A complex dynamics solution method for adiabatic acceleration calorimetry

By using differential equal conversion rate method and dynamic compensation effect in the adiabatic acceleration calorimeter, combined with the improved sparrow search algorithm, the difficulty in selecting models and multi-parameter solutions in the adiabatic acceleration calorimeter are solved. The problem of this paper improves the accuracy of thermal hazard assessment and the reliability of chemical safety assessment.

CN114694763BActive Publication Date: 2024-11-19CHINA JILIANG UNIV
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202210374372.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-04-11
Publication Date
2024-11-19
Estimated Expiration
2042-04-11

AI Technical Summary

Technical Problem

In the thermodynamic calculation, existing adiabatic acceleration calorimeters have problems such as improper model selection and multi-parameter solution that are prone to local optimality, resulting in insufficient accuracy and reliability of thermal risk assessment.

Method used

The adiabatic calorimetric reaction model identification method based on differential equal conversion rate method and kinetic compensation effect is adopted, and the objective function is optimized and solved globally with the improved sparrow search algorithm to obtain complex thermodynamic parameters.

Benefits of technology

Through the analysis of multiple sets of ARC calorimetry experimental data, the error caused by single experimental data is avoided, the accuracy of substance thermal risk assessment is improved, the difficulty in model selection and local optimal problems are effectively solved, and the chemical safety assessment is enhanced reliability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114694763B_ABST
    Figure CN114694763B_ABST
Patent Text Reader

Abstract

The present invention discloses a complex dynamics solution method for adiabatic acceleration calorimetry. The present invention first solves multiple groups of adiabatic calorimetry experimental data of different thermal inertia factors based on the differential equal conversion rate method; secondly, the adiabatic calorimetry reaction model is identified by combining the solution results of the differential equal conversion rate method and the kinetic compensation effect; then, activation energy modeling is performed based on the identified reaction model, and an objective function is established at the same time; finally, the objective function is intelligently optimized based on the improved sparrow search algorithm to obtain complex thermodynamic parameters. The present invention can effectively solve the problems of difficult model selection and easy local optimization of multi-parameter solutions in adiabatic data processing, and can accurately solve complex kinetic parameters, which has a positive effect on improving the reliability and efficiency of adiabatic acceleration calorimetry as a thermal science analysis instrument in chemical safety assessment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to adiabatic acceleration calorimetry and a thermodynamic solution method thereof for chemical process safety, and more specifically to a complex dynamics solution method for adiabatic acceleration calorimetry. Background Art

[0002] Adiabatic Accelerating Rate Calorimeter (ARC) simulates the adiabatic environment when hazardous chemicals decompose and release heat, and records the temperature, time, pressure and other data of the sample reaction process in real time. Researchers analyze the calorimetric data through numerical calculation methods, calculate the thermodynamics of the tested samples, and then predict the thermal hazard of the substances. Therefore, adiabatic accelerating rate calorimeter is widely used in the field of chemical safety thermal hazard assessment. However, no matter what the purpose of using the adiabatic accelerating rate calorimeter is, the thermodynamic calculation of adiabatic calorimetric data is the basis of all ARC related research work.

[0003] Looking back at the development of ARC, the thermodynamics of adiabatic calorimetric data has always been limited to the use of model fitting methods based on a single calorimetric experiment. Since Townsend and Tou developed the first ARC in 1980, linear model fitting methods have been used to calculate the thermodynamic parameters of adiabatic data. [1] In this case, only the thermodynamic parameters of the simple n-order reaction model can be calculated. In 1995, Kossoy introduced the method of nonlinear model fitting into adiabatic data processing. [2] , the thermodynamic parameters of complex reactions can be calculated using nonlinear fitting methods. However, whether it is linear fitting or nonlinear fitting, minimizing the residual sum of squares between experimental values ​​and calculated values ​​is the key point of model fitting. In order to use the model fitting method reasonably, there are two main aspects that need to be carefully considered.

