A method, system, device and medium for monitoring and warning of runaway temperature in a DMTO reactor

Abstract

The present invention relates to the technical field of runaway temperature monitoring and early warning for DMTO reactors. Based on Bayesian regression to reconstruct the principal component monitoring model, specifically, it is a method, system, device and medium for runaway temperature monitoring and early warning in DMTO reactors. After the monitoring data is reconstructed by the Bayesian regression algorithm, it is input into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model. The first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model output the on-line monitoring value T1 2 and the on-line monitoring value T2 2 value. By comparing a continuous number of on-line monitoring values T1 2 and the limit value T1 2 lim , a continuous number of on-line monitoring values T2 2 and the limit value T2 2 lim , thereby monitoring the temperature in the DMTO reactor, realizing early warning of runaway temperature, analyzing faults, assisting in the production decision-making of the DMTO process, and improving the detection efficiency of the runaway temperature problem of the DMTO reactor.

Description

Technical Field

[0001] The present invention relates to the technical field of runaway temperature monitoring and early warning for DMTO reactors, and reconstructs a principal component monitoring model based on Bayesian regression. Specifically, it is a method, system, device and medium for monitoring and early warning of runaway temperature in DMTO reactors. Background Technique

[0002] Methanol to olefins (DMTO) is an important C1 chemical process. Using methanol synthesized from coal as raw material, through a fluidized bed reaction form similar to a catalytic cracking unit, it produces light olefins. As a new way to synthesize light olefins from methanol, DMTO has great practical significance for balancing the supply and demand of light olefins and promoting national energy security.

[0003] As one of the core equipment of the DMTO process, controlling the internal temperature of the DMTO reactor to be stable between 460 and 480 °C is beneficial to ensure catalyst activity, olefin conversion rate, light olefin yield, etc. Due to reasons such as a decrease in the recycle gas volume, an excessive radial temperature difference in the bed layer, changes in raw material properties and components, and instrument failures, the temperature inside the DMTO reactor rises rapidly in a short time, which will cause changes in the internal structure of the reactor, a decrease in catalyst activity, resulting in a decrease in product yield and unqualified quality. In severe cases, it will cause the reactor to explode, forming a major safety accident. The existing methods to prevent reactor runaway temperature mainly include setting up a cold gas bypass and modifying the reactor to add runaway temperature control equipment, etc. These methods all require certain modifications to the existing DMTO reactor. Summary of the Invention

[0004] Aiming at the problem of runaway temperature in DMTO reactors in the prior art, the present invention provides a method, system, device and medium for monitoring and early warning of runaway temperature in DMTO reactors. By modeling based on historical data, real-time data monitoring, early warning, and analyzing faults, it ensures the stable operation of the DMTO reactor.

[0005] The present invention is realized through the following technical solutions:

[0006] A method for monitoring and early warning of runaway temperature in a DMTO reactor includes the following steps:

[0007] Obtain a number of on-line monitoring parameters in the DMTO process, and after preprocessing the number of on-line monitoring parameters, input them into a number of Bayesian regression models respectively. The number of Bayesian regression models correspondingly output prediction values of the number of on-line monitoring parameters, and determine a number of on-line monitoring deviation values by corresponding the number of on-line monitoring parameters with the prediction values of the number of on-line monitoring parameters;

[0008] Input the number of on-line monitoring deviation values into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor respectively, and correspondingly output the on-line monitoring value T12 and the online monitoring value T2 2 ;

[0009] Obtain historical monitoring parameters, preprocess the historical monitoring parameters to train several Bayesian regression models, reconstruct variables using the predicted values and historical monitoring parameter values of several Bayesian regression models, and train the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor for the reconstructed variables, and output to obtain the limit value T1 2 lim and the limit value T2 2 lim ;

[0010] Compare several consecutive online monitoring values T1 2 with the limit value T1 2 lim and several consecutive online monitoring values T2 2 with the limit value T1 2 lim respectively. When several consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , no prompt is made; when several consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and several consecutive online monitoring values T2 2 are less than the limit value T2 2 lim , the warning state is displayed; when several consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and several consecutive online monitoring values T2 2 are greater than the limit value T2 2 lim , the alarm state is displayed, and the cause of the failure is determined through the principal component contribution plot.

[0011] Preferably, the method steps for preprocessing the online monitoring parameters are as follows:

[0012] Perform data cleaning on the online monitoring parameters, supplement or delete abnormal reaction data and missing reaction data;

[0013] Perform normalization processing on the online monitoring parameters;

[0014] The method steps for preprocessing the historical monitoring parameters are as follows:

[0015] Perform data cleaning on the historical monitoring parameters, and delete abnormal reaction data and missing reaction data;

[0016] Normalize historical monitoring parameters.

[0017] Preferably, the modeling data of the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model are respectively reconstructed by the deviation between the online monitoring parameter prediction values of the Bayesian regression model and the preprocessed historical monitoring parameters.

