Spiral electric radiant tube and PID (Proportion Integration Differentiation) control system thereof based on neural network

Abstract

A spiral electric radiant tube comprises a spiral electric heating wire, a spiral magnetic column, a magnetic disk and a radiant tube sleeve. A cavity structure is arranged in the radiant tube sleeve, a plurality of spiral magnetic columns are arranged in the cavity structure, a plurality of magnetic discs are arranged on the spiral magnetic columns in a sleeving mode, the spiral magnetic columns are fixedly installed in the radiant tube sleeve through the magnetic discs, and the spiral electric heating wire is evenly wound in the axial direction of the spiral magnetic columns. According to the method, the spiral electric radiant tube is accurately controlled through a model combining a single neuron model and a PID control model, a Delta learning rule is adopted, and the function relationship between the weighting coefficient and the input, output and output deviation of the neuron is considered; when the Delta learning rule is adopted, the connection weight can be adjusted according to the difference between the actual output and the expected output of the neuron, and the weight coefficient can be adjusted online according to the learning capability and the self-adaptability of the neural network, so that the method has a more stable tracking characteristic and better robustness, and has higher self-adaptability and higher control precision.

Description

Technical Field

[0001] This invention relates to the field of industrial electric heating technology, and in particular to a spiral electric radiant tube and its PID control system based on a neural network. Background Technology

[0002] Electric radiant tubes are electric heating devices used in heating and heat treatment processes, widely applied in industrial furnaces, heat treatment equipment, and other fields requiring high-temperature heating. They convert electrical energy into heat energy and transfer it to the object being heated via radiation. Traditional electric radiant tubes are mostly gas-fired, which presents many problems in industrial applications, such as high exhaust gas temperature, uneven tube wall temperature, and large amounts of combustion product emissions. Although technical optimization measures have been proposed for gas-fired radiant tubes, NOx emissions remain a persistent problem due to the use of coal gas.

[0003] To address the aforementioned issues, the current solution employs a cage-type electric radiant heating system, primarily composed of a radiant tube sleeve, a ceramic body, and heating wires. The ceramic body mainly serves as a support for the heating wires. During operation, the surface temperature of the heating wires can reach 1200℃. Electric radiant heating not only improves heating efficiency but also addresses the root cause of NOx emissions.

[0004] For example, patent document with patent number 201620491175.1 discloses an energy-saving and high-efficiency electric radiation tube, including a radiator assembly, a heating element, a first insulating element, a second insulating element, and electrode lead-out rods. The radiator assembly includes several radiator tubes and connectors. The radiator tubes are connected by several connectors to form a hollow pipe. The two ends of the hollow pipe are respectively fixed on the first insulating element and the second insulating element. The heating element is disposed in the hollow pipe and one end of the heating element is connected to two electrode lead-out rods passing through the first insulating element.

[0005] The above technical solution involves cutting the radiant tube into several short radiant tube coils, and then connecting the radiant tube coils with several connectors to form a hollow radiant tube outer tube pipe. The heating element inside remains unchanged. During the heating process, the heating element radiates heat to the outside through the hollow pipe, which prevents heat from concentrating inside the radiant tube and affecting the service life of the heating element. It can also improve the radiation efficiency of the radiant tube.

[0006] However, the electric radiation tube in the above technical solution experiences significant thermal stress within its high-temperature resistant ceramic material during operation, which also leads to a shorter service life.

[0007] Meanwhile, due to the characteristics of the billet heating temperature control system in the annealing furnace, such as nonlinearity, time-varying nature, large inertia, large lag, and unidirectional heating, traditional PID control algorithms cannot accurately control nonlinearity. This is because the parameter adjustment process is cumbersome and the operating conditions are unstable, and the system parameters do not have the ability to adapt.

[0008] Therefore, it is necessary to improve the existing technology to overcome the above-mentioned defects. Summary of the Invention

[0009] The purpose of this invention is to provide a spiral electric radiation tube and its neural network-based PID control system to solve the problems existing in the prior art.