[0004] The first point: find a suitable reaction model. Model fitting method is a mathematical processing method based on the model. If the model is not selected correctly, the kinetic parameters calculated by the model fitting method will have no practical significance. In order to solve this problem, some researchers have conducted corresponding research on reaction model identification. Kossoy once pointed out that the correct selection of reaction model is very important for the thermal hazard assessment of materials. [3] In his article, the identification of reaction models mainly depends on the analyst's personal experience, and no universal method is proposed. Dong Ze proposed a method to identify autocatalytic reaction models based on the strength of autocatalysis. [4] , this method is helpful for the selection of autocatalytic models. However, there is still no universal and reliable method for reaction model identification.

[0005] The second point: avoid local extreme values. In the process of using the least squares method to fit the reaction model to solve the kinetics, because the reaction model is complex and there are many optimization parameters, it is necessary to avoid the occurrence of local extreme values. In order to solve this problem, some researchers use optimization algorithms when calculating parameters. Guo Zichao used the chaotic particle swarm algorithm to optimize the kinetic parameters when processing adiabatic calorimetric data. [5] Kuang used convolutional neural networks to predict kinetic parameters from thermogravimetric analyzer data. [6] In addition, according to the recommendations of the International Society for Thermal Analysis Dynamics on kinetic solutions, trying different initial values ​​is also a way to avoid local optimality. [7] .

[0006] The model fitting method based on single calorimetric experimental data is always unreliable, but as far as the current adiabatic calorimetric data processing method is concerned, the thermodynamic calculation is still performed using the model fitting method based on the single experimental data. Therefore, in view of the problems of inappropriate model selection leading to errors in the existing adiabatic accelerating calorimetry kinetics and the easy fall into local optimum in multi-parameter solutions, a complex kinetic calculation method for adiabatic accelerating calorimetry was proposed and studied. This method can be used to obtain the reaction model of the substance and solve the thermodynamics of complex thermal decomposition reactions, thus improving the level of thermal analysis kinetics research and the accuracy of thermal hazard assessment of chemicals.

[0007] References

[0008] [1]DI Townsend,JCTou.Thermal hazard evaluation by an acceleratingrate calorimeter.Thermochimica Acta,1980,37(1):1-30.

[0009] [2]AAKossoy,EYKoludarova.Specific features of kinetics evaluation in calorimetric studies of runaway reactions,J.Loss Prev.Process Ind.8(1995)229-235.

[0010] [3]A.Kossoy,Y.Akhmetshin.Identification of kinetic models for the assessment of reaction hazards,Process Saf.Prog.26(2007)209-220.

[0011] [4]Z.Dong,BBXue,LPChen,WHChen,Astudy of classifyingautocatalytic strength with adiabatic conditions,J.Therm.Anal.Calorim.137(2019)217-227.

[0012] [5]ZCGuo,LPChen,WHChen.Estimation of kinetic parameters from adiabatic calorimetric data by a hybrid particle swarm optimization method,Chem.Eng.Res.Des.122(2017)273-279.

[0013] [6]DYKuang,B.Xu.Predicting kinetic triplets using a 1dconvolutional neural network,Thermochim.Acta 669(2018)8-15.

[0014] [7]S.Vyazovkin,AKBurnham,JMCriado,LAPérez-Maqueda,C.Popescu,N.Sbirrazzuoli,ICTAC Kinetic Committee recommendations for performing kineticcomputations on thermal analysis data,Thermochim.Acta 520(2011)1-19. Summary of the invention

[0015] The present invention aims at the deficiencies of the prior art and provides a complex dynamics solution method for adiabatic acceleration calorimetry.

[0016] The technical solution of the present invention is:

[0017] Solve multiple sets of adiabatic calorimetry experimental data with different thermal inertia factors based on differential equal conversion rate method;

[0018] The adiabatic calorimetric reaction model is identified by combining the solution of differential equal conversion rate method and kinetic compensation effect;

[0019] Activation energy modeling is performed based on the identified reaction model, and the objective function is established at the same time;

[0020] Based on the improved sparrow search algorithm, the objective function is intelligently optimized to obtain complex thermodynamic parameters.

[0021] Compared with the existing ARC thermodynamic calculation method, the beneficial effects of the present invention are:

[0022] 1. Using multiple sets of ARC calorimetric experimental data for kinetic analysis avoids the risks brought by single calorimetric experimental data analysis and improves the accuracy of thermal hazard assessment of materials.

