A small pyrolysis gasification incinerator combustion control method and system

Abstract

The application discloses a small pyrolysis gasification incinerator combustion control method and system, the method comprising: collecting relevant parameters of the pyrolysis gasification incineration process in each sampling period as a sample set; constructing a parameter prediction model based on an improved LSTM neural network; iteratively controlling the parameter prediction model using the sample set of each sampling period; and predicting the relevant control parameters using the parameter prediction model determined after the iterative control, to predict the optimal primary air door opening degree, secondary air door opening degree and air induction fan frequency under the real-time combustion state. The parameter prediction model constructed by the application can better obtain the dynamic relationship between the key parameters, and the model iterative control can dynamically adjust the model according to the working conditions, ensuring the reliability of the parameter prediction value, ensuring that the harmful substances are fully burned and decomposed, having high control precision and good long-term prediction ability, and realizing the automatic control of the garbage pyrolysis gasification incineration process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of waste incineration, in particular to a small-scale pyrolysis gasification incinerator combustion control method and system. BACKGROUND

[0002] The currently widely used is the waste pyrolysis gasification incineration technology. Waste pyrolysis gasification incineration refers to the process of decomposing unstable organic matter in waste into combustible gas, tar and residue under high temperature and oxygen deficiency. The combustible gas after pyrolysis gasification is fully combusted at high temperature and oxygen to realize waste harmless and resource utilization. When the temperature, pressure and oxygen content of the pyrolysis gasification incinerator are not within the optimal control range, not only the harmful substances in the waste cannot be fully decomposed, but also the working efficiency and safety of the whole system are affected. Therefore, the selection of the optimal control parameters of the pyrolysis gasification incinerator is of great significance to the stable operation of the whole pyrolysis furnace system.

[0003] The existing waste incineration control method includes dividing the control mode according to the error change rate of the control parameters obtained by the sensing terminal as the judgment condition, using the empirical value as the division limit, using the fuzzy control method in a certain range, using the traditional PID control method in a large or small range, and using the open loop control method in the case of over-limit, and the temperature output of the incineration temperature is realized by time-sharing control. However, the limitation of the finite level of fuzzy rules makes the fuzzy control prone to errors, and the time-sharing control method fails to establish an accurate mathematical model for the process parameters and control parameters of the waste incineration system, resulting in low control accuracy. Not only does it fail to solve the problem of insufficient decomposition of harmful substances, but also it reduces the working efficiency of the whole waste incineration system.

[0004] Therefore, how to solve the problems of low control accuracy, insufficient decomposition of harmful substances, low system working efficiency and safety during waste pyrolysis gasification incineration is a technical problem to be solved by those skilled in the art. SUMMARY

[0005] Therefore, the purpose of the present application is to provide a small-scale pyrolysis gasification incinerator combustion control method and system, which can ensure the reliability of the parameter prediction value, ensure that the harmful substances are fully burned and decomposed, have high control accuracy, good long-term prediction ability and strong anti-interference ability. The specific scheme is as follows:

[0006] A small-scale pyrolysis gasification incinerator combustion control method, comprising:

[0007] Collecting the related parameters of the pyrolysis gasification incineration process in each sampling period as a sample set;

[0008] Constructing a parameter prediction model based on an improved LSTM neural network;

[0009] iteratively control the parameter prediction model by using the sample set of each sampling period;

[0010] The parameter prediction model determined after the iterative control is used to predict the related control parameters, and the optimal primary air damper opening, secondary air damper opening and induced draft fan frequency under the real-time combustion state are predicted.

[0011] Preferably, in the small-scale pyrolysis gasification incinerator combustion control method provided by the embodiment of the present application, after collecting the related parameters of the pyrolysis gasification incineration process in each sampling period, the method further comprises:

[0012] The parameter data is normalized, and the primary air damper opening, secondary air damper opening and induced draft fan frequency in the current time parameter data are taken as the labels of the input data of the previous n seconds.

[0013] Preferably, in the small-scale pyrolysis gasification incinerator combustion control method provided by the embodiment of the present application, the collected related parameters include a combustion chamber temperature and pressure, a secondary combustion chamber temperature and pressure, a primary combustion chamber outlet flue gas oxygen content, a secondary combustion chamber outlet flue gas oxygen content, a primary air damper opening, a secondary air damper opening and an induced draft fan frequency; wherein,

[0014] The primary combustion chamber temperature and pressure, the secondary combustion chamber temperature and pressure, the primary combustion chamber outlet flue gas oxygen content, the secondary combustion chamber outlet flue gas oxygen content, the primary air damper opening, the secondary air damper opening and the induced draft fan frequency are taken as the input of the parameter prediction model; and the primary air damper opening, the secondary air damper opening and the induced draft fan frequency are taken as the output of the parameter prediction model.

[0015] Preferably, in the small-scale pyrolysis gasification incinerator combustion control method provided by the embodiment of the present application, the first layer of the parameter prediction model is a fully connected layer, which is used to obtain input data; the second layer is a one-dimensional convolution layer, which is used to extract the spatial features of the input data; the third layer and the fourth layer are LSTM layers, which are used to extract the time sequence features of the input data; the fifth layer is a Dropout layer; and the sixth layer is a fully connected layer, which is used for dimension transformation and outputs a label vector.