[0010] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0011] A spiral electric radiant tube includes a spiral heating wire, a spiral magnetic column, a magnetic disk, and a radiant tube sleeve. The radiant tube sleeve has a cavity structure inside, and a plurality of spiral magnetic columns are arranged inside the cavity structure. A plurality of magnetic disks are sleeved on the spiral magnetic columns, and the spiral magnetic columns are fixedly installed inside the radiant tube sleeve through the magnetic disks. The spiral heating wire is uniformly wound along the axial direction of the spiral magnetic columns.

[0012] Furthermore, the spiral magnetic column is a support column for the spiral heating wire, and its diameter is slightly smaller than the spiral inner diameter of the spiral heating wire, which facilitates the assembly between the spiral magnetic column and the spiral heating wire.

[0013] The disk has evenly spaced holes along its circumference, with the hole diameter slightly larger than the outer diameter of the spiral heating wire, which facilitates the assembly of the disk with the spiral heating wire and the spiral magnetic column.

[0014] Furthermore, the spiral heating wire is divided into two groups, which are connected in parallel, and the spiral heating wires within the same group are connected in series.

[0015] Furthermore, the radiant tube sleeve is the same as the sleeve used for a 70-80KW gas radiant tube in a quenching furnace.

[0016] Furthermore, both the spiral magnetic column and the disk are made of high-temperature resistant alumina material.

[0017] A neural network-based PID control system for a spiral radiating tube precisely controls the spiral radiating tube by combining a single neuron model with a PID control model.

[0018] The parameters included in the model combining the single neuron model and the PID control model are as follows:

[0019] s v Set the temperature;

[0020] p v The sensor collects temperature data.

[0021] K: System output gain;

[0022] Δu: Increment of system output control quantity;

[0023] u: The control quantity ultimately passed to the execution unit;

[0024] z -1 : Represents system latency;

[0025] Among them, s v and p v The system inputs x1, x2, and x3 of the single neuron model are obtained through corresponding converters, and the values ​​of the three weights w1, w2, and w3 are adjusted through a learning algorithm, thereby realizing the self-adjustment of system parameters in the control system.

[0026] Furthermore, the system inputs x1, x2, and x3 at time k are respectively:

[0027] x1(k) = e(k) - e(k-1)

[0028] x2(k)=e(k)

[0029] x3(k)=e(k)-2e(k-1)+e(k-2) (1)

[0030] Where e(k) is the deviation between the set value and the system output value:

[0031] e(k) = s v -p v (2)

[0032] The system output increment Δu is:

[0033]

[0034] u(k)=Δu+u(k-1) (3)

[0035] Where K is the system gain, w i (k) represents the weighting coefficient.

[0036] Furthermore, the PID control system employs a Delta learning rule and takes into account the weighting coefficient w. i (k) should have a functional relationship with the neuron's input, output, and output bias. When using the Delta learning rule, the connection weights can be adjusted based on the difference between the neuron's actual output and its expected output.

[0037] Δw ij =a(di -y i )·x i (t)

[0038] Where, Δw ij To increase the weight, d i y is the expected output of neuron i. i 'i' is the actual output of neuron 'i', and 'a' is the learning speed.

[0039] Furthermore, when the PID control system is working, the stability and correlation of the system are related to the parameters. After screening the parameters used in actual production, the relevant parameters of the system are adjusted accordingly based on the following conditions:

[0040] 1) The initial weighting coefficients w1(0), w2(0), and w3(0) will not affect the stability of the system operation, so the values ​​can be chosen arbitrarily;

[0041] 2) The adaptive gain K of the system can be arbitrarily chosen at first, and then corrected according to the simulation response curve until the most suitable K value is found.

