Solution filling control method and system

Abstract

The present invention discloses a solution filling control method and system, relating to the field of quantum refrigeration control technology. The method comprises obtaining the peak tunneling current of a quantum resonant tunneling diode during the refrigerant filling process; inputting the peak tunneling current into a Schrödinger-Poisson coupling solver, calculating the electron tunneling probability based on Fermi's golden rule, and outputting the scalar value of the instantaneous molecular cohesion at the refrigerant molecular level by inverting the exponential relationship between the barrier width and the peak tunneling current; constructing a filling exergy loss functional in Hilbert space based on the entropy production gradient tensor, and using a variational method to solve for the optimal opening proportional factor in combination with a real-time load parameter vector; and adjusting the flow cross-sectional area of ​​the electromagnetic filling valve channel in real time based on the optimal opening proportional factor. The present invention constructs a filling exergy loss functional in Hilbert space based on the entropy production gradient tensor and uses a variational method to solve for the optimal opening proportional factor, thereby minimizing dynamic entropy production and maximizing energy efficiency during the filling process.

Description

Technical Field

[0001] The present invention relates to the field of quantum refrigeration control technology, in particular to a solution filling control method and system. Background Art

[0002] Refrigerant charge control plays a key role in HVAC and refrigeration systems. Related technologies have evolved from traditional mechanical valve regulation to sensor-based feedback control, such as pressure, temperature, or flow monitoring. In recent years, the integration of quantum sensing and computational physics has gradually emerged, such as using the tunneling effect to capture molecular behavior and using numerical models to predict charge dynamics. These advances have driven the refinement of control algorithms, including optimization strategies that combine thermodynamic equations and statistical physics, to improve system efficiency and stability.

[0003] However, existing technologies still have significant shortcomings. The most significant problem is the inability to capture dynamic interactions at the molecular level in real time, such as changes in the instantaneous cohesion between refrigerant molecules. This results in limited control accuracy and an inability to accurately respond to microscopic disturbances, which in turn leads to inefficient charging processes, increased entropy production losses, and decreased system stability. Specifically, macro-parameter control methods ignore molecular behavior at the quantum level, making it difficult to achieve adaptive optimization of valve openings, limiting the overall energy efficiency improvement space. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a solution filling control method to solve the problem of insufficient accuracy in real-time modeling and optimization control of instantaneous cohesion at the molecular level during refrigerant filling.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a solution filling control method, which includes obtaining the tunnel current peak of a quantum resonant tunneling diode during the refrigerant filling process; inputting the tunnel current peak into a Schrödinger-Poisson coupling solver, calculating the electron tunneling probability based on Fermi's golden rule, and outputting the instantaneous molecular cohesion scalar value at the refrigerant molecular level by inverting the exponential relationship between the barrier width and the tunnel current peak; inputting the instantaneous molecular cohesion scalar value into a non-Markov quantum master equation, coupling the Lévy noise driving field to construct a generalized Langevin equation, and generating an entropy production gradient tensor; constructing a filling exergy loss functional in the Hilbert space based on the entropy production gradient tensor, and using a variational method to solve the optimal opening proportional factor in combination with a real-time load parameter vector; and adjusting the flow cross-sectional area of ​​the electromagnetic filling valve channel in real time according to the optimal opening proportional factor.

[0008] As a preferred embodiment of the solution filling control method of the present invention, the steps of obtaining the tunnel current peak value of the quantum resonant tunneling diode during the refrigerant filling process are as follows:

[0009] A heterojunction structure array consisting of AlGaAs barrier layers and GaAs quantum well layers periodically alternating with each other is deposited on the inner wall surface of the refrigerant flow channel;

[0010] Applying a forward bias voltage within a preset bias voltage range to the heterojunction structure array to capture the electron tunneling current waveform caused by the interaction between the refrigerant molecules and the quantum well barrier;

[0011] The global amplitude maximum in the electron tunneling current waveform is extracted as the tunneling current peak.

[0012] As a preferred embodiment of the solution filling control method of the present invention, the Schrödinger-Poisson coupling solver is a physical model that uses self-consistent iteration to calculate the interaction between electron wave functions and electrostatic potential;

[0013] The input of the coupled solver is the tunnel current peak value, and the output is the electron transmission probability density distribution;

[0014] The specific steps of outputting the instantaneous molecular cohesion scalar value of the refrigerant molecular level are as follows:

[0015] The electron transmission probability density distribution is converted into a potential barrier penetration function based on Fermi's golden rule;

[0016] The change in effective barrier width is inverted according to the mapping relationship between the barrier penetration function and the tunnel current peak value. The binding energy strength between refrigerant molecules is calculated using the Landau quantum droplet theory, and the instantaneous molecular cohesion scalar value representing the instantaneous intermolecular force is output.