[0016] Preferably, in the small-scale pyrolysis gasification incinerator combustion control method provided by the embodiment of the present application, the method further comprises:

[0017] According to the prediction result, the primary air damper opening, the secondary air damper opening and the induced draft fan frequency are automatically adjusted;

[0018] If after the adjustment, the pressure and the oxygen content are within the optimal control range, and the primary combustion chamber and the secondary combustion chamber do not reach the optimal temperature control range, then the second burner is started, and the primary combustion chamber and the secondary combustion chamber are heated to the optimal temperature control range, and then the second burner is turned off.

[0019] If the first combustion chamber and the second combustion chamber still do not reach the optimal temperature control range after the second burner is turned on, the first burner is turned on for temperature rising until the first combustion chamber and the second combustion chamber reach the optimal temperature control range, and then the first burner and the second burner are turned off.

[0020] Preferably, in the small-scale pyrolysis gasification incinerator combustion control method provided by the embodiment of the present application, the method further comprises the following steps of:

[0021] If the first combustion chamber and the second combustion chamber still do not reach the optimal temperature control range after the first burner and the second burner are turned on, whether the garbage residue is sufficient is taken as a judgment condition for relevant control; if the garbage residue is sufficient, the automatic feeding device is controlled to automatically feed the first combustion chamber; and if the garbage residue is insufficient, manual control is performed.

[0022] Preferably, in the small-scale pyrolysis gasification incinerator combustion control method provided by the embodiment of the present application, the parameter prediction model is iteratively controlled, and the method comprises the following steps of:

[0023] A model update signal is automatically generated when the sampling amount in the current sampling period reaches a preset data amount;

[0024] The model update signal is received, and a sample set corresponding to the current sampling period is used to update the parameter prediction model;

[0025] Based on the actual observation value of the relevant parameter, the prediction value of the updated parameter prediction model in the current sampling period and the prediction value of the parameter prediction model in the last sampling period, a root mean square error is calculated to determine whether the updated parameter prediction model in the current sampling period or the parameter prediction model in the last sampling period is used for prediction.

[0026] Preferably, in the small-scale pyrolysis gasification incinerator combustion control method provided by the embodiment of the present application, the determination of whether the updated parameter prediction model in the current sampling period or the parameter prediction model in the last sampling period is used for prediction comprises the following steps of:

[0027] If the root mean square error of the prediction value of the updated parameter prediction model in the current sampling period and the actual observation value is less than the root mean square error of the prediction value of the parameter prediction model in the last sampling period and the actual observation value, it is determined that the updated parameter prediction model in the current sampling period is used for prediction;

[0028] If the root mean square error of the prediction value of the updated parameter prediction model in the current sampling period and the actual observation value is greater than or equal to the root mean square error of the prediction value of the parameter prediction model in the last sampling period and the actual observation value, it is determined that the parameter prediction model in the last sampling period is used for prediction.

[0029] The embodiment of the present application also provides a garbage pyrolysis gasification incineration system, comprising: a control center integrating a Windows system and a soft PLC, and a small pyrolysis gasification incineration furnace connected with the control center.

[0030] The Windows system is used for building a parameter prediction model based on an improved LSTM neural network and generating a corresponding SCL file.

[0031] The soft PLC is used for collecting relevant parameters of the pyrolysis gasification incineration process in each sampling period as a sample set, and is also used for calling the SCL file generated by the Windows system and converting into a model updating FB function block and a model prediction FB function block; wherein,

[0032] The model updating FB function block is used for performing iterative control on the parameter prediction model by using the sample set of each sampling period.

[0033] The model prediction FB function block is used for predicting relevant control parameters by using the parameter prediction model determined after the iterative control, and predicting optimal primary air door opening degree, secondary air door opening degree and induced draft fan frequency under a real-time combustion state.

[0034] Preferably, in the garbage pyrolysis gasification incineration system provided by the embodiment of the present application, the soft PLC is also used for automatically adjusting the primary air door opening degree, the secondary air door opening degree and the induced draft fan frequency according to the prediction result; if after the adjustment, the pressure and the oxygen content are within the optimal control range, and the primary combustion chamber and the secondary combustion chamber do not reach the optimal temperature control range, then the second burner is started, the primary combustion chamber and the secondary combustion chamber are heated to the optimal temperature control range, and then the second burner is stopped; if after the second burner is started, the primary combustion chamber and the secondary combustion chamber still do not reach the optimal temperature control range, then the first burner is started, and the heating is continued until the temperature of the primary combustion chamber and the secondary combustion chamber reaches the optimal temperature control range, and then the first burner and the second burner are stopped respectively.

[0035] As can be seen from the above technical solution, the small pyrolysis gasification incineration furnace combustion control method provided by the present application comprises: collecting relevant parameters of the pyrolysis gasification incineration process in each sampling period as a sample set; building a parameter prediction model based on an improved LSTM neural network; performing iterative control on the parameter prediction model by using the sample set of each sampling period; predicting relevant control parameters by using the parameter prediction model determined after the iterative control, and predicting optimal primary air door opening degree, secondary air door opening degree and induced draft fan frequency under a real-time combustion state.

