Data set construction method for particle size distribution neural network inversion algorithm

Abstract

The application provides a dataset construction method for a particle size distribution neural network inversion algorithm, comprising the following steps: obtaining part of basic data; calculating upper and lower limits and a center particle size of a particle size measurement range, and further obtaining a light energy distribution coefficient matrix; assuming that particle sizes meet Rosin-Rammler distribution, determining upper and lower limits of characteristic parameters, and generating a first sequence of the characteristic parameters; calculating a first particle size distribution vector and a first light energy distribution vector, so as to generate a unimodal distribution dataset; generating a second sequence of the characteristic parameters in a uniform distribution manner within the upper and lower limits of the characteristic parameters; obtaining a second and third particle size distribution vector, and further obtaining a second and third light energy distribution vector; and generating a second sequence of the characteristic parameters in a uniform distribution manner within the upper and lower limits of the characteristic parameters. The beneficial effect of the application is that, through accurate calculation and reasonable parameter setting, the generated dataset can accurately reflect the real particle size distribution, and is helpful to improve the learning efficiency and prediction accuracy of the neural network model.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of computer, and particularly relates to a data set construction method for a particle size distribution neural network inversion algorithm. BACKGROUND

[0002] Particles generally refer to geometric bodies with specific shapes and sizes ranging from nanometers to millimeters. One of the most important physical properties of particles is particle size distribution, which is closely related to many fields such as combustion, firefighting, metallurgy, petroleum, chemical industry, materials, medicine, and environmental protection. Therefore, it is very important to measure the particle size distribution. Laser particle size analyzers are particle size measuring instruments based on light scattering principle, which need to obtain the particle size distribution from the scattered light energy distribution measured by multiple annular photodetectors. Essentially, it is to solve the first kind of Fredholm integral, which belongs to a typical ill-posed problem and has poor noise resistance. Common inversion algorithms include nonlinear iterative algorithm, least squares iterative algorithm, singular value decomposition algorithm, and BP neural network algorithm. At present, the commonly used algorithms have problems such as low inversion accuracy, long calculation time, and unstable calculation. With the rapid development of artificial intelligence technology, deep learning network models have strong nonlinear regression ability, generalization ability, and parallel computing ability, and have been widely applied. If the neural network model is applied to particle size inversion, a data set containing a large number of particle size distributions and corresponding light energy distributions is needed to supervise the training and testing of the network model. The data set should have full coverage of the measurement range, scalability, and support for the strong generalization ability of the network model. In order to solve the above problems and promote the development of this research field, it is of great significance to construct a data set for the particle size distribution neural network inversion algorithm. SUMMARY

[0003] To solve the above technical problems, the present application provides a data set construction method for a particle size distribution neural network inversion algorithm, which is particularly suitable for constructing a data set for a particle size distribution neural network inversion algorithm according to the principle of a laser particle size analyzer, structural parameters, and a specific particle size distribution function.

[0004] The technical solution adopted by the present application is: in the first aspect, a data set construction method for a particle size distribution neural network inversion algorithm is provided, comprising:

[0005] Obtaining the laser wavelength, Fourier lens focal length, and inner and outer radii of each ring of the multi-element photodetector from the laser particle size analyzer;

[0006] Calculating the upper and lower limits and the center particle size of the particle size measurement range corresponding to each ring of the multi-element photodetector based on the obtained data, and further calculating the light energy distribution coefficient matrix of the laser particle size analyzer based on the results;

[0007] Assuming that the particle size satisfies the Rosin-Rammler distribution, the upper limit and the lower limit of the characteristic parameter are determined in combination with the center particle size corresponding to each ring of the detector, and a first sequence of the characteristic parameter is generated within the interval;

[0008] A first particle size distribution vector is calculated according to the first sequence of the characteristic parameter and the upper and lower limits of the particle size measurement range corresponding to each ring of the detector, and a corresponding first light energy distribution vector is further calculated based on the result, so as to generate a unimodal distribution data set;

[0009] A second sequence of the characteristic parameter is generated in a uniform distribution manner within the upper limit and the lower limit of the characteristic parameter;

[0010] A second particle size distribution vector and a third particle size distribution vector are calculated according to the second sequence of the characteristic parameter and the upper and lower limits of the particle size measurement range corresponding to each ring of the detector, and a corresponding second light energy distribution vector and a third light energy distribution vector are further calculated based on the result;

[0011] A mixing ratio factor is set, and the multi-modal distribution data set is converted into a mixed distribution data set after being generated.

