Powder feeding rate control method for radio frequency plasma gasification separation

Abstract

A powder feeding rate control method for radio frequency plasma gasification separation comprises the steps that the powder feeding rate of powder to be separated is detected, a powder feeding rate error and a first-order derivative of the powder feeding rate error serve as input values of a neural network, and lumped uncertainty influencing the powder feeding rate is output after learning of the neural network. On the basis of lumped uncertainty, the self-adaptive superhelix strategy and the fractional order nonsingular terminal sliding mode function are combined to be used for a control law of a powder feeding motor; the rotating speed of a powder feeding motor is adjusted in real time by monitoring the actual powder feeding rate in real time, the actual powder feeding rate is fed back to a control layer of the powder feeding motor in real time, the rotating speed of a servo motor of a powder feeder is controlled online, and online detection and closed-loop control over the powder feeding rate are achieved; accurate feeding of the powder to be separated is improved, uniformity, stability and accuracy of powder conveying are guaranteed, and further improvement of metal separation and purification precision is facilitated.

Description

Technical Field

[0001] The invention relates to the field of separation and purification, and in particular to a powder feeding rate control method for radio frequency plasma gasification separation. Background Art

[0002] Radio frequency plasma spheroidization technology is suitable for the separation and purification of impurities in high melting point materials (such as ceramics, tungsten, molybdenum, tantalum, niobium and other metal materials, high melting point composite materials); radio frequency plasma spheroidization technology uses plasma to heat the raw materials at high temperature, so that the high melting point metal is melted by heat, and the low melting point impurities are gasified and burned. The molten high melting point metal quickly solidifies under the combined action of surface tension and extremely high temperature gradient to form spherical powder, which is then separated from the impurities. The advantages of radio frequency plasma separation and purification are as follows: no electrode pollution, high material purity; large plasma flame volume, high temperature, large processing capacity; can be applied to a variety of working environments such as inert atmosphere, oxidizing atmosphere, reducing atmosphere, etc. The above many characteristics make radio frequency plasma one of the most cutting-edge choices for industrial material production.

[0003] During the RF plasma gasification separation process, the raw material powder mixed with impurities needs to be rotated by a stepper motor to drive the roller powder feeder. The upper powder inlet of the roller powder feeder is connected to the powder storage barrel. The stepper motor drives the powder feeding shaft in the roller powder feeder to rotate. The powder in the groove on the surface of the powder feeding shaft enters the conveying pipeline through the powder outlet, and is blown into the RF plasma through the air supply system for heating, separation and purification. The powder feeder needs to provide a stable amount of powder to be separated for the high-temperature plasma torch according to the powder feeding rate requirements, which is also one of the core technologies of RF induction plasma separation and purification. In the process of the powder feeder conveying the powder to be separated, the uniformity, stability and accuracy of the powder conveying will directly affect the accuracy of metal separation and purification (especially for materials used in the aerospace field with high purity requirements). Traditional powder feeder control system methods such as PID cannot meet the high-performance powder feeding control requirements and are in urgent need of improvement.

[0004] The present invention provides a powder feeding rate control method for radio frequency plasma gasification separation to solve the above technical problems. Summary of the invention

[0005] A powder feeding rate control method for radio frequency plasma gasification separation includes:

[0006] S1: The photoelectric sensor senses the powder delivery rate v in the transparent conveying pipeline to obtain the set powder delivery rate v r The powder feeding rate error compared to the actual powder feeding rate v;

[0007] S2: Online estimation of lumped uncertainty through RBF neural network; In the actual powder feeding control process of RF plasma gasification separation, there are many factors that affect the powder feeding accuracy (such as sensor error, friction of the powder feeding execution mechanical mechanism, structural gap, execution response lag time difference, etc.). These interference factors, namely lumped uncertainty, seriously affect the stability and accuracy of the control system. The lumped uncertainty can be estimated online through RBF neural network;

[0008] S3: Based on the lumped uncertainty estimated by the RBF neural network, the adaptive superhelical strategy and the fractional-order non-singular terminal sliding mode function are combined to obtain the control law of the powder feeding motor, and the speed of the powder feeding motor is controlled in real time.

