Hybrid heat pump energy efficiency self-adaptive optimization method

Abstract

The invention provides a hybrid heat pump energy efficiency self-adaptive optimization method, which relates to the technical field of heat pump energy-saving control, and comprises the following steps: acquiring a circulation period parameter of a main loop refrigerant and a bypass pressure parameter of a bypass loop refrigerant in a hybrid heat pump system, and calculating the residence time of the main loop refrigerant based on the circulation period parameter; and a critical point of working condition switching is accurately predicted by constructing a dual buffer matrix. Before the critical point is detected, the pulse valve set opening degree sequence in the bypass loop is adjusted in advance, a reverse compensation channel is formed, and therefore refrigerant inertia migration in the main loop is effectively restrained during working condition switching, and instantaneous pressure fluctuation is reduced. Thus, the energy efficiency stability and the adjustment response accuracy of the hybrid heat pump system under the dynamic working condition can be improved, the COP of the system can be rapidly and stably achieved, and therefore the overall operation efficiency of the system is improved, and the operation energy consumption is reduced.

Description

Technical Field

[0001] This invention relates to the field of heat pump energy-saving control technology, and more specifically, to a method for adaptive optimization of hybrid heat pump energy efficiency. Background Technology

[0002] With the escalating global energy crisis and the continuous implementation of energy conservation and environmental protection policies, heat pump technology, due to its high efficiency and energy saving characteristics, has become one of the core technologies widely used in heating, cooling, and hot water preparation. After decades of development, traditional heat pump technology has evolved from a single refrigeration cycle into multifunctional composite systems, such as air source heat pumps, water source heat pumps, and ground source heat pumps. In recent years, to further improve the operating efficiency and applicability of heat pump systems, hybrid heat pump (HHP) technology has emerged. This technology, based on traditional heat pump systems, combines multi-loop control, variable frequency regulation, and bypass loop optimization to achieve efficient energy transfer and flexible adaptability under dynamic operating conditions. In hybrid heat pump systems, the coordinated operation of the main loop and bypass loop is key to achieving efficient and stable system operation. However, complex changes in operating conditions place higher demands on the system's control accuracy and response speed. Especially during operating condition switching, the refrigerant's inertial migration and pressure fluctuations often become bottlenecks affecting the system's Coefficient of Performance (COP).

[0003] While existing hybrid heat pump technology has improved system stability to some extent through optimized design of the main and bypass loops, several shortcomings remain. First, under complex dynamic conditions, the refrigerant residence time is difficult to predict accurately, resulting in poor transient response performance during operating condition switching and a high risk of uncontrolled refrigerant flow fluctuations in the main loop. Second, traditional bypass loop control strategies typically employ fixed logic regulation, failing to fully consider the dynamic interaction characteristics between the main and bypass loops during operating condition switching. This single regulation method cannot effectively suppress inertial pressure fluctuations during operating condition switching, easily leading to a non-linear decrease in system energy efficiency. Furthermore, during hybrid heat pump operation, the adjustment of the pulse valve assembly typically lags behind changes in operating conditions, resulting in insufficient judgment and pre-adjustment capabilities for critical operating conditions, further reducing system stability and economy. Therefore, how to construct a high-precision predictive model based on real-time dynamic parameters to achieve accurate judgment and pre-adjustment during operating condition switching has become a pressing technical challenge for improving the energy efficiency of hybrid heat pump systems. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention is proposed. This invention provides an adaptive optimization method for the energy efficiency of hybrid heat pumps, which to some extent solves the problem of COP fluctuations caused by slow heat exchanger transition response during frequent operating condition switching in hybrid heat pump systems.

[0005] According to one aspect of the present invention, a hybrid heat pump energy efficiency adaptive optimization method is provided, comprising:

[0006] Obtain the circulation cycle parameters of the refrigerant in the main loop and the bypass pressure parameters of the refrigerant in the bypass loop of the hybrid heat pump system, and calculate the residence time of the refrigerant in the main loop based on the circulation cycle parameters.