[0012] Further, the upper and lower limits of the particle size measurement range and the center particle size corresponding to each ring of the multi-element photoelectric detector are respectively solved by formula j=1,2,…,K;l=1,2 and , wherein d j,l represents the upper and lower limits of the particle size measurement range corresponding to each ring of the detector, d j represents the center particle size corresponding to each ring of the detector, λ represents the laser wavelength of the laser particle size analyzer, f represents the focal length of the Fourier lens, (r i,1 ,r i,2 ), i=1,2,…,K represents the inner and outer radii of each ring of the multi-element photoelectric detector.

[0013] Further, the light energy distribution coefficient matrix of the laser particle size analyzer is solved according to formula j=1,2,…,K, wherein t i,j represents each element in the light energy distribution coefficient matrix of the laser particle size analyzer, J0 and J1 are the 0th and 1st Bessel functions; X j,i,1 =πd j r i,1 / λf, X j,i,2 =πd j r i,2 / λf.

[0014] Further, the first sequence of characteristic parameters comprises a first sequence of particle size and a first sequence of distribution parameters, wherein the first sequence of particle size and the first sequence of distribution parameters satisfy the arithmetic progression rule or the parabolic function rule.

[0015] Further, the first particle size distribution vector and the first light energy distribution vector are solved by the following equations

[0016] and represents the first particle size distribution vector, represents the first light energy distribution vector, De m represents any value in the first sequence of particle size, N m represents any value in the first sequence of distribution parameters.

[0017] Further, the second sequence of characteristic parameters comprises a second sequence of particle size, a third sequence of particle size, a second sequence of distribution parameters and a third sequence of distribution parameters, wherein the second sequence of particle size and the third sequence of particle size have no same value, and the second sequence of distribution parameters and the third sequence of distribution parameters have no same value.

[0018] Further, the multi-peak distribution dataset is solved by the following equation l = (m-1) x P + p, p = 1, 2, …, P, wherein μ p represents the mixing ratio factor, represents the second particle size distribution vector, represents the third particle size distribution vector, represents the second light energy distribution vector, represents the third light energy distribution vector; is the multi-peak distribution dataset, (W M ,E M ) + (W 1 ,E 1 ) + (W 2 ,E 2 ) is the mixed distribution dataset.

[0019] In a third aspect, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the data set construction method for the particle size distribution neural network inversion algorithm provided by the present disclosure.

[0020] ​In a fourth aspect, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause a computer to execute the data set construction method for the particle size distribution neural network inversion algorithm provided in the present disclosure.

[0021] In a fifth aspect, a computer program product is provided, comprising computer programs / instructions for executing the data set construction method for the particle size distribution neural network inversion algorithm provided in the present disclosure when executed by a processor.

[0022] The present application has the advantages and positive effects that: due to the above technical solution, through accurate calculation and reasonable parameter setting, the generated data set can more accurately reflect the real particle size distribution, which helps to improve the learning efficiency and prediction accuracy of the neural network model; by generating unimodal, multimodal and mixed distribution data sets, the trained model can handle more complex particle size distribution scenarios, enhancing the generalization ability and practicality of the model; the design rules of the feature parameter sequence are defined to ensure that the generated data has certain diversity and rationality. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 is a flowchart of the data set construction method for the particle size distribution neural network inversion algorithm according to an embodiment of the present application

[0024] Figure 2 is a schematic diagram of the first particle size sequence according to the parabolic law discrete value according to an embodiment of the present application

[0025] Figure 3 is a narrow unimodal R-R distribution schematic diagram according to an embodiment of the present application

[0026] Figure 4 is a narrow unimodal particle size distribution corresponding to the light energy distribution schematic diagram according to an embodiment of the present application

