Polysilicon reduction furnace control method based on positive mechanism model multi-control variables

Abstract

The present application relates to a multi-control variable control method of a polycrystalline silicon reduction furnace based on a positive mechanism model, comprising the following steps: S1, collecting and calculating the reaction state, i.e. the real-time deposition rate and the heat loss rate of the reaction; S2, inferring the control parameters: determining the control parameters in the PID control process based on the difference between the real-time deposition rate curve and the target deposition rate curve, wherein the real-time deposition rate curve is generated based on the real-time deposition rate of the reaction progress sequence, and the target deposition rate curve is a preset curve; S3, performing the PID control process: using the control parameters in step S2 to perform PID control on the hydrogen input and the reaction current; the present method determines the optimal deposition rate curve through data analysis, and calculates the deposition rate in real time during the control process. The current heat proportion in the furnace is calculated by using the thermodynamic formula, and the hydrogen and the current are precisely regulated and controlled by combining the proportion with the deposition rate difference, which significantly improves the stability and accuracy of the production process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of polycrystalline silicon reduction control, and particularly relates to a polycrystalline silicon reduction furnace control method based on a positive mechanism model and multiple control variables. BACKGROUND

[0002] Polycrystalline silicon is the core raw material for the manufacture of solar photovoltaic devices, and the optimization and efficiency improvement of its production process have a profound impact on the development of the entire renewable energy industry. In the current global energy transition and low-carbon economy context, the photovoltaic industry is rapidly expanding, and the demand for polycrystalline silicon is also surging. Therefore, improving the production efficiency and quality of polycrystalline silicon while reducing its production cost is a key issue in the industry.

[0003] The improved Siemens method is currently the most commonly used method for producing polycrystalline silicon. It uses chemical vapor deposition (CVD) technology to react hydrogen and trichlorosilane (SiHCl3) at high temperatures to deposit silicon on a silicon rod, forming polycrystalline silicon. This process needs to be carried out in a high-temperature reduction furnace, and accurate control of the furnace temperature is crucial to ensure the quality and production efficiency of polycrystalline silicon. However, due to technical and equipment limitations, it is not practical to install sensors inside the furnace, and sensors installed outside the furnace cannot accurately measure the temperature, which poses a great challenge to temperature control.

[0004] Traditional control methods usually rely on the experience and observation of operators, who judge the reaction inside the furnace by observing power curves, tail gas temperatures, voltage curves, etc., and then adjust hydrogen, current, etc. based on manual experience. However, these curves are all apparent reaction curves of the reduction furnace and are not core parameter indicators. Moreover, manual observation is subjective and error-prone, which can lead to unstable and inaccurate control, affecting the yield and quality of polycrystalline silicon. SUMMARY

[0005] In view of the fact that existing polycrystalline silicon reduction control processes rely on apparent curves such as voltage, current, and power, the control process is unstable, and it is difficult to reliably control the polycrystalline silicon yield, quality, and single power consumption of the reduction process, the present application provides a polycrystalline silicon reduction furnace control method based on a positive mechanism model and multiple control variables.

[0006] The technical solution of the present application provides a polycrystalline silicon reduction furnace control method based on a positive mechanism model and multiple control variables, comprising the following steps.

[0007] S1, collect and calculate the reaction state: the reaction state includes the real-time deposition rate ΔM of the reaction and the heat loss rate ΔQ of the reaction; the real-time deposition rate ΔM is the change rate of the mass M of the polycrystalline silicon rod in the reduction furnace, and the heat loss rate ΔQ is the ratio of the heat loss Q of the reaction to the total energy E input;

[0008] S2, inferring control parameters: determining control parameters in a PID control process based on a difference between a real-time deposition rate curve and a target deposition rate curve, the control parameters including a proportional adjustment parameter P, an integral adjustment parameter I, and a differential adjustment parameter D; wherein the real-time deposition rate curve is generated based on real-time deposition rates AM of a reaction progress sequence, and the target deposition rate curve is a preset curve;

[0009] S3, performing a PID control process: performing PID control on the hydrogen input quantity and the reaction current of the reaction using the control parameters in step S2; hydrogen input adjustment quantity ; reaction current adjustment quantity ; k p1 , k i1 , k d1 , k p2 , k i2 , k d2 are control coefficients; wherein at least one of k p2 , k i2 , k d2 has a preset value set, and its value is related to the heat loss AQ of the reaction.