[0017] As a preferred embodiment of the solution filling control method of the present invention, the Lévy noise driving field refers to a non-Gaussian random process with an α-stable distribution characteristic;

[0018] The specific steps of generating the entropy production gradient tensor are as follows:

[0019] The instantaneous molecular cohesion scalar value is temporally convolved with the historical molecular cohesion scalar value sequence to construct a non-Markovian quantum master equation with memory effect.

[0020] Fractional-order derivatives are used to perform non-local quantum correlation correction on the dissipative term in the non-Markovian quantum master equation, and the gradient distribution of the corrected quantum master equation in the Hilbert space is solved to output the entropy production gradient tensor.

[0021] As a preferred solution filling control method of the present invention, the specific steps of constructing the filling exergy loss functional are as follows:

[0022] Taking the entropy production gradient tensor as the basis function, the integrable function space is constructed in the Hilbert space;

[0023] Collect ambient temperature sensor data and convert it into temperature field distribution in imaginary time coordinates, synchronously obtain compressor speed signal and map it into speed pulse sequence in imaginary time coordinates;

[0024] In the integrable function space, the temperature field distribution and the speed pulse sequence are embedded in the functional integral kernel function, the functional integral expression is established and the charging exergy loss functional is output.

[0025] As a preferred solution filling control method of the present invention, wherein: the optimal opening proportional factor is solved by using the variational method in combination with the real-time load parameter vector, the specific steps are as follows:

[0026] Collect the measured values ​​of the temperature difference between the inlet and outlet of the condenser and the saturation temperature of the evaporator as the real-time load parameter vector;

[0027] The real-time load parameter vector is embedded into the Lagrange multiplier term through linear transformation to form the constraint condition of the charging exergy loss functional;

[0028] Initialize the conjugate gradient descent search direction based on the weak differentiability of Sobolev space;

[0029] The exergy loss functional extreme point is solved iteratively, and the optimal opening proportional factor is output when the gradient norm reaches the convergence threshold.

[0030] As a preferred solution filling control method of the present invention, wherein: the real-time adjustment of the flow cross-sectional area of ​​the electromagnetic filling valve channel is carried out in the following specific steps:

[0031] The optimal opening proportional factor is converted into a pulse width modulation signal duty cycle to drive the permanent magnet linear motor to drive the conical valve core to produce axial displacement;

[0032] The cross-sectional area variation curve is calculated according to the axial displacement, and the cross-sectional area of ​​the refrigerant flowing through the channel is continuously adjusted.

[0033] In a second aspect, the present invention provides a solution filling control system, comprising a current detection module, a quantum computing module, a noise modeling module, a functional optimization module and a valve control module; the current detection module is used to obtain the tunnel current peak of the quantum resonant tunneling diode during the refrigerant filling process; the quantum computing module is used to input the tunnel current peak into the Schrödinger-Poisson coupling solver, calculate the electron tunneling probability based on Fermi's golden rule, and output the instantaneous molecular cohesion scalar value at the refrigerant molecular level by inverting the exponential relationship between the barrier width and the tunnel current peak; the noise modeling module is used to input the instantaneous molecular cohesion scalar value into the non-Markov quantum master equation, couple the Lévy noise driving field to construct the generalized Langevin equation, and generate an entropy production gradient tensor; the functional optimization module is used to construct a charging exergy loss functional in the Hilbert space based on the entropy production gradient tensor, and use the variational method to solve the optimal opening proportional factor in combination with the real-time load parameter vector; the valve control module is used to adjust the flow cross-sectional area of ​​the electromagnetic filling valve channel in real time according to the optimal opening proportional factor.

[0034] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the solution filling control method described in the first aspect of the present invention is implemented.

[0035] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the solution filling control method described in the first aspect of the present invention is implemented.

[0036] The beneficial effects of the present invention are as follows: based on the entropy production gradient tensor, a charging exergy loss functional is constructed in the Hilbert space and the variational method is used to solve the optimal opening proportional factor, thereby achieving the minimization of dynamic entropy production and maximization of energy efficiency in the charging process. Specifically, by embedding the entropy production gradient tensor as a basis function into the functional integral kernel function, combined with the constraints of the real-time load parameter vector, variational optimization is performed in the integrable function space to generate an accurate opening proportional factor, ensuring that the refrigerant flow rate matches the load demand in real time, significantly reducing smoke losses and improving the thermodynamic efficiency of the system. In addition, through fractional-order derivatives and linear transformation processing, the robustness of the calculation framework is enhanced, avoiding oscillation instability in the non-equilibrium state, and making the charging process have high-precision, low-energy consumption and continuous adjustment characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 Flowchart of the solution filling control method.