[0023] 2. A method for identifying adiabatic calorimetric reaction models that combines differential equiconversion rate with kinetic compensation is proposed. The reaction model f(α) can be identified, which effectively solves the problem of difficult model selection when fitting the calculation kinetics.

[0024] 3. When solving complex dynamic parameters, traditional search methods have the problems of low efficiency and low accuracy. The present invention utilizes the global optimization capability of the improved sparrow search during parameter calculation to optimize and solve the complex dynamic parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 Reaction architecture diagram;

[0026] Figure 2 Flow chart of the application of differential equal conversion method to ARC calorimetric data;

[0027] Figure 3 Flow chart of reaction model identification based on adiabatic data;

[0028] Figure 4 Temperature-time curves of 15% BTBPC / tolueneacetic acid with different thermal inertness factors;

[0029] Figure 5 Negative inverse curves of temperature rise rate and temperature of 15% BTBPC / tolueneacetic acid with different thermal inertia factors;

[0030] Figure 6 E α and A α f(α) is a graph of the conversion rate α;

[0031] Figure 7 Kinetic compensation effect curve between activation energy E and pre-exponential factor lnA;

[0032] Figure 8 Response model fitting effect diagram;

[0033] Fig. 9 Activation energy fitting effect diagram;

[0034] Fig.10 Reproduce the temperature-time graph. DETAILED DESCRIPTION

[0035] The present invention proposes a method for solving complex dynamics of adiabatic calorimetry, the idea of ​​which is as follows:

[0036] Firstly, in order to address the risk of solving the kinetics based on a single experimental data in the existing adiabatic accelerating calorimetry, it is proposed to solve the thermal decomposition kinetic parameters based on multiple groups (generally more than 3 groups) of experimental data with different thermal inertia factors using the differential equiconversion rate method, so as to avoid the errors and risks caused by solving the kinetic parameters by fitting the single experimental data model in the traditional adiabatic data analysis.

[0037] Secondly, in view of the problem that the inappropriate model in the solution of the existing adiabatic accelerating calorimetry kinetic parameters causes errors in the solution results, a method for reaction model identification in adiabatic accelerating calorimetry data processing is proposed. Through the model identification results, the reaction model of the material can be obtained, but all the kinetic parameters need to be solved to predict the thermal hazard of the material, so further research is needed. The activation energy model is built through the identified reaction model, and the objective function is established based on the least squares principle using the activation energy theoretical value and the activation energy calculated value obtained by the differential conversion rate method. At the same time, considering the complexity of the mathematical model of the objective function, combined with the global optimization ability of the intelligent algorithm, the improved sparrow search algorithm that integrates Cauchy mutation and reverse learning is applied to the optimization of kinetic parameters, and finally the complex thermodynamics of adiabatic calorimetry can be calculated.

[0038] The technical principle of the present invention is as follows:

[0039] Under ideal adiabatic conditions, all the heat generated by the sample's self-reaction is used to increase its own temperature. However, when using an adiabatic accelerating calorimeter to test samples, a spherical metal container resistant to high temperature and high pressure is used, such as Figure 1 As shown in the figure, the sample pool itself has a fixed mass and heat capacity, which means that part of the heat released by the reaction will be absorbed by the sample pool. In order to describe the relationship between the heat absorbed by the sample pool and the total heat released by the sample, relevant scholars proposed the thermal inertia factor As shown in formula (1).

[0040]

[0041] In formula (1), m b is the mass of the sample pool, in g; c b is the specific heat capacity of the sample cell, in J / (g·K); m s is the sample mass, in g; c sis the specific heat capacity of the sample, in J / (g·K).

[0042] Kinetic analysis generally includes model-free method and model fitting method. The model-free method is further divided into integral method and differential method. The most commonly used differential equal conversion rate method is shown in formula (2).

[0043]

[0044] In formula (2), α is the conversion rate; i represents different temperature programs; R is the universal gas constant, 8.314 J / mol; dα / dt is the reaction rate; f(α) is the reaction model; A α is the pre-exponential factor, s -1 ; E α is the calculated activation energy, J / mol; T α,i is the sample temperature corresponding to the conversion rate α under a certain temperature program i. By linear fitting formula (2), the slope is used to obtain E α The intercept is ln[A α f(α)].