[0010] Preferably, in the S1 collecting and calculating reaction state step, the mass M of the polysilicon rod is:

[0011]

[0012] wherein m is the number of current loops of the polysilicon reduction furnace, M i is the mass of the polysilicon rod of the i-th phase loop, calculated by the following formula:

[0013]

[0014] wherein p2 is the resistivity of the silicon rod, L is the total length of the silicon rod, n i is the logarithm of the silicon rod of the i-th phase loop calculated in the reduction furnace; I i is the current of the i-th phase loop; U i is the voltage of the i-th phase loop; p is the density of the polysilicon rod.

[0015] Preferably, in the S1 collecting and calculating reaction state step, the mass M or deposition rate AM of the polysilicon rod is subjected to Gaussian filtering processing.

[0016] Preferably, in the S1 collecting and calculating reaction state step, the heat loss Q of the reaction is the heat of the tail gas, and the total energy E is the input electric energy:

[0017]

[0018] wherein P is the input electric power, and T is the time difference.

[0019] Preferably, in the reaction state collecting and calculating step S1, the heat loss Q of the reaction is calculated as follows:

[0020] According to the chemical reaction equation in the production process, the heat formula of each chemical reaction in the reaction process can be obtained as follows:

[0021]

[0022]

[0023]

[0024]

[0025]

[0026] Wherein:

[0027] Cp is the specific heat capacity of TCS, is the average mass flow of TCS in the corresponding time, is the mass of the silicon rod at t2, is the mass of the silicon rod at t1, is the relative atomic mass of elemental silicon, is the relative molecular mass of silicon tetrachloride, is the relative molecular mass of trichlorosilane, is the temperature difference between the tail gas temperature and the feed temperature in the corresponding time; Cp is the specific heat capacity of gaseous hydrogen chloride, Cp is the specific heat capacity of gaseous silicon tetrachloride, Cp is the specific heat capacity of hydrogen, is the average mass flow of hydrogen in the corresponding time, is the relative molecular mass of hydrogen chloride.

[0028] Preferably, in the control parameter inferring step S2, the target deposition rate curve is synthesized based on the historical deposition rate curve that meets the preset requirement of single power consumption qualification.

[0029] Preferably, in the control parameter inferring step S2, the polysilicon reduction process is divided into several stages, and in each stage, the subsegment of the historical deposition rate curve that is locally optimal in single power consumption in the stage is selected, and then the subsegments obtained in each stage are spliced to synthesize the target deposition rate curve.

[0030] Preferably, in the control parameter inferring step S2, the optimal deposition rate curve segment at each time is sequentially connected to obtain the final target deposition rate curve.

[0031] Preferably, in the S2 step of inferring the control parameter, the target deposition rate curve is , the real-time deposition rate record is , the current time is , the historical data of the previous time step is , the proportional adjustment parameter is , the integral adjustment parameter is , and the differential adjustment parameter is .

[0032] Preferably, in the S3 step of performing the PID control process, the hydrogen input adjustment amount is ; the reaction current adjustment amount is .

[0033] For the hydrogen input adjustment amount, the proportional adjustment coefficient is , the integral adjustment coefficient is , and the differential adjustment coefficient is .

[0034] For the reaction current adjustment amount, the proportional adjustment coefficient is , and the integral adjustment coefficient is , when is between 40% and 50%, , when is less than 40%, , when is greater than 50%, .

[0035] The present application provides a new control strategy with deposition rate calculated from current and voltage as the control target. The method determines the optimal deposition rate curve through data analysis, and calculates the deposition rate in real time during the control process. The current heat proportion in the furnace is calculated using thermodynamic formula, and this proportion is combined with the deposition rate difference to input into the PID control module for precise regulation of hydrogen and current. This automatic control strategy adjusts both current and hydrogen, rather than adjusting a single variable. The stability and accuracy of the production process are significantly improved, thereby ultimately improving the yield and quality of polysilicon and ensuring a small power consumption. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is the flowchart of the polysilicon reduction furnace control method based on the forward mechanism model and multiple control variables of the present application.