[0042] Furthermore, the derivation process of the transfer function of the PID control system is as follows:

[0043] S1: Assuming the external temperature of the quenching furnace is T2, the increase in heat dissipation per unit time is:

[0044] ΔQ0=λA[(T1-T2)-(T0-T2)]=λA(T1-T0)

[0045] in,

[0046] ΔQ0: Heat dissipation, J;

[0047] λ: heat transfer coefficient;

[0048] A: Heat transfer area, m 2 ;

[0049] T0: Furnace temperature before heating, in K;

[0050] T1: Furnace temperature after heating, in K;

[0051] T2: External temperature of the quenching furnace, K;

[0052] S2: Heat generated by the heating wire per unit time ΔQ i Subtracting the heat dissipated from the furnace wall surface ΔQ0 equals the change in thermal energy Q inside the furnace:

[0053] Q = CM(T1-T0) + Q0

[0054]

[0055] in,

[0056] C: Specific heat capacity of the heating wire;

[0057] M: Mass of heating wire, kg;

[0058] S3: The nonlinear relationship between the heat generated by the heating wire per unit time and the voltage U:

[0059]

[0060] S4: When the electric radiant tube reaches equilibrium, the heat generated by the heating wire is equal to the heat dissipated from the furnace wall, i.e., Q. i =Q0, linearize the expression in S3 at thermal equilibrium:

[0061]

[0062] S5: Obtain the change in heat generated by increasing the constant voltage on the heating wire:

[0063] ΔQ i =K u ΔU

[0064] S6: Heat balance equation for heating wire:

[0065]

[0066] S7: Eliminating intermediate variables, the heat balance equation for the heating wire:

[0067]

[0068] S8: Let ΔT = T1 - T0, time constant T = MC / λA, K = K u / λA is the transfer coefficient of the electric radiation tube; under zero initial conditions, performing a Laplace transform on both sides of the above equation yields the transfer function between the heating wire temperature and the control voltage change:

[0069]

[0070] In summary, the present invention has the following beneficial effects:

[0071] A new type of electric radiant tube replaces the gas radiant tube. Compared with the traditional squirrel-cage electric radiant tube, the new electric radiant tube has a simpler structure, and the spiral magnetic column and disk have less thermal stress in high-temperature environments. In terms of the control system, a single neuron PID control model is selected, which makes the control more accurate and ensures the overall response to the process. Attached Figure Description

[0072] Figure 1This is a schematic diagram of the spiral-shaped electric radiation tube described in this invention.

[0073] Figure 2 This is a schematic diagram of the spiral-shaped electric radiation tube described in this invention.

[0074] Figure 3 This is a schematic diagram of the spiral electric radiation tube described in this invention.

[0075] Figure 4 This is a schematic diagram of the PID control system for a spiral electric radiation tube based on a neural network as described in this invention. Detailed Implementation

[0076] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below with reference to the figures and specific embodiments.

[0077] like Figures 1 to 4 As shown, the present invention proposes a 55KW electric radiant tube and control system for gas-to-electric conversion in a quenching furnace, comprising a spiral heating wire 1, a spiral magnetic column 2 for fixing and supporting the spiral heating wire, a disk 3 for supporting the spiral heating wire and the spiral magnetic column, and a radiant tube sleeve 4 fitted onto the spiral magnetic column and the disk, thereby realizing a control system for continuous adjustment and control of the electric radiant tube. The spiral heating wire 1 is fixed to the spiral magnetic column 2 by winding, and the spiral magnetic column 2 is supported and fixed by the disk 3. The spiral heating wire 1 is uniformly wound along the axial direction of the spiral magnetic column 2.

[0078] The spiral magnetic column 2 is a support column for the spiral heating wire 1. Its diameter is slightly smaller than the spiral inner diameter of the spiral heating wire 1, which facilitates the assembly between the spiral magnetic column 2 and the spiral heating wire.

[0079] The disk 3 has uniformly spaced holes along its circumference, with the hole diameter slightly larger than the outer diameter of the spiral heating wire 1, which facilitates the assembly of the disk 3 with the spiral heating wire 1 and the spiral magnetic post 2.

[0080] The spiral heating wire 1 consists of 6 wires, which are divided into two groups of 3 wires each. The two groups of heating wires are connected in parallel, and the 3 heating wires in the same group are connected in series.

[0081] The control system that enables continuous adjustment of the electric radiation tube adopts a PID control algorithm based on neural networks, thereby solving the shortcomings of traditional PID control algorithms, such as inability to accurately operate nonlinear systems and poor system parameter adaptive capability.

[0082] The aforementioned radiant tube sleeve 4 is a sleeve for a 70-80KW gas radiant tube used in a quenching furnace.