[0036] The combustion control method for the small pyrolysis gasification incinerator provided by this invention can better obtain the dynamic relationship between key parameters by constructing a parameter prediction model based on an improved LSTM neural network. Furthermore, the model can be dynamically adjusted according to the operating conditions to ensure the reliability of the parameter prediction values ​​and guarantee that harmful substances are fully combusted and decomposed. It has high control accuracy, good long-term prediction ability and strong anti-interference ability.

[0037] Furthermore, this invention also provides a corresponding waste pyrolysis gasification incineration system for the combustion control method of a small pyrolysis gasification incinerator, further making the above method more practical, and the system has corresponding advantages. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0039] Figure 1 A flowchart of a combustion control method for a small pyrolysis gasification incinerator provided in an embodiment of the present invention;

[0040] Figure 2 This is a schematic diagram of the structure of a small-scale pyrolysis gasification incinerator provided in an embodiment of the present invention;

[0041] Figure 3 This is a flowchart of the parameter prediction model construction and training provided in an embodiment of the present invention;

[0042] Figure 4 The following is a flowchart of the iterative control of the parameter prediction model provided in an embodiment of the present invention;

[0043] Figure 5 The following is a flowchart of the automatic control process for real-time parameter prediction provided in an embodiment of the present invention;

[0044] Figure 6 This is a schematic diagram of the control center structure of a waste pyrolysis gasification incineration system provided in an embodiment of the present invention;

[0045] Figure 7 The flowchart shows the model prediction and update program call for the waste pyrolysis gasification incineration system provided in this embodiment of the invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] This invention provides a combustion control method for a small-scale pyrolysis gasification incinerator, such as... Figure 1 As shown, it includes the following steps:

[0048] S101. Collect relevant parameters of the pyrolysis gasification and incineration process within each sampling period as a sample set;

[0049] S102. Construct a parameter prediction model based on an improved LSTM neural network;

[0050] S103. Iteratively control the parameter prediction model using the sample set of each sampling period;

[0051] S104. The parameter prediction model determined by iterative control is used to predict the relevant control parameters, and the optimal primary damper opening, secondary damper opening and induced draft fan frequency under real-time combustion conditions are predicted.

[0052] It should be noted that the control center for implementing the combustion control method of the small pyrolysis gasification incinerator provided by the present invention is an integration of Windows system and soft PLC. The control center is set in the waste pyrolysis gasification incineration system, which can improve the timeliness of data transmission and speed up the response.

[0053] In the combustion control method for a small pyrolysis gasification incinerator provided in this embodiment of the invention, by collecting a large amount of data under normal combustion conditions as a sample set, and constructing a corresponding parameter prediction model based on an improved LSTM neural network, the dynamic relationship between key parameters can be better obtained. Furthermore, during online real-time control, iterative control of the model allows for dynamic adjustment based on operating conditions, dynamically matching real-time conditions, and completing data prediction. This provides the optimal primary damper opening, secondary damper opening, and induced draft fan frequency under real-time combustion conditions, enabling the system to operate automatically and stably within the optimal temperature, pressure, and oxygen content control range. This ensures that harmful substances are fully combusted and decomposed, exhibiting high control accuracy, good long-term prediction capability, and strong anti-interference ability. It can effectively describe the dynamic characteristics of the system and has high utilization value. This invention solves the problems of nonlinear and strongly coupled control between system process parameters and control parameters, as well as the challenges of variable real-time operating conditions, realizing automated control of combustion in a small pyrolysis gasification incinerator.

[0054] It is necessary to understand that, such as Figure 2As shown, the main structure of the small pyrolysis gasification incinerator of the present invention includes: a primary combustion chamber 1, a primary combustion chamber temperature sensor 2, a primary combustion chamber pressure sensor 3, a primary air damper 4, a secondary air damper 5, a secondary combustion chamber temperature sensor 6, a secondary combustion chamber pressure sensor 7, a secondary combustion chamber 8, a waste heat utilization module 9, a tail gas treatment module 10, an induced draft fan 11, a secondary combustion chamber outlet flue gas oxygen sensor 12, a blower 13, a No. 2 burner 14, a primary combustion chamber outlet flue gas oxygen sensor 15, a No. 1 burner 16, and an automatic feeding system 17.

[0055] The working process of the above-mentioned pyrolysis gasification incinerator is roughly as follows: After the automatic feeding system 17 feeds the waste into the primary combustion chamber 1, the waste undergoes high-temperature, oxygen-deficient pyrolysis reduction combustion in the primary combustion chamber 1. The combustible gas produced by pyrolysis gasification is discharged into the secondary combustion chamber 8, while the slag is discharged from the bottom of the furnace. The combustible gas produced by pyrolysis gasification undergoes sufficient high-temperature, oxygen-excessive combustion in the secondary combustion chamber 8, and harmful substances are fully decomposed. Subsequently, it is discharged into the waste heat utilization module and the tail gas treatment module for further processing.