[0037] Figure 2 is the average value curve data of the deposition rate obtained by grouping the historical data in the present application. DETAILED DESCRIPTION

[0038] The present application will be described in detail below in conjunction with the accompanying drawings and specific embodiments. In the present description, the size proportions of the drawings do not represent actual size proportions, which are only used to reflect the relative positional relationship and connection relationship between components. Components with the same name or the same reference numeral represent similar or identical structures, and are only for illustrative purposes.

[0039] In the reduction control process of polysilicon, the process parameters are positively fed back to realize the optimized control of the quality of polysilicon and the power consumption by combining the deposition rate and the variable reaction heat. As shown in FIG. 1, on the basis of a preset process scheme, the general control flow in one control cycle includes the following steps. Figure 1

[0040] S1, collecting and calculating the reaction state. The reaction state includes the real-time deposition rate AM of the reaction and the heat loss rate AQ of the reaction. The real-time deposition rate AM is the change rate of the mass M of the polysilicon rod in the reduction furnace. The heat loss rate AQ is the ratio of the heat loss Q of the reaction to the total energy E input. The heat loss mainly refers to the heat carried away by the tail gas.

[0041] S2, inferring the control parameters. The control parameters in the PID control process are determined based on the difference between the real-time deposition rate curve and the target deposition rate curve. The control parameters include the proportional adjustment parameter P, the integral adjustment parameter I and the differential adjustment parameter D. The real-time deposition rate curve is generated based on the real-time deposition rate AM of the reaction process sequence, and the target deposition rate curve is a preset curve.

[0042] S3, performing the PID control process. The hydrogen input amount and the reaction current of the reaction are PID controlled by using the control parameters in step S2. The hydrogen input adjustment amount ; the reaction current adjustment amount . k p1 , k i1 , k d1 , k p2 , k i2 , k d2 are control coefficients. At least one of k p2 , k i2 , k d2 has a preset value set, and the value thereof is related to the heat loss rate AQ of the reaction.

[0043] The following is one of the specific embodiments of the above control flow.

[0044] S1, collecting and calculating the reaction state.

[0045] ​During the operation of the reduction furnace, the amount of feedstock and the voltage and current of the reaction are generally controlled by a preset process scheme. The temperature, flow rate, and power-related parameters are usually collected and recorded in real time. On this basis, the deposition rate and heat loss of the reaction can be inferred.

[0046] Specifically, the deposition rate ΔM is the change in the mass M of the polysilicon rod over time, i.e. The mass M of the polysilicon in the reduction furnace cannot be directly calculated, but it can be estimated more accurately based on the collected parameters.

[0047] As the cross-sectional area of the polysilicon increases during deposition and growth, the resistance decreases, and the electrical characteristics of the reduction furnace change accordingly. However, according to experience, the resistivity of polysilicon does not change much at a certain temperature. To ensure the quality of the output material, the polysilicon in the furnace is generally controlled at around 1050°C. Therefore, during the reduction of polysilicon, the resistivity of the polysilicon rod can be considered to be essentially constant, so the volume V of the polysilicon rod in the single-phase loop in the reduction furnace can be preliminarily inferred as i .

[0048]

[0049] where ρ2 is the resistivity of the silicon rod, which can be considered a constant value. When heated to 1050°C (1050°C is the optimal temperature for the reduction reaction of polysilicon), the resistivity is , and the temperature change is not large, so the resistivity is assumed to be constant. In the formula, 1000 is a dimensional difference, which degenerates to 1 when based on the International System of Units. L is the total length of the silicon rod, including the length of the beam at the top connection of the silicon rod, which can be determined before the reaction begins.

[0050] For a certain reduction furnace, in order to fully utilize the space of the reduction furnace and improve the utilization rate, multiple current loops can be set up in the same reduction furnace, and each current loop is a single-phase loop as described below. In each single-phase loop, multiple polysilicon rods can also be connected in series to achieve simultaneous deposition.

[0051] n i is the number of silicon rods in the single-phase loop of the reduction furnace; I i is the current of the i-th single-phase loop (A); U i is the voltage of the i-th single-phase loop (V).