[0045] In the differential equal conversion rate method, the reaction rate equation is generally written in the differential form shown in formula (3). In formula (3), t is the reaction time; T is the sample temperature; and A is the pre-exponential factor.

[0046]

[0047] Generally speaking, the ARC calorimetric experiment can only obtain one set of experimental data, which cannot meet the use conditions of the differential equal conversion rate method. However, in the ARC experiment, based on the different thermal decomposition paths of the same sample under different thermal inertia factors, the differential equal conversion rate method can be used to calculate the thermodynamics according to formula (5) using calorimetric data of more than or equal to 3 sets of the same sample with different thermal inertia factors.

[0048] The temperature is converted into conversion rate according to formula (4), thereby satisfying the conditions of the differential equal conversion rate method.

[0049]

[0050] In formula (4), α is the conversion rate, T is the sample temperature at any time during the reaction stage, T0 is the starting decomposition temperature during the reaction stage, and T f is the reaction end temperature.

[0051] The process of calculating the thermodynamics of ARC calorimetric data based on the differential equal conversion rate method is as follows Figure 2 shown.

[0052]

[0053] The subscript in formula (5) Represents different thermal inertia factors, Representing a group The sample temperature corresponding to the conversion rate α, T onset represents the initial decomposition temperature, T final represents the reaction end temperature, and f(α) represents the reaction model.

[0054] Differential conversion rate method can only calculate A α The value of f(α) cannot be obtained, and the specific form of the reaction model f(α) cannot be obtained. In order to solve A α and f(α), it is necessary to combine the method of dynamic compensation effect for parameter decoupling.

[0055] After making corresponding transformations to formula (4), we can obtain formula (5). In formula (5), when the reaction model f(α) changes, it can be proved from the mathematical formula that there is an obvious kinetic compensation effect.

[0056]

[0057] By choosing different reaction models and parameters, the value of [(dα / dt) / f(α)] at each given α can be calculated. Then, using [(dα / dt) / f(α)] and The linear fit between can get the slope and intercept. The slope represents The intercept represents the value of Then, based on different reaction models and different thermal inertia factors, and The values ​​of the dynamic compensation parameters a and b can be obtained by linear fitting through formula (6).

[0058]

[0059] Finally, after the dynamic compensation parameters are calculated, use lnA α With E α The dynamic compensation effect and the E obtained by the differential conversion rate method α Value, combined with formula (7), lnA α The value of can be calculated.

[0060] ln α =aE α +b (7)

[0061] At this point, the kinetic parameters obtained by the differential equal conversion rate method are ln[A α f(α)] decoupling is completed. After the kinetic parameter decoupling is completed, the adiabatic calorimetric reaction model identification is performed.

[0062] According to the calculated ln[A α f(α)] after removing the logarithm, we get A α f(α), and then A decoupled from the kinetic parameters α , the value of f(α) can be calculated based on formula (8).

[0063] f(α)=[A α f(α)] / A α (8)

[0064] On the basis of the solved value of f(α), based on the commonly used kinetic model of self-reactive substances and combined with formula (9), nonlinear fitting is performed according to R 2 To evaluate the fitting effect, R 2 The closer it is to 1, the better the fitting effect. The reaction model corresponding to the best fitting degree is the reaction model of the corresponding substance. This method can be used to identify the reaction model of the substance to be tested. Figure 3 Identify the process for a reaction model based on adiabatic data.

[0065]

[0066] Where N is the total number of data, f(α) is the actual calculated value, and f(α) fit are the values ​​calculated for different response models, and SS is the residual sum of squares.

[0067] For complex decomposition reactions, they are always composed of several reactions, and each reaction contributes differently to the overall reaction. Equation (10) is the reaction rate expression for the two-step reaction.

[0068]

[0069] In equation (10), subscripts 1 and 2 represent two reaction steps, and K1(T) and K2(T) are Arrhenius constants.

[0070]

[0071]

[0072] Taking the logarithm of both sides of equation (10), we obtain equation (13).