[0083] The spiral magnetic column 2 and the disk 3 are both made of high-temperature resistant alumina material. The addition of special materials ensures that the spiral magnetic column 2 and the disk 3 have excellent thermal shock resistance and high temperature resistance.

[0084] The effective cooperation between the spiral heating wire 1, the spiral magnetic column 2, and the disk 3, compared to the traditional squirrel-cage type electric radiation tube, results in less thermal stress generated by the high-temperature resistant materials spiral magnetic column 2 and disk 3 under high-temperature conditions, allowing for longer service life and thus improving the service life of the electric radiation tube.

[0085] A comparison of the thermal stress generated in a conventional squirrel-cage type electric radiation tube and the electric radiation tube of this invention under high-temperature conditions is as follows:

[0086] Using CFD simulation, physical models of a 55KW traditional squirrel-cage electric radiation tube and the electric radiation tube of this invention were established respectively. The thermal stress of the ceramic material of the two electric radiation tubes under high temperature environment was obtained through Fluent simulation.

[0087] Traditional squirrel-cage type electric radiant tube: It adopts U-shaped heating wires with a length of 1738mm. To achieve the rated power, 36 U-shaped heating wires are used, with a diameter of 8mm. The heating wire material is Shougang HRE material. A ceramic body support is used to support the U-shaped heating wires. The ceramic body support consists of a ceramic body shaft and ceramic plates. The ceramic body shaft is 1800mm long, the ceramic plate is 270mm in diameter, and the ceramic plate is 25mm thick. The radiant tube sleeve is made of 2535 steel. The voltage of a single radiant tube is 220V. When the furnace top temperature is set to 985℃ and the billet temperature to 900℃ according to the actual environment of the annealing furnace, the simulated average thermal stress of the ceramic body is 0.44~0.73×106Pa.

[0088] This invention patent describes an electric radiant tube that uses a spiral heating wire with an axial length of 948 mm. To achieve the rated power, six spiral heating wires are used, each with a diameter of 7.2 mm. The heating wire material is Shougang HRE material. A spiral magnetic column and a disk are used to support the spiral heating wire. The spiral magnetic column is 1080 mm long, and the disk has a diameter of 250 mm. The radiant tube sleeve is made of 2535 steel. The voltage of a single radiant tube is 220 V. Simulated average thermal stress of the spiral magnetic column is 0.09–0.31 × 10⁶ Pa, based on the actual environment of the annealing furnace with a furnace top temperature of 985℃ and a billet temperature of 900℃.

[0089] Simulation results show that, under the same high-temperature environment, the electric radiation tube of this invention, due to the effective combination of the spiral heating wire, spiral magnetic column, and disk, generates less thermal stress inside the high-temperature resistant ceramic material than the traditional squirrel-cage electric radiation tube, thus extending the service life of the electric radiation tube.

[0090] The spiral heating wire 1 can be arranged according to the internal cavity volume of the radiant tube sleeve 4 to make full use of the internal space of the radiant tube sleeve 4, and the number of spiral heating wires 1 is not specifically limited. In this embodiment, it is recommended to set two sets of spiral heating wires 1. The spiral heating wires 1 are arranged in parallel with the heating wires in series within each set. This can reduce the maximum temperature of the heating wires and extend the service life of the heating wires without changing the output power of the radiant tube.

[0091] A comparison of the heating wire temperature at the same power between the heating wire of the series-connected radiant tube and the radiant tube of this invention is as follows:

[0092] Using CFD simulation, physical models of an electric radiant tube with all heating wires connected in series and the electric radiant tube of the present invention were established respectively. Then, the temperature distribution of the heating wires of the two electric radiant tubes under the working environment was obtained by Fluent simulation.

[0093] All heating wires are connected in series: the heating wires are spiral heating wires, with a total of 6 wires, a diameter of 7.2mm, and an axial length of 948mm; the spiral magnetic column is 1000mm long and 30mm in diameter; the magnetic disk has a diameter of 250mm and a thickness of 40mm; the radiant tube sleeve is made of 2535 steel; the voltage across the radiant tube is 220V; the heating wires are grouped together and connected in series; when the furnace top temperature is set to 985℃ and the billet temperature to 900℃ according to the actual environment of the annealing furnace, the simulated highest temperature of the heating wire is 1381~1389℃, the lowest temperature is 1196~1223℃, and the average temperature is 1313~1327℃.