[0052] In order to obtain the total mass of the polysilicon rods in the reduction furnace, the masses of the polysilicon rods in all single-phase loops need to be added up, and we determine the mass M i of the polysilicon rod in each single-phase loop, which obviously satisfies Where ρ is the density of the polycrystalline silicon rod, which is generally considered a constant value, and we ignore density differences caused by the quality of the polycrystalline silicon raw material. Depending on the specific design of the reduction furnace, it has a fixed number of loops m, for example, m=6 in the embodiment of this application. Then the mass M of the polycrystalline silicon rod is:

[0053]

[0054] The first derivative of the polysilicon quality with respect to time, calculated based on the collected data, is the desired deposition rate. .

[0055] In actual calculations, the deposition rate can be discretized as the ratio of the deposition rate to the time step within a certain time step. For example, if the time step is 20 minutes, then:

[0056]

[0057] Among them, M 0min For the current quality of polycrystalline silicon rods, M -20min The mass of the silicon rod was measured 20 minutes ago.

[0058] Preferably, the quality of the polycrystalline silicon rod is mainly calculated from the voltage and current values ​​collected in the reduction furnace. However, due to noise interference during the data acquisition process, the fluctuation range of the relevant data is relatively large, and measurement noise exists. Therefore, in the process of calculating the deposition rate, it is necessary to filter the polycrystalline silicon quality or deposition rate to ensure the smoothness of the results and to ensure the stability and rapid convergence of the subsequent PID control process. The noise reduction process can apply Gaussian filtering to process the polycrystalline silicon quality curve or deposition rate curve.

[0059]

[0060] in It is the distance between the data point and the center of the Gaussian filter. It is the standard deviation of the Gaussian distribution, which determines the filter width. A larger standard deviation results in a larger standard deviation. A higher value results in a wider filter, which in turn produces a smoother effect on the data. The selection of [a] requires obtaining their autocorrelation coefficients from historical data, and these autocorrelation coefficients are used to determine [the appropriate selection].

[0061] Furthermore, the key to determining the heat loss rate ΔQ lies in determining the heat loss Q of the reaction and the total energy E.

[0062] The total energy input of the reaction is relatively singular, mainly in the form of electrical energy.

[0063]

[0064] where P is the power and T is the time difference. The specific formula is , in J

[0065] is the average power supply during the reduction reaction from t1 to t2, in Watts (W), .

[0066] The heat loss Q is mainly from the tail gas. In the reaction process, the heat of the tail gas temperature, i.e. the heat carried away by output, can be calculated through the chemical reaction equation and thermodynamic formula. The chemical reaction equation needed is:

[0067]

[0068]

[0069] According to the above chemical reaction equation in the production process, the heat formula of each chemical reaction in the reaction process can be obtained as follows:

[0070]

[0071]

[0072]

[0073]

[0074] where:

[0075] is the specific heat capacity of TCS (trichlorosilane), which is ;

[0076] is the average mass flow of TCS in the corresponding time, in ;

[0077] is the mass of the silicon rod at t2, and the calculation method is referred to the previous step

[0078] is the mass of the silicon rod at t1, and the calculation method is referred to the previous step;

[0079] is the relative atomic mass of elemental silicon, which is ;

[0080] is the relative molecular mass of silicon tetrachloride, which is ;​​

[0081] M is the relative molecular mass of trichlorosilane, and is valued at l;

[0082] T is the temperature of the exhaust gas minus the temperature of the feedstock over the corresponding time;

[0083] Cp is the specific heat capacity of gaseous hydrogen chloride, and is valued at ;

[0084] Cv is the specific heat capacity of gaseous silicon tetrachloride, and is valued at ;

[0085] Ch is the specific heat capacity of hydrogen, and is valued at ;

[0086] Qh is the average mass flow of hydrogen over the corresponding time, and is valued in ;

[0087] Mh is the relative molecular mass of hydrogen chloride, and is valued at ;

[0088] The total heat carried away by the exhaust gas is obtained by adding all the above heat formulas together:

[0089]

[0090] The heat loss rate at this moment is:

[0091]

[0092] S2, deduce the control parameters.