[0018] Preferably, the range of the training data of the first reconstructed PCA monitoring model is the historical data set when the temperature in the DMTO reactor is between 460°C and 480°C.

[0019] Preferably, the range of the training data of the second reconstructed PCA monitoring model is the historical data set when the temperature in the DMTO reactor is between 460°C and 490°C.

[0020] Preferably, a number of consecutive online monitoring values T1 2 and the limit value T1 2 lim and a number of consecutive online monitoring values T2 2 and the limit value T1 2 lim When making numerical comparisons respectively, when a number of consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , and no prompt is made, a number of online monitoring values T1 2 and a number of online monitoring values T2 2 are respectively converted into historical monitoring parameters, and after being preprocessed, they are input into the Bayesian regression model for training and variable reconstruction, and then input into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor respectively, and the output is updated to obtain the limit value T1 2 lim and the limit value T2 2 lim .

[0021] A temperature runaway monitoring and warning system in a DMTO reactor, comprising:

[0022] A determination module, configured to obtain a number of online monitoring parameters in the DMTO process, and after preprocessing the number of online monitoring parameters, input them into a number of Bayesian regression models respectively. The number of Bayesian regression models respectively output prediction values of the number of online monitoring parameters, and determine a number of online monitoring deviation values through the corresponding relationship between the number of online monitoring parameters and the prediction values of the number of online monitoring parameters;

[0023] A first output module, configured to input the number of online monitoring deviation values into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor respectively, and output the online monitoring value T1 correspondingly 2and the online monitoring value T2 2 ;

[0024] A second output module, configured to obtain historical monitoring parameters, preprocess the historical monitoring parameters to train a plurality of Bayesian regression models, reconstruct variables using the predicted values and historical monitoring parameter values of the plurality of Bayesian regression models, and respectively train a first reconstructed PCA monitoring model and a second reconstructed PCA monitoring model in the DMTO reactor, and output to obtain the limit value T1 2 lim and the limit value T2 2 lim ;

[0025] A comparison module, configured to numerically compare a plurality of consecutive online monitoring values T1 2 with the limit value T1 2 lim and a plurality of consecutive online monitoring values T2 2 with the limit value T1 2 lim respectively. When a plurality of consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , no prompt is made; when a plurality of consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and a plurality of consecutive online monitoring values T2 2 are less than the limit value T2 2 lim , a warning state is displayed; when a plurality of consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and a plurality of consecutive online monitoring values T2 2 are greater than the limit value T2 2 lim , an alarm state is displayed, and the cause of the fault is determined through the principal component contribution plot.

[0026] Preferably, it further includes an update module, configured to numerically compare a plurality of consecutive online monitoring values T1 2 with the limit value T1 2 lim and a plurality of consecutive online monitoring values T2 2 with the limit value T1 2 lim respectively. When a plurality of consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , and no prompt is made, a plurality of online monitoring values T1 2 and a plurality of online monitoring values T2 2They are respectively converted into historical monitoring parameters, and after preprocessing, they are input into a Bayesian regression model for training and variable reconstruction, and then respectively input into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor, and the output is updated to obtain the limit value T1 2 lim and the limit value T2 2 lim .

[0027] A runaway temperature monitoring and warning device in a DMTO reactor, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the runaway temperature monitoring and warning method in the above-mentioned DMTO reactor are implemented.

[0028] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method steps of a runaway temperature monitoring and warning method in a DMTO reactor as described above are implemented.

[0029] Compared with the prior art, the present invention has the following beneficial technical effects:

[0030] The present invention provides a runaway temperature monitoring and warning method in a DMTO reactor. After reconstructing the monitoring data through a Bayesian regression algorithm, it is input into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model. The first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model output the online monitoring values T1 2 and the online monitoring value T2 2 . By comparing a continuous number of online monitoring values T1 2 and the limit value T1 2 lim , a continuous number of online monitoring values T2 2 and the limit value T2 2 lim , and then monitoring the temperature in the DMTO reactor, realizing early warning of runaway temperature, analyzing faults, assisting in DMTO process production decisions, and improving the detection efficiency of the runaway temperature problem in the DMTO reactor.