[0039] Figure 2 Flowchart for quantum resonant tunneling current detection and molecular cohesion calculation.

[0040] Figure 3 Construct a flow chart for entropy-generating gradient tensor generation and functionals.

[0041] Figure 4 This is a flow chart of the interaction relationship between system modules. DETAILED DESCRIPTION

[0042] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0043] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0044] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0045] Reference Figures 1 to 4 , is an embodiment of the present invention, which provides a solution filling control method, comprising the following steps:

[0046] S1: Obtain the tunnel current peak of the quantum resonant tunneling diode during the refrigerant charging process.

[0047] S1.1: Deposit a heterojunction structure array consisting of AlGaAs barrier layers and GaAs quantum well layers periodically alternating on the inner wall surface of the refrigerant flow channel.

[0048] Specifically, a wet chemical cleaning method is used to remove surface contaminants and oxides to ensure surface flatness and cleanliness. Molecular beam epitaxy technology is used to grow a layer of AlGaAs material on the inner wall surface of the cleaned refrigerant flow channel. The thickness is controlled in the nanoscale range to form a barrier layer.

[0049] On the deposited AlGaAs barrier layer, a layer of GaAs material is grown using molecular beam epitaxy technology, with the thickness controlled within the nanoscale range to form a quantum well layer; the process of depositing the AlGaAs barrier layer and the GaAs quantum well layer is repeated alternately to form a heterojunction structure array.

[0050] For example, multiple periods are deposited, each period including an AlGaAs barrier layer and a GaAs quantum well layer.

[0051] S1.2: Apply a forward bias voltage within a preset bias voltage range to the heterojunction structure array to capture the electron tunneling current waveform caused by the interaction between the refrigerant molecules and the quantum well barrier.

[0052] Specifically, a programmable DC power supply is connected to the electrode interface of the heterojunction structure array, and the forward bias voltage range is set to a preset value; the forward bias voltage is applied point by point within the preset bias voltage range in a step-by-step manner, and a stable voltage output is obtained after each application.

[0053] Among them, the preset bias voltage range refers to the voltage interval pre-set according to the material band structure of the quantum resonant tunneling diode and the ionization characteristics of the refrigerant molecules, which is used to ensure that the electron tunneling effect can be stably excited and measured.

[0054] Use a high-sensitivity current probe to monitor the current response of the heterojunction structure array in real time;

[0055] When refrigerant molecules flow through the quantum well barrier, electron tunneling events are triggered, and the current probe captures and stores the corresponding current-time waveform data.

[0056] It should be noted that the quantum well barrier refers to the energy barrier formed by the AlGaAs barrier layer in the heterojunction structure array composed of periodically alternating AlGaAs barrier layers and GaAs quantum well layers, which is used to limit the movement of electrons in the GaAs quantum well layer and enhance the electron tunneling effect when the refrigerant molecules interact with the quantum well.

[0057] S1.3: Extract the global amplitude maximum in the electron tunneling current waveform as the tunneling current peak value.

[0058] Specifically, the time series current data is read from the stored electron tunneling current waveform.

[0059] The peak search algorithm in digital signal processing is used to scan the waveform data points of the time series current data and identify the global amplitude maximum.

[0060] The detected global amplitude maximum is recorded and output as the tunneling current peak value.

[0061] S2: Input the tunneling current peak into the Schrödinger-Poisson coupling solver, calculate the electron tunneling probability based on Fermi's golden rule, and output the instantaneous molecular cohesion scalar value at the refrigerant molecular level by inverting the exponential relationship between the barrier width and the tunneling current peak.

[0062] Among them, the Schrödinger-Poisson coupling solver refers to a physical model that uses self-consistent iteration to calculate the interaction between electron wave functions and electrostatic potential; the input of the coupling solver is the tunnel current peak, and the output is the electron transmission probability density distribution.

[0063] S2.1: Convert the electron transmission probability density distribution into a barrier penetration function based on Fermi's golden rule.