[0093] The feedback control parameters required in the PID control process include the proportional adjustment parameter P, the integral adjustment parameter I, and the differential adjustment parameter D. The control parameters are derived from the process difference, and are usually determined by the real-time deposition rate and the ideal deposition rate curve determined in advance. Therefore, a certain means is needed to determine the ideal target deposition rate curve.

[0094] Simply put, we select the deposition rate curve that meets the requirements from a large amount of historical data. In the screening process of the present embodiment, the deposition rate curves corresponding to the reaction process abnormalities, the discharge density rate not meeting the requirements, the reaction process overtime, etc. are removed, and then the target deposition rate curve is selected or synthesized based on the remaining deposition rate curves. The purpose is to retain the characteristics of the historical deposition rate curve with excellent single power consumption in the target deposition rate curve as much as possible.

[0095] The target deposition rate curve can be determined as a globally optimal deposition rate curve in the entire reduction process, but considering the stage-by-stage nature of the polysilicon reduction process, the global optimality does not guarantee optimality in each stage. In actual data, there is generally a cross between the deposition rate curves with better power consumption, which means that the stages with higher local deposition rate corresponding to different reduction processes can be different from each other, and thus there is room for optimization.

[0096] Firstly, the polysilicon reduction process is generally divided into five different stages, namely, the feeding stage, the stable material stage, the reducing material stage, the peak stage, and the flat material stage. The reaction state in the reduction furnace is different in the five processes, and thus the entire reduction process can be divided into several stages according to the corresponding stages of the polysilicon reduction process, and the locally optimal historical deposition rate curve in each stage is selected in the stage, and the criterion is that the average deposition rate in the stage is maximum, and then the segments obtained in each stage are spliced to form the target deposition rate curve, wherein the adjacent segments can be appropriately smoothed during splicing to avoid the influence of the sharp connection on the stability of the control process.

[0097] In addition, the target deposition rate curve can also be preferably selected without considering the stage-by-stage nature of the specific reduction process, but based on the data characteristics. In short, considering that the historical deposition rate curves cross each other, after obtaining the historical target curves meeting the conditions, the optimal deposition rate curve at each time is segmented and connected in turn to obtain the final target deposition rate curve, which is equivalent to obtaining the maximum deposition rate envelope of all the historical deposition rate curves meeting the conditions.

[0098] Based on the target deposition rate curve, the control parameters can be determined as follows.

[0099] Suppose the target deposition rate curve is , the real-time deposition rate record is , the current time is , the historical data of the previous time step is , the time step in this embodiment is 10 min, and the target difference Ay is obtained t8 =Y t8 -y t8 , Ay t9 =Y t9 -y t9 , then the proportional adjustment parameter P is defined as Ay t8 , the integral adjustment parameter I is defined as Ay t9 , and the differential adjustment parameter D is defined as Ay t8 - Ay t9 .

[0100] The reason for choosing the deposition rate as the control target is that through analysis of historical data, it can be found that the power consumption has better correlation with the deposition rate, so compared with other control quantities, the effect of energy saving control by controlling the deposition rate should be better. Generally, the control target curve uses power or voltage data that can be directly obtained through sensors. However, both curves are affected by the furnace itself and the operator's control, and have poor convergence. In order to achieve process convergence, the applicant needs to shorten the global length of the control process to prevent global divergence. An example of a similar approach is that in the applicant's individual historical scheme, there is a technical scheme that divides the process into different stages during the entire global reduction process and uses different control strategies, the main purpose of which is to avoid the problem of control divergence of the representative quantity.

[0101] In order to determine whether the deposition rate can be used as a control target curve, the obtained historical production data is first screened to remove outliers and divided into high, medium and low groups according to the high and low of the single power consumption. In the exploration process, the three groups are taken from the top 30%, middle 40% and bottom 30% of the descending order respectively. The mean values of voltage, power and deposition rate of all data in the high, medium and low groups are calculated, and the change curve of the mean values in the reaction process is as shown in Figure 2 It can be found that the red line of high power consumption has a lower mean deposition rate, and the blue line of low power consumption has a higher deposition rate, in other words, the mean deposition rate of data with lower power consumption tends to be higher. Further research shows that the larger the trace of the covariance of the voltage, power and deposition rate curves of all data (covariance trace ), the higher the correlation and the better the convergence, and the original data is more sensitive, which is not provided here). The results show that the trace of the covariance between the deposition rate and the single power consumption is the largest, on the one hand, indicating that the deposition rate has better correlation with the single power consumption, which conforms to the initial judgment, and on the other hand, due to the large absolute value of the deposition rate, it provides more adjustment space and is more suitable for application in PID control process, so the deposition rate is more suitable as the control target curve of this project.