[0027] Figure 5 is a wide unimodal R-R distribution schematic diagram according to an embodiment of the present application

[0028] Figure 6 is a wide unimodal particle size distribution corresponding to the light energy distribution schematic diagram according to an embodiment of the present application

[0029] Figure 7 is a non-obvious bimodal particle size distribution schematic diagram composed of a certain proportion of unimodal R-R distribution according to an embodiment of the present application

[0030] Figure 8 is a non-obvious bimodal particle size distribution corresponding to the light energy distribution schematic diagram composed of a certain proportion of unimodal R-R distribution according to an embodiment of the present application

[0031] Figure 9is a schematic diagram of an obvious bimodal particle size distribution formed by mixing a unimodal R-R distribution according to an embodiment of the present application according to a certain proportion

[0032] Figure 10 is a schematic diagram of the light energy distribution corresponding to the obvious bimodal particle size distribution formed by mixing a unimodal R-R distribution according to an embodiment of the present application according to a certain proportion DETAILED DESCRIPTION

[0033] The present disclosure will be described more fully hereinafter with reference to the accompanying drawings, in which example embodiments of the present disclosure are described. The technical solutions in the embodiments of the present disclosure will be described clearly and completely below with reference to the accompanying drawings of the embodiments of the present disclosure. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present disclosure.

[0034] As Figure 1 indicated, the present application provides a data set construction method for a particle size distribution neural network inversion algorithm, comprising:

[0035] S100, obtaining laser wavelength, Fourier lens focal length and inner and outer radii of each ring of a multi-element photoelectric detector by a laser particle size analyzer;

[0036] S200, calculating the upper and lower limits and the center particle size of the particle size measurement range corresponding to each ring of the multi-element photoelectric detector according to the obtained data, and further calculating a light energy distribution coefficient matrix of the laser particle size analyzer based on the results;

[0037] Based on system parameters such as laser wavelength λ, Fourier lens focal length f and inner and outer radii (r i,1 ,r i,2 ) of each ring of the multi-element photoelectric detector, i = 1, 2, …, K, the upper and lower limits of the particle size measurement range corresponding to each ring of the detector are calculated according to the formula, and then the center particle size corresponding to each ring is calculated according to the formula. Based on the center particle size d j of each ring of the detector and the above system parameters, the elements in the light energy distribution coefficient matrix T are calculated according to the Fraunhofer diffraction theory using the formula, so as to obtain the light energy distribution coefficient matrix T.

[0038] S300, assuming that the particle size satisfies the Rosin-Rammler distribution, determining the upper and lower limits of the characteristic parameters in combination with the center particle size corresponding to each ring of the detector, and generating a first sequence of the characteristic parameters within the interval;

[0039] Rosin-Rammler distribution is a statistical model used to describe particle size distribution, which is common in fields such as materials science and mineral processing engineering. It is a continuous probability distribution, particularly suitable for describing particle groups with a wide distribution. The upper and lower limits of the characteristic parameters are set according to the central particle size dj and RR distribution characteristics corresponding to the multivariate detector. The characteristic parameters are the particle size De and the distribution parameter N. Within the upper and lower limits of the characteristic parameters (that is, [De min ,De max ]、[N min ,N max ] interval) generates the first sequence of characteristic parameters according to certain rules, that is, the sequence of the first particle size De and the sequence of the first distribution parameter N.

[0040] S400, calculating a first particle size distribution vector based on the first sequence of characteristic parameters and the upper and lower limits of the particle size measurement range corresponding to each ring of the detector, and further calculating a corresponding first light energy distribution vector based on the result, thereby generating a unimodal distribution data set;

[0041] Based on the first sequence of characteristic parameters m and N m And the upper and lower limits of particle size d corresponding to each ring of the detector j,l The first particle size distribution vector is obtained by applying the formula m=1,2,…,M, get the first particle size distribution vector Then, combined with the light energy distribution coefficient matrix T of the laser particle size analyzer, the formula is obtained: The corresponding first light energy distribution vector m=1,2,…,M, according to the above results That is, the RR unimodal distribution data set.