[0064] Specifically, the electron transmission probability density distribution output by the Schrödinger-Poisson coupling solver is read; the electron transmission probability density distribution represents a function of the probability of an electron passing through a potential barrier as the energy changes;

[0065] The Fermi Golden Rule formula is:

[0066] ;

[0067] Where, is the electron tunneling rate, is the reduced Planck constant (Dirac constant), Represents the initial state and final state In the perturbative Hamiltonian The transition matrix element under , characterizes the coupling strength between quantum states, Represents the final state The electron energy, is the final state density, which describes the number of quantum states that can be reached within a unit energy interval. Indicates the arc value of a complete circle (that is, the arc corresponding to 360°), represents the final state of the electron, represents the initial state of the electron, represents the square of the modulus;

[0068] Substitute the electron transmission probability density distribution into the final state density and calculate the incident electron energy based on Fermi's golden rule. Related tunneling rate , generating the barrier penetration function , the expression is:

[0069] ;

[0070] Where, represents the reference tunneling rate, which is usually the free electron tunneling rate in the absence of a potential barrier;

[0071] S2.2: Based on the mapping relationship between the barrier penetration function and the tunnel current peak, the change in the effective barrier width is inverted, the binding energy strength between refrigerant molecules is calculated using the Landau quantum droplet theory, and the instantaneous molecular cohesion scalar value representing the instantaneous intermolecular force is output.

[0072] Specifically, establish the tunnel current peak and barrier penetration function The exponential relationship model is expressed as:

[0073] ;

[0074] ;

[0075] Where, represents the attenuation coefficient, represents the effective mass of the electron, represents the barrier height, represents the current barrier width, It is proportional to the tunnel current peak value Barrier penetration Directly proportional relationship;

[0076] Fitting by nonlinear least squares method and The exponential relationship curve of the effective barrier width is solved inversely. , the expression is:

[0077] ;

[0078] Where, represents the initial barrier width when not affected by molecules;

[0079] Based on Landau quantum drop theory, the change in barrier width is Mapping is the change in the potential energy of the interaction between refrigerant molecules, and the expression is:

[0080] ;

[0081] Where, represents the change in the potential energy of intermolecular interaction, represents the surface tension coefficient of the droplet;

[0082] The droplet surface tension coefficient is It can be calculated through molecular dynamics simulation combined with known refrigerant physical properties. The example value range is 1×10⁻³ to 1×10⁻¹ joule / m².

[0083] The binding energy strength is defined as the negative value of the change in the potential energy of intermolecular interaction.

[0084] Normalize the binding energy intensity to a dimensionless scalar value and record the dimensionless scalar value as the instantaneous molecular cohesion scalar value of the refrigerant molecular level at the current moment .

[0085] S3: The instantaneous molecular cohesion scalar value is input into the non-Markov quantum master equation, coupled with the Lévy noise driving field to construct the generalized Langevin equation and generate the entropy production gradient tensor.

[0086] The Lévy noise driving field refers to a non-Gaussian random process with α-stable distribution characteristics.

[0087] S3.1: Perform a time convolution operation on the instantaneous molecular cohesion scalar value and the historical molecular cohesion scalar value sequence to construct a non-Markov quantum master equation with memory effect.

[0088] Specifically, the instantaneous molecular cohesion scalar value at the current moment is obtained;

[0089] Retrieve the historical molecular cohesion scalar value sequence within the preset time window length (stored in ascending order of timestamp);

[0090] The memory kernel function is defined as an exponentially decaying cosine kernel function, and the expression is:

[0091] ;

[0092] Where, represents the kernel function value, represents the time delay variable, represents the base of natural logarithms, represents the dissipation rate, represents the characteristic frequency, is the cosine function, reflecting the oscillation characteristics of the kernel function;

[0093] It should be noted that and Obtained through vibration spectrum fitting of refrigerant molecules, the example value range is:

[0094] ;

[0095] The instantaneous molecular cohesion scalar value at the current moment and the historical molecular cohesion scalar value sequence are convolved using an exponentially decaying cosine kernel function. The expression is:

[0096] ;

[0097] Where, is a non-Markov term (dimensionless), which represents the cumulative effect of historical molecular cohesion on the current quantum state. Indicates the moment, Indicates time The scalar value of the instantaneous molecular cohesion (dimensionless), represents the time differential, the integral variable;

[0098] Adding a convolution term to the standard quantum master equation , the expression is:

[0099] ;

[0100] Where, represents the density matrix, Represents the density matrix About time The explicit rate of change of , which represents the explicit evolution of the standard quantum master equation, Represents an imaginary unit , represents the partial differential, Represents the density matrix Partial differentials (infinitesimal changes), represents the partial differential of time (infinitesimal increments), represents the Hamiltonian, Indicates the exchange , represents the dissipative term;

[0101] S3.2: Use fractional derivatives to perform non-local quantum correlation correction on the dissipative term in the non-Markovian quantum master equation, solve the gradient distribution of the corrected quantum master equation in Hilbert space, and output the entropy production gradient tensor.

