Control method of flow meter for silicon carbide growing equipment based on PVT method

Abstract

The invention discloses a control method of a flowmeter for equipment for growing silicon carbide based on a PVT method, which comprises a gas flow data acquisition module used for acquiring a real-time gas flow value acquired by a digital gas flowmeter; a gas analog signal noise reduction module; a gas analog signal feature extraction module; a gas digital signal feature extraction module; a different-mode gas feature fusion module; a discretization module; a differential module; and the control result generation module is used for enabling the differential feature matrix to pass through a classifier to obtain a classification result, and the classification result is used for representing whether the gas flow is increased or not. The method has the beneficial effects that the gas flow in the furnace chamber can be accurately monitored, the flow value of the gas needing to be supplemented into the furnace chamber in real time can be accurately obtained after collection and analysis, the real-time information collected by the control scheme is accurate, and the performance is reliable.

Description

Technical Field

[0001] The present invention relates to the technical field of flow meter control, and in particular to a control method for a flow meter used in a silicon carbide growing device based on a PVT method. Background Art

[0002] The PVT method, namely the atmosphere thermal chemical vapor deposition method, is a technology for preparing very pure and very large single crystal silicon materials. Its basic principle is to deposit single crystal silicon materials from a solid source through chemical reactions and transmission processes under conditions of high temperature, high pressure and inert atmosphere. At present, the equipment using the PVT method on the market all have flow meters for monitoring gas flow. The flow meter is an important component of the PVT equipment and is used to control the flow or replenishment of gas in the furnace chamber. Therefore, it is necessary to develop a control method for a flow meter for silicon carbide growth equipment based on the PVT method that is suitable for silicon carbide growth. After searching, no technical solution identical to the present invention was found. Summary of the invention

[0003] The main technical problem solved by the present invention is to provide a control method for a flow meter used in a silicon carbide growth device based on a PVT method suitable for silicon carbide growth, thereby solving one or more of the above-mentioned prior art problems.

[0004] In order to solve the above technical problems, a technical solution adopted by the present invention is: a control method for a flow meter for a silicon carbide growing device based on a PVT method, the innovation of which is: comprising:

[0005] A gas flow data acquisition module, which is used to obtain the real-time gas flow value collected by the digital gas flow meter and the real-time gas flow analog signal collected by the analog gas flow meter;

[0006] A gas simulation signal denoising module, which is used to pass the gas flow simulation signal through a denoiser based on an autoencoder to obtain a denoised gas flow simulation signal;

[0007] A gas simulation signal feature extraction module, which is used to obtain a gas flow simulation feature vector by using the de-noised gas flow simulation signal as a convolutional neural network model of a filter;

[0008] A gas digital signal feature extraction module, which is used to arrange the real-time gas flow value into an input vector and then pass it through a multi-scale neighborhood feature extraction module to obtain a multi-scale gas flow digital feature vector;

[0009] A heteromodal gas feature fusion module, which is used to use a Gaussian density map to fuse the gas flow simulation feature vector and the multi-scale gas flow digital feature vector to obtain a gas fusion Gaussian density map;

[0010] A discretization module, which is used to perform Gaussian discretization on the Gaussian distribution of each position in the gas fusion Gaussian density map to obtain a gas flow characteristic matrix;

[0011] A difference module, which is used to calculate a gas flow characteristic matrix;

[0012] A control result generation module is used to pass the differential feature matrix through a classifier to obtain a classification result, and the classification result is used to indicate whether to increase the gas flow rate.

[0013] In some embodiments, the control result generation module includes

[0014] A classification unit, the classification unit is used to use the classifier to process the corrected differential feature matrix to obtain the classification result;

[0015] A compensation unit, the compensation unit is used to perform feature compensation on the differential feature matrix based on the probability value obtained by pre-classifying the differential feature matrix through the classifier to obtain a corrected differential feature matrix;

[0016] The compensation unit performs feature compensation on the differential feature matrix using the following formula to obtain a corrected differential feature matrix, the formula being: 1 =N n ·P n-1 ·h -n·p , where N is the probability value obtained by pre-classifying the differential feature matrix P through the classifier.