[0007] A dual buffer matrix is ​​constructed based on the residence time and the bypass pressure parameter, and the critical point of operating condition switching is predicted by the dual buffer matrix.

[0008] Before the critical point is detected, the opening sequence of the pulse valve group in the bypass circuit is pre-adjusted. The opening sequence of the pulse valve group is opposite to the fluctuation trend of the double buffer matrix, forming a reverse compensation channel.

[0009] When a change in operating condition occurs, the instantaneous pressure fluctuation generated by the reverse compensation channel suppresses the inertial migration of refrigerant in the main circuit, and the duration of the compensation channel is adjusted according to the buffer coefficient of the main circuit to achieve rapid stabilization of the system COP.

[0010] Furthermore, the dual buffer matrix includes a main loop buffer coefficient and a bypass adjustment coefficient;

[0011] The dwell time is divided into M time series nodes according to a preset time interval, and the corresponding bypass pressure parameter is recorded at each time series node to generate a time series state matrix.

[0012] The main loop buffer coefficient is obtained by exponentially smoothing the residence time sequence in the time-series state matrix; the bypass adjustment coefficient is obtained by differential operation on the bypass pressure parameter sequence.

[0013] Furthermore, differential operations are performed on the bypass pressure parameter sequence. By downsampling the bypass pressure sequence to a 2-second interval, differential operations are performed to obtain the cumulative values ​​of positive and negative pressure gradients. These values ​​are then compared with the reference pressure values ​​to obtain preliminary coefficients. Finally, cubic spline interpolation is used for smoothing to obtain the final bypass adjustment coefficient sequence.

[0014] Furthermore, predicting the critical point of operating condition switching using the dual buffer matrix includes the following steps:

[0015] The sequence complexity S1 of the main loop buffer coefficient and the sequence complexity S2 of the bypass adjustment coefficient are calculated using sliding entropy.

[0016] Construct a phase space trajectory based on the sequence complexity S1 and the sequence complexity S2, and calculate the Lyapunov exponent of the phase space trajectory;

[0017] When the Lyapunov index changes from negative to positive and the duration exceeds a preset period, it is determined that the system is about to experience a switching inflection point.

[0018] Furthermore, the Lyapunov exponent is calculated as follows:

[0019]

[0020] in, For the overall Lyapunov index of the system, Let P0 be the total number of times it moves along the phase space trajectory. The number of effective nearest neighbors for the m-th reference point. Let be the distance constraint function. and Let be the sequence complexity coordinates of the i-th evolution point. and The sequence complexity coordinates are used as a reference evolution point. The initial distance, The number of nearest neighbors selected. The nearest neighbor weight function, Let be the hypersphere radius.

[0021] Furthermore, the pulse valve group opening sequence converts the predicted fluctuation trend into a compensation sequence with opposite phase, and adjusts it according to the adaptive step size mapping rule and the high and low frequency valve ratio to form an actual valve control quantity that actively suppresses system pressure fluctuations.

[0022] Furthermore, the generation of the compensation sequence includes:

[0023] The predicted fluctuation sequence is analyzed to separate the instantaneous amplitude and instantaneous phase information from the analyzed signal, and then the phase is reversed.

[0024] The predicted wave sequence is decomposed into subsequences of different frequency bands. The energy proportion of each subsequence is calculated and an energy density distribution curve is constructed.

[0025] The frequency band filtering boundary is automatically determined based on the inflection point characteristics of the curve. When the energy density of the subsequence is greater than the value corresponding to the inflection point, the frequency band component is retained.

[0026] The retained frequency band sub-sequences are reconstructed with the corresponding anti-phase information, and a piecewise continuous gain function is constructed based on the characteristic frequencies of the frequency response curve.

[0027] By utilizing the local statistical properties of the sequence, adaptive smoothing and amplitude normalization are performed to generate a signal sequence with the same amplitude but opposite phase as the original sequence.