[0056] In specific implementation, in the combustion control method for the small pyrolysis gasification incinerator provided in the embodiments of the present invention, the relevant parameters collected in step S101 may include the temperature and pressure of the primary combustion chamber 1, the temperature and pressure of the secondary combustion chamber 8, the oxygen content of the flue gas at the outlet of the primary combustion chamber 1, the oxygen content of the flue gas at the outlet of the secondary combustion chamber 8, the opening degree of the primary air damper 4, the opening degree of the secondary air damper 5, and the frequency of the induced draft fan 11. When executing step S102, the temperature and pressure of the primary combustion chamber 1, the temperature and pressure of the secondary combustion chamber 8, the oxygen content of the flue gas at the outlet of the primary combustion chamber 1, the oxygen content of the flue gas at the outlet of the secondary combustion chamber 8, the opening degree of the primary air damper 4, the opening degree of the secondary air damper 5, and the frequency of the induced draft fan 11 are used as inputs to the parameter prediction model; the opening degree of the primary air damper 4, the opening degree of the secondary air damper 5, and the frequency of the induced draft fan 11 are used as outputs to the parameter prediction model. This invention selects the optimal prediction interval length, sequence length, and prediction step size to construct the model. The constructed parameter prediction model can extract the data features of key parameters, obtain the dynamic relationship between each key parameter, and then establish the dynamic relationship between the temperature and pressure of the primary combustion chamber 1, the temperature and pressure of the secondary combustion chamber 8, the oxygen content of the flue gas at the outlet of the primary combustion chamber 1, the oxygen content of the flue gas at the outlet of the secondary combustion chamber 8, the opening degree of the primary air damper 4, the opening degree of the secondary air damper 5, and the frequency of the induced draft fan 11.

[0057] In specific implementation, in the combustion control method of the small pyrolysis gasification incinerator provided in the embodiments of the present invention, after collecting the relevant parameters of the pyrolysis gasification incineration process in each sampling cycle, the method further includes: normalizing the parameter data, and using the primary damper opening, secondary damper opening and induced draft fan frequency in the parameter data at the current moment as the labels of the input data for the previous n seconds; where n is a positive integer.

[0058] like Figure 3As shown, data acquisition begins first, followed by normalization, transformation, and segmentation. The process and control parameters of the waste pyrolysis gasification incineration system are collected as identification data. When the sampled data volume reaches the required modeling volume P, the system automatically starts building a parameter prediction model using the current P parameter data as the sample set. Simultaneously, the sampled data volume is reset to zero and counting restarts, constituting one sampling cycle, which is repeated continuously. Afterwards, for easier calculation, the data needs to be normalized, transforming all data to the range [0,1]. The transformation formula is as follows:

[0059] ;

[0060] Where y is the normalized data, and x is the original data. max x min These are the maximum and minimum values ​​of the original dataset, respectively.

[0061] During the data transformation process, the time series problem is transformed into a supervised learning problem. The opening degree of the primary air damper 4, the opening degree of the secondary air damper 5, and the frequency of the induced draft fan 11 that need to be predicted are used as labels. At the same time, the time step is set to n. The opening degree of the primary air damper 4, the opening degree of the secondary air damper 5, and the frequency of the induced draft fan 11 at the current time (t) are the labels of the input data in the previous n seconds, i.e., (t-1)...(tn).

[0062] Next, the sample set is split into a training set, a validation set, and a test set. The training set is used to train the model after the network structure is determined. The validation set is used to verify the accuracy of the model after training and to prevent overfitting. The test set is used to test the accuracy of the model after testing.

[0063] In specific implementation, in the combustion control method for a small-scale pyrolysis gasification incinerator provided in this embodiment of the invention, the first layer of the parameter prediction model is a fully connected layer used to acquire input data; the second layer is a one-dimensional convolutional layer used to extract spatial features of the input data; the third and fourth layers are LSTM layers used to extract temporal features of the input data; the fifth layer is a Dropout layer used to temporarily discard some neural network units from the network during the training process of the deep learning model, with a probability of randomly updating parameters set to 0.5, reducing the possibility of data overfitting; the sixth layer is a fully connected layer used to perform dimensionality transformation and output label vectors. This parameter prediction model is optimized based on the traditional LSTM neural network structure, extracting the spatial and temporal features of key parameters of the waste pyrolysis gasification incineration system, and can better obtain the dynamic relationship between various key parameters. In practical applications, this parameter prediction model can be a Sequential model built in Keras.

[0064] During the training of the parameter prediction model, this invention sets the batch size to 64 bytes and the epoch iteration count to 100. The training set data is converted to the standard input data format and input into the pre-built model for training. The number of iterations is adjusted promptly based on the training result curve, and the optimal number of iterations is selected to complete the model training. After training, the validation set data is converted to the standard input data format and input into the model to verify its performance. Based on the validation results, parameters such as the number of neurons in each layer, the objective function, and the optimizer are adjusted. The validated model and weights are then saved as an HDF5 file.

[0065] The saved HDF5 file can then be called and loaded. The test set data is converted to the standard format for model input data and input into the model to test its performance. The model's accuracy is then assessed to determine if it meets the control accuracy target. If not, the parameters are readjusted, and the model is repeatedly trained, validated, and tested until it achieves the control accuracy target. Once the trained and tested model meets the control accuracy target, the model and weights that meet the accuracy requirements are saved as an HDF5 file, completing the model building process.