[0017] In some implementations, the gas flow value includes a first gas flow value, a second gas flow value, and a third gas flow value.

[0018] In some implementations, the gas flow simulation signal includes a first gas flow simulation signal, a second gas flow simulation signal, and a third gas flow simulation signal.

[0019] In some implementations, the gas flow simulation feature vector includes a first gas flow simulation feature vector, a second gas flow simulation feature vector, and a third gas flow simulation feature vector.

[0020] In some implementations, the gas flow digital feature vector includes a first gas flow digital feature vector, a second gas flow digital feature vector, and a third gas flow digital feature vector.

[0021] In some embodiments, the gas fusion Gaussian density map is a fusion Gaussian density map of nitrogen, the first argon gas, and the second argon gas.

[0022] In some embodiments, the gas flow characteristic matrix includes flow characteristic matrices of nitrogen gas, the first argon gas, and the second argon gas.

[0023] The beneficial effects of the present invention are as follows: the technical solution can accurately monitor the gas flow in the furnace chamber, and after collection and analysis, the flow value of the gas that needs to be replenished in the furnace chamber in real time can be accurately obtained. The real-time information collected by the control solution of the technical solution is accurate and the performance is reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work, among which:

[0025] Figure 1 The present invention is a DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] Embodiment 1

[0028] like Figure 1 As shown, the embodiment of the present invention includes:

[0029] This embodiment is an embodiment of growing conductive silicon carbide;

[0030] A control method for a flow meter for a silicon carbide growing device based on a PVT method, comprising:

[0031] A gas flow data acquisition module, the gas flow data acquisition module is used to obtain the real-time nitrogen gas flow value, the first argon gas flow value and the second argon gas flow value collected by the digital gas flow meter and the real-time nitrogen gas flow, the first argon gas flow and the second argon gas flow analog signals collected by the analog gas flow meter;

[0032] A gas simulation signal denoising module, the gas simulation signal denoising module is used to pass the nitrogen gas flow, the first argon gas flow and the second argon gas flow analog signals through a denoiser based on an autoencoder to obtain denoised nitrogen gas flow, the first argon gas flow and the second argon gas flow analog signals;

[0033] A gas simulation signal feature extraction module, the gas simulation signal feature extraction module is used to obtain the nitrogen gas flow, the first argon gas flow and the second argon gas flow simulation signal simulation feature vectors after using the de-noised nitrogen gas flow, the first argon gas flow and the second argon gas flow simulation signal simulation signals as the convolutional neural network model of the filter;

[0034] A gas digital signal feature extraction module, which is used to arrange the real-time nitrogen gas flow value, the first argon gas flow value and the second argon gas flow analog signal value as an input vector and then pass it through a multi-scale neighborhood feature extraction module to obtain a multi-scale nitrogen gas flow, the first argon gas flow and the second argon gas flow analog signal digital feature vector;

[0035] A heteromodal gas feature fusion module, which is used to use a Gaussian density map to fuse the nitrogen gas flow, the first argon gas flow and the second argon gas flow analog signal simulation feature vectors and the multi-scale nitrogen gas flow, the first argon gas flow and the second argon gas flow analog signal digital feature vectors to obtain a gas fusion Gaussian density map;

[0036] A discretization module, the discretization module is used to perform Gaussian discretization on the Gaussian distribution of each position in the gas fusion Gaussian density map to obtain a characteristic matrix of nitrogen gas flow, a first argon gas flow, and a second argon gas flow simulation signal;

[0037] A differential module, the differential module is used to calculate a characteristic matrix of a nitrogen gas flow rate, a first argon gas flow rate, and a second argon gas flow rate analog signal;

[0038] A control result generation module is used to pass the differential feature matrix through a classifier to obtain a classification result, and the classification result is used to indicate whether to increase the flow rate of nitrogen, the first argon gas or the second argon gas.