[0102] Specifically, for the dissipative term Applying Caputo-type fractional derivatives (Value range ), generates the modified dissipation term, which is expressed as:

[0103] ;

[0104] Where, represents the corrected dissipation term, represents the correlation strength coefficient (dimensionless), represents the Caputo-type fractional derivative operator, represents the Caputo-type fractional derivative applied to the operator in brackets;

[0105] Among them, the correlation strength coefficient It can be obtained by molecular dynamics simulation combined with refrigerant physical property parameter calculation, and the example value range is 0.1 to 1.0.

[0106] Generates Lévy noise sequences with α-stable distribution characteristics , construct the generalized Langevin equation, the expression is:

[0107] ;

[0108] Where, is the density matrix About time The total rate of change represents the total evolution of the generalized Langevin equation, represents the Lévy noise driving field, which has an α-stable distribution characteristic, represents the Hadamard product (element-wise multiplication), Represents the density matrix The differential of represents the differential of time;

[0109] Numerically solve the generalized Langevin equation (such as the Runge-Kutta method) and calculate the gradient of the entropy production rate in Hilbert space, which is expressed as:

[0110] ;

[0111] Where, represents the entropy production gradient, represents the trace operation of the matrix, represents the logarithm operation of the density matrix, represents the entropy production rate;

[0112] Expand the entropy production rate gradient into a second-order tensor and record it as the entropy production gradient tensor. The expression is:

[0113] ;

[0114] Where, represents the matrix element of the entropy production gradient tensor, and The density matrix The row and column indexes of represents the entropy production rate The second-order partial differential of and Both represent density matrices The matrix elements of Represents the density matrix elements The partial differential of Represents the density matrix elements The partial differential of .

[0115] It should be noted that the matrix elements of the entropy generation gradient tensor are specific components of the entropy generation gradient tensor, representing the second-order partial derivative values ​​of the entropy generation rate with respect to different elements of the density matrix. All matrix elements constitute the complete entropy generation gradient tensor.

[0116] S4: Based on the entropy production gradient tensor, a charging exergy loss functional is constructed in the Hilbert space. Combined with the real-time load parameter vector, the variational method is used to solve the optimal opening proportional factor.

[0117] S4.1: Use the entropy production gradient tensor as a basis function to construct an integrable function space in Hilbert space.

[0118] Specifically, each specific component of the entropy production gradient tensor is considered as a basis function in the Hilbert space;

[0119] Define the integrable function space consisting of the linear combination of all entropy production gradient tensor components, expressed as:

[0120] ;

[0121] Where, represents the space of integrable functions (a subset of Hilbert space), Represents the index of the tensor component and The summation operation covers all matrix elements of the entropy production gradient tensor, is the expansion coefficient (complex or real number), which represents the weight of each basis function in the combination. Indicates the conditional separator in the set definition, indicating "the following conditions are met". Representation function of Norm integration ensures that the function is square integrable, Generally refers to any function in the integrable function space, Representation function The square of the modulus, represents the measure (integral differential) in Hilbert space, is infinite, indicating that the integral value must be finite (convergent), Represents the functional integral symbol.

[0122] S4.2: Collect ambient temperature sensor data and convert it into temperature field distribution in imaginary time coordinates, and synchronously obtain the compressor speed signal and map it into a speed pulse sequence in imaginary time coordinates.

[0123] Specifically, read the real-time measurement value of the ambient temperature sensor and map the real time coordinate to the imaginary time coordinate ;

[0124] Temperature data in imaginary time coordinates Reparameterize to form a continuous temperature field distribution ;

[0125] Get the pulse signal of the compressor speed sensor and map the speed pulse sequence to the imaginary time coordinate , output discrete pulse train ,in is the pulse index.

[0126] S4.3: In the integrable function space, embed the temperature field distribution and the speed pulse sequence into the functional integral kernel function, establish the functional integral expression and output the charging exergy loss functional.