[0073]

[0074] Substitute the expressions of K1(T) and K2(T) into equation (13), take the partial derivatives of both sides of equation (13) with respect to (-1 / T), and then rearrange to obtain the theoretical equation (14) of activation energy E with respect to conversion rate α and temperature T.

[0075]

[0076] It can be seen from equation (14) that once the reaction model f(α) is determined, if the conversion rate α and temperature T are known, the activation energy E can be obtained. α The total activation energy value E α It is always between the activation energies E1 and E2 of the two reaction steps.

[0077] Let γ = A1 / A2, and transform formula (14) to obtain formula (15).

[0078]

[0079] The activation energy has been calculated by analyzing the adiabatic calorimetric data of four groups with different thermal inertia factors using the differential equal conversion rate method. Combined with the theoretically derived activation energy E of equation (14), equation (16) is obtained.

[0080]

[0081] In formula (16), E exp represents the calculated activation energy, E cal Represents the theoretical value of activation energy. According to the principle of least squares method, when the sum of squares of the residuals between the calculated data and the theoretical data is minimized, the parameters to be solved can be obtained.

[0082] The parameter value obtained by optimizing F1 γ , E1, E2 are substituted into equation (10) as known values. Based on equation (17), nonlinear fitting can be performed to calculate the remaining kinetic parameters A1, A2, n1, n2, and m.

[0083]

[0084] In equation (17), (dα / dt) exp is the measured value of the reaction rate, (dα / dt) cal is the theoretical calculated value of the reaction rate.

[0085] After the above derivation process, the objective functions F1 and F2 have been established. However, due to the complexity of the objective function mathematical model and the existence of nonlinear problems, it is difficult to optimize and solve it with the traditional optimization fitting method, so it is necessary to use the improved sparrow search algorithm that integrates Cauchy mutation and reverse learning to optimize and solve the objective function.

[0086] This embodiment is mainly divided into four steps:

[0087] Step 1: Carry out multiple sets of adiabatic accelerated calorimetry HWS mode experiments with different masses on the same sample to obtain multiple sets of adiabatic calorimetry data with different thermal inertia factors, convert the temperature into conversion rate, meet the conditions of the differential equal conversion rate method, and use the differential equal conversion rate method to solve E(α) and ln[A α f(α)].

[0088] Step 2: Decouple the kinetic parameters based on the kinetic compensation effect and the differential equal conversion rate method to obtain the f(α) value, and identify the reaction model using the least squares method.

[0089] Step 3: Based on the identified reaction model of the self-reactive substance, activation energy modeling is carried out based on the reaction model, and then two objective functions F1 and F2 are established by combining the experimental value of activation energy obtained by the differential equal conversion rate method with the reaction rate equation.

[0090] Step 4: Optimize and solve the objective function based on the global optimization capability of the improved sparrow search algorithm, and reproduce the temperature-time curve based on the kinetic parameters obtained.

[0091] 15% BTBPC / toluene-acetic acid solution was selected as the sample, and four groups of HWS mode experiments of 15% BTBPC / toluene-acetic acid samples with different thermal inertia factors were carried out by adiabatic accelerating calorimeter, with sample masses of 3g, 4g, 5g, and 6g respectively. The statistical information of the thermodynamic parameter results of the four groups of experiments is shown in Table 1. Figure 4 The temperature-time curves of 15% BTBPC / toluene-acetic acid with different thermal inertness factors. Figure 5 It is a curve of temperature rise rate and negative inverse of temperature.

[0092] Table 1 Statistics of thermodynamic parameters of 15% BTBPC / toluene-acetic acid

[0093]

[0094] The sample temperature data were screened at a certain step length to obtain 4 groups of screened sample temperatures. The adiabatic temperature rise, reaction rate and temperature rise rate of the sample decomposition exothermic reaction were calculated based on the screened temperature data.

[0095] The screened temperature is converted into a conversion rate according to formula (18), thereby satisfying the condition of the differential equal conversion rate formula (2).

[0096]

[0097] Where α is the conversion rate, T is the sample temperature at any time during the reaction, and T onset is the reaction starting temperature, T final is the reaction end temperature.