[0028] Furthermore, the piecewise continuous gain function obtains the characteristic frequency by performing singular value decomposition on the frequency response curve, and then constructs the fundamental function in different frequency ranges using exponential decay, cosine transition, and inverse proportional forms respectively.

[0029] By combining the least squares method to optimize parameters, smooth transition, and normalize, a gain control function with continuous derivative characteristics across the entire frequency range is finally formed.

[0030] Furthermore, the instantaneous amplitude and instantaneous phase information are separated, and an analytical signal is obtained by performing a Fourier transform on the original signal. Then, the envelope is extracted using polar coordinate transformation, the phase is extracted using the four-quadrant arctangent function, and the result is obtained after phase jump correction and adaptive curvature filtering.

[0031] Furthermore, the inertial migration of refrigerant in the main circuit is suppressed by detecting pressure fluctuations through the main circuit temperature and pressure sensor. When the pressure threshold is exceeded, the reverse compensation channel is activated. The opening degree of the pulse valve group and the compensation duration are dynamically adjusted based on the buffer coefficient and refrigerant state. The response characteristics of the compensation channel are optimized according to the COP change trend and refrigerant flow characteristics.

[0032] Compared with existing technologies, the hybrid heat pump energy efficiency adaptive optimization method provided by this invention obtains the circulation cycle parameters of the refrigerant in the main loop and the bypass pressure parameters of the refrigerant in the bypass loop of the hybrid heat pump system. Based on the circulation cycle parameters, it calculates the residence time of the refrigerant in the main loop and accurately predicts the critical point of operating condition switching by constructing a dual buffer matrix. Before detecting the critical point, the opening sequence of the pulse valve group in the bypass loop is pre-adjusted to form a reverse compensation channel, thereby effectively suppressing the inertial migration of refrigerant in the main loop and reducing instantaneous pressure fluctuations during operating condition switching. This improves the energy efficiency stability and adjustment response accuracy of the hybrid heat pump system under dynamic operating conditions, helps to achieve rapid stabilization of the system's COP, thereby improving the overall operating efficiency of the system and reducing operating energy consumption. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0034] Figure 1This is a flowchart of a hybrid heat pump energy efficiency adaptive optimization method according to an embodiment of the present invention.

[0035] Figure 2 This is a schematic diagram of the energy density distribution curve in the hybrid heat pump energy efficiency adaptive optimization method according to an embodiment of the present invention. Detailed Implementation

[0036] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0037] Figure 1 This is a flowchart of a hybrid heat pump energy efficiency adaptive optimization method according to an embodiment of the present invention. Figure 1 As shown, the adaptive optimization method for hybrid heat pump energy efficiency includes:

[0038] S1: Obtain the circulation cycle parameters of the refrigerant in the main circuit and the bypass pressure parameters of the refrigerant in the bypass circuit of the hybrid heat pump system, and calculate the residence time of the refrigerant in the main circuit based on the circulation cycle parameters;

[0039] Obtaining the circulation cycle parameters of the refrigerant in the main loop and the bypass pressure parameters of the refrigerant in the bypass loop of a hybrid heat pump system specifically includes: setting a first temperature sensor at the condenser outlet of the main loop and a second temperature sensor at the evaporator outlet of the main loop; calculating the temperature fluctuation period based on the temperature data collected by the first and second temperature sensors, and using the temperature fluctuation period as the circulation cycle parameter; setting a first pressure sensor and a second pressure sensor at the inlet and outlet of the bypass loop respectively, and using the pressure difference collected by the first and second pressure sensors as the bypass pressure parameter; wherein, the specific process of calculating the residence time of the refrigerant in the main loop based on the circulation cycle parameter is as follows: continuously collecting N circulation cycle parameters within a preset time period, calculating the average of the N circulation cycle parameters, and dividing the average by 2 to obtain the residence time of the refrigerant in the main loop, where N is an integer greater than 10; the residence time characterizes the actual residence time of the refrigerant in the heat exchanger and is used for subsequent evaluation of the system response characteristics.