[0066] It should be noted that, due to the complex and variable operating states of the waste pyrolysis gasification incineration system, the parameter prediction model for the current operating state is not applicable to the entire pyrolysis gasification incineration process. Therefore, this invention performs iterative control on the parameter prediction model during step S103. In specific implementation, step S103 iteratively controls the parameter prediction model, and the specific steps may include: automatically generating a model update signal when the sampling amount in the current sampling period reaches a preset data amount P; receiving the model update signal and updating the parameter prediction model using the sample set corresponding to the current sampling period; calculating the root mean square error based on the actual observed values ​​of relevant parameters, the predicted values ​​of the currently updated parameter prediction model, and the predicted values ​​of the parameter prediction model in the previous sampling period, and determining whether to use the currently updated parameter prediction model or the parameter prediction model in the previous sampling period for prediction.

[0067] Specifically, such as Figure 4 As shown, after the sampling amount in the current sampling period reaches the modeling data amount P, the parameter prediction model is rebuilt using the latest sample set, following the model building steps described above. The root mean square error is calculated based on the actual observed values ​​of the relevant parameters in step M, the predicted values ​​of the current updated parameter prediction model in step M, and the predicted values ​​of the parameter prediction model in step M of the previous sampling period. This determines whether to use the current updated parameter prediction model for parameter prediction.

[0068] In specific implementation, in the combustion control method for the small pyrolysis gasification incinerator provided in the embodiments of the present invention, the step of determining whether to use the currently updated parameter prediction model or the parameter prediction model of the previous sampling period for prediction may include:

[0069] If the root mean square error between the predicted value and the actual observation value of the updated parameter prediction model is less than the root mean square error between the predicted value and the actual observation value of the parameter prediction model in the previous sampling period, then the updated parameter prediction model will be used for prediction.

[0070] If the root mean square error between the predicted value and the actual observation value of the updated parameter prediction model is greater than or equal to the root mean square error between the predicted value and the actual observation value of the parameter prediction model of the previous sampling period, then the parameter prediction model of the previous sampling period will be used for prediction.

[0071] In other words, such as Figure 4 As shown, when the root mean square error (RMSE) between the predicted and actual observations of the currently updated parameter prediction model is smaller than that of the parameter prediction model in the previous sampling period, the currently updated parameter prediction model is confirmed to be used. A loading signal for the currently updated parameter prediction model is generated, and the loading and parameter prediction of the currently updated parameter prediction model are automatically completed, realizing the model iteration function and dynamically matching the real-time operating conditions of the pyrolysis gasification incineration system. When the RMSE between the currently updated parameter prediction model and the actual observations is larger than or equal to that of the parameter prediction model in the previous sampling period, the currently updated parameter prediction model is not used for prediction, and the system automatically loads the parameter prediction model from the previous sampling period to complete the parameter prediction function.

[0072] like Figure 5 As shown, the system collects real-time data on the temperature and pressure of combustion chamber 1, the temperature and pressure of combustion chamber 8, the oxygen content of the flue gas at the outlets of combustion chamber 1 and 8, the opening degree of primary damper 4, the opening degree of secondary damper 5, and the frequency of induced draft fan 11. This data is then processed and used as input to a parameter prediction model to predict the optimal values ​​for the opening degree of primary damper 4, the opening degree of secondary damper 5, and the frequency of induced draft fan 11 in the next control cycle. Based on the root mean square error calculation, it determines whether to load the currently updated parameter model, automatically predicting and adjusting the opening degree of primary damper 4, the opening degree of secondary damper 5, and the frequency of induced draft fan 11, ensuring the entire system operates stably within the optimal temperature, pressure, and oxygen content control range.

[0073] Furthermore, in specific implementations, in the combustion control method for the small-scale pyrolysis gasification incinerator provided in the embodiments of the present invention, such as... Figure 5 As shown, it may also include: automatically adjusting the opening of the primary damper, the opening of the secondary damper, and the frequency of the induced draft fan based on the prediction results.

[0074] If, after adjustment, the pressure and oxygen content are within the optimal control range, but the primary combustion chamber 1 and the secondary combustion chamber 8 have not reached the optimal temperature control range, then start burner #2 14, heat the primary combustion chamber 1 and the secondary combustion chamber 8 to the optimal temperature control range, and then turn off burner #2 14.

[0075] If the primary combustion chamber 1 and the secondary combustion chamber 8 still do not reach the optimal temperature control range after starting burner #2 14, then start burner #1 16 and continue to heat up until the temperature of primary combustion chamber 1 and the secondary combustion chamber 8 reach the optimal temperature control range, then turn off burner #1 16 and burner #2 14 respectively.

[0076] If the primary combustion chamber 1 and secondary combustion chamber 8 still do not reach the optimal temperature control range after starting burner 16 and burner 24, the control will be based on whether there is enough waste residue. If there is enough waste residue, the automatic feeding device 17 will automatically replenish the primary combustion chamber. If there is insufficient waste residue, the system will switch to manual control mode, and the fans, burners and other system equipment will be manually controlled, and the system will be manually shut down.