[0031] Furthermore, first, abnormal reaction data and missing reaction data are deleted and normalized, which improves the utilization efficiency of online monitoring parameters and historical monitoring parameters, and ensures the accuracy of fault analysis in the early warning and alarm states. The normalized online monitoring parameters and historical monitoring parameters are respectively input into a number of Bayesian regression models after preprocessing, and the monitoring data is reconstructed based on the deviation between the predicted value obtained by the Bayesian regression model and the reaction data monitoring value. Description of the Drawings

[0032] Figure 1It is the flow chart of the method for monitoring and warning of runaway temperature in the DMTO reactor of the present invention;

[0033] Figure 2 It is the flow chart of the system for monitoring and warning of runaway temperature in the DMTO reactor of the present invention;

[0034] Figure 3 It is the flow block diagram of the method for monitoring and warning of runaway temperature in the DMTO reactor of the present invention;

[0035] Figure 4 It is the equipment diagram for monitoring and warning of runaway temperature in the DMTO reactor of the present invention;

[0036] Figure 5 They are the predicted values and monitored values of the Bayesian regression algorithm for the temperature of the DMTO reactor in the embodiments of the present invention;

[0037] Figure 6 They are the percentage distributions of the relative errors of each regression algorithm for all variables in the embodiments of the present invention;

[0038] Figure 7 They are the on-line monitored values T1 2 of the present invention;

[0039] Figure 8 They are the on-line monitored values T2 2 of the present invention;

[0040] Figure 9 They are the actual runaway temperature effect diagrams in the DMTO reactor of the embodiments of the present invention. Detailed implementation manners

[0041] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to 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 the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0042] It should be noted that the terms "first", "second", etc. in the description, claims and the above-mentioned drawings of the present invention are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present invention described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0043] The present invention will be further described in detail below with reference to the accompanying drawings:

[0044] See Figure 1 、 Figure 3 and Figure 4 The present invention provides a method for monitoring and early warning of runaway temperature in a DMTO reactor, which models according to historical data, monitors real-time data, gives early warnings, and analyzes faults to ensure the stable operation of the DMTO reactor.

[0045] Specifically, the method for monitoring and early warning of runaway temperature in the DMTO reactor includes the following steps:

[0046] Obtain a number of on-line monitoring parameters in the DMTO process, and after preprocessing the number of on-line monitoring parameters, input them into a number of Bayesian regression models respectively. The number of Bayesian regression models correspondingly output the predicted values of the number of on-line monitoring parameters, and determine a number of on-line monitoring deviation values by corresponding the number of on-line monitoring parameters with the predicted values of the number of on-line monitoring parameters;

[0047] Input the number of on-line monitoring deviation values into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor respectively, and correspondingly output to obtain the on-line monitoring value T1 2 and the on-line monitoring value T2 2 ;

[0048] Obtain historical monitoring parameters, preprocess the historical monitoring parameters to train a number of Bayesian regression models, reconstruct variables using the predicted values and historical monitoring parameter values of the number of Bayesian regression models, and train the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor with the reconstructed variables respectively, and output to obtain the limit value T1 2 lim and the limit value T2 2 lim ;

[0049] The continuous number of on-line monitoring values T1 2With the limit value T1 2 lim And a number of consecutive on - line monitoring values T2 2 With the limit value T1 2 lim Respectively conduct numerical comparisons. When a number of consecutive on - line monitoring values T1 2 Are less than the limit value T1 2 lim , no prompt is made; when a number of consecutive on - line monitoring values T1 2 Are greater than the limit value T1 2 lim , and a number of consecutive on - line monitoring values T2 2 Are less than the limit value T2 2 lim , the warning state is displayed; when a number of consecutive on - line monitoring values T1 2 Are greater than the limit value T1 2 lim , and a number of consecutive on - line monitoring values T2 2 Are greater than the limit value T2 2 lim , the alarm state is displayed, and the cause of the fault is determined through the principal component contribution diagram.

[0050] Specifically, the method steps of pre - processing on - line monitoring parameters are as follows:

[0051] Conduct data cleaning on on - line monitoring parameters, supplement or delete abnormal reaction data and missing reaction data;

[0052] Conduct normalization processing on on - line monitoring parameters;

[0053] The method steps of pre - processing historical monitoring parameters are as follows:

[0054] Conduct data cleaning on historical monitoring parameters, and delete abnormal reaction data and missing reaction data;

[0055] Conduct normalization processing on historical monitoring parameters.

[0056] In the present invention, the modeling data of the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model are respectively reconstructed by the deviation between the predicted value of on - line monitoring parameters of the Bayesian regression model and the pre - processed historical monitoring parameters.

[0057] The selection range of the training data of the first reconstructed PCA monitoring model is the historical data set when the temperature in the DMTO reactor is between 460 °C and 480 °C.

[0058] The selection range of the training data of the second reconstructed PCA monitoring model is the historical data set when the temperature in the DMTO reactor is between 460 °C and 490 °C.

[0059] A number of Bayesian regression models are trained with preprocessed historical monitoring data, and the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model respectively output T1 2 lim and T2 2 lim limit values.

[0060] Specifically, when a number of consecutive online monitoring values T1 2 are numerically compared with the limit value T1 2 lim and a number of consecutive online monitoring values T2 2 are numerically compared with the limit value T1 2 lim respectively, when a number of consecutive online monitoring values T1 2 are less than the limit value T1 2 lim and no prompt is given, a number of online monitoring values T1 2 and a number of online monitoring values T2 2 are respectively converted into historical monitoring parameters, and after preprocessing, they are input into the Bayesian regression model for training and variable reconstruction, and then input into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor respectively, and the output is updated to obtain the limit value T1 2 lim and the limit value T2 2 lim .