[0040] S2: Construct a dual buffer matrix based on the residence time and the bypass pressure parameter. The dual buffer matrix includes a main loop buffer coefficient and a bypass adjustment coefficient. Predict the critical point of operating condition switching through the dual buffer matrix.

[0041] Constructing a dual buffer matrix based on the residence time and the bypass pressure parameter specifically includes: dividing the residence time into M time series nodes according to a preset time interval, recording the corresponding bypass pressure parameter at each time series node, and generating an M×2 time series state matrix, where M is an integer greater than 20; performing exponential smoothing on the residence time series in the time series state matrix to obtain the main loop buffer coefficient, performing differential operation on the bypass pressure parameter series to obtain the bypass adjustment coefficient, and forming a dual buffer matrix with the main loop buffer coefficient and the bypass adjustment coefficient; wherein, the specific process of predicting the critical point of operating condition switching through the dual buffer matrix is ​​as follows: The sequence complexity S1 of the main loop buffer coefficient and the sequence complexity S2 of the bypass adjustment coefficient are calculated using sliding entropy. A phase space trajectory is constructed based on the sequence complexity S1 and the sequence complexity S2. The Lyapunov exponent of the phase space trajectory is calculated. When the Lyapunov exponent changes from negative to positive and the duration exceeds a preset period, it is determined that the system is about to experience a switching inflection point. The sliding entropy is calculated using a sliding window of length L, where L ranges from 1 / 5 to 1 / 3 of the sequence length. The critical point of the switching inflection point refers to the time node before the system switches from heating mode to cooling mode or from cooling mode to heating mode.

[0042] The specific process of obtaining the bypass adjustment coefficient by performing differential operation on the bypass pressure parameter sequence is as follows: First, the collected bypass pressure parameter sequence is downsampled at equal intervals, with a time interval of 2 seconds, to obtain the downsampled pressure sequence; then, a first-order differential operation is performed on the downsampled pressure sequence. When the pressure difference between two adjacent pressure sampling points is greater than 0, it is recorded as a positive pressure gradient; when the pressure difference between two adjacent pressure sampling points is less than 0, it is recorded as a negative pressure gradient; next, the cumulative values ​​of the positive and negative pressure gradients within one sampling period are calculated. If the cumulative value is positive, the bypass adjustment coefficient for that sampling period is recorded as the ratio of the cumulative value to the reference pressure value; if the cumulative value is negative, the bypass adjustment coefficient for that sampling period is recorded as the opposite of the ratio of the cumulative value to the reference pressure value; wherein, the reference pressure value is the calibration pressure value of the system under stable operating conditions; finally, the calculated bypass adjustment coefficient sequence is smoothed using the cubic spline interpolation method to eliminate high-frequency fluctuations introduced by downsampling, resulting in the final bypass adjustment coefficient sequence.

[0043] The specific process for calculating the Lyapunov exponent of the phase space trajectory is as follows: First, the sequence complexity S1 and the sequence complexity S2 are mapped onto a two-dimensional plane to construct the phase space trajectory. A reference point P0 is selected on the phase space trajectory, and a hypersphere region with radius r is established with P0 as the center, where r is 1 / 10 of the maximum range of the phase space. Then, the k trajectory points closest to the reference point P0 are selected within the hypersphere region to form a nearest neighbor set, and the initial distance d0 between each nearest neighbor point and the reference point P0 is recorded. Next, the evolution of each point in the nearest neighbor set is tracked over n time steps. The trajectory is tracked, and when the i-th time step is reached, the distance di between each evolution point and the reference evolution point is calculated. If the distance di is greater than twice the value r, the tracking process for that point is stopped. Then, the distance growth rate ln(di / d0) is calculated for all tracked points. A linear regression is performed on the distance growth rate against the time step i, and the slope of the regression line is the local Lyapunov exponent corresponding to the reference point P0. Finally, the reference point P0 is moved along the phase space trajectory, and the above calculation process is repeated. The average of all local Lyapunov exponents is taken as the overall Lyapunov exponent of the system. The calculation of the Lyapunov exponent is shown in the following formula:

[0044] in, For the overall Lyapunov index of the system, Let P0 be the total number of times it moves along the phase space trajectory. The number of effective nearest neighbors for the m-th reference point. Let φ(d_i,r) be the distance constraint function and H(2r-d_i), where H is the heaviside step function used to stop tracking when d_i>2r. and Let be the sequence complexity coordinates of the i-th evolution point. and The sequence complexity coordinates are used as a reference evolution point. The initial distance, The number of nearest neighbors selected. The nearest neighbor weight function, ψ(k,i)=(k-i+1) / k, is used to assign higher weights to trajectory points that are closer to the reference point. Let r be the hypersphere radius and r = R_max / 10, where R_max is the maximum range of the phase space.

[0045] It should be noted that in calculating the Lyapunov exponent, a reference point P0 is first selected in the phase space based on sequence complexities S1 and S2. Using this point as the center, the hypersphere radius r is determined according to the maximum range of the phase space. Within this hypersphere, k nearest neighbors are selected, and the initial distance d0 between each neighbor and the reference point is recorded. The sequence complexity coordinates of these neighbors are all distributed within the hypersphere surrounding the reference point. Next, the evolution of these k neighbors is tracked within each time step. When the evolution reaches the i-th time step, the distance di between the new sequence complexity coordinates of each neighbor and the reference evolution point is calculated. Simultaneously, the distance is used to... The constraint function φ(d_i,r) determines whether the hypersphere region is exceeded. If the distance is greater than 2r, the tracking of that point is stopped. Then, the distance growth rate ln(di / d0) of each completed tracking nearest neighbor point is multiplied by its corresponding nearest neighbor weight function ψ(k,i) and the sum is averaged to obtain the local Lyapunov exponent of the reference point. Finally, the reference point is moved along the phase space trajectory, and the above calculation process is repeated to obtain multiple local Lyapunov exponents. The arithmetic mean of these local exponents is taken to obtain the overall Lyapunov exponent λ_sys of the system. The sign of this exponent can be used to determine whether the system is in a critical state of operating condition switching.

[0046] S3: Before detecting the critical point, the opening sequence of the pulse valve group in the bypass circuit is pre-adjusted. The opening sequence of the pulse valve group is opposite to the fluctuation trend of the double buffer matrix, forming a reverse compensation channel.

[0047] Before detecting the critical point, the specific process of pre-adjusting the pulse valve group opening sequence in the bypass circuit is as follows: First, perform wavelet transform on the double buffer matrix to obtain the fluctuation characteristics at different scales, and select the fluctuation coefficient corresponding to the maximum energy scale as the main fluctuation trend; then, push the main fluctuation trend forward by N sampling points on the time axis, where N is an integer greater than 5, to obtain the predicted fluctuation sequence; next, generate a compensation sequence with equal amplitude and opposite phase based on the amplitude and frequency characteristics of the predicted fluctuation sequence; then, map the compensation sequence to the opening range of the pulse valve group, specifically, when the compensation sequence is positive, the corresponding... The opening of the pulse valve group decreases according to an adaptive step size. When the compensation sequence is negative, the opening of the corresponding pulse valve group increases according to an adaptive step size, wherein the adaptive step size is proportional to the local rate of change of the compensation sequence. At the same time, the pulse valve group includes a high-frequency regulating valve and a low-frequency regulating valve connected in series. By adjusting the opening ratio of the two valves, the compensation effect of different frequency components is achieved. The high-frequency regulating valve mainly responds to the rapid fluctuation component, and the low-frequency regulating valve mainly responds to the slow change component. Finally, based on the actual response characteristics of the system, the adaptive step size and opening ratio parameters are dynamically optimized to ensure that the formed reverse compensation channel can effectively suppress pressure fluctuations during the switching of operating conditions.