[0102]

[0103] S3, performing a PID control process.

[0104] The hydrogen input and reaction current of the reaction are controlled by PID control using the control parameters in step S2. The hydrogen input adjustment amount ; the reaction current adjustment amount .

[0105] Among them, for the hydrogen input adjustment amount, the proportional adjustment coefficient , the integral adjustment coefficient , and the differential adjustment coefficient The value of the reaction current adjustment amount can be set , The value is preferably classified and adjusted according to the heat loss rate, when between 40% and 50%, when less than 40%, when greater than 50%, .

[0106] The present description describes a new control strategy that takes the deposition rate calculated from current and voltage as the control target. This method first determines the optimal deposition rate curve through data analysis, and calculates the deposition rate in real time during control. Then, the current heat ratio value in the furnace is calculated using thermodynamic formulas, and this ratio value is combined with the deposition rate difference to input into the PID control module, which precisely regulates hydrogen and current. This automated control strategy adjusts both current and hydrogen, rather than a single variable, significantly improving the stability and accuracy of the production process, thereby ultimately improving the yield and quality of polysilicon.

[0107] This control method has significant technical advantages in polysilicon production: 1. Precise control: This method can accurately determine the growth weight of the silicon rod through real-time calculation, achieving precise control. In addition, the control cycle is shortened to every 10 minutes, significantly improving control frequency and density. 2. High adaptability: The deposition rate is calculated from real-time current and voltage data, so it can more accurately reflect the growth of the silicon rod in the furnace. Compared to traditional control methods, which usually rely on initial observation of power and later observation of voltage, this method can track the same indicator throughout the process, making it more adaptable. 3. Improved automation level: This method controls both current and hydrogen, forming a double closed loop, and through thermodynamic formulas, the current regulation ratio can be better obtained, achieving precise control of the current. This method reduces reliance on human experience, making control more automated. 4. More reasonable selection of control targets: Mathematical statistical analysis method is used to quantify the correlation between target curves by calculating the trace of covariance, providing a scientific basis for control target selection. This method is objective and quantitative, avoiding control failure caused by unreasonable selection of control targets.

[0108] The above merely describes the preferred embodiments of the present application, and does not limit the scope of the present application. Without departing from the design spirit of the present application, various modifications and improvements to the technical solutions of the present application made by those of ordinary skill in the art shall fall within the protection scope determined by the claims of the present application.

Claims

1. A control method for a polycrystalline silicon reduction furnace based on a positive mechanism model with multiple control variables, characterized in that, Includes the following steps: S1. Collect and calculate the reaction state: The reaction state includes the real-time deposition rate ΔM and the heat loss rate ΔQ of the reaction; the real-time deposition rate ΔM is the rate of change of the mass M of the polycrystalline silicon rod in the reduction furnace, and the heat loss rate ΔQ is the ratio of the heat loss Q of the reaction to the total energy input E. S2. Inferring control parameters: The control parameters in the PID control process are determined based on the difference between the real-time deposition rate curve and the target deposition rate curve. The control parameters include the proportional control parameter P, the integral control parameter I, and the differential control parameter D. The real-time deposition rate curve is generated based on the real-time deposition rate ΔM of the reaction process sequence, and the target deposition rate curve is a preset curve. S3. Execute the PID control process: Use the control parameters from step S2 to perform PID control on the hydrogen input and reaction current; adjust the hydrogen input... ; Reaction current adjustment amount ;k p1 k i1 k d1 k p2 k i2 k d2 Here, k is the control coefficient; p2 k i2 k d2 At least one of them has a preset set of values, and its value is related to the heat loss rate ΔQ of the reaction; Integral adjustment coefficient ,when Between 40% and 50%, ,when Less than 40%, ,when Greater than 50%, .