[0077] Based on the same inventive concept, this invention also provides a waste pyrolysis gasification incineration system. Since the principle of this waste pyrolysis gasification incineration system is similar to the aforementioned combustion control method for a small pyrolysis gasification incinerator, the implementation of this waste pyrolysis gasification incineration system can refer to the implementation of the combustion control method for a small pyrolysis gasification incinerator. Repeated details will not be repeated.

[0078] In specific implementation, the waste pyrolysis gasification incineration system provided in this embodiment of the invention, such as... Figure 6 As shown, it specifically includes: a control center that integrates a Windows system and a soft PLC, and a small pyrolysis gasification incinerator connected to the control center;

[0079] Windows system, used to build parameter prediction models based on improved LSTM neural networks and generate corresponding SCL files;

[0080] The soft PLC is used to collect relevant parameters of the pyrolysis gasification and incineration process as a sample set within each sampling cycle. It is also used to call SCL files generated by the Windows system and convert them into model update (FB) and model prediction (FB) function blocks.

[0081] The Model Update FB function block is used to iteratively control the parameter prediction model using the sample set of each sampling period;

[0082] The Model Prediction (FB) function block is used to predict relevant control parameters using a parameter prediction model determined after iterative control, predicting the optimal primary damper opening, secondary damper opening, and induced draft fan frequency under real-time combustion conditions.

[0083] In the waste pyrolysis gasification incineration system provided in this embodiment of the invention, a Windows system integrated with a soft PLC is used to replace the traditional long-distance communication between the host computer and the traditional PLC. This solves the problems of high resource consumption and high load rate of traditional PLCs, improves the timeliness of data transmission, and speeds up the system response. It can ensure the stable operation of the entire pyrolysis gasification incineration system. Furthermore, through the interaction of the model update FB function block and the model prediction FB function block, the dynamic relationship between various key parameters can be better obtained. The parameter prediction model can be iteratively controlled online, and the real-time operating conditions of the pyrolysis gasification incineration system can be dynamically matched to achieve full automation of the waste pyrolysis gasification incineration process. The key parameters of the system can be predicted in real time and adjusted to the optimal control range, improving the system control accuracy, ensuring the full decomposition of harmful substances, and realizing the stable, efficient and automatic operation of the waste pyrolysis gasification incineration system.

[0084] Specifically, such as Figure 7 As shown, the Windows system is used to write model prediction and model update programs, build a parameter prediction model based on an improved LSTM neural network, and generate corresponding DLL and SCL files. The soft PLC collects real-time data of key system parameters through the DI / DO and AI / AO modules, and transmits it to the Windows system via internal instructions. In addition, the PLC calls the SCL file generated by the Windows system to generate and call the model prediction FB function block and the model update FB function block to complete the loading and running of the control program. This realizes the PLC's functions of calculation, control, storage, and programming, and transmits data with the Windows system through internal instructions to achieve automated control of the entire waste pyrolysis gasification incineration system.

[0085] It should be noted that the waste pyrolysis gasification incineration system of the present invention has two control modes: manual control and automatic control. When the system is in manual control mode, relevant system parameters can be manually adjusted via buttons to control the operation of the waste pyrolysis gasification incineration system. When the system enters automatic control mode, the system automatically collects relevant parameters for prediction and automatically adjusts the system parameters to the optimal control range based on the prediction results to ensure the stable operation of the entire system. The manual control mode and the automatic control mode can be switched manually.

[0086] During initial feeding, the system is in manual control mode. The startup procedure includes: manually turning on the main power supply to the pyrolysis gasification incinerator, starting the blower 13 and induced draft fan 11, and activating burners 16 and 2 to preheat primary combustion chamber 1 and secondary combustion chamber 8; after preheating, the automatic feeding system begins feeding material into primary combustion chamber 1. At this time, the entire pyrolysis gasification incineration system maintains a slightly negative pressure state; the primary air damper 4 and secondary air damper 5 control the oxygen content in the two combustion chambers by adjusting the air intake of primary combustion chamber 1 and secondary combustion chamber 8, ensuring that primary combustion chamber 1 is in a sub-oxygen state and secondary combustion chamber 8 is in a super-oxygen state. When the temperature of primary combustion chamber 1 reaches the optimal control requirement, the system automatically enters automatic control mode.

[0087] Once the system automatically enters automatic control mode, burners #1 (16) and #2 (14) automatically shut down. During each sampling cycle, the system collects real-time data of relevant parameters, inputs it into the established parameter prediction model, completes real-time prediction, and adjusts relevant system parameters in real-time based on the prediction results to ensure stable operation of the pyrolysis gasification incineration system within the optimal control range.

[0088] If the temperature of combustion chamber 1 and combustion chamber 8 cannot reach the optimal control range and there is insufficient waste material when the system is in automatic control mode, the system will automatically switch to manual control mode, and the blowers, burners and other system equipment will be manually controlled, and the system will be manually shut down.

[0089] In this invention, when the opening degree of the primary air damper 4, the opening degree of the secondary air damper 5, and the frequency of the induced draft fan 11 reach the optimal control range, the temperature and pressure of the primary combustion chamber 1, the temperature and pressure of the secondary combustion chamber 8, the oxygen content of the flue gas at the outlet of the primary combustion chamber 1, and the oxygen content of the flue gas at the outlet of the secondary combustion chamber 8 are also within the optimal control range. At this time, the waste put into the system can be fully pyrolyzed, gasified, and incinerated, and harmful substances can be fully decomposed, so that the entire system reaches a stable and efficient operating state.