[0127] Specifically, define the kernel function , the expression is:

[0128] ;

[0129] Where, is the gradient operator of the continuous temperature field distribution, which represents the spatial rate of change of temperature with imaginary time. represents the square of the gradient modulus, reflecting the intensity of temperature field inhomogeneity. is the integral of the square of the gradient modulus over imaginary time, representing the overall temperature fluctuation, is the cumulative sum of all pulses, reflecting the cumulative effect of the compressor speed. represents the weight coefficient of the temperature field gradient term, Indicates the weight coefficient of the speed pulse term;

[0130] It should be noted that Obtained through quantum thermodynamics simulation combined with refrigerant heat transfer characteristics calculation, the example value range is 0.1 to 1.0. This is based on compressor power spectrum analysis and refrigerant flow characteristic calibration, with example values ​​ranging from 0.05 to 0.5.

[0131] Combine the kernel function with the basis function of the integrable function space to output the exergy loss functional , the expression is:

[0132] ;

[0133] Where, represents a linear combination of the entropy production gradient tensors.

[0134] S4.4: Collect the measured values ​​of the condenser inlet and outlet temperature difference and the evaporator saturation temperature as the real-time load parameter vector.

[0135] Specifically, read the condenser inlet temperature and outlet temperature, and calculate the condenser temperature difference ;

[0136] Read the evaporator pressure sensor data and check the evaporator saturation temperature through the refrigerant saturation temperature table ;

[0137] Define the real-time load parameter vector: .

[0138] S4.5: The real-time load parameter vector is embedded into the Lagrange multiplier term through linear transformation to form the constraint condition of the charging exergy loss functional.

[0139] Specifically, define the transformation matrix and the bias vector ;

[0140] It should be noted that the transformation matrix is ​​a 2×2 real matrix used to linearly map the real-time load parameter vector to the constraint space. Indicates the condenser temperature difference The contribution weight to the first dimension of the constraint, Indicates the evaporator saturation temperature The contribution weight to the first dimension of the constraint, Indicates the condenser temperature difference The contribution weight to the second dimension of the constraint, Indicates the evaporator saturation temperature The contribution weight to the second dimension of the constraint, Represents the reference offset of the first dimension of the constraint condition, which is used to compensate for the inherent deviation of the system. The reference offset of the second dimension of the constraint condition is used to compensate for the inherent deviation of the system;

[0141] It should also be noted that the corresponding contribution weights are obtained by fitting the historical operating data of the refrigeration unit using the least squares method or by analytical calculation based on the thermodynamic equation. The example value range is 、 、 、 The reference offset is determined by statistical analysis of parameter deviations under steady-state conditions.

[0142] Compute the constraint vector: ;

[0143] Will As the charge exergy loss functional The equality constraints of are used to construct the Lagrange multiplier terms.

[0144] S4.6: Initializing the conjugate gradient descent search direction based on weak differentiability in Sobolev space.

[0145] Specifically, in Sobolev space, the exergy loss functional is confirmed The first-order weak derivative of exists;

[0146] Calculate the gradient of the charging exergy loss functional In the initial expansion coefficient The value at which to set the initial search direction , the expression is:

[0147] ;

[0148] S4.7: Iteratively solve the extreme point of the charge exergy loss functional and output the optimal opening scale factor when the gradient norm reaches the convergence threshold.

[0149] Specifically, when setting the convergence threshold for the iterative solution, the numerical characteristics of the exergy loss functional and the required computational accuracy are considered. By analyzing the decay characteristics of the functional's gradient norm during the iteration process, the minimum gradient norm threshold required for the optimization result to meet the engineering accuracy requirements is determined as the convergence threshold. The convergence threshold serves as the basis for determining whether the exergy loss functional has reached a converged state. Example values ​​for the convergence threshold can range from 10 to the power of -6 to 10 to the power of -8 to ensure accuracy and computational efficiency of the optimization process while avoiding premature termination or over-computation.

[0150] Furthermore, at the start of the iterative solution process, after initializing the conjugate gradient descent search direction, the following loop is executed: a one-dimensional line search is performed along the current search direction to determine the updated step size; the expansion coefficient is adjusted using the step size; the gradient of the exergy loss functional is recalculated; the new gradient norm is tested to see if it is less than a preset convergence threshold. If so, the iteration is terminated; if not, the search direction is updated and the next iteration is continued. This loop continues until the gradient norm falls below the convergence threshold, at which point the converged expansion coefficient is mapped to the optimal aperture scaling factor output.

[0151] S5: Real-time adjustment of the flow cross-sectional area of ​​the electromagnetic filling valve channel according to the optimal opening proportional factor.

[0152] S5.1: Convert the optimal opening proportional factor into a pulse width modulation signal duty cycle to drive the permanent magnet linear motor to drive the conical valve core to generate axial displacement.