[0061] Input the monitoring data corresponding to the latest DMTO temperature between 460 - 480 °C into the first reconstructed PCA model to replace the monitoring data with a longer time in the original model. Input the monitoring data corresponding to the latest DMTO temperature between 460 - 490 °C into the second reconstructed PCA model to replace the monitoring data with a longer time in the original model.

[0062] Embodiment

[0063] This embodiment provides a method for monitoring and warning of runaway temperature in a DMTO reactor based on a Bayesian regression reconstructed principal component monitoring model, including:

[0064] S101: Obtain multiple online monitoring parameters in the DMTO process. The multiple online monitoring parameters in the DMTO process are shown in the following table:

[0065]

[0066] S102: Substitute the multiple online monitoring parameters obtained in the DMTO process into the corresponding 22 Bayesian regression models one by one, and the corresponding 22 Bayesian regression models output the predicted values of 22 online monitoring parameters in the DMTO process.

[0067] Among them, the 22 Bayesian regression models are established by using historical monitoring parameters as training data for training.

[0068] Preprocess the online monitoring parameters and historical monitoring parameters. The data preprocessing includes deleting abnormal response data, supplementing missing response data, and data normalization.

[0069] Establish a Bayesian regression model: y i |x i = β1 + β2x i + u i , i = 1, 2......, n.

[0070] Among them, y i represents the i-th observed value of the dependent variable; x i represents the i-th observed value of the independent variable; μ i represents the i-th disturbance term; β1 and β2 are unknown parameters. Assume that μ1,..., μ n are independent of each other and follow N(0, σ2).

[0071] Determine the maximum likelihood estimator: To make the probability generated by the model parameters estimated by the sample the largest, the method of maximum likelihood estimation is selected. Therefore, to obtain the unknown parameters β1, β2, and σ2 in the model, the likelihood function is:

[0072]

[0073] Joint posterior density:

[0074]

[0075] Therefore, the maximum likelihood estimator can be obtained as:

[0076]

[0077] Among them, The unbiased estimator of σ2 is:

[0078]

[0079] Integrate with respect to σ to obtain the two-dimensional posterior density. When x and y are known, the joint distribution of β1 and β2 follows a bivariate student t-distribution.

[0080] Based on a new online monitoring parameter * and the historical Bayesian regression model, the predicted value of a specific monitoring parameter can be obtained

[0081] S103: Calculate the deviation between the predicted value and the online monitoring value, such asFigure 5 as shown

[0082]

[0083] Δ is the deviation value between the predicted value and the on-line monitoring value.

[0084] S104: Substitute the obtained monitoring variable deviation into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model, and the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model respectively output T1 2 and T2 2 values. If the T1 2 value does not exceed the T1 2 lim limit, no prompt is made; if the T1 2 value exceeds the T1 2 lim limit, and the T2 2 value does not exceed the T2 2 lim limit, the warning state is displayed; if the T1 2 value exceeds the T1 2 lim limit, and the T2 2 value exceeds the T2 2 lim limit, the alarm state is displayed.

[0085] where the T1 2 lim limit and the T2 2 lim limit are obtained by training with historical monitoring parameters.

[0086] Use the data matrix reconstructed from the deviation between the actual value of the historical monitoring parameter and the predicted value after Bayesian regression as the training data for the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model.

[0087] The first reconstructed PCA monitoring model selects historical monitoring parameters with the DMTO reaction temperature between 460 and 480 °C as the training data set X1.

[0088] Perform standardization processing on the training data set:

[0089]

[0090] where, x1 i is the original value of the variable, u1 i is the mean value of the variable, σ1 i is the standard deviation of the variable, m is the number of monitoring parameters, and n is the number of monitoring points.

[0091] Solve the covariance matrix:

[0092]

[0093] Among them, cov1(X) is the covariance matrix of X1.

[0094] Solve the eigenvalues of the covariance matrix and arrange them from largest to smallest:

[0095] cov(X1)p1 i = λ1 i p1 i i = 1, 2, ……, 2(mn + 4)

[0096] λ11 ≥ λ12 ≥ Λ ≥ λ1 2(mn+4)

[0097]

[0098] Among them, λ1 i is the eigenvalue, p1 i is the eigenvector, ∧1 is the eigenvalue matrix, and P1 is the load matrix.

[0099]

[0100] Among them, k is the number of principal components with a principal component contribution rate greater than 85%.

[0101] The original matrix X1 can be expressed after principal component transformation as:

[0102] X1 = t11p11 T + t12p12 T + …… + t1 k p1 k T + E1

[0103] Among them, t1 i is the score vector, and E1 is the residual matrix.

