Conductive pump wireless driving control method and device based on deep learning

Abstract

The invention provides a conductive pump wireless driving control method and device based on deep learning, and relates to the technical field of fluid transmission. The driving device comprises a microwave transmitting assembly, a microwave receiving assembly, a conductive pump and a control unit; the microwave transmitting assembly comprises a microwave generator, a power amplifier and a directional antenna; using a deep learning method to obtain pose control parameters, changing the polarity and transmission path of the microwave emission assembly, controlling the circuit to open and close, and adjusting the output flow of the conductive pump; the polarization mismatch loss can be regulated and controlled in real time, and the energy transmission efficiency is maximized. A traditional wired driving mode is broken through, wireless driving of the conductive pump is achieved, the size and the mass can be reduced, the output flow precision is improved, the device is suitable for the fields of conductive pump systems, medical equipment, industrial fluid control and the like, and a new thought is brought to driving design and application of the conductive pump.

Description

Technical Field

[0001] The present invention relates to the field of fluid transmission technology, and in particular to a deep learning-based wireless drive control method and device for a conductivity pump. Background Art

[0002] In cutting-edge technologies like microfluidic chips and implantable medical devices, precise fluid control plays a crucial role in device performance. Traditional conductivity pumps often utilize wired power supply drives. However, these external wires significantly increase the size of the device, severely hindering its development towards miniaturization and micro-miniaturization. Furthermore, wiring is extremely difficult in enclosed spaces or applications requiring frequent movement, such as implantable medical devices or sealed reaction chambers, significantly limiting the device's installation and flexibility. Furthermore, long-term use often leads to wire aging, causing equipment failures, increasing maintenance costs, and decreasing reliability, making it difficult to meet the demands of modern precision fluid control scenarios.

[0003] Although wireless energy transmission technologies such as electromagnetic induction and magnetic resonance have been gradually applied to the field of device driving in an attempt to solve the limitations of wired driving, these technologies generally have technical bottlenecks such as short transmission distance and low energy transmission efficiency. In practical applications, they cannot stably and efficiently provide sufficient energy for conductivity pumps.

[0004] In terms of power control, existing conductivity pump systems primarily rely on manually manipulated voltage parameter settings or open-loop voltage modulation. This control approach not only suffers from poor control accuracy, making it difficult to precisely regulate fluid flow and pressure, but also suffers from significant hysteresis. In fields such as microfluidic chemical analysis and biomedical drug delivery, where fluid manipulation precision and response speed are critical, these performance shortcomings severely restrict their application effectiveness and potential. Therefore, the development of fluid control systems with wireless actuation, adaptive regulation, and high dynamic response characteristics has become an urgent need to overcome current technological bottlenecks and promote innovation and development in related fields. Summary of the Invention

[0005] To address the shortcomings of the existing technology, the present invention provides a deep learning-based wireless drive control method and device for a conductivity pump. A microwave transmitter transmits a microwave signal of a specific frequency, and a microwave receiving module built into the conductivity pump converts the microwave energy into electrical energy to drive the conductivity pump, thereby resolving problems such as spatial limitations in a wired environment. The control unit adjusts the multi-parameter spatial position of the microwave receiving component to change the polarity coupling degree and path loss ratio between the transmitting and receiving components, thereby precisely regulating the received power and controlling the output pressure and flow of the conductivity pump. This method has important engineering practical significance.

[0006] In the far field, antennas radiate transverse electromagnetic waves, meaning the electric and magnetic field components are perpendicular to each other and to the direction of propagation. Based on the intensity of the electric field at a certain location at different times, it can be divided into linear polarization, circular polarization, and elliptical polarization. The polarization direction is related to the spatial position of the antenna. When the polarization direction of the receiving component is the same as that of the transmitting component, the receiving efficiency is the highest; when there is an angle, the received power will be significantly reduced due to polarization mismatch, and its efficiency attenuation follows the rule μ = μ. max cos 2 By dynamically adjusting the polarization direction, the mismatch loss can be suppressed to the theoretical minimum, thereby achieving global optimization of energy transmission efficiency.

[0007] During microwave transmission, the path loss formula is L(d) = k0 + k1d -n +k2e -αd , where d is the distance, n is the path loss exponent, and α is the attenuation coefficient. It can be seen that energy transmission efficiency is nonlinear. Traditional voltage regulation can only compensate for first-order linear errors in a fixed spatial configuration. Displacement modulation, by reconstructing spatial coupling parameters in real time, can directly correct higher-order nonlinear terms and significantly reduce its steady-state error.