[0042] S500, generating a second sequence of characteristic parameters in a uniformly distributed manner within the upper limit and lower limit of the characteristic parameters;

[0043] In the characteristic parameter [De min ,De max ]、[N min ,N max ] interval according to the uniform distribution to generate the second sequence of characteristic parameters, including the second particle size De 1 Sequence, third particle size De 2 Sequence, second distribution parameter N 1 Sequence and third distribution parameter N 2 sequence.

[0044] S600, calculating the second particle size distribution vector and the third particle size distribution vector according to the second sequence of characteristic parameters and the upper and lower limits of the particle size measurement range corresponding to each ring of the detector, and further calculating the corresponding second light energy distribution vector and the third light energy distribution vector based on the results;

[0045] Based on the first sequence of characteristic parameters and and the upper and lower limits of the particle size corresponding to each ring of the detector j,l The second particle size distribution vector m=1,2,…,MM and the third particle size distribution vector m=1,2,…,MM are obtained by applying the formula and the third particle size distribution vector The second light energy distribution vector m=1,2,…,M and the third light energy distribution vector m=1,2,…,M corresponding to the second particle size distribution vector and the third particle size distribution vector are obtained by combining the light energy distribution coefficient matrix T of the laser particle size analyzer with the formula

[0046] S700, set the mixing ratio factor, generate the multi-peak distribution data set and convert it into a mixed distribution data set.

[0047] Set the mixing ratio factor μ: {μ p ,p=1,2,…,P}, generate the multi-peak distribution data set l=1,2,…,P×MM, convert it into a mixed distribution data set (W M ,E M )+(W 1 ,E 1 )+(W 2 ,E 2 ). Through accurate calculation and reasonable parameter setting, the generated data set can more accurately reflect the real particle size distribution, which helps to improve the learning efficiency and prediction accuracy of the neural network model; by generating single-peak, multi-peak and mixed distribution data sets, the training model can handle more complex particle size distribution scenarios, enhancing the generalization ability and practicality of the model.

[0048] In an embodiment, the upper and lower limits of the particle size measurement range and the center particle size corresponding to each ring of the multi-element photoelectric detector are solved by the formula l=1,2 and , where d j,l represents the upper and lower limits of the particle size measurement range corresponding to each ring of the detector, d j represents the center particle size corresponding to each ring of the detector, λ represents the wavelength of the laser particle size analyzer, f represents the focal length of the Fourier lens, and (r i,1 ,ri,2 ), i = 1, 2, …, K represent the inner and outer radii of each ring of the multi-element photodetector.

[0049] In an embodiment, the laser particle size analyzer light energy distribution coefficient matrix is according to the formula Solve for j = 1, 2, …, K, where t i,j represents each element in the laser particle size analyzer light energy distribution coefficient matrix, J0, J1 are the 0th and 1st Bessel functions; X j,i,1 = πd j r i,1 / λf, X j,i,2 = πd j r i,2 / λf.

[0050] In an embodiment, as shown in Figure 2 , the first sequence of characteristic parameters includes a first particle size sequence and a first distribution parameter sequence, wherein the first particle size sequence and the first distribution parameter sequence need to satisfy the arithmetic progression rule or the parabolic function rule.

[0051] The design rules of the sequence of characteristic parameters are defined to ensure that the generated data has a certain diversity and rationality.

[0052] In an embodiment, the first particle size distribution vector and the first light energy distribution vector are respectively solved by the formulas

[0053] and , wherein represents the first particle size distribution vector, represents the first light energy distribution vector, De m represents any value in the first particle size sequence, N m represents any value in the first distribution parameter sequence.

[0054] In an embodiment, the second sequence of characteristic parameters includes a second particle size sequence, a third particle size sequence, a second distribution parameter sequence, and a third distribution parameter sequence, wherein there are no identical values in the second particle size sequence and the third particle size sequence, and there are no identical values in the second distribution parameter sequence and the third distribution parameter sequence.

[0055] In an embodiment, the multi-peak distribution data set is solved by the formula l = (m-1) × P + p, p = 1, 2, …, P, wherein μ p represents the mixing ratio factor, represents the second particle size distribution vector, represents the third particle size distribution vector, represents the second light energy distribution vector, represents the third light energy distribution vector; is a multimodal distribution data set, (W M ,E M )+(W 1 ,E 1 )+(W 2 ,E 2 ) is a mixed distribution dataset.