[0104] Generate the control limit T1 2 lim :

[0105]

[0106] Among them, a is the test level, with a value of 0.95, and F k,n-1,a is the critical value of the F-distribution under the condition that the test level is a and the degrees of freedom are k, n - 1.

[0107] The real-time T1 value of the on-line monitoring parameter 2 value:

[0108]

[0109] is t1 i the variance of

[0110] The second reconstructed PCA monitoring model selects the historical monitoring parameters of the DMTO reaction temperature between 460 and 490 °C as the training data set X2. The training process is the same as that of the first reconstructed PCA monitoring model and will not be elaborated here.

[0111] Select the principal components with a principal component contribution rate greater than 85% to construct the first reconstructed PCA monitoring model.

[0112] S105: If in the alarm state, determine the cause of the fault through the principal component contribution plot method.

[0113] S106: Substitute the real-time online monitoring parameters of the DMTO parameters as training data into the Bayesian regression model to update the Bayesian regression model.

[0114] Take the monitoring parameters of the DMTO reactor temperature between 460 and 480 °C as the training data of the first reconstructed PCA monitoring model to update the first reconstructed PCA monitoring model;

[0115] Take the monitoring parameters of the DMTO reactor temperature between 460 and 490 °C as the training data of the second reconstructed PCA monitoring model to update the second reconstructed PCA monitoring model.

[0116] According to Figure 6 shown, the percentage distribution of the relative errors of all variables for each regression algorithm; the online monitoring effect is as Figure 7 , Figure 8 and Figure 9 shown.

[0117] As can be seen from the above technical solutions, a method for monitoring and warning of runaway temperature in a DMTO reactor provided by the present invention, the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model respectively output T1 2 and T2 2 values. If the T1 2 value does not exceed the T1 2 lim limit value, no prompt is made; if the T1 2 value exceeds the T1 2 lim limit value, and the T2 2 value does not exceed the T2 2 lim limit value, the warning state is displayed; if the T1 2 value exceeds the T1 2 lim limit value, and the T2 2 value exceeds the T2 2 limThe limit value is used to display the alarm status. Meanwhile, the Bayesian regression model, the first reconstructed PCA monitoring model, and the second reconstructed PCA monitoring model are updated according to the latest online monitoring parameters, adapting to the actual situation, making the monitoring and control more accurate and adaptable, and ensuring the stable operation of the DMTO process.

[0118] According to Figure 2 As shown, the present invention also provides a runaway temperature monitoring and warning system in a DMTO reactor, including a determination module, a first output module, a second output module, a comparison module, and an update module;

[0119] The determination module is configured to obtain a plurality of online monitoring parameters during the DMTO process, and respectively input the plurality of online monitoring parameters after preprocessing into the Bayesian regression model. The Bayesian regression model correspondingly outputs the predicted values of the plurality of online monitoring parameters, and determines a plurality of online monitoring deviation values according to the corresponding relationship between the plurality of online monitoring parameters and the predicted values of the plurality of online monitoring parameters;

[0120] The first output module is configured to respectively input the plurality of online monitoring deviation values into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor, and correspondingly output the online monitoring value T1 2 and the online monitoring value T2 2 ;

[0121] The second output module is configured to obtain historical monitoring parameters, preprocess and train a plurality of Bayesian regression models with the historical monitoring parameters, reconstruct variables using the predicted values of the plurality of Bayesian regression models and the historical monitoring parameter values, and respectively train the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor with the reconstructed variables, and output the limit value T1 2 lim and the limit value T2 2 lim ;

[0122] The comparison module is configured to numerically compare a plurality of consecutive online monitoring values T1 2 with the limit value T1 2 lim and a plurality of consecutive online monitoring values T2 2 with the limit value T1 2 lim respectively. When a plurality of consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , no prompt is made; when a plurality of consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and a plurality of consecutive online monitoring values T2 2 are less than the limit value T22 lim When it is, the warning state is displayed; when a number of consecutive on-line monitoring values T1 2 are greater than the limit value T1 2 lim , and a number of consecutive on-line monitoring values T2 2 are greater than the limit value T2 2 lim When it is, the alarm state is displayed, and the cause of the fault is determined through the principal component contribution diagram;

[0123] An update module is used for a number of consecutive on-line monitoring values T1 2 and the limit value T1 2 lim as well as a number of consecutive on-line monitoring values T2 2 and the limit value T1 2 lim When numerical comparisons are made respectively, when a number of consecutive on-line monitoring values T1 2 are less than the limit value T1 2 lim , when no prompt is made, a number of on-line monitoring values T1 2 and a number of on-line monitoring values T2 2 are respectively converted into historical monitoring parameters, and after preprocessing, they are input into the Bayesian regression model for training and variable reconstruction, and are respectively input into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor, and the updated output obtains the limit value T1 2 lim and the limit value T2 2 lim .