[0008] Specifically, on the one hand, the present invention provides a wireless drive control method for a conductivity pump based on deep learning, which comprises the following steps:

[0009] S1. Connect the microwave transmitting assembly to the power supply and perform accuracy detection through the power detection unit installed on the microwave receiving assembly;

[0010] S2. Prefilling the microfluidic channel with charged ionic liquid, measuring the electrode pair impedance of the conductivity pump, and setting the target flow rate Q0 of the microcontroller; setting the initial angle of the polarized antenna of the microwave transmitting component and the initial distance between the microwave receiving component and the microwave transmitting component;

[0011] S3. Mechanically connect the dual-axis stepper motor of the control unit to the polarized antenna of the microwave transmitting assembly. The control unit drives the dual-axis stepper motor to drive the polarized antenna to rotate and scan clockwise. The power detection unit collects the received signal strength indication value in real time and records the angle-power mean correspondence relationship data set {P(α)|α∈[0°,360°)}. The control unit controls the polarized antenna to adjust to the angle α corresponding to the maximum received power in the data set. max ;

[0012] S4. Determine whether the angle α is reached max After the corresponding accuracy range is reached, the control unit sends an opening electrical signal to the relay switch, and the microwave receiving component is connected to the conductivity pump to drive the conductivity pump to work;

[0013] S5. During the injection process, the dynamic parameter flow rate Q1 of the fluid in the microfluidic channel is continuously monitored. The microcontroller compares the dynamic parameter flow rate Q1 with the preset target value Q0 and calculates the relative error: ΔQ = |(Q1-Q0) / Q0| × 100%. When ΔQ exceeds the threshold range, closed-loop control is initiated. When ΔQ is within the threshold range, the process proceeds to step S6.

[0014] S6. If the flow fluctuation rate is less than the threshold, it is determined that the system has entered a stable state. The current angle α1 of the polarized antenna and the distance L1 between the microwave receiving component and the microwave transmitting component are recorded and used as initial parameters for the same working condition.

[0015] S7. After entering the stable state, if the ΔQ fluctuation rate is greater than the threshold, the control unit sends a closing electrical signal to the relay switch and repeats steps S3 to S6.

[0016] Preferably, in step S3, the control unit performs Gaussian filtering on the angle-power data set and determines the angle α corresponding to the maximum received power by using an extreme value detection algorithm. max ; Drive the dual-axis stepper motor to adjust the angle of the polarized antenna of the microwave transmitting component to angle α max The corresponding accuracy range α max ±0.25°.

[0017] Preferably, in step S4, the microwave receiving component is connected to the conductivity pump and the specific steps of driving the conductivity pump to work are as follows: after the microwave is received by the coupling antenna, the high-frequency microwave energy is rectified into pulsating DC by a bridge rectifier after impedance matching and high-frequency filtering through the input filter network, and the filtering outputs a smooth DC voltage to the oscillation control unit. The oscillation control unit sets the pulse parameters and limits the current before driving the primary coil of the step-up transformer to form a high-frequency square wave AC signal. The high-frequency voltage after the transformer boosts the voltage is input to the three-stage voltage multiplier, which outputs high-voltage DC in a step-by-step manner to drive the micro conductivity pump to work, so that the charged ionic liquid continues to pass through the microfluidic channel.

[0018] Preferably, the power detection unit performs precision detection in step S1 as follows: the power detection unit verifies whether the output power precision error of the microwave transmitting component at the frequency point is less than ±5%, if so, proceeds to the next step; otherwise, replaces the polarized antenna of the microwave receiving component or performs equipment inspection.

[0019] Preferably, in step S3 , the control unit drives the dual-axis stepping motor to perform a 720° clockwise rotation scan at an angular velocity of θ=15° / s.

[0020] Preferably, in step S4, the microwave passes through the input filter network composed of L1 and C1 for impedance matching and high-frequency filtering, and then the high-frequency microwave energy is rectified into pulsating DC by a bridge rectifier composed of diodes D1-D4, and then filtered by L2 and C2 to output a smooth DC voltage to supply the oscillation control unit. The oscillation control unit sets the pulse parameters through R1, R2 and C4, and drives the primary coil of the step-up transformer after current limiting by R3 to form a high-frequency square wave AC signal.