[0094] The electric radiant tube of this invention comprises: six spiral heating wires with a diameter of 7.5 mm and an axial length of 1660 mm; a spiral magnetic column with a length of 1720 mm and a diameter of 30 mm; a magnetic disk with a diameter of 270 mm and a thickness of 30 mm; a radiant tube sleeve made of 2535 steel; a voltage of 220V across the tube; two groups of heating wires connected in parallel between the groups and in series within each group; and simulations obtained by setting the furnace top temperature to 985℃ and the billet temperature to 900℃ based on the actual environment of the annealing furnace.

[0095] The minimum temperature of the heating element is 1085-1134℃, and the average temperature of the heating element is 1150-1173℃.

[0096] Simulation results show that the connection method of the heating wire in this invention results in a heating wire temperature significantly lower than that of traditional series-connected heating wires, with a temperature difference of approximately 150°C. The heating wire's high-temperature resistance limit is 1400°C, while the traditional connection method keeps the heating wire temperature around 1300°C, with a maximum temperature of 1380°C. The heating wire connection method in this invention can extend the service life of the heating wire, thereby increasing the service life of the radiant tube.

[0097] The neural network-based PID control algorithm described above uses a combination of a single neuron model and a traditional PID control model to achieve precise control of a 55KW electric radiation tube.

[0098] The parameters included in the model combining the single-neuron model and the traditional PID control model mainly include:

[0099] s v : Set temperature, °C;

[0100] p v The sensor collects the temperature, in °C.

[0101] K: System output gain;

[0102] Δu: Increment of system output control quantity;

[0103] u: The control quantity ultimately passed to the execution unit;

[0104] z -1 : Represents system latency;

[0105] Among them, s v and p v The system inputs x1, x2, and x3 of the single neuron model are obtained through corresponding converters, and the values ​​of the three weights w1, w2, and w3 are adjusted through a learning algorithm, thereby realizing the self-adjustment of system parameters in the control system.

[0106] According to the present invention, an electric radiant tube and its control system are characterized in that:

[0107] The system inputs x1, x2, and x3 at time k are respectively:

[0108] x1(k) = e(k) - e(k-1)

[0109] x2(k)=e(k)

[0110] x3(k)=e(k)-2e(k-1)+e(k-2) (1)

[0111] Where e(k) is the deviation between the set value and the system output value:

[0112] e(k) = sv -p v (2)

[0113] The system output increment Δu is:

[0114]

[0115] u(k)=Δu+u(k-1) (3)

[0116] Where K is the system gain, w i (k) represents the weighting coefficient.

[0117] The system adopts the Delta learning rule and takes into account the weighting coefficient w. i (k) should have a functional relationship with the neuron's input, output, and output bias. When using the Delta learning rule, the connection weights can be adjusted based on the difference between the neuron's actual output and its expected output.

[0118] Δw ij =a·(d i -y i )·x i (t)

[0119] Where, Δw ij To increase the weight, d i y is the expected output of neuron i. i 'i' is the actual output of neuron 'i', and 'a' is the learning speed.

[0120] In actual operation, the stability and correlation of a single-neuron PID control system are related to parameters such as K. After screening parameters used in actual production, adjustments are made to the relevant system parameters based on the following conditions:

[0121] (1) The initial weighting coefficients w1(0), w2(0), and w3(0) will not have any impact on the stability of the system operation. Therefore, the values ​​can be chosen arbitrarily.

[0122] (2) The value of the adaptive gain K of the system can be arbitrarily chosen first, and then corrected according to the simulation response curve until the most suitable value of K is found.

[0123] The heat transfer in the quenching furnace mainly includes heat conduction and heat radiation between the heating wire and the furnace lining, and between the furnace lining and the furnace chamber. There is a certain delay in temperature change inside the quenching furnace.

[0124] The weights can be adjusted online based on the learning ability and adaptability of the neural network, thereby improving the robustness of some nonlinear systems.