[0124] In the present invention, the monitoring parameters when the temperature of the DMTO reactor is between 460 and 480 °C are used as the training data of the first reconstructed PCA monitoring model to update the first reconstructed PCA monitoring model;

[0125] The monitoring parameters when the temperature of the DMTO reactor is between 460 and 490 °C are used as the training data of the second reconstructed PCA monitoring model to update the second reconstructed PCA monitoring model.

[0126] The present invention also provides a runaway temperature monitoring and warning device in a DMTO reactor, including: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a runaway temperature monitoring and warning program in a DMTO reactor.

[0127] When the processor executes the computer program, it implements the steps of the above-mentioned method for monitoring and warning of runaway temperature in the DMTO reactor. For example, it obtains a number of on-line monitoring parameters during the DMTO process, and after preprocessing the number of on-line monitoring parameters, it inputs them into the Bayesian regression model respectively. The Bayesian regression model outputs the predicted values of a number of on-line monitoring parameters correspondingly, and determines a number of on-line monitoring deviation values by corresponding the number of on-line monitoring parameters with the predicted values of the number of on-line monitoring parameters.

[0128] Input the number of on-line monitoring deviation values into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor respectively, and output the on-line monitoring value T1 correspondingly. 2 And the on-line monitoring value T2 2 ;

[0129] Obtain historical monitoring parameters, preprocess the historical monitoring parameters to train a number of Bayesian regression models, reconstruct variables using the predicted values of the number of Bayesian regression models and the historical monitoring parameter values, and train the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor with the reconstructed variables respectively, and output the limit value T1. 2 lim And the limit value T2 2 lim ;

[0130] Compare the continuous number of on-line monitoring values T1 2 with the limit value T1 2 lim and compare the continuous number of on-line monitoring values T2 2 with the limit value T1 2 lim numerically. When the continuous number of on-line monitoring values T1 2 is less than the limit value T1 2 lim , no prompt is made; when the continuous number of on-line monitoring values T1 2 is greater than the limit value T1 2 lim , and the continuous number of on-line monitoring values T2 2 is less than the limit value T2 2 lim , the warning status is displayed; when the continuous number of on-line monitoring values T1 2 is greater than the limit value T1 2 lim , and the continuous number of on-line monitoring values T2 2 is greater than the limit value T2 2 lim , the alarm status is displayed, and the cause of the fault is determined through the principal component contribution plot.

[0131] Alternatively, when the processor executes the computer program, it realizes the functions of each module in the above system. For example: a determination module, configured to obtain a plurality of on-line monitoring parameters in the DMTO process, and respectively input the plurality of on-line monitoring parameters into a Bayesian regression model after preprocessing. The Bayesian regression model correspondingly outputs a plurality of on-line monitoring parameter prediction values, and determines a plurality of on-line monitoring deviation values corresponding to the plurality of on-line monitoring parameters and the plurality of on-line monitoring parameter prediction values;

[0132] A first output module, configured to respectively input the plurality of on-line monitoring deviation values into a first reconstructed PCA monitoring model and a second reconstructed PCA monitoring model in the DMTO reactor, and correspondingly output to obtain an on-line monitoring value T1 2 and an on-line monitoring value T2 2 ;

[0133] A second output module, configured to obtain historical monitoring parameters, preprocess the historical monitoring parameters to train a plurality of Bayesian regression models, reconstruct variables by using the prediction values of the plurality of Bayesian regression models and the historical monitoring parameter values, and respectively train the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor with the reconstructed variables, and output to obtain a limit value T1 2 lim and a limit value T2 2 lim ;

[0134] A comparison module, configured to respectively perform numerical comparisons on a plurality of consecutive on-line monitoring values T1 2 and the limit value T1 2 lim and a plurality of consecutive on-line monitoring values T2 2 and the limit value T1 2 lim When a plurality of consecutive on-line monitoring values T1 2 are less than the limit value T1 2 lim , no prompt is made; when a plurality of consecutive on-line monitoring values T1 2 are greater than the limit value T1 2 lim , and a plurality of consecutive on-line monitoring values T2 2 are less than the limit value T2 2 lim , a warning state is displayed; when a plurality of consecutive on-line monitoring values T1 2 are greater than the limit value T1 2 lim , and a plurality of consecutive on-line monitoring values T2 2 are greater than the limit value T2 2 lim , an alarm state is displayed, and the cause of the fault is determined through a principal component contribution plot;

[0135] An update module for a number of consecutive online monitoring values T1 2 and the limit value T1 2 lim as well as a number of consecutive online monitoring values T2 2 and the limit value T1 2 lim When making numerical comparisons respectively, when a number of consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , and no prompt is made, a number of online monitoring values T1 2 and a number of online monitoring values T2 2 are respectively converted into historical monitoring parameters, and after preprocessing, they are input into the Bayesian regression model for training and variable reconstruction, and are respectively input into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor, and the updated output obtains the limit value T1 2 lim and the limit value T2 2 lim .