[0039] The control result generating module includes

[0040] A classification unit, the classification unit is used to use the classifier to process the corrected differential feature matrix to obtain the classification result;

[0041] A compensation unit, the compensation unit is used to perform feature compensation on the differential feature matrix based on the probability value obtained by pre-classifying the differential feature matrix through the classifier to obtain a corrected differential feature matrix;

[0042] The compensation unit performs feature compensation on the differential feature matrix using the following formula to obtain a corrected differential feature matrix, the formula being: 1 =N n ·P n-1 ·h -n·p , where N is the probability value obtained by pre-classifying the differential feature matrix P through the classifier.

[0043] Embodiment 2

[0044] like Figure 1 As shown, the embodiment of the present invention includes:

[0045] This embodiment is an embodiment of growing conductive silicon carbide;

[0046] A control method for a flow meter for a silicon carbide growing device based on a PVT method, comprising:

[0047] A gas flow data acquisition module, the gas flow data acquisition module is used to obtain the real-time hydrogen gas flow value, the first argon gas flow value and the second argon gas flow value collected by the digital gas flow meter and the real-time hydrogen gas flow, the first argon gas flow and the second argon gas flow analog signals collected by the analog gas flow meter;

[0048] A gas simulation signal denoising module, the gas simulation signal denoising module is used to pass the hydrogen gas flow, the first argon gas flow and the second argon gas flow analog signals through a denoiser based on an autoencoder to obtain denoised hydrogen gas flow, the first argon gas flow and the second argon gas flow analog signals;

[0049] A gas simulation signal feature extraction module, the gas simulation signal feature extraction module is used to obtain the hydrogen gas flow, the first argon gas flow and the second argon gas flow simulation signal simulation feature vectors after using the de-noised hydrogen gas flow, the first argon gas flow and the second argon gas flow simulation signal simulation signals as the convolutional neural network model of the filter;

[0050] A gas digital signal feature extraction module, which is used to arrange the real-time hydrogen gas flow value, the first argon gas flow value and the second argon gas flow analog signal value as an input vector and then pass it through a multi-scale neighborhood feature extraction module to obtain a multi-scale hydrogen gas flow, the first argon gas flow and the second argon gas flow analog signal digital feature vector;

[0051] A heteromodal gas feature fusion module, which is used to use a Gaussian density map to fuse the hydrogen gas flow, the first argon gas flow and the second argon gas flow analog signal simulation feature vectors and the multi-scale hydrogen gas flow, the first argon gas flow and the second argon gas flow analog signal digital feature vectors to obtain a gas fusion Gaussian density map;

[0052] A discretization module, the discretization module is used to perform Gaussian discretization on the Gaussian distribution of each position in the gas fusion Gaussian density map to obtain a characteristic matrix of a hydrogen gas flow rate, a first argon gas flow rate, and a second argon gas flow rate simulation signal;

[0053] A differential module, the differential module is used to calculate the characteristic matrix of the hydrogen gas flow, the first argon gas flow and the second argon gas flow analog signal;

[0054] A control result generation module is used to pass the differential feature matrix through a classifier to obtain a classification result, and the classification result is used to indicate whether to increase the flow rate of hydrogen, the first argon gas or the second argon gas.

[0055] The control result generating module includes

[0056] A classification unit, the classification unit is used to use the classifier to process the corrected differential feature matrix to obtain the classification result;

[0057] A compensation unit, the compensation unit is used to perform feature compensation on the differential feature matrix based on the probability value obtained by pre-classifying the differential feature matrix through the classifier to obtain a corrected differential feature matrix;

[0058] The compensation unit performs feature compensation on the differential feature matrix using the following formula to obtain a corrected differential feature matrix, the formula being: 1 =N n ·P n-1 ·h -n·p , where N is the probability value obtained by pre-classifying the differential feature matrix P through the classifier.