[0056] In a specific embodiment, Figures 3-10 As shown, the wavelength of the laser particle size analyzer is λ = 632.8 nm. The focal length f of the Fourier lens has four optional values, namely 300 mm, 500 mm, 800 mm and 1000 mm. Here, f = 300 mm is used. The multi-element detector has a total of 32 semicircular ring detector units, that is, K = 32. The inner and outer radius data of the detector unit are obtained, and the light energy distribution coefficient matrix T can be calculated according to the formula 32×32 The upper and lower limits of the particle size corresponding to each detector ring are calculated according to the formula based on the inner and outer radius values ​​of the detection unit, and the first characteristic size parameter De in the RR distribution is determined accordingly. min and De max According to the RR distribution characteristics and the particle size distribution in actual applications, the first distribution characteristic parameter N of the RR distribution is determined. min and N max Set the number of data pairs in the unimodal distribution dataset, M (this can be increased or decreased based on model training, e.g., M = 20,000), and follow the steps in S400 to construct the unimodal distribution dataset. Set the mixing scale factors μ = {0.2, 0.4, 0.5, 0.6, 0.8}, i.e., P = 5 and MM = 20,000, and follow the steps in S500-S700 to construct the mixed distribution dataset. After these operations, the dataset now contains 20,000 + 20,000 + 5 × 20,000 = 140,000 data pairs. Figures 2-8 Visualization examples of particle size distribution and light energy distribution are given, including narrow unimodal RR distribution, broad unimodal RR distribution, obvious bimodal distribution and non-obvious bimodal distribution. Figure 3 Narrow unimodal RR distribution with characteristic parameters De = 70.01, N = 10.83, Figure 4 The light energy distribution corresponding to the narrow single-peak particle size distribution with characteristic parameters De=70.01 and N=10.83 is Figure 5 A broad unimodal RR distribution with characteristic parameters De = 35.33 and N = 1.6, Figure 6 The light energy distribution corresponding to the broad single-peak particle size distribution with characteristic parameters De=35.33 and N=1.6, Figure 7The non-obvious bimodal particle size distribution composed of the unimodal R-R distribution with characteristic parameters De=35.17, N=4.17 and De=53.8, N=1.3 mixed according to 0.8:0.2, Figure 8 The light energy distribution corresponding to the non-obvious bimodal particle size distribution composed of the unimodal R-R distribution with characteristic parameters De=35.17, N=4.17 and De=53.8, N=1.3 mixed according to 0.8:0.2, Figure 9 The obvious bimodal particle size distribution composed of the unimodal R-R distribution with characteristic parameters De=70.17, N=10.19 and De=41.93, N=7.77 mixed according to 0.5:0.5, Figure 10 The light energy distribution corresponding to the obvious bimodal particle size distribution composed of the unimodal R-R distribution with characteristic parameters De=70.17, N=10.19 and De=41.93, N=7.77 mixed according to 0.5:0.5.

[0057] Based on the embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0058] An electronic device includes at least one processor; and a memory connected with the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method for uniform coding of park equipment assets collected by Internet of Things provided by the present disclosure.

[0059] The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not intended to limit implementations of the present disclosure described and / or claimed in this document.

[0060] A non-transitory computer readable storage medium storing computer instructions, wherein the computer instructions are used to make a computer execute the method for uniform coding of park equipment assets collected by Internet of Things provided by the present disclosure.

[0061] The various implementations in this disclosure can be implemented in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a special-purpose standard product (ASSP), a system on a chip (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0062] A computer program product comprising computer programs / instructions for implementing the method of the present disclosure.

[0063] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces the functions / operations specified in the flowcharts and / or the block diagrams. The program code can be entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine or server, or entirely on a remote machine or server.

[0064] In the context of the present disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0065] The above detailed description of the embodiments of the present application is only preferred embodiments of the present application, and should not be considered as limiting the scope of the present application. Any equivalent changes and improvements made according to the scope of the present application should still belong to the patent coverage of the present application.