[0125] The derivation process of the transfer function of the PID control system is as follows:

[0126] S1: Assuming the external temperature of the quenching furnace is T2, the increase in heat dissipation per unit time is:

[0127] ΔQ0=λA[[(T1-T2)-(T0-T2)]=λA(T1-T0)

[0128] in,

[0129] ΔQ0: Heat dissipation, J;

[0130] λ: heat transfer coefficient;

[0131] A: Heat transfer area, m2;

[0132] T0: Furnace temperature before heating, in K;

[0133] T1: Furnace temperature after heating, in K;

[0134] T2: External temperature of the quenching furnace (K);

[0135] S2: Heat generated by the heating wire per unit time ΔQ i Subtracting the heat dissipated from the furnace wall surface ΔQ0 equals the change in thermal energy Q inside the furnace:

[0136] Q = CM(T1-T0) + Q0

[0137]

[0138] in,

[0139] C: Specific heat capacity of the heating wire;

[0140] M: Mass of heating wire, kg;

[0141] S3: The nonlinear relationship between the heat generated by the heating wire per unit time and the voltage U:

[0142]

[0143] S4: When the electric radiant tube reaches equilibrium, the heat generated by the heating wire is equal to the heat dissipated from the furnace wall, i.e., Q. i =Q0, linearize the expression in S3 at thermal equilibrium:

[0144]

[0145] S5: Obtain the change in heat generated by increasing the constant voltage on the heating wire:

[0146] ΔQ i =K u ΔU

[0147] S6: Heat balance equation for heating wire:

[0148]

[0149] S7: Eliminating intermediate variables, the heat balance equation for the heating wire:

[0150]

[0151] S8: Let ΔT = T1 - T0, time constant T = MC / λA, K = K u / λA is the transfer coefficient of the electric radiation tube. Under zero initial conditions, performing a Laplace transform on both sides of the above equation yields the transfer function between the heating wire temperature and the control voltage change:

[0152]

[0153] The single-neuron PID controller described above has more robust tracking characteristics and better robustness, as well as strong adaptability and high control accuracy.

[0154] In this document, the terms "upper," "lower," "front," "back," "left," "right," "top," "bottom," "inner," "outer," "vertical," and "horizontal," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only used for the clarity of expressing the technical solution and for the convenience of description, and therefore should not be construed as limiting the present invention.

[0155] In this document, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, which includes not only the elements listed but also other elements not expressly listed.

[0156] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A spiral electric radiant tube, comprising a spiral heating wire (1), a spiral magnetic column (2), a magnetic disk (3), and a radiant tube sleeve (4); characterized in that, The interior of the radiant tube sleeve (4) has a cavity structure, and a number of spiral magnetic columns (2) are provided inside the cavity structure. A number of disks (3) are sleeved on the spiral magnetic columns (2), and the spiral magnetic columns (2) are fixedly installed inside the radiant tube sleeve (4) through the disks (3). The spiral heating wire (1) is uniformly wound along the axial direction of the spiral magnetic column (2).

2. The spiral radiant tube according to claim 1, characterized in that, The spiral magnetic column (2) is a support column for the spiral heating wire (1). Its diameter is slightly smaller than the spiral inner diameter of the spiral heating wire (1), which facilitates the assembly between the spiral magnetic column (2) and the spiral heating wire (1). The disk (3) has holes evenly spaced along the circumference, with the hole diameter slightly larger than the outer diameter of the spiral heating wire (1), which facilitates the assembly of the disk (3) with the spiral heating wire (1) and the spiral magnetic column (2).

3. The spiral radiant tube according to claim 1, characterized in that, At least one set of the spiral heating wire (1) is provided.

4. The spiral electric radiation tube according to claim 3, characterized in that, The spiral heating wire (1) is divided into two groups. The two groups of spiral heating wires (1) are connected in parallel, and the spiral heating wires (1) within the same group are connected in series.

5. The spiral radiant tube according to claim 1, characterized in that, The radiant tube sleeve (4) is the sleeve of a 70-80KW gas radiant tube in a quenching furnace.