[0098] After calculation by differential conversion rate method, we get E α and A α f(α). Set the polynomial order of the Savitzky-Golay filter algorithm to 3 and the frame length to 19. α and A α f(α) is filtered. Figure 6 E after filtering α and A α Plot of f(α) versus conversion α.

[0099] according to Figure 6 As shown, α is in the range of 0.2-0.8, E α The maximum value is 143.586KJ / mol, and the minimum value is 102.276KJ / mol. α As the change of α is obvious, the thermal decomposition reaction is a complex reaction.

[0100] Then, under the conditions of different reaction models and different thermal inertia factors, the linear fitting is performed by formula (6): and The kinetic compensation parameters a and b were obtained as 0.00031346 KJ / mol and -8.1786, respectively. Figure 7 This is the kinetic compensation effect curve between the activation energy E and the pre-exponential factor lnA. It can be seen that there is a good linear relationship between lnA and E.

[0101] After obtaining the kinetic compensation parameters a and b, the activation energy E is obtained by combining the differential equal conversion rate method based on formula (7): α , we can obtain lnA α .

[0102] Solve for lnA α Then, the value of f(α) can be solved by combining formula (8). Select the common reaction models in Table 2 and perform the least squares fitting with f(α) in turn.

[0103] Table 2 Common reaction models of self-reactive substances

[0104]

[0105] According to the fitting results, it can be seen that the Autocatalytic reaction model has the best fitting effect, so it can be determined that the reaction model of 15% BTBPC / toluene-acetic acid is an autocatalytic model.

[0106] According to the expression of the autocatalytic reaction model, the expressions of f1(α) and f2(α) can be obtained as shown in equations (19) and (20).

[0107]

[0108]

[0109] Substituting the expressions of f1(α) and f2(α) in (19) and (20) into formula (14) yields formula (21).

[0110]

[0111] Next, equation (21) is rearranged to obtain equation (22). In equation (22), β = n2 - n1.

[0112]

[0113] However, both the numerator and the denominator of the fraction in formula (22) need to be solved for the parameters A1 and A2, which is not conducive to the solution. Therefore, by setting γ = A1 / A2, formula (22) needs to be further rearranged to obtain formula (23).

[0114]

[0115] The kinetic parameters to be optimized for the autocatalytic decomposition reaction were changed from 7 (A1, E1, n1, A2, E2, n2, m) to 4 (n1, n2, A1, A2), and then (n1, n2, A1, A2) to (n1, n1+β, γA2, A2), where β and γ are constants. In this way, through a series of mathematical derivations, the dimension of the parameters to be optimized was reduced to reduce the possibility of falling into the optimum.

[0116] Finally, the objective function is optimized and solved based on the improved sparrow algorithm ISSA. Algorithm parameter settings: population number N = 1000, maximum number of iterations is 200, warning value ST = 0.6, discoverer ratio PD = 0.1, scout ratio SD = 0.3. First, the algorithm is used to optimize the parameters of the first objective function F1. The activation energy fitting effect is as follows: Fig. 9 As shown. Substitute the optimized parameters (γ, E1, E2, m, β) as known values ​​into formula (10), optimize the parameters of the second objective function F2 according to formula (16), and solve (n1, A2). In this way, the kinetic parameter group to be solved (E1, E2, A1, A2, m, n1, n2) has been obtained. The calculation results of the kinetic parameters of 15% BTBPC / toluene-acetic acid are shown in Table 3.

[0117] Table 3 Calculation results of kinetic parameters of 15% BTBPC / toluene-acetic acid

[0118]

[0119] Fig.10The temperature-time curves reproduced based on the optimized kinetic parameter set are shown in Figure 2. From the multiple sets of temperature-time reproduction curves, the reproduced curves are highly consistent with the experimental curves, indicating that the kinetic parameters obtained are accurate.

[0120] In summary, the present invention can effectively solve the problems of difficult model selection and easy local optimization in multi-parameter solution in adiabatic data processing, and further expand the data analysis theory of adiabatic accelerating calorimeter, which has a positive effect on improving the reliability and efficiency of ARC as a thermal science analysis instrument in chemical safety assessment.

