculate the time ratio using the GPC algorithm: Use the difference between the upward and downward temperatures, the historical operation data, and the blowing time calculated previously as inputs, and use the GPC algorithm again to calculate the ratio of the upblow time to the downblow time for the next cycle. The GPC algorithm will optimize the allocation of the upblow time and the downblow time according to the change trend and target of the temperature difference to achieve the purpose of making the upward and downward temperatures more balanced; Step 5: Determine the upblow and downblow times for the next cycle Calculate the sum of the upblow and downblow times: According to the process requirements and operation cycle of the gasifier, determine a fixed time cycle. At the same time, subtract the fixed double-up time and the purging time, and then subtract the blowing time for the next cycle calculated previously to obtain the sum of the upblow and downblow times, where the double-up time is fixed at 10 s and the purging time is fixed at 1 s or 2 s; Determine the upblow and downblow times according to the time ratio: Allocate the sum of the upblow and downblow times according to the ratio of the upblow time to the downblow time calculated previously to determine the upblow time and the downblow time for the next cycle; Step 6: Implement control and monitor feedback Execute the control instruction: Convert the calculated blowing time, upblow time, and downblow time for the next cycle into control instructions and send them to the actuator of the gasifier; Monitor and feedback in real time: During the operation of the gasifier, continuously monitor the changes in parameters such as the upward temperature and downward temperature in real time, and compare and analyze the actual operation data with the target values.

3. The automatic optimization intelligent system of a gasifier according to claim 1, characterized in that: The specific working steps of the feed time control module are as follows: Step A: Data collection and preparation Temperature data acquisition: The actual value of the upward temperature is collected in real time through temperature sensors installed at specific positions of the gasifier. Setpoint determination: According to factors such as the production process requirements of the gasifier, the target gas quality, and the safe operating range of the equipment, a suitable setpoint for the upward temperature is preset in advance. Determination of gear time range: According to the operating experience and historical data of the gasifier, the feeding time is divided into several gears, and the corresponding time range for each gear is determined. Step B. Calculate the temperature deviation Calculate the difference: Compare the actual value of the upward temperature collected with the setpoint, and calculate the difference between the two, that is, temperature deviation = actual value of upward temperature - setpoint of upward temperature; then, according to the positive or negative of the calculated temperature deviation, determine whether the upward temperature is higher or lower than the setpoint. Step C. Select the fuzzy control algorithm By establishing a fuzzy rule base, the input variables such as temperature deviation and deviation change rate are fuzzified, then reasoning is carried out according to the fuzzy rules, and finally the reasoning result is defuzzified to obtain the adjustment amount of the feeding time. Step D. Calculate the adjustment amount of the feeding time Combine with the current feeding time: Combine the calculated adjustment amount of the feeding time with the current feeding time to obtain a new given value of the feeding time. Limit within the gear time range: Compare the new given value of the feeding time with the selected gear time range to ensure that it is within this range. If it exceeds the gear time range, it needs to be adjusted to the nearest appropriate gear. Step E. Implement control and feedback adjustment Execute the adjustment of the feeding time: Convert the determined given value of the feeding time into a control signal and send it to the feeding equipment; during the feeding process, continuously monitor the change of the upward temperature in real time, compare the new actual temperature value with the setpoint, and calculate the temperature deviation again.

4. An automatic optimization intelligent system for a gasifier according to claim 1, characterized in that: The long short-term memory network algorithm is introduced into the carbon layer height control module to establish a carbon layer height prediction model. The specific steps for predicting the change trend of the carbon layer height in advance are as follows: Step Ⅰ. Data collection and preprocessing Data collection: Collect historical data related to the carbon layer height. These data should include multiple features related to the change of the carbon layer height. The specific features include feeding time, blowing time, upblow / downblow steam volume, upward temperature, downward temperature, grate speed, and record the corresponding carbon layer height value at the same time. Data preprocessing: Clean the collected data to remove noise, outliers, and missing values; then normalize all feature data and carbon layer height data and scale them to a fixed range; subsequently, divide the processed data into a training set, a validation set, and a test set according to a certain ratio, and the specific ratio is 7:2:

1. Step Ⅱ. Construct the LSTM model Import necessary libraries: In Python, use the deep learning framework TensorFlow to construct the LSTM model. Step Ⅲ. Model training Convert the training set data into a format suitable for input to the LSTM model, i.e., a three-dimensional tensor, and then use the sliding window method to generate input sequences and corresponding target values; use the prepared training data to train the LSTM model, set appropriate number of training epochs and batch size. During the training process, the model will continuously adjust the parameters to minimize the loss function. At the same time, use the validation set data to validate the model and monitor the loss value of the validation set to prevent overfitting of the model. Step Ⅳ: Model Evaluation After training, use the test set data to evaluate the model and calculate the loss value and other evaluation metrics of the model on the test set. Step Ⅴ: Predict the Change Trend of the Carbon Layer Height Input real-time data: In practical applications, process the real-time collected feature data according to the same preprocessing method and then input it into the trained LSTM model. Make predictions: The model will output the predicted carbon layer height values. Analyze the change trend of the carbon layer height based on the prediction results of multiple consecutive time steps.

5. The automatic optimization intelligent system of a gasifier according to claim 1, characterized in that: When the load of the gasifier increases, the intelligent steam flow regulation module automatically increases the steam flow; when the pressure in the furnace fluctuates abnormally, it timely adjusts the steam flow to maintain the stable progress of the reaction in the furnace.

6. The automatic optimization intelligent system of a gasifier according to claim 1, characterized in that: The expert system optimization module evaluates and optimizes the control strategy of the system according to different coal characteristics, adjusts the blowing intensity and the ratio of the up-blow / down-blow steam volume. When encountering complex working conditions or abnormal situations, it provides control suggestions according to the knowledge and inference rules in the knowledge base to assist the self-optimization algorithm in making decisions.

Images