[0153] Specifically, the optimal opening proportional factor is obtained, a proportional relationship between the optimal opening proportional factor and the duty cycle of the pulse width modulation signal is established by a linear conversion method, and a pulse waveform corresponding to the duty cycle value is output;

[0154] Input the pulse waveform into the driver of the permanent magnet linear motor to drive the motor output shaft to generate axial displacement;

[0155] The motor output shaft is rigidly connected to the conical valve core, transmitting the axial displacement to the conical valve core (the displacement is proportional to the duty cycle).

[0156] S5.2: Calculate the cross-sectional area change curve based on the axial displacement and continuously adjust the flow cross-sectional area of ​​the refrigerant flowing through the channel.

[0157] Specifically, based on the fixed cone angle of the conical valve core (60 degrees for example), the mathematical relationship between the axial displacement and the cross-sectional area of ​​the annular diameter is calculated (the cross-sectional area increases nonlinearly with the displacement, approximately in a square relationship);

[0158] The instantaneous cross-sectional area value is dynamically calculated based on the current axial displacement, and the annular gap area is continuously adjusted through the axial movement of the tapered valve core;

[0159] Monitor the cross-sectional area change curve in real time to ensure that it is consistent with the requirements of the optimal opening proportional factor (for example, the optimal proportional factor of 0.5 corresponds to a 50% opening of the cross-sectional area).

[0160] This embodiment also provides a solution filling control system, comprising: a current detection module, a quantum computing module, a noise modeling module, a functional optimization module, and a valve control module; the current detection module is used to obtain the tunneling current peak of the quantum resonant tunneling diode during the refrigerant filling process; the quantum computing module is used to input the tunneling current peak into a Schrödinger-Poisson coupling solver, calculate the electron tunneling probability based on Fermi's golden rule, and output the instantaneous molecular cohesion scalar value at the refrigerant molecular level by inverting the exponential relationship between the barrier width and the tunneling current peak; the noise modeling module is used to input the instantaneous molecular cohesion scalar value into a non-Markov quantum master equation, couple the Lévy noise driving field to construct a generalized Langevin equation, and generate an entropy production gradient tensor; the functional optimization module is used to construct a charging exergy loss functional in Hilbert space based on the entropy production gradient tensor, and use a variational method to solve the optimal opening proportional factor in combination with a real-time load parameter vector; the valve control module is used to adjust the flow cross-sectional area of ​​the electromagnetic filling valve channel in real time according to the optimal opening proportional factor.

[0161] This embodiment also provides a computer device suitable for the solution filling control method, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the solution filling control method proposed in the above embodiment.

[0162] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0163] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the solution filling control method proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0164] In summary, the present invention constructs a charging exergy loss functional in the Hilbert space based on the entropy production gradient tensor and uses the variational method to solve the optimal opening proportional factor, thereby achieving dynamic entropy generation minimization and energy efficiency maximization in the charging process. Specifically, by embedding the entropy production gradient tensor as a basis function into the functional integral kernel function, combined with the constraints of the real-time load parameter vector, variational optimization is performed in the integrable function space to generate an accurate opening proportional factor, ensuring that the refrigerant flow rate matches the load demand in real time, significantly reducing smoke losses and improving the thermodynamic efficiency of the system. In addition, through fractional-order derivatives and linear transformation processing, the robustness of the calculation framework is enhanced, avoiding oscillation instability in the non-equilibrium state, and making the charging process have high-precision, low-energy consumption and continuous adjustment characteristics.