[0021] Preferably, the volatility threshold in step S6 is 1%, and the volatility calculation formula is:

[0022]

[0023] Where k is the time point, N is the number of sampling cycles, Q1(i) is the flow rate at the i-th time point, is the average flow parameter of N cycles.

[0024] Preferably, the specific process of closed-loop control in step S5 is:

[0025] The microcontroller drives the motor to drive the microwave emitting component to perform intermittent reciprocating motion at a uniform distance in a straight line, continuously monitoring the flow rate Q1 in the microfluidic channel. The detection algorithm drives the microwave emitting component to move to the distance range corresponding to the required flow rate Q0, and uses a collaborative control model of deep reinforcement learning and adaptive temporal convolutional network to obtain specific control parameters. The specific steps are as follows:

[0026] S51. Use reinforcement learning algorithm to define Markov decision process, optimize relative distance, and design state space formula as s k =[Q1(kT:k),L(k),ΔQ(k),V(k)], Q1(kT:k),L(k),ΔQ(k),V(k) are the historical flow rate, the relative distance between the microwave transmitting component and the microwave receiving component, the flow rate error and the fluctuation rate respectively; take the action space a k =ΔL(k)∈[-3cm,3cm], the microcontroller drives the dual-axis stepper motor to drive the microwave emission component to move, and the reward function r is defined k =-(ω1·|ΔQ(k)|)+ω2·|ΔL(k)|+ω3·Ι, where Ι is the violation penalty term, ω1, ω2, and ω3 are the weights of the flow error, the distance error between the microwave transmitting component and the microwave receiving component, and the weight of the violation penalty term, respectively, to obtain the optimal moving distance ΔL DRL (k);

[0027] S52. Use the proximal strategy to optimize the strategy network parameter φ:

[0028] A(s,a) is the advantage function, a is the action parameter, s is the state parameter, ε is the shear threshold, θ0 is the initial angle of the polarized antenna of the microwave transmitting component, E is the expectation of all state parameters s and action parameters a, and a temporal convolutional network is designed to predict traffic error. The input is the historical flow {Q1(kT),...,Q1(k)} and the historical distance variable {ΔL(kT),...,ΔL(k)};

[0029] S53, residual connection and gating mechanism predict the compensation amount ΔL based on Output = tanh(W1*z)⊙σ(W2*z) c (k), we get the moving distance ΔL(k)=ΔL DRL (k)+ΔL c (k).

[0030] Preferably, ΔL(k) is used as the distance adjustment amount. When ΔL(k) ≥ 0, the dual-axis stepper motor drives the microwave transmitting component to move ΔL(k) toward the microwave receiving component through the slide rail; when ΔL(k) ≤ 0, the microwave transmitting component is driven away from the receiving component by |ΔL(k)|; after the movement is executed, the actual distance L(k) = L(k-1) + ΔL(k) is updated.

[0031] Preferably, it comprises a microwave transmitting component, a microwave receiving component, a conductivity pump and a control unit;

[0032] The microwave transmitting component includes a microwave generator, a power amplifier and a directional antenna; the frequency range of the microwave transmitting component is 1-10GHz; the microwave receiving component includes a coupling antenna, a resonant matching network, a rectifier bridge, a filter, an RC network, a boost control chip, a current limiting resistor, a transformer, a voltage doubler rectifier module and a load interface; the conductivity pump includes a microfluidic channel and an electrode pair; the control unit includes a relay switch, a flow sensor, a power detection unit, a microcontroller, an acquisition board, a dual-axis stepper motor and a turntable, which changes the polarity of the microwave transmitting component and the transmission path loss through posture modulation, controls the circuit to open and close, and adjusts the output flow of the conductivity pump.

[0033] Compared with the prior art, the advantages of the present invention are as follows:

[0034] (1) The present invention transmits microwave signals of a specific frequency through a microwave transmitter, and the microwave receiving module built into the conductivity pump converts the microwave energy into electrical energy to drive the conductivity pump to work, thereby solving the problem of space limitations in a wired environment;

[0035] (2) The method of the present invention can realize long-distance wireless energy transmission: Based on the principle that the antenna structure resonates with the specific frequency of microwaves to efficiently receive energy and the polarization characteristics of the transverse electromagnetic waves radiated by the antenna in the far field, the present invention adjusts the spatial position of the microwave transmitting component through the control unit to achieve dynamic matching of the polarization direction and maximize the energy transmission efficiency.