[0048] The specific process for generating a compensation sequence with equal amplitude but opposite phase is as follows: First, perform a Hilbert transform on the predicted wave sequence to obtain its analytic signal. Separate the instantaneous amplitude and instantaneous phase information from the analytic signal. After obtaining the instantaneous phase, add π radians to it to achieve phase reversal. Then, perform wavelet packet decomposition on the predicted wave sequence to obtain sub-sequences of different frequency bands. Calculate the energy proportion of each sub-sequence and construct an energy density distribution curve. Based on the inflection point characteristics of the curve, automatically determine the frequency band selection boundary. When the energy density of a sub-sequence is greater than the value corresponding to the inflection point, retain the frequency band component. Finally, compare the retained frequency band sub-sequences with the corresponding anti-phase information. The reconstruction employs a frequency band adaptive gain control method. A piecewise continuous gain function is constructed based on the characteristic frequencies of the frequency response curve, ensuring continuous differentiability of the gain at the characteristic frequency points to guarantee that the reconstructed sequence has the same amplitude as the original sequence. Subsequently, an adaptive smoothing operator is constructed using the local statistical characteristics of the sequence. The length of the smoothing window is determined by the position of the first zero of the sequence's autocorrelation function. The smoothing intensity is dynamically adjusted based on the changing trend of the local variance to achieve adaptive smoothing of the sequence. Finally, the processed sequence is mapped and normalized using the amplitude range of the original sequence, ensuring that its amplitude distribution characteristics are completely consistent with the original sequence, thus obtaining the final compensated sequence.

[0049] The specific process for separating instantaneous amplitude and instantaneous phase information from the analytic signal is as follows: First, the original signal is transformed to the frequency domain using Fourier transform. After setting the negative frequency components to zero, an inverse Fourier transform is performed to obtain the analytic signal. This analytic signal contains the complete frequency characteristics of the original signal and eliminates spectral aliasing. Then, the envelope information is extracted from the analytic signal using polar coordinate transformation. The instantaneous amplitude sequence is obtained by calculating the magnitude of the signal at each time point. Next, the phase information is extracted from the analytic signal using the four-quadrant arctangent function. The principal phase sequence is calculated based on the ratio of the real and imaginary parts of the signal in the complex plane. Subsequently, the principal phase sequence is subjected to cumulative differential processing. When a phase jump between adjacent sampling points is detected to exceed two-thirds of a cycle, the phase jump is corrected by accumulating or subtracting the phase of a complete cycle. Finally, an adaptive filtering method based on local curvature is used to smooth the instantaneous amplitude sequence and the continuous phase sequence. The cutoff frequency of the filter is dynamically adjusted according to the local variation characteristics of the signal, thereby obtaining smooth and physically meaningful instantaneous amplitude and phase information.

[0050] The specific process for calculating the energy proportion and constructing the energy density distribution curve for each subsequence is as follows: First, perform joint time-frequency analysis on each subsequence, using a windowed short-time Fourier transform to calculate its time spectrum. A Gaussian window is chosen as the window function, and the window length is determined based on the square root of the sequence length. The overlap rate between adjacent windows is set to 50%. Then, calculate the energy distribution of each subsequence in each frequency band. Specifically, integrate the time spectrum along the time dimension to obtain the frequency-energy distribution curve, and simultaneously calculate the total energy value of the subsequence. Next, divide the energy value of each frequency band by the total energy to obtain the normalized energy proportion. Then, perform a normalized energy proportion analysis on the energy proportions of all subsequences. The energy density distribution curve is constructed by sorting and storing the corresponding frequency information. Then, the kernel density estimation method is used to construct the energy density distribution curve. The kernel function is the Epanechnikov kernel, and the bandwidth parameter is adaptively determined by the Silverman criterion. The frequency is used as the independent variable, and the corresponding energy proportion is used as the dependent variable. Finally, the derivative curve is obtained by differentiating the energy density distribution curve. The zero point of the derivative curve corresponds to the local extreme point of the original curve. The inflection point of the energy density distribution curve is identified by analyzing the sign change characteristics of the derivative curve. The frequency and energy proportion information corresponding to the inflection point are extracted as the basis for subsequent frequency band selection.