[0165] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A solution filling control method, characterized in that: include, Obtain the tunnel current peak of the quantum resonant tunneling diode during the refrigerant charging process; The tunneling current peak is input into the Schrödinger-Poisson coupling solver, and the electron tunneling probability is calculated based on Fermi's golden rule. The exponential relationship between the barrier width and the tunneling current peak is inverted to output the instantaneous molecular cohesion scalar value at the refrigerant molecular level. The instantaneous molecular cohesion scalar value is input into the non-Markov quantum master equation, coupled with the Lévy noise driving field to construct the generalized Langevin equation and generate the entropy production gradient tensor; Based on the entropy production gradient tensor, a charging exergy loss functional is constructed in the Hilbert space. Combined with the real-time load parameter vector, the optimal opening proportional factor is solved by the variational method. The flow cross-sectional area of ​​the electromagnetic filling valve channel is adjusted in real time according to the optimal opening proportional factor; The specific steps of obtaining the tunnel current peak value of the quantum resonant tunneling diode during the refrigerant charging process are as follows: A heterojunction structure array consisting of AlGaAs barrier layers and GaAs quantum well layers periodically alternating with each other is deposited on the inner wall surface of the refrigerant flow channel; Applying a forward bias voltage within a preset bias voltage range to the heterojunction structure array to capture the electron tunneling current waveform caused by the interaction between the refrigerant molecules and the quantum well barrier; Extract the global maximum amplitude value in the electron tunneling current waveform as the tunneling current peak value; The Schrödinger-Poisson coupling solver is a physical model that uses self-consistent iteration to calculate the interaction between electron wave functions and electrostatic potential; The input of the coupled solver is the tunnel current peak value, and the output is the electron transmission probability density distribution; The specific steps of outputting the instantaneous molecular cohesion scalar value of the refrigerant molecular level are as follows: The electron transmission probability density distribution is converted into a potential barrier penetration function based on Fermi's golden rule; The effective barrier width variation is inverted based on the mapping relationship between the barrier penetration function and the tunnel current peak value. The binding energy strength between refrigerant molecules is calculated using the Landau quantum drop theory, and the instantaneous molecular cohesion scalar value representing the instantaneous intermolecular force is output. The Lévy noise driving field refers to a non-Gaussian random process with an α-stable distribution characteristic; The specific steps of generating the entropy production gradient tensor are as follows: The instantaneous molecular cohesion scalar value is temporally convolved with the historical molecular cohesion scalar value sequence to construct a non-Markovian quantum master equation with memory effect. Fractional-order derivatives are used to perform non-local quantum correlation correction on the dissipative term in the non-Markovian quantum master equation, and the gradient distribution of the corrected quantum master equation in the Hilbert space is solved to output the entropy production gradient tensor.

2. The solution filling control method according to claim 1, wherein: The specific steps of constructing the charge exergy loss functional are as follows: Taking the entropy production gradient tensor as the basis function, the integrable function space is constructed in the Hilbert space; Collect ambient temperature sensor data and convert it into temperature field distribution in imaginary time coordinates, synchronously obtain compressor speed signal and map it into speed pulse sequence in imaginary time coordinates; In the integrable function space, the temperature field distribution and the speed pulse sequence are embedded in the functional integral kernel function, the functional integral expression is established and the charging exergy loss functional is output.

3. The solution filling control method according to claim 2, wherein: The above method combines the real-time load parameter vector and uses the variational method to solve the optimal opening proportional factor. The specific steps are as follows: Collect the measured values ​​of the temperature difference between the inlet and outlet of the condenser and the saturation temperature of the evaporator as the real-time load parameter vector; The real-time load parameter vector is embedded into the Lagrange multiplier term through linear transformation to form the constraint condition of the charging exergy loss functional; Initialize the conjugate gradient descent search direction based on the weak differentiability of Sobolev space; The exergy loss functional extreme point is solved iteratively, and the optimal opening proportional factor is output when the gradient norm reaches the convergence threshold.

4. The solution filling control method according to claim 3, wherein: The specific steps of real-time adjustment of the flow cross-sectional area of ​​the electromagnetic filling valve channel are as follows: The optimal opening proportional factor is converted into a pulse width modulation signal duty cycle to drive the permanent magnet linear motor to drive the conical valve core to produce axial displacement; The cross-sectional area variation curve is calculated according to the axial displacement, and the cross-sectional area of ​​the refrigerant flowing through the channel is continuously adjusted.

5. A solution filling control system, based on the solution filling control method according to any one of claims 1 to 4, characterized in that: Including current detection module, quantum computing module, noise modeling module, functional optimization module and valve control module; The current detection module is used to obtain the tunnel current peak value of the quantum resonant tunneling diode during the refrigerant charging process; The quantum computing module is used to input the tunneling current peak into the Schrödinger-Poisson coupling solver, calculate the electron tunneling probability based on Fermi's golden rule, and output the scalar value of the instantaneous molecular cohesion at the refrigerant molecular level by inverting the exponential relationship between the barrier width and the tunneling current peak; The noise modeling module is used to input the instantaneous molecular cohesion scalar value into the non-Markov quantum master equation, couple the Lévy noise driving field to construct the generalized Langevin equation, and generate the entropy production gradient tensor; The functional optimization module is used to construct a charging exergy loss functional in the Hilbert space based on the entropy production gradient tensor, and to solve the optimal opening proportional factor using a variational method in combination with the real-time load parameter vector; The valve control module is used to adjust the flow cross-sectional area of ​​the electromagnetic filling valve channel in real time according to the optimal opening proportional factor.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the solution filling control method according to any one of claims 1 to 5 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the solution filling control method according to any one of claims 1 to 5 are implemented.

Images