[0036] (3) The control unit changes the polarity coupling degree and path loss ratio between the transmitting component and the receiving component by adjusting the multi-parameters of the spatial position of the microwave receiving component, thereby accurately regulating the receiving power. The wireless drive control method and device of the conductivity pump based on deep learning integrates polarization matching, path loss correction and load feedback control, and controls the output pressure and flow of the conductivity pump through multi-parameter joint optimization (such as displacement, angle, and impedance matching), which has important practical engineering significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is an overall schematic diagram of the device of the present invention;

[0038] Figure 2 This is a schematic structural diagram of a microwave receiving assembly according to the present invention;

[0039] Figure 3 Schematic diagram of the adaptive control process of the present invention;

[0040] Figure 4 It is a schematic block diagram of the overall structure of the present invention;

[0041] Figure 5 It is a schematic diagram of the overall process of the present invention. DETAILED DESCRIPTION

[0042] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

[0043] Specifically, on the one hand, the present invention provides a conductivity pump driving device based on wireless energy transmission technology, such as Figure 1 As shown, it includes a microwave transmitting component, a microwave receiving component, a conductivity pump and a control unit.

[0044] The microwave transmitting component includes a microwave generator, a power amplifier and a directional antenna; the frequency range of the microwave transmitting component is 1-10GHz; the microwave receiving component includes a coupling antenna, a resonant matching network, a rectifier bridge, a filter, an RC network, a boost control chip, a current limiting resistor, a transformer, a voltage doubler rectifier module and a load interface; the conductivity pump includes a microfluidic channel and an electrode pair; the control unit includes a relay switch, a flow sensor, a pressure sensor, a power detection unit, a microcontroller, an acquisition board, a dual-axis stepper motor and a turntable, which changes the polarity of the microwave receiving component and the transmission path loss through posture modulation, and controls the circuit to open and close and adjust the output flow of the conductivity pump.

[0045] like Figure 2 and Figure 4 As shown, in the microwave receiving component, the coupling antenna is specifically the coupling antenna E2, the inductor L1 and the capacitor C1 form a resonant matching network, the diodes D1-D4 form a rectifier bridge, the inductor L2 and the capacitor C2 form a filter, the resistors R1, R2 and the capacitor C4 form an RC network, the boost control chip is an 8-pin boost control chip, and the current limiting resistor R3, the transformer, the diodes D10-D14, the capacitors C5-C9, and the C6-C8 form a voltage doubling rectifier module.

[0046] On the other hand, the present invention provides a wireless drive control method for a conductivity pump based on deep learning, such as Figure 3 and Figure 5 As shown, it includes the following steps:

[0047] S1. Install dual stepper motors in a cross-axis configuration on the guide rails. A turntable is installed above the motors and physically connected to the antenna in the transmitting assembly, securing the microwave receiving assembly. Power is supplied to the microwave transmitting assembly, and an accuracy test is performed using a power detection unit installed on the microwave receiving assembly. In a specific embodiment, the accuracy test performed by the power detection unit specifically verifies whether the output power accuracy error of the microwave transmitting assembly at a specific frequency point is less than ±5%. If so, the next step is performed. Otherwise, the polarized antenna of the microwave receiving assembly is replaced or the equipment is inspected. The frequency range of the microwave transmitting assembly is 1-10 GHz, with the 2.45 GHz ISM (Industrial, Scientific, and Medical Band) being preferred.

[0048] S2. Use charged ionic liquid with a carbon source to pre-fill the microfluidic channel, measure the electrode pair impedance of the conductivity pump, set the target flow rate Q0 of the microcontroller; and set the initial angle of the polarized antenna of the microwave transmitting component and the distance between it and the microwave receiving component.

[0049] S3. Mechanically connect the dual-axis stepper motor of the control unit to the polarized antenna of the microwave transmitting assembly. The control unit drives the dual-axis stepper motor to drive the polarized antenna to rotate and scan clockwise. The power detection unit collects the received signal strength indication value in real time and records the angle-power mean correspondence relationship data set {P(α)|α∈[0°,360°)}. The control unit controls the polarized antenna to adjust to the angle α corresponding to the maximum received power in the data set. max .

[0050] Specifically, the control unit drives the dual-axis stepper motor to perform a 720° clockwise rotation scan at an angular velocity of θ = 15° / s. After the control unit performs Gaussian filtering on the angle-power data set, it determines the angle α corresponding to the maximum received power through the extreme value detection algorithm. max; Drive the dual-axis stepper motor to adjust the angle of the polarized antenna of the microwave transmitting component to angle α max The corresponding accuracy range α max ±0.25°.