[0059] The beneficial effects of the present invention are as follows: the technical solution can accurately monitor the gas flow in the furnace chamber, and after collection and analysis, the flow value of the gas that needs to be replenished in the furnace chamber in real time can be accurately obtained. The real-time information collected by the control solution of the technical solution is accurate and the performance is reliable.

[0060] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A control method for a flow meter for a silicon carbide growing device based on a PVT method, characterized in that: include A gas flow data acquisition module, which is used to obtain the real-time gas flow value collected by the digital gas flow meter and the real-time gas flow analog signal collected by the analog gas flow meter; A gas simulation signal denoising module, which is used to pass the gas flow simulation signal through a denoiser based on an autoencoder to obtain a denoised gas flow simulation signal; A gas simulation signal feature extraction module, which is used to obtain a gas flow simulation feature vector by using the de-noised gas flow simulation signal as a convolutional neural network model of a filter; A gas digital signal feature extraction module, which is used to arrange the real-time gas flow value into an input vector and then pass it through a multi-scale neighborhood feature extraction module to obtain a multi-scale gas flow digital feature vector; A heteromodal gas feature fusion module, which is used to use a Gaussian density map to fuse the gas flow simulation feature vector and the multi-scale gas flow digital feature vector to obtain a gas fusion Gaussian density map; A discretization module, which is used to perform Gaussian discretization on the Gaussian distribution of each position in the gas fusion Gaussian density map to obtain a gas flow characteristic matrix; A difference module, which is used to calculate a gas flow characteristic matrix; A control result generation module is used to pass the differential feature matrix through a classifier to obtain a classification result, and the classification result is used to indicate whether to increase the gas flow rate.

2. The control method of a flow meter for a silicon carbide growing device based on a PVT method according to claim 1, characterized in that: The control result generating module includes A classification unit, the classification unit is used to use the classifier to process the corrected differential feature matrix to obtain the classification result; A compensation unit, the compensation unit is used to perform feature compensation on the differential feature matrix based on the probability value obtained by pre-classifying the differential feature matrix through the classifier to obtain a corrected differential feature matrix; The compensation unit performs feature compensation on the differential feature matrix using the following formula to obtain a corrected differential feature matrix, the formula being: P1 = N n ·P n-1 ·h -n·p , where N is the probability value obtained by pre-classifying the differential feature matrix P through the classifier.

3. The control method of a flow meter for a silicon carbide growing device based on a PVT method according to claim 1, characterized in that: The gas flow value includes a first gas flow value, a second gas flow value and a third gas flow value.

4. The control method of a flow meter for a silicon carbide growing device based on a PVT method according to claim 1, characterized in that: The gas flow simulation signal includes a first gas flow simulation signal, a second gas flow simulation signal and a third gas flow simulation signal.

5. The control method of a flow meter for a silicon carbide growing device based on a PVT method according to claim 1, characterized in that: The gas flow simulation feature vector includes a first gas flow simulation feature vector, a second gas flow simulation feature vector and a third gas flow simulation feature vector.

6. The control method of a flow meter for a silicon carbide growing device based on a PVT method according to claim 1, characterized in that: The gas flow digital feature vector includes a first gas flow digital feature vector, a second gas flow digital feature vector and a third gas flow digital feature vector.

7. The control method of a flow meter for a silicon carbide growing device based on a PVT method according to claim 1, characterized in that: The gas fusion Gaussian density map is a fusion Gaussian density map of nitrogen, the first argon gas and the second argon gas.

8. The control method of a flow meter for a silicon carbide growing device based on a PVT method according to claim 1, characterized in that: The gas flow characteristic matrix includes flow characteristic matrices of nitrogen, the first argon gas and the second argon gas.

Images