[0090] In specific implementation, in the waste pyrolysis gasification incineration system provided in this embodiment of the invention, the soft PLC can also be used to automatically adjust the opening of the primary air damper, the opening of the secondary air damper, and the frequency of the induced draft fan based on the prediction results. If, after adjustment, the pressure and oxygen content are within the optimal control range, but the primary combustion chamber 1 and the secondary combustion chamber 8 have not reached the optimal temperature control range, then burner #2 14 is turned on to raise the temperature of primary combustion chamber 1 and the secondary combustion chamber 8 to the optimal temperature control range, and then burner #2 14 is turned off. If, after turning on burner #2 14, primary combustion chamber 1 and the secondary combustion chamber 8 still have not reached the optimal temperature control range, then burner #1 is turned on. Burner 16 continuously heats up until the temperatures of the primary combustion chamber 1 and the secondary combustion chamber 8 reach the optimal temperature control range, then shuts down burner 16 and burner 2 respectively. It can also be used to determine whether there is enough waste residue if the primary combustion chamber 1 and the secondary combustion chamber 8 still do not reach the optimal temperature control range after turning on burner 16 and burner 2. If there is enough waste residue, the automatic feeding device 17 is controlled to automatically replenish the primary combustion chamber. If there is insufficient waste residue, the system switches to manual control mode, and the fans, burners and other system equipment are manually controlled, and the system is manually shut down.

[0091] For more detailed information on the working process of each of the above modules, please refer to the relevant content disclosed in the foregoing embodiments, which will not be repeated here.

[0092] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0093] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0094] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0095] In summary, the combustion control method for a small-scale pyrolysis gasification incinerator provided by this invention includes: collecting relevant parameters of the pyrolysis gasification incineration process as a sample set within each sampling period; constructing a parameter prediction model based on an improved LSTM neural network; iteratively controlling the parameter prediction model using the sample set of each sampling period; and predicting relevant control parameters using the parameter prediction model determined after iterative control, thereby predicting the optimal primary damper opening, secondary damper opening, and induced draft fan frequency under real-time combustion conditions. This method, by constructing a parameter prediction model based on an improved LSTM neural network, can better obtain the dynamic relationships between key parameters, and the iterative control of the model can be dynamically adjusted according to the operating conditions, ensuring the reliability of the parameter prediction values ​​and guaranteeing the complete combustion and decomposition of harmful substances. It has high control accuracy, good long-term prediction capability, and strong anti-interference capability. Furthermore, this invention also provides a corresponding waste pyrolysis gasification incineration system for the combustion control method of the small-scale pyrolysis gasification incinerator, further enhancing the practicality of the above method. This system has corresponding advantages.

[0096] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0097] The combustion control method and system for a small pyrolysis gasification incinerator provided by the present invention have been described in detail above. Specific examples have been used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of ​​the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A small pyrolysis gasification incinerator combustion control method characterized by, The method comprises the following steps: Collecting relevant parameters of the pyrolysis gasification incineration process in each sampling period as a sample set; Selecting the optimal prediction interval length, sequence length and prediction step to construct a parameter prediction model based on an improved LSTM neural network; the parameter prediction model is used to extract the data features of key parameters, obtain the dynamic relationship between the key parameters, and establish a dynamic relationship between the primary combustion chamber temperature and pressure, the secondary combustion chamber temperature and pressure, the primary combustion chamber outlet flue gas oxygen content, the secondary combustion chamber outlet flue gas oxygen content, the primary air damper opening, the secondary air damper opening and the induced draft fan frequency; Using the sample set of each sampling period to perform iterative control on the parameter prediction model, including automatically generating a model update signal when the sampling amount in the current sampling period reaches the preset data amount; Receiving the model update signal, updating the parameter prediction model using the sample set corresponding to the current sampling period; based on the actual observation value of the relevant parameters, the prediction value of the current updated parameter prediction model and the prediction value of the parameter prediction model of the last sampling period, performing root mean square error calculation to determine whether to use the current updated parameter prediction model or the parameter prediction model of the last sampling period for prediction; Using the parameter prediction model determined after iterative control to predict the relevant control parameters, and predicting the optimal primary air damper opening, secondary air damper opening and induced draft fan frequency under the real-time combustion state.

2. The small-scale pyrolysis gasification incinerator combustion control method according to claim 1, characterized by, After collecting the relevant parameters of the pyrolysis gasification incineration process in each sampling period, the method further comprises the following steps: Normalizing the parameter data, and taking the primary air damper opening, the secondary air damper opening and the induced draft fan frequency in the current time parameter data as the labels of the input data of the previous n seconds.

3. The small-scale pyrolysis gasification incinerator combustion control method according to claim 2, characterized by, The relevant parameters collected include the primary combustion chamber temperature and pressure, the secondary combustion chamber temperature and pressure, the primary combustion chamber outlet flue gas oxygen content, the secondary combustion chamber outlet flue gas oxygen content, the primary air damper opening, the secondary air damper opening and the induced draft fan frequency; wherein, The primary combustion chamber temperature and pressure, the secondary combustion chamber temperature and pressure, the primary combustion chamber outlet flue gas oxygen content, the secondary combustion chamber outlet flue gas oxygen content, the primary air damper opening, the secondary air damper opening and the induced draft fan frequency are used as the input of the parameter prediction model; the primary air damper opening, the secondary air damper opening and the induced draft fan frequency are used as the output of the parameter prediction model.