[0009] Preferably, in the powder feeding rate control method for radio frequency plasma gasification separation, the input value of the RBF neural network in step S2 is (The input value of the RBF neural network is defined as e, The transposed matrix is ​​composed of: e is the powder feeding rate error, v is the r is the set target powder feeding rate, v is the actual measured powder feeding rate, and the network output value is f(x), as follows:

[0010] e=v r -v;

[0011]

[0012] f(x)=W T h(x)+ε;

[0013] In the formula, j represents the jth node of the hidden layer of the network, x is the input value of the neural network, and h(x) is the value obtained by h j The vector composed of W is the ideal network weight, W T represents the transposed matrix of the ideal network weights, ε is the network approximation error, k j is a positive scalar representing the width of the Gaussian basis function, l j is the center vector of the hidden layer.

[0014] Preferably, the powder feeding rate control method of the radio frequency plasma gasification separation, the operation method of step S3 is as follows:

[0015] S 3.1, the fractional-order non-singular terminal sliding mode function is:

[0016]

[0017] In the formula, 0<b1, b2, C1, c2<1, a1, a2 are positive parameters to be adjusted, (b1, b2 are fractions) is the Riemann-Liouville fractional calculus function, and

[0018] S 3.2, in order to further improve the convergence speed and robustness of the fractional-order non-singular terminal sliding mode and reduce chattering, the adaptive super-helical method is used to further optimize the non-singular terminal sliding mode function:

[0019]

[0020] In the formula, η1 and η2 are adaptive parameters, and the upper and lower limits of η1 and η2 are set manually or empirical values ​​are obtained from the system.

[0021]

[0022] In the formula, t represents the running time of the controller that controls the motor speed;

[0023]

[0024] ηη2=θ3ηη1;

[0025] Among them, θ1, θ2 (θ2>θ1), θ3, Δ1 are constant parameters, and s is a fractional-order non-singular terminal sliding mode function; it can be seen that when the tracking performance is good and η1 is within the design range, the convergence speed increases from θ1 to θ2. Therefore, while suppressing the influence of noise, the control performance can be maximized; the values ​​of θ1 and θ2 should be reasonably selected according to the motor performance. The larger the value, the greater the set convergence speed (the powder feeding motor can adjust the powder feeding speed in a shorter time), but the motor has a certain tolerance limit. If the set convergence speed is too large, it will lead to execution difficulties and motor damage. Risks;

[0026] S 3.3: Based on the calculation results of the above steps, the fractional-order non-singular terminal sliding mode control law based on RBF neural network and adaptive superhelix is ​​obtained as follows:

[0027]

[0028] Preferably, in the powder feeding rate control method for radio frequency plasma gasification separation, the adaptive law of the ideal weight W is as follows:

[0029]

[0030] Where τ is the learning rate, s is the fractional-order non-singular terminal sliding mode function, and h(x) is the function of h j Vector composed of.

[0031] Here’s how it works:

[0032] First detect the actual powder feeding rate, and calculate the powder feeding rate error e and its first-order derivative Then e, The transposed matrix data of is used as the input value of the neural network. After the neural network operation and learning, the lumped uncertainty of the powder feeding rate control system is output. The setting is the same as e, The relevant intermediate function s is obtained, and in order to improve the convergence of s (make s quickly approach 0), the adaptive superhelix strategy is combined with the non-singular terminal sliding mode function. While suppressing the influence of noise, the control performance can be maximized. The fractional-order non-singular terminal sliding mode control law based on RBF neural network and adaptive superhelix is ​​obtained, and the speed of the powder feeding motor is controlled in real time by using this control law.