[0051] S4. Determine whether the angle α is reached max After the corresponding accuracy range is reached, the control unit sends an opening electrical signal to the relay switch, the microwave receiving component is connected to the conductivity pump, and the microwave is received by the coupling antenna. After impedance matching and high-frequency filtering are performed through the input filter network, the high-frequency microwave energy is rectified into pulsating DC through the bridge rectifier, and the filtering outputs a smooth DC voltage to the oscillation control unit. The oscillation control unit sets the pulse parameters and drives the primary coil of the step-up transformer after current limiting to form a high-frequency square wave AC signal. The high-frequency voltage after the transformer step-up is input to the three-stage voltage multiplier, and the high-voltage DC is output step by step, driving the micro conductivity pump to work and continuously injecting carbon source liquid containing charged ions into the microbial reaction chamber.

[0052] Specifically, the microwave passes through the input filter network composed of L1 and C1 for impedance matching and high-frequency filtering, and then the high-frequency microwave energy is rectified into pulsating DC through the bridge rectifier composed of diodes D1-D4. The smooth DC voltage is then filtered and output by L2 and C2 to supply the oscillation control unit. The oscillation control unit sets the pulse parameters through R1, R2 and C4, and drives the primary coil of the step-up transformer after current limiting by R3 to form a high-frequency square wave AC signal.

[0053] S5. During the injection process, the dynamic parameter flow rate Q1 of the fluid in the microfluidic channel is continuously monitored. The microcontroller compares the dynamic parameter flow rate Q1 with the preset target value Q0 and calculates the relative error: ΔQ = |(Q1-Q0) / Q0|×100%. When ΔQ exceeds the threshold range, closed-loop control is started. When ΔQ is within the threshold range, step S6 is entered.

[0054] In a specific embodiment, the specific process of closed-loop control is:

[0055] The microcontroller drives the motor to drive the microwave emitting component to perform intermittent reciprocating motion at a uniform distance in a straight line. The steady-state flow rate Q1 in the microfluidic channel is continuously monitored. The detection algorithm drives the microwave emitting component to move to the distance range corresponding to the required flow rate Q0. A collaborative control model of deep reinforcement learning and adaptive temporal convolutional network is used. The specific steps are as follows:

[0056] S51, use the reinforcement learning algorithm (DRL) to define the Markov decision process, optimize the relative distance, and design the state space formula as s k=[Q1(kT:k),L(k),ΔQ(k),V(k)], Q1(kT:k),L(k),ΔQ(k),V(k) are the historical flow rate, the relative distance between the microwave transmitting component and the microwave receiving component, the flow rate error and the fluctuation rate respectively; take the action space a k =ΔL(k)∈[-3cm,3cm], the microcontroller drives the dual-axis stepper motor to drive the microwave emission component to move, and the reward function r is defined k =-(ω1·|ΔQ(k)|)+ω2·|ΔL(k)|+ω3·Ι, where Ι is the violation penalty term, ω1, ω2, and ω3 are the weights of the flow error, the distance error between the microwave transmitting component and the microwave receiving component, and the weight of the violation penalty term, respectively, to obtain the optimal moving distance ΔL DRL (k).

[0057] S52. Use the proximal strategy to optimize the strategy network parameter φ:

[0058] A(s,a) is the advantage function, a is the action parameter, s is the state parameter, ε is the shear threshold, θ0 is the initial angle of the polarized antenna of the microwave transmitting component, E is the expectation of all state parameters s and action parameters a, and a temporal convolutional network is designed to predict traffic error. The input is the historical traffic {Q1(kT),......,Q1(k)} and the historical distance variable {ΔL(kT),......,ΔL(k)}; at the same time, the expansion causal convolution layer is designed l is the expansion coefficient level, d is the convolution kernel size; the final prediction value ΔL DRL (k).

[0059] S53, residual connection and gating mechanism predict the compensation amount ΔL based on Output = tanh(W1*z)⊙σ(W2*z) c (k), we get the moving distance ΔL(k)=ΔL DRL (k)+ΔL c (k).

[0060] Preferably, ΔL(k) is used as the distance adjustment amount. When ΔL(k) ≥ 0, the dual-axis stepper motor drives the microwave transmitting component to move ΔL(k) toward the microwave receiving component through the slide rail; when ΔL(k) ≤ 0, the microwave transmitting component is driven away from the receiving component by |ΔL(k)|; after the movement is executed, the actual distance L(k) = L(k-1) + ΔL(k) is updated.