4. The small-scale pyrolysis gasification incinerator combustion control method according to claim 3, characterized by, The first layer of the parameter prediction model is a fully connected layer for obtaining input data; the second layer is a one-dimensional convolution layer for extracting spatial features of the input data; the third layer and the fourth layer are LSTM layers for extracting time sequence features of the input data; the fifth layer is a Dropout layer; and the sixth layer is a fully connected layer for dimension transformation and outputting a label vector.

5. The small-scale pyrolysis gasification incinerator combustion control method according to claim 4, characterized by, The method further comprises the following steps: According to the prediction result, automatically adjusting the primary air damper opening, the secondary air damper opening and the induced draft fan frequency; If after adjustment, the pressure and the oxygen content are within the optimal control range, but the primary combustion chamber and the secondary combustion chamber do not reach the optimal temperature control range, then the second burner is started, and the primary combustion chamber and the secondary combustion chamber are heated to the optimal temperature control range, and then the second burner is stopped. If the first combustion chamber and the second combustion chamber still do not reach the optimal temperature control range after the second burner is started, the first burner is started to continue to increase the temperature until the first combustion chamber and the second combustion chamber reach the optimal temperature control range, and then the first burner and the second burner are turned off.

6. The small-scale pyrolysis gasification incinerator combustion control method according to claim 5, characterized by, Also include: If the first combustion chamber and the second combustion chamber still do not reach the optimal temperature control range after the first burner and the second burner are started, whether the garbage residue is sufficient is used as a judgment condition for related control; if the garbage residue is sufficient, the automatic feeding device is controlled to automatically supplement the first combustion chamber; if the garbage residue is insufficient, manual control is performed.

7. The small-scale pyrolysis gasification incinerator combustion control method according to claim 1, characterized by, Determine to use the current updated parameter prediction model or the parameter prediction model of the last sampling period for prediction, including: If the root mean square error of the predicted value and the actual observation value of the current updated parameter prediction model is less than the root mean square error of the predicted value and the actual observation value of the parameter prediction model of the last sampling period, it is determined to use the current updated parameter prediction model for prediction; If the root mean square error of the predicted value and the actual observation value of the current updated parameter prediction model is greater than or equal to the root mean square error of the predicted value and the actual observation value of the parameter prediction model of the last sampling period, it is determined to use the parameter prediction model of the last sampling period for prediction.

8. A waste pyrolysis gasification incineration system, characterized by, Including: A control center integrating a Windows system and a soft PLC, and a small-scale pyrolysis gasification incinerator connected with the control center; The Windows system is used to select the optimal prediction interval length, sequence length and prediction step distance to build a parameter prediction model based on an improved LSTM neural network, and generate a corresponding SCL file; the parameter prediction model is used to extract the data features of key parameters, obtain the dynamic relationship between each key parameter, and establish the dynamic relationship of the first combustion chamber temperature and pressure, the second combustion chamber temperature and pressure, the first combustion chamber outlet flue gas oxygen content, the second combustion chamber outlet flue gas oxygen content, the primary air door opening, the secondary air door opening and the induced draft fan frequency; The soft PLC is used to collect related parameters of the pyrolysis gasification incineration process in each sampling period as a sample set, and is also used to call the SCL file generated by the Windows system to convert into a model update FB function block and a model prediction FB function block; wherein, The model update FB function block is used to iteratively control the parameter prediction model using the sample set of each sampling period, including: generating a model update signal automatically when the sampling amount in the current sampling period reaches a preset data amount; receiving the model update signal, updating the parameter prediction model using the sample set corresponding to the current sampling period; calculating the root mean square error based on the actual observation value of the related parameters, the predicted value of the current updated parameter prediction model and the predicted value of the parameter prediction model of the last sampling period, and determining to use the current updated parameter prediction model or the parameter prediction model of the last sampling period for prediction. The model prediction FB function block is configured to predict the relevant control parameters by using the parameters determined after the iterative control, and to predict the optimal primary air damper opening, the secondary air damper opening and the induced draft fan frequency under the real-time combustion state.

9. The waste pyrolysis gasification incineration system according to claim 8, characterized by, The soft PLC is further configured to automatically adjust the primary air damper opening, the secondary air damper opening and the induced draft fan frequency according to the prediction result; if the pressure and the oxygen content are within the optimal control range after the adjustment, and the primary combustion chamber and the secondary combustion chamber do not reach the optimal temperature control range, then the second burner is started, the primary combustion chamber and the secondary combustion chamber are heated to the optimal temperature control range, and then the second burner is stopped; if the primary combustion chamber and the secondary combustion chamber still do not reach the optimal temperature control range after the second burner is started, then the first burner is started, and the heating is continued until the temperature of the primary combustion chamber and the secondary combustion chamber reaches the optimal temperature control range, and then the first burner and the second burner are stopped, respectively.

Images