[0033] The advantages are as follows: RBF neural network, adaptive superhelix strategy, fractional-order non-singular terminal sliding mode function are combined to monitor the actual powder feeding efficiency and output lumped uncertainty in real time. The adaptive superhelix strategy is used to optimize the fractional-order non-singular terminal sliding mode function to improve the function convergence speed while improving the system robustness and reducing jitter. The speed of the powder feeding motor (DC motor or AC motor) is controlled in real time by u, so that the actual powder feeding rate is infinitely close to the ideal powder feeding rate. According to the set ideal powder feeding rate requirements, the high-temperature plasma torch is provided with stable powder to be separated, so that the powder to be separated and purified is evenly, stably and accurately transported, ensuring the high purity state of the material after separation. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The specific implementation is further described below in conjunction with the accompanying drawings, wherein:

[0035] Figure 1 is a schematic diagram of the photoelectric detection principle of the powder flow involved in the specific implementation example 1 of the present invention;

[0036] Figure 2 It is a schematic diagram of a photoelectric powder feeding rate control system according to a specific implementation example 1 of the present invention;

[0037] The following specific implementation manner will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0038] A powder feeding rate control method for radio frequency plasma gasification separation includes:

[0039] like Figure 2 As shown, the powder to be separated is transported to the RF plasma separation and purification system through the powder feeding motor (60 servo 400W set) purchased from Luwei Technology. The commonly used roller shaft speed is about 12r / min, the carrier gas flow rate is about 7.7L / min, the conveyed powder particle size is 250 mesh, and the corresponding powder conveying rate is about 5.9g / min. However, in order to further improve the accuracy and stability of the powder conveying rate and improve the purity of the product after separation and purification, the speed of the powder feeding motor can be adjusted in real time by the following methods, as follows:

[0040] S1: The powder delivery rate v in the transparent conveying pipeline is sensed by a photoelectric sensor (of course, it can also be monitored in real time by other means) to obtain the set powder delivery rate v r The powder feeding rate error compared to the actual powder feeding rate v;

[0041] The light emitted by the light-emitting tube passes through a transparent tube with powder airflow and is received by a photoelectric sensor. The light intensity received by the photoelectric sensor is related to the flow rate of powder in the transparent tube. The greater the powder flow rate, the greater the light attenuation, and the smaller the light intensity obtained by the photoelectric sensor. In this way, after experimental calibration, the powder conveying rate flowing through the transparent tube can be calculated based on the output signal of the photoelectric sensor.

[0042] S2: Online estimation of aggregate uncertainty through RBF neural network;

[0043] S3: Based on the lumped uncertainty estimated by the RBF neural network, the adaptive superhelical strategy and the fractional-order non-singular terminal sliding mode function are combined to obtain the control law of the powder feeding motor, and the speed of the powder feeding motor is controlled in real time.

[0044] Preferably, in the powder feeding rate control method for radio frequency plasma gasification separation, the input value of the RBF neural network in step S2 is (The input value of the RBF neural network is defined as e, The transposed matrix is ​​composed of, e is the powder feeding rate error, V r is the set target powder feeding rate, v is the actual measured powder feeding rate, and the network output value is f(x), as follows:

[0045] e=v r -v;

[0046]

[0047] f(x)=W T h(x)+ε;

[0048] Among them, j represents the jth node of the hidden layer of the network, x is the input value of the neural network, and h(x) is the value obtained by h j The vector composed of W is the ideal network weight, W T represents the transposed matrix of the ideal network weights, ε is the network approximation error, k i is a positive scalar representing the width of the Gaussian basis function, l j is the center vector of the hidden layer.

[0049] Preferably, the powder feeding rate control method of the radio frequency plasma gasification separation, the operation method of step S3 is as follows:

[0050] S 3.1, the fractional-order non-singular terminal sliding mode function is:

[0051]

[0052] a1=1.9,a2=1.5,b1=0.1,b2=0.5,c1=0.86,c2=0.99 are the positive parameters to be adjusted, D 0.1 , D-0.5 is the Riemann-Liouville fractional calculus function and,

[0053] S 3.2, in order to further improve the convergence speed and robustness of the fractional-order non-singular terminal sliding mode and reduce chattering, the adaptive super-helical method is used to further optimize the non-singular terminal sliding mode function:

[0054]

[0055] In the formula, η1 and η2 are adaptive parameters, and the upper and lower limits of η1 and η2 are set manually or empirical values ​​are obtained from the system.