[0061] S6. If the fluctuation rate in any k-th period within N sampling periods is less than the fluctuation rate threshold, it is determined that the system has entered a stable state. The current angle α1 of the polarized antenna and the distance L1 between the microwave receiving component and the microwave transmitting component are recorded and used as initial parameters for the same working condition.

[0062] In a specific embodiment, the volatility threshold in step S6 is 1%, and the volatility calculation formula is:

[0063]

[0064] Where k is the time point, N is the number of sampling cycles, Q1(i) is the flow rate at the i-th time point, is the average flow parameter of N cycles.

[0065] S7. After entering the stable state, if the ΔQ fluctuation rate is greater than the threshold, the control unit sends a closing electrical signal to the relay switch and repeats steps S3 to S6. Specific embodiments

[0067] This embodiment demonstrates the application of the microwave-driven conductivity pump of the present invention in a microbial culture system to achieve continuous and precise injection of a carbon source, thereby dynamically regulating the metabolic state of the microorganisms and improving culture efficiency and product consistency.

[0068] A dual stepper motor is installed on the guide rail in a cross-axis manner, and a turntable is installed above the motor, which is physically connected to the antenna in the transmitting component to fix the microwave receiving component.

[0069] First, connect the microwave transmitting component to a power source and use a spectrum analyzer to configure the directional antenna's beamwidth to a 30° half-power angle. Use the power detection unit installed on the microwave receiving component to verify that the transmitter's output power accuracy at the frequency point is <±5%. If not, replace the antenna or perform an equipment inspection.

[0070] The microchannel was prefilled with a charged ionic liquid containing a carbon source, the electrode pair impedance of the conductivity pump was measured, and the target flow rate Q0 = 6 μL / min was set for the microcontroller; the initial angle and distance were set according to empirical formulas; and the maximum output power of the microwave emission component was limited to 120 W and the electrode alarm threshold I = 500 mA.

[0071] The stepper motor is mechanically connected to the polarized antenna of the microwave transmitting assembly. The control unit drives the dual-axis stepper motor to perform a 720° clockwise rotation scan at an angular velocity of θ = 15° / s. The power detection unit collects the received signal strength indicator value in real time and records the angle-power mean correspondence data set {P(α)|α∈[0°,360°)}. The control unit performs Gaussian filtering on the angle-power data set and determines the angle α corresponding to the maximum received power using an extreme value detection algorithm. max; Drive the stepper motor to adjust the polarized antenna to α max ±0.25° accuracy range.

[0072] Upon determining that the coupling angle range has been reached, the control unit sends an electrical signal to the relay switch, connecting the microwave receiving assembly to the conductivity pump. The microwaves are received by the coupling antenna and then pass through the input filter network formed by L1 and C1 for impedance matching and high-frequency filtering. A bridge rectifier consisting of Schottky diodes D1-D4 rectifies the high-frequency microwave energy into pulsating DC, which is then filtered by L2 and C2 to output a smooth DC voltage. This DC voltage is supplied to the oscillator control unit based on a timing chip (NE555). This unit sets the pulse parameters using R1, R2, and C4. After current limiting by R3, it drives the primary coil of a step-up transformer, generating a high-frequency square wave AC signal. The stepped-up high-frequency voltage is then fed into a three-stage voltage multiplier consisting of D10-D14, C5-C9, and C6-C8. This voltage is then added to the output of a high-voltage DC signal, ultimately driving the microconductivity pump, which continuously injects the charged ionic liquid carbon source into the microbial reactor.

[0073] Flow and pressure sensors continuously monitor the dynamic parameters of the fluid within the microchannel (flow rate Q1) at a 60Hz sampling rate. The microcontroller compares the real-time data with the preset target value and calculates the relative error: ΔQ = |(Q1-Q0) / Q0| × 100%. When the threshold is exceeded, closed-loop control is initiated.

[0074] After steady state, if the ΔQ fluctuation is greater than 1%, the control unit drives the stepper motor to perform a 720° clockwise rotation scan at an angular velocity of θ = 15° / s. The stepper motor is mechanically connected to the polarized antenna of the microwave transmitting component; the received signal strength indicator value is collected in real time through the power detection unit, and the angle-power correspondence data set {P(α)|α∈[0°, 360°)} is recorded.