[0136] Exemplarily, the computer program can be divided into one or more modules / units. The one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program in the runaway temperature monitoring and warning device of the DMTO reactor. For example, the computer program can be divided into a determination module, a first output module, a second output module, a comparison module, and an update module. The specific functions of each module are as follows: The determination module is used to obtain a number of online monitoring parameters in the DMTO process, and after preprocessing the number of online monitoring parameters, input them into the Bayesian regression model respectively. The Bayesian regression model correspondingly outputs the predicted values of the number of online monitoring parameters, and determines a number of online monitoring deviation values by corresponding the number of online monitoring parameters with the predicted values of the number of online monitoring parameters;

[0137] The first output module is used to input the number of online monitoring deviation values into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor respectively, and correspondingly outputs to obtain the online monitoring value T1 2 and the online monitoring value T2 2 ;

[0138] The second output module is used to obtain historical monitoring parameters, preprocess the historical monitoring parameters to train a number of Bayesian regression models, reconstruct variables using the predicted values and historical monitoring parameter values of the number of Bayesian regression models, and train the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor respectively, and output the limit value T1 2 lim and the limit value T2 2 lim ;

[0139] The comparison module is used to numerically compare a number of consecutive online monitoring values T1 2 with the limit value T1 2 lim and a number of consecutive online monitoring values T2 2 with the limit value T1 2 lim respectively. When a number of consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , no prompt is made; when a number of consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and a number of consecutive online monitoring values T2 2 are less than the limit value T2 2 lim , the warning status is displayed; when a number of consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and a number of consecutive online monitoring values T2 2 are greater than the limit value T2 2 lim , the alarm status is displayed, and the cause of the fault is determined through the principal component contribution plot;

[0140] The update module is used for a number of consecutive online monitoring values T1 2 compared with the limit value T1 2 lim and a number of consecutive online monitoring values T2 2 compared with the limit value T1 2 lim respectively. When a number of consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , and no prompt is made, a number of online monitoring values T1 2 and a number of online monitoring values T2 2They are respectively converted into historical monitoring parameters, preprocessed, and then input into a Bayesian regression model for training and variable reconstruction. They are respectively input into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor, and the output is updated to obtain the limit value T1. 2 lim and the limit value T2. 2 lim .

[0141] The runaway temperature monitoring and warning device in the DMTO reactor can be computing devices such as a desktop computer, a notebook, a palm computer, and a cloud server. The runaway temperature monitoring and warning in the DMTO reactor may include, but are not limited to, a processor and a memory.

[0142] The processor can be a Central Processing Unit (CPU), or can also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The processor is the control center of the runaway temperature monitoring and warning device in the DMTO reactor, and connects various parts of the runaway temperature monitoring and warning device in the entire DMTO reactor through various interfaces and lines.

[0143] The memory can be used to store the computer program and / or module. The processor realizes various functions of the runaway temperature monitoring and warning device in the DMTO reactor by running or executing the computer program and / or module stored in the memory, and by calling the data stored in the memory.

[0144] The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, phone book, etc.), etc. In addition, the memory can include high-speed random access memory, and can also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0145] The present invention also provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the method for monitoring and warning of runaway temperature in a DMTO reactor are implemented.

[0146] If the modules / units integrated in the runaway temperature monitoring and warning device in the DMTO reactor are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0147] Based on such understanding, all or part of the processes in the above method of the present invention can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the method for monitoring and warning of runaway temperature in the DMTO reactor can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc.

[0148] The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0149] It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0150] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent replacements can still be made to the specific implementation manners of the present invention, and any modification or equivalent replacement without departing from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A method for monitoring and warning of runaway temperature in a DMTO reactor, characterized in that, It includes the following steps: Obtain multiple on-line monitoring parameters in the DMTO process, and respectively input the pre-processed on-line monitoring parameters into multiple Bayesian regression models. The Bayesian regression models correspondingly output multiple on-line monitoring parameter prediction values, and determine multiple on-line monitoring deviation values through the corresponding relationship between the on-line monitoring parameters and the on-line monitoring parameter prediction values; Input the online monitoring deviation values into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor respectively, and correspondingly output the online monitoring values T1 2 and the online monitoring value T2 2 ; Obtain historical monitoring parameters, preprocess the historical monitoring parameters to train multiple Bayesian regression models, use the predicted values of the online monitoring parameters and the historical monitoring parameters of the Bayesian regression models to reconstruct variables, and train the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor with the reconstructed variables respectively, and output to obtain the limit value T1 2 lim and the limit value T2 2 lim ; The selection range of the training data of the first reconstructed PCA monitoring model is the historical data set when the temperature in the DMTO reactor is between 460°C and 480°C; The selection range of the training data of the second reconstructed PCA monitoring model is the historical data set when the temperature in the DMTO reactor is between 460°C and 490°C; A series of consecutive online monitoring values T1 2 and the limit value T1 2 lim as well as a series of consecutive online monitoring values T2 2 and the limit value T2 2 lim are respectively compared numerically. When a series of consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , no prompt is made; when a series of consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and a series of consecutive online monitoring values T2 2 are less than the limit value T2 2 lim , a warning status is displayed; when a series of consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and a series of consecutive online monitoring values T2 2 are greater than the limit value T2 2 lim , an alarm status is displayed, and the cause of the fault is determined through the principal component contribution plot.