6. The spiral radiant tube according to claim 1, characterized in that, The spiral magnetic column (2) and the disk (3) are both made of high-temperature resistant alumina material.

7. A PID control system for a spiral radiating tube based on a neural network, characterized in that, A model combining a single-neuron model and a PID control model is used to precisely control a spiral electric radiation tube. The parameters included in the model combining the single neuron model and the PID control model are as follows: s v Set the temperature; p v The sensor collects temperature data. K: System output gain; Δu: Increment of system output control quantity; u: The control quantity ultimately passed to the execution unit; z -1 : Represents system latency; Among them, s v and p v The system inputs x1, x2, and x3 of the single neuron model are obtained through the corresponding converters, and the values ​​of the three weights w1, w2, and w3 are adjusted through the learning algorithm, thereby realizing the self-adjustment of system parameters in the control system.

8. The PID control system for a spiral radiating tube based on a neural network according to claim 7, characterized in that, The system inputs x1, x2, and x3 at time k are respectively: x1(k) = e(k) - e(k-1) x2(k)=e(k) x3(k)=e(k)-2e(k-1)+e(k-2)(1) Where e(k) is the deviation between the set value and the system output value: e(k)=s v -p v (2) System output increment Δ u for: in (k)= Δ u+u(k-1)(3) Where K is the system gain, w i (k) represents the weighting coefficient.

9. The PID control system for a spiral radiating tube based on a neural network according to claim 7, characterized in that, The PID control system employs the Delta learning rule and takes into account the weighting coefficient w. i (k) should have a functional relationship with the neuron's input, output, and output bias. When using the Delta learning rule, the connection weights can be adjusted based on the difference between the neuron's actual output and its expected output. Δw ij =a·(d i -y i )·x i (t) Where, Δw ij To increase the weight, d i y is the expected output of neuron i. i 'i' is the actual output of neuron 'i', and 'a' is the learning speed.

10. The PID control system for a spiral radiating tube based on a neural network according to claim 7, characterized in that, When the PID control system is working, the stability and correlation of the system are related to the parameters. After screening the parameters used in actual production, the relevant parameters of the system are adjusted accordingly based on the following conditions: 1) The initial weighting coefficients w1(0), w2(0), and w3(0) will not affect the stability of the system operation, so the values ​​can be chosen arbitrarily; 2) The adaptive gain K of the system can be arbitrarily chosen at first, and then corrected according to the simulation response curve until the most suitable K value is found.

11. The PID control system for a spiral radiating tube based on a neural network according to claim 6, characterized in that, The derivation process of the transfer function of the PID control system is as follows: S1: Assuming the external temperature of the quenching furnace is T2, the increase in heat dissipation per unit time is: ΔQ0=λA[(T1-T2)-(T0-T2)]=λA(T1-T0) in, ΔQ0: Heat dissipation, J; λ: heat transfer coefficient; A: Heat transfer area, m 2 ; T0: Furnace temperature before heating, in K; T1: Furnace temperature after heating, in K; T2: External temperature of the quenching furnace, K; S2: Heat generated by the heating wire per unit time ΔQ i Subtracting the heat dissipated from the furnace wall surface ΔQ0 equals the change in thermal energy Q inside the furnace: Q = CM(T1-T0) + Q0 in, C: Specific heat capacity of the heating wire; M: Mass of heating wire, kg; S3: The nonlinear relationship between the heat generated by the heating wire per unit time and the voltage U: S4: When the electric radiant tube reaches equilibrium, the heat generated by the heating wire is equal to the heat dissipated from the furnace wall, i.e., Q. i =Q0, linearize the equation in S3 at thermal equilibrium: S5: Obtain the change in heat generated by increasing the constant voltage on the heating wire: ΔQ i =K u ΔU S6: Heat balance equation for heating wire: S7: Eliminating intermediate variables, the heat balance equation for the heating wire: S8: Let ΔT = T1 - T0, time constant T = MC / λA, K = K u / λA is the transfer coefficient of the electric radiation tube; under zero initial conditions, performing a Laplace transform on both sides of the above equation yields the transfer function between the heating wire temperature and the control voltage change:

Images