2. The polycrystalline silicon reduction furnace control method based on a positive mechanism model with multiple control variables as described in claim 1, characterized in that, In the step S1 of collecting and calculating the reaction state, the mass M of the polycrystalline silicon rod is: Where m is the number of current loops in the polysilicon reduction furnace, M i The mass of the polysilicon rod in the i-th phase circuit is calculated using the following formula: Where ρ2 is the resistivity of the silicon rod, L is the total length of the silicon rod, and n i I is the number of silicon rod pairs in the i-th phase circuit calculated in the reduction furnace; i U is the current in the i-th phase loop; i ρ is the voltage of the i-th phase circuit; ρ is the density of the polycrystalline silicon rod.

3. The polycrystalline silicon reduction furnace control method based on a positive mechanism model with multiple control variables as described in claim 2, characterized in that, In the step S1 of collecting and calculating the reaction state, the mass M or deposition rate ΔM of the polycrystalline silicon rod is subjected to Gaussian filtering.

4. The polycrystalline silicon reduction furnace control method based on a forward mechanism model with multiple control variables as described in claim 1, characterized in that, In the S1 step of collecting and calculating the reaction state, the heat loss Q of the reaction is the heat of the exhaust gas, and the total energy E is the input electrical energy. E=PT Where P is the input electrical power and T is the time difference.

5. The polycrystalline silicon reduction furnace control method based on a positive mechanism model with multiple control variables as described in claim 4, characterized in that, In the step S1 of collecting and calculating the reaction state, the heat loss Q of the reaction is calculated as follows: Based on the chemical reaction equations in the production process, the heat formulas for each chemical reaction in the process can be obtained as follows: in: The specific heat capacity of TCS This represents the average mass flow rate of the TCS over the corresponding time period. Let t2 be the mass of the silicon rod. Let be the mass of the silicon rod at time t1. This represents the relative atomic mass of elemental silicon. is the relative molecular mass of silicon tetrachloride. This represents the relative molecular mass of trichlorosilane. The exhaust gas temperature minus the feed temperature within the corresponding time period; The specific heat capacity of gaseous hydrogen chloride. The specific heat capacity of gaseous silicon tetrachloride, This refers to the specific heat capacity of hydrogen. This represents the average mass flow rate of hydrogen over the corresponding time period. is the relative molecular mass of hydrogen chloride.

6. The polycrystalline silicon reduction furnace control method based on a positive mechanism model with multiple control variables as described in claim 1, characterized in that, In the S2 inference control parameter step, the target deposition rate curve is synthesized based on historical deposition rate curves that meet preset requirements for single-power consumption.

7. The polycrystalline silicon reduction furnace control method based on a positive mechanism model with multiple control variables as described in claim 6, characterized in that, In the S2 inference control parameter step, the polysilicon reduction process is divided into several stages. Within each stage, a segment of the historical deposition rate curve with the best local single power consumption within that stage is selected, and then the segments obtained in each stage are spliced ​​together to synthesize the target deposition rate curve.

8. The polycrystalline silicon reduction furnace control method based on a positive mechanism model with multiple control variables as described in claim 6, characterized in that, In the S2 inference control parameter step, the optimal deposition rate curve at each moment is segmented and sequentially connected to obtain the final target deposition rate curve.

9. The polycrystalline silicon reduction furnace control method based on a positive mechanism model with multiple control variables as described in any one of claims 6-8, characterized in that, In the S2 inference control parameter step, the target deposition rate curve is: The real-time deposition rate was recorded as The current time is Historical data with a previous time step is The proportional adjustment parameter P=Y t8 -y t8 Integral adjustment parameter I=Y t9 -y t9 and differential adjustment parameter D=(Y t8 -y t8 )-(Y t9 -y t9 ), The subscript indicates the value on the corresponding time curve.

10. The polysilicon reduction furnace control method based on a positive mechanism model with multiple control variables as described in claim 9, characterized in that, In the S3 step of executing the PID control process, the hydrogen input adjustment amount... ; Reaction current adjustment amount ; Among them, the proportional adjustment coefficient is used for adjusting the amount of hydrogen input. Integral adjustment coefficient Differential adjustment coefficient .

Images