2. The method for monitoring and warning of runaway temperature in a DMTO reactor according to claim 1, characterized in that, The method steps of pre-processing the on-line monitoring parameters are as follows: Perform data cleaning on the on-line monitoring parameters, supplement or delete abnormal reaction data and missing reaction data; Perform normalization processing on the on-line monitoring parameters; The method steps of pre-processing the historical monitoring parameters are as follows: Perform data cleaning on the historical monitoring parameters, and delete abnormal reaction data and missing reaction data; Perform normalization processing on the historical monitoring parameters.

3. The method for monitoring and warning of runaway temperature in a DMTO reactor according to claim 1, characterized in that, Multiple consecutive online monitoring values T1 2 and the limit value T1 2 lim as well as multiple consecutive online monitoring values T2 2 and the limit value T2 2 lim When making numerical comparisons respectively, when multiple consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , without any prompt, multiple online monitoring values T1 2 and several online monitoring values T2 2 are respectively converted into historical monitoring parameters, and after preprocessing, they are input into the Bayesian regression model for training and variable reconstruction, and then respectively input into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor, and the output is updated to obtain the limit value T1 2 lim and the limit value T2 2 lim .

4. A system for monitoring and warning of runaway temperature in a DMTO reactor, characterized in that, It includes: A determination module, configured to obtain multiple on-line monitoring parameters in the DMTO process, and respectively input the pre-processed on-line monitoring parameters into multiple Bayesian regression models. The Bayesian regression models correspondingly output multiple on-line monitoring parameter prediction values, and determine multiple on-line monitoring deviation values through the corresponding relationship between the on-line monitoring parameters and the on-line monitoring parameter prediction values; The first output module is configured to input the online monitoring deviation values into a first reconstructed PCA monitoring model and a second reconstructed PCA monitoring model in the DMTO reactor respectively, and correspondingly output the online monitoring value T1 2 and the online monitoring value T2 2 ; A second output module, configured to obtain historical monitoring parameters, preprocess the historical monitoring parameters to train multiple Bayesian regression models, reconstruct variables by using the predicted values and historical monitoring parameter values of the Bayesian regression models, train a first reconstructed PCA monitoring model and a second reconstructed PCA monitoring model in a DMTO reactor respectively with the reconstructed variables, and output a limit value T1 2 lim and a limit value T2 2 lim ; A comparison module for comparing multiple consecutive online monitoring values T1 2 with a limit value T1 2 lim and multiple consecutive online monitoring values T2 2 with a limit value T2 2 lim respectively for numerical comparison. When multiple consecutive online monitoring values T1 2 are less than the limit value T1 2 lim , no prompt is made; when multiple consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and multiple consecutive online monitoring values T2 2 are less than the limit value T2 2 lim , a warning status is displayed; when multiple consecutive online monitoring values T1 2 are greater than the limit value T1 2 lim , and multiple consecutive online monitoring values T2 2 are greater than the limit value T2 2 lim , an alarm status is displayed, and the cause of the fault is determined through a principal component contribution plot.

5. The system for monitoring and warning of runaway temperature in a DMTO reactor according to claim 4, characterized in that, It also includes an update module for a continuous series of online monitoring values T1 2 and the limit value T1 2 lim as well as a continuous series of online monitoring values T2 2 and the limit value T2 2 lim When performing numerical comparisons respectively, when a continuous series of online monitoring values T1 2 is less than the limit value T1 2 lim , and no prompt is made, the multiple online monitoring values T1 2 and the multiple online monitoring values T2 2 are respectively converted into historical monitoring parameters, and after preprocessing, they are input into the Bayesian regression model for training and variable reconstruction and are respectively input into the first reconstructed PCA monitoring model and the second reconstructed PCA monitoring model in the DMTO reactor, and the updated output obtains the limit value T1 2 lim and the limit value T2 2 lim .

6. A device for monitoring and warning of runaway temperature in a DMTO reactor, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for monitoring and warning of runaway temperature in the DMTO reactor according to any one of claims 1 to 3.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of a method for monitoring and warning of runaway temperature in a DMTO reactor according to any one of claims 1-3.

Images