[0056]

[0057] Where t represents the running time of the controller for controlling the motor speed; optional, 3≦η1≦20,

[0058]

[0059] η2=θ3η1;

[0060] Among them, θ1=30,θ2=60,θ3=1.5,Δ1=0.5, and s is a fractional-order non-singular terminal sliding mode function. It can be seen that when the tracking performance is good and η1 is within the design range, the convergence speed increases from θ1 to θ2. Therefore, the control performance can be maximized while suppressing the influence of noise.

[0061] S 3.3: Based on the calculation results of the above steps, the fractional-order non-singular terminal sliding mode control law based on RBF neural network and adaptive superhelix is ​​obtained as follows:

[0062]

[0063] Furthermore, the adaptive law of the ideal weight W is as follows:

[0064]

[0065] Where τ is the learning rate and is set to 10, s is the fractional-order non-singular terminal sliding mode function, and h(x) is the function given by h j Vector composed of.

[0066] By using the above powder feeding rate control method, the powder feeding rate can be strictly controlled around the target rate. The experimental data are as follows: r =5.90g / min:

[0067]

[0068]

[0069] It can be seen from the above table that the speed of the roller shaft (powder feeding shaft) driven by the reducer-motor is not linearly related to the actually measured powder feeding rate, so real-time monitoring and real-time adjustment of the motor speed are particularly necessary for accurately controlling the powder feeding amount; the above method is used to monitor the actual powder feeding rate in real time and through the RBF neural network, adaptive superhelical strategy, and fractional-order non-singular terminal sliding mode function, the speed of the powder feeding motor can be effectively controlled in real time and the powder feeding rate can be accurately adjusted.

[0070] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for controlling the powder feeding rate of radio frequency plasma gasification separation, characterized in that: The steps are as follows, S1: Measure the actual powder feeding rate v of the powder to be separated and purified, and obtain the set powder feeding rate v r The powder feeding rate error compared to the actual powder feeding rate v; S2: Online estimation of the aggregate uncertainty affecting the powder feeding rate of the separated and purified powders through a neural network; S3: Based on the lumped uncertainty estimated by the neural network, the adaptive superhelical strategy and the fractional-order non-singular terminal sliding mode function are combined to obtain the control law of the powder feeding motor, and the speed of the powder feeding motor is controlled in real time.

2. The powder feeding rate control method for radio frequency plasma gasification separation as claimed in claim 1, characterized in that: The input value of the neural network in step S2 is e, The transposed matrix of e is the powder feeding rate error, v r is the set target powder feeding rate, v is the actual measured powder feeding rate, and the network output value is f(x), as follows: e=v r -v; f(x)=W T h(x)+ε; Where j represents the jth node in the hidden layer of the network, x is the input value of the neural network, and h(x) is the value obtained by h j The vector composed of W is the ideal network weight, W T represents the transposed matrix of the ideal network weights, ε is the network approximation error, k j is a positive scalar representing the width of the Gaussian function, l j is the center vector of the hidden layer.

3. The powder feeding rate control method for radio frequency plasma gasification separation as claimed in claim 2, characterized in that: The operation method of step S3 is as follows, S 3.1, the fractional-order non-singular terminal sliding mode function is: where \(0 < b_1, b_2, c_1, c_2 < 1\), and \(a_1, a_2\) are positive parameters to be tuned. is the Riemann-Liouville fractional calculus function, and S 3.

2. The non-singular terminal sliding mode function is optimized by using the adaptive super-twisting strategy: In the formula, η1 and η2 are adaptive parameters, and the upper and lower limits of η1 and η2 are set manually or the empirical values ​​of η1 and η2 are obtained from the system. t represents the running time of the controller that controls the motor speed; η2=θ3η1; Among them, θ1, θ2 (θ2>θ1), θ3, Δ1 are constant parameters, and s is a fractional-order non-singular terminal sliding mode function; S 3.3: Based on the calculation results of the above steps, the fractional-order non-singular terminal sliding mode control law based on neural network and adaptive superhelix is ​​obtained:

4. The powder feeding rate control method for radio frequency plasma gasification separation as claimed in claim 3, characterized in that: The adaptive law of the ideal weight W is as follows: Where τ is the learning rate, s is the fractional-order non-singular terminal sliding mode function, and h(x) is the function of h j Vector composed of.

Images