Claims

1. A method for constructing a data set for a neural network inversion algorithm for particle size distribution, characterized in that: include: Obtain the laser wavelength, the focal length of the Fourier lens and the inner and outer radii of each ring of the multi-element photoelectric detector from the laser particle size analyzer; The upper and lower limits of the particle size measurement range and the center particle size corresponding to each ring of the multi-element photoelectric detector are calculated based on the acquired data, and the light energy distribution coefficient matrix of the laser particle size analyzer is further calculated based on the results; Assuming that the particle size satisfies the Rosin-Rammler distribution, the upper and lower limits of the characteristic parameters are determined in combination with the central particle size corresponding to each ring of the detector, and the first sequence of characteristic parameters is generated within the interval; A first particle size distribution vector is calculated based on the first sequence of characteristic parameters and the upper and lower limits of the particle size measurement range corresponding to each ring of the detector, and a corresponding first light energy distribution vector is further calculated based on the result, thereby generating a unimodal distribution data set; generating a second sequence of characteristic parameters in a uniformly distributed manner within an upper limit and a lower limit of the characteristic parameters; A second particle size distribution vector and a third particle size distribution vector are calculated based on the second sequence of characteristic parameters and the upper and lower limits of the particle size measurement range corresponding to each ring of the detector, and a corresponding second light energy distribution vector and a third light energy distribution vector are further calculated based on the results; Set the mixing scale factor, generate a multimodal distribution data set, and convert it into a mixed distribution data set.

2. The construction method according to claim 1, wherein: The upper and lower limits of the particle size measurement range and the central particle size corresponding to each ring of the multi-element photoelectric detector are respectively expressed by the formula and Solve, where d j,l Represents the upper and lower limits of the particle size measurement range corresponding to each ring of the detector, d j represents the central particle size corresponding to each ring of the detector, λ represents the laser wavelength of the laser particle size analyzer, f represents the focal length of the Fourier lens, (r i,1 ,r i,2 ), i = 1, 2, ..., K represents the inner and outer radii of each ring of the multi-element photodetector.

3. The construction method according to claim 2, wherein: The light energy distribution coefficient matrix of the laser particle size analyzer is based on the formula Solve, where t i,j represents the elements in the laser particle size analyzer light energy distribution coefficient matrix, J0 and J1 are 0th and 1st order Bessel functions; X j,i,1 =πd j r i,1 / λf,X j,i,2 =πd j r i,2 / λf.

4. The construction method according to claim 1, wherein: The first sequence of characteristic parameters includes a first particle size sequence and a first distribution parameter sequence, wherein the first particle size sequence and the first distribution parameter sequence need to satisfy an arithmetic progression law or a parabolic function law.

5. The construction method according to claim 4, characterized in that: The first particle size distribution vector and the first light energy distribution vector are respectively expressed by and Solve the equation, represents the first particle size distribution vector, Represents the first light energy distribution vector, De m Represents any value in the first particle size sequence, N m represents any value in the sequence of first distribution parameters.

6. The construction method according to claim 1, wherein: The second sequence of characteristic parameters includes a second particle size sequence, a third particle size sequence, a second distribution parameter sequence and a third distribution parameter sequence, wherein the second particle size sequence and the third particle size sequence do not have the same value, and the second distribution parameter sequence and the third distribution parameter sequence do not have the same value.

7. The construction method according to claim 6, characterized in that: Multimodal distribution data sets are obtained by l=(m-1)×P+p,p=1,2,…,P to solve, where μ p represents the mixing scale factor, represents the second particle size distribution vector, represents the third particle size distribution vector, represents the second light energy distribution vector, represents the third light energy distribution vector; For a multimodal distribution dataset, It is a mixed distribution dataset.

8. An electronic device comprising: at least one processor; as well as a memory communicatively connected to at least one processor; wherein, The memory stores instructions that can be executed by at least one processor. The instructions are executed by the at least one processor so that the at least one processor can perform the method according to any one of claims 1 to 7.

9. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause a computer to execute the method according to any one of claims 1 to 7.

10. A computer program product comprising a computer program / instructions, which, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.

Images