[0075] After the control unit performs Gaussian filtering on the angle-power data set, it determines the angle α corresponding to the maximum received power through the extreme value detection algorithm. max ; Drive the stepper motor to adjust the polarized antenna to α max The accuracy range is ±0.25°; on this basis, repeat the distance adjustment and steady-state determination.

[0076] During operation, if the current exceeds the alarm threshold, the system will automatically power off.

[0077] The embodiments described above are merely descriptions of preferred implementations of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should fall within the scope of protection determined by the claims of the present invention.

Claims

1. A wireless drive control method for a conductivity pump based on deep learning, characterized by: It includes the following steps: S1. Connect the microwave transmitting assembly to the power supply and perform accuracy detection through the power detection unit installed on the microwave receiving assembly; S2. Prefilling the microfluidic channel with charged ionic liquid, measuring the electrode pair impedance of the conductivity pump, and setting the target flow rate Q0 of the microcontroller; setting the initial angle of the polarized antenna of the microwave transmitting component and the initial distance between the microwave receiving component and the microwave transmitting component; S3. Mechanically connect the dual-axis stepper motor of the control unit to the polarized antenna of the microwave transmitting assembly. The control unit drives the dual-axis stepper motor to drive the polarized antenna to rotate and scan clockwise. The power detection unit collects the received signal strength indication value in real time and records the angle-power mean correspondence relationship data set {P(α)|α∈[0°,360°)}. The control unit controls the polarized antenna to adjust to the angle α corresponding to the maximum received power in the data set. max ; S4. Determine whether the angle α is reached max After the corresponding accuracy range is reached, the control unit sends an opening electrical signal to the relay switch, and the microwave receiving component is connected to the conductivity pump to drive the conductivity pump to work; S5. During the injection process, the dynamic parameter flow rate Q1 of the fluid in the microfluidic channel is continuously monitored. The microcontroller compares the dynamic parameter flow rate Q1 with the preset target value Q0 and calculates the relative error: ΔQ = |(Q1-Q0) / Q0| × 100%. When ΔQ exceeds the threshold range, closed-loop control is initiated. When ΔQ is within the threshold range, the process proceeds to step S6. S6. If the flow fluctuation rate is less than the threshold, it is determined that the system has entered a stable state. The current angle α1 of the polarized antenna and the distance L1 between the microwave receiving component and the microwave transmitting component are recorded and used as initial parameters for the same working condition. S7. After entering the stable state, if the ΔQ fluctuation rate is greater than the threshold, the control unit sends a closing electrical signal to the relay switch and repeats steps S3 to S6.

2. The deep learning-based wireless drive control method for a conductivity pump according to claim 1, characterized in that: In step S3, the control unit performs Gaussian filtering on the angle-power data set and determines the angle α corresponding to the maximum received power by using an extreme value detection algorithm. max ; Drive the dual-axis stepper motor to adjust the angle of the polarized antenna of the microwave transmitting component to angle α max The corresponding accuracy range α max ±0.25°.

3. The deep learning-based wireless drive control method for a conductivity pump according to claim 1, characterized in that: In step S4, the microwave receiving component is connected to the conductivity pump and the specific steps of driving the conductivity pump to work are as follows: after the microwave is received by the coupling antenna, the high-frequency microwave energy is rectified into pulsating DC by the bridge rectifier after impedance matching and high-frequency filtering through the input filter network, and the filtering outputs a smooth DC voltage to the oscillation control unit. The oscillation control unit sets the pulse parameters and limits the current before driving the primary coil of the step-up transformer to form a high-frequency square wave AC signal. The high-frequency voltage after the transformer boosts the voltage is input to the three-stage voltage multiplier, which outputs high-voltage DC in a step-by-step manner to drive the micro conductivity pump to work, so that the charged ionic liquid continues to pass through the microfluidic channel.

4. The deep learning-based wireless drive control method for a conductivity pump according to claim 1, characterized in that: In step S1, the power detection unit performs precision detection as follows: the power detection unit verifies whether the output power precision error of the microwave transmitting component at the frequency point is less than ±5%. If so, the next step is entered; otherwise, the polarized antenna of the microwave receiving component is replaced or the equipment is inspected.

5. The deep learning-based wireless drive control method for a conductivity pump according to claim 2, characterized in that: In step S3 , the control unit drives the biaxial stepping motor to perform a 720° clockwise rotation scan at an angular velocity of θ=15° / s.