Claims

1. A complex dynamics solution method for adiabatic acceleration calorimetry, characterized by: Solve multiple sets of adiabatic calorimetry experimental data with different thermal inertia factors based on differential equal conversion rate method; The adiabatic calorimetric reaction model is identified by combining the solution of differential equal conversion rate method and kinetic compensation effect; Activation energy modeling is performed based on the identified reaction model, and the objective function is established at the same time; Based on the improved sparrow search algorithm, the objective function is intelligently optimized to obtain complex thermodynamic parameters; The method for solving adiabatic calorimetry experimental data of multiple groups of different thermal inertia factors based on differential equal conversion rate is specifically: Conduct adiabatic calorimetry experiments on the same sample with different thermal inertia factors, filter the temperature data according to a fixed step size, and solve it by combining differential and other conversion rate methods: The subscript Represents different thermal inertia factors, Represents a group of thermal inertia factors The sample temperature corresponding to the conversion rate α, T onset represents the initial decomposition temperature, T final represents the reaction end temperature, R represents the universal gas constant, f(α) represents the reaction model, A α represents the pre-exponential factor; The slope E is obtained by linear fitting of the above equation. α , represents the calculated activation energy, and the intercept is ln[A α f(α)]; The adiabatic calorimetric reaction model identification is specifically: Select different reaction models f(α) and calculate the value of [(dα / dt) / f(α)] at each given α; then, use The linear fit between the two gives the slope and intercept, where the slope represents the thermal inertia factor. Lower activation energy The intercept represents the thermal inertia factor Pre-index factor The value of Secondly, based on different reaction models f(α), the corresponding and For parameter groups Perform linear fitting to obtain the values ​​of dynamic compensation parameters a and b; E calculated by combining the differential equal conversion rate method α , get A α ; lnA α =aE α +b Then according to A α The value of f(α) is calculated to obtain f(α) cal ; f(a) cal =[A α f(a)] / A α Finally, f(α) is analyzed based on different reaction models. cal Perform least squares fitting, and the reaction model with the best fitting effect is the reaction model of the substance to be determined; Where N is the total number of data, f(α) cal is the actual calculated value, f(α) fit is the mathematical expression of different reaction models, SS is the residual sum of squares; The activation energy modeling is performed based on the identified reaction model, and the objective function is established as follows: Based on the identified reaction model, the activation energy theoretical formula E(α,T) is obtained by taking the logarithm of the reaction rate equation and then taking the partial derivative of the negative inverse of temperature; Among them, A1 and A2 are two pre-exponential factors of complex reactions, and E1 and E2 are two activation energies of complex reactions; Let γ = A1 / A2, and then rearrange the activation energy theory formula to obtain the following formula: The reorganized activation energy formula contains the parameters γ, E1, and E2 to be solved; Combining the activation energy theoretical formula with the activation energy calculated value E α , establish the objective function F1; F1=∑(E a -AND cal ) 2 Where E α represents the calculated activation energy, E cal represents the theoretical value of activation energy; The intelligent optimization of the objective function based on the improved sparrow search algorithm to obtain complex thermodynamic parameters is specifically: After optimizing the objective function F1 based on the improved sparrow search algorithm, the parameters γ, E1, and E2 are obtained, and these parameters are substituted into the reaction rate equation as known values; the objective function F2 is established by fitting the reaction rate equation using the least square method; F2=∑[(dα / dt) exp -(dα / dt) cal ] 2 Where (dα / dt) exp is the measured value of the reaction rate, (dα / dt) cal is the theoretical calculated value of the reaction rate; The remaining kinetic parameters obtained by optimizing the objective function F2 include: two pre-exponential factors A1, A2, and three reaction orders n1, n2, and m.

Citation Information

Patent Citations

  • A method for obtaining pyrolysis kinetic parameters and a mechanism function in multiple heating modes

    CN108985006A

  • Differential AGV trajectory tracking control method based on sparrow search algorithm

    CN112000096A

  • WSN node positioning method based on ranging optimization and improved sparrow search algorithm

    CN112714397A

  • Prediction method based on CEEMD and improved SSA-LSSVM

    CN113836801A