6. The deep learning-based wireless drive control method for a conductivity pump according to claim 3, characterized in that: In step S4, the microwave passes through the input filter network composed of L1 and C1 for impedance matching and high-frequency filtering, and then the high-frequency microwave energy is rectified into pulsating DC by the bridge rectifier composed of diodes D1-D4. The smooth DC voltage is then filtered and output by L2 and C2 to supply the oscillation control unit. The oscillation control unit sets the pulse parameters through R1, R2 and C4, and drives the primary coil of the step-up transformer after current limiting by R3 to form a high-frequency square wave AC signal.

7. The deep learning-based wireless drive control method for a conductivity pump according to claim 3, characterized in that: In step S6, the volatility threshold is 1%, and the volatility calculation formula is: Where k is the time point, N is the number of sampling cycles, Q1(i) is the flow rate at the i-th time point, is the average flow parameter of N cycles.

8. The deep learning-based wireless drive control method for a conductivity pump according to claim 1, characterized in that: The specific process of closed-loop control in step S5 is: The microcontroller drives the motor to drive the microwave emitting component to perform intermittent reciprocating motion at a uniform distance in a straight line, continuously monitoring the flow rate Q1 in the microfluidic channel. The detection algorithm drives the microwave emitting component to move to the distance range corresponding to the required flow rate Q0, and uses a collaborative control model of deep reinforcement learning and adaptive temporal convolutional network to obtain specific control parameters. The specific steps are as follows: S51. Use reinforcement learning algorithm to define Markov decision process, optimize relative distance, and design state space formula as s k =[Q1(kT:k),L(k),ΔQ(k),V(k)], Q1(kT:k),L(k),ΔQ(k),V(k) are the historical flow rate, the relative distance between the microwave transmitting component and the microwave receiving component, the flow rate error and the fluctuation rate respectively; take the action space a k =ΔL(k)∈[-3cm,3cm], the microcontroller drives the dual-axis stepper motor to drive the microwave emission component to move, and the reward function r is defined k =-(ω1·|ΔQ(k)|)+ω2·|ΔL(k)|+ω3·Ι, where Ι is the violation penalty term, ω1, ω2, and ω3 are the weights of the flow error, the distance error between the microwave transmitting component and the microwave receiving component, and the weight of the violation penalty term, respectively, to obtain the optimal moving distance ΔL DRL (k); S52. Use the proximal strategy to optimize the strategy network parameter φ: A(s,a) is the advantage function, a is the action parameter, s is the state parameter, ε is the shear threshold, θ0 is the initial angle of the polarized antenna of the microwave transmitting component, E is the expectation of all state parameters s and action parameters a, and a temporal convolutional network is designed to predict traffic error. The input is the historical flow {Q1(kT),...,Q1(k)} and the historical distance variable {ΔL(kT),...,ΔL(k)}; S53, residual connection and gating mechanism predict the compensation amount ΔL based on Output = tanh(W1*z)⊙σ(W2*z) c (k), we get the moving distance ΔL(k)=ΔL DRL (k)+ΔL c (k).

9. The deep learning-based wireless drive control method for a conductivity pump according to claim 8, characterized in that: Taking ΔL(k) as the distance adjustment value, when ΔL(k)≥0, the dual-axis stepper motor drives the microwave transmitting assembly to move ΔL(k) toward the microwave receiving assembly through the slide rail; when ΔL(k)≤0, the microwave transmitting assembly is driven away from the receiving assembly by |ΔL(k)|; after the movement is executed, the actual distance L(k)=L(k-1)+ΔL(k) is updated.

10. A conductance pump driving device based on wireless energy transmission technology for use in the deep learning-based conductance pump wireless driving control method according to claim 1, characterized in that: It includes a microwave transmitting component, a microwave receiving component, a conductivity pump and a control unit; The microwave transmitting component includes a microwave generator, a power amplifier and a directional antenna; the frequency range of the microwave transmitting component is 1-10GHz; the microwave receiving component includes a coupling antenna, a resonant matching network, a rectifier bridge, a filter, an RC network, a boost control chip, a current limiting resistor, a transformer, a voltage doubler rectifier module and a load interface; the conductivity pump includes a microfluidic channel and an electrode pair; the control unit includes a relay switch, a flow sensor, a power detection unit, a microcontroller, an acquisition board, a dual-axis stepper motor and a turntable, which changes the polarity of the microwave transmitting component and the transmission path loss through posture modulation, controls the circuit to open and close, and adjusts the output flow of the conductivity pump.

Images