[0051] Furthermore, the specific process of constructing a piecewise continuous gain function based on the characteristic frequencies of the frequency response curve is as follows: First, singular value decomposition is performed on the frequency response curve to extract its main eigenvectors. Based on the amplitude distribution of the eigenvectors, the set of inherent frequency points of the system is identified. The frequency points are sorted in descending order of amplitude to obtain the main characteristic frequency f1 and the secondary characteristic frequency f2. Then, a transition function is constructed between adjacent characteristic frequencies using cubic Hermitian interpolation. The interpolation function must satisfy the continuity constraints of the function value and the first derivative at the characteristic frequency points. Specifically, when the frequency f is less than f1, the gain function adopts the exponential decay form G(f) = exp(-αf / f1), where α is the decay coefficient. When the frequency f is between f1 and f2, the gain function adopts the cosine transition form G(f) = exp(-αf / f1). =0.5[1+cos(π(f-f1) / (f2-f1))], when the frequency f is greater than f2, the gain function adopts the inverse proportional form G(f)=β / (f-f2+γ), where β and γ are shape parameters; then, the parameter values ​​in the gain function are determined by using the amplitude-frequency characteristics of the frequency response curve, and the values ​​of α, β and γ are obtained by fitting and optimizing the actual response data using the least squares method; subsequently, a smooth transition interval is introduced near the characteristic frequency point, and a Bezier curve is used to achieve a smooth connection of the gain function in the transition interval, ensuring the continuity of the derivative of the gain function throughout the frequency range; finally, the gain function is normalized according to the dynamic response characteristics of the system, so that its maximum gain value does not exceed the expected range, thus obtaining the final piecewise continuous gain function. As shown in the following formula:

[0052] in, is the optimized gain function, is the normalization coefficient and η ∈ (0, 1], is the attenuation coefficient in the low-frequency band, is the input frequency, is the main characteristic frequency, is the secondary characteristic frequency, is the half-width of the transition interval, and are the cubic Hermite interpolation coefficients, and are the shape parameters in the high-frequency band.

[0053] It should be noted that the specific process of obtaining the values of α, β, and γ through fitting and optimizing the actual response data by the least squares method is as follows: First, normalize the measured frequency response data, and map all gain values to the interval [0, 1] based on the maximum gain value; then select at least 10 frequency sampling points and their corresponding actual gain values in the low-frequency band f < f1, and use the natural logarithm transformation to convert the exponential fitting problem into a linear fitting problem, and preliminarily determine the value of α by solving the linear equations; then select at least 15 frequency sampling points and their corresponding actual gain values in the high-frequency band f > f2, and use the Newton iteration method to jointly solve the initial values of β and γ, where the initial value of β is selected as the product of the gain value and frequency at the maximum frequency point, and the initial value of γ is selected as 5% of f2; subsequently, construct the overall objective function, which is the sum of the squares of the differences between the theoretical gain values and the actual gain values at each frequency point, and use the gradient descent method to jointly optimize α, β, and γ. The step size in the optimization process is determined by linear search, and the iteration stops when the relative change rate of the objective function is less than 0.1%; finally, calculate the first-order derivatives of the gain function values at the characteristic frequency points f1 and f2 based on the obtained parameters, and fine-tune the values of α and β to make the derivatives meet the continuity requirements, while keeping the value of γ unchanged to maintain the attenuation characteristics in the high-frequency band.

[0054] S4: When the operating condition switch occurs, suppress the inertial migration of the refrigerant in the main circuit based on the instantaneous pressure fluctuation generated by the reverse compensation channel, and adjust the duration of the compensation channel according to the buffer coefficient of the main circuit to achieve the rapid stability of the system COP.

[0055] First, instantaneous refrigerant state parameters are acquired using temperature and pressure sensors in the main circuit. The refrigerant mass flow rate and specific enthalpy of the main circuit are then calculated. When the pressure fluctuation amplitude exceeds the pressure threshold at the time of operating condition switching, the reverse compensation channel is immediately activated. Next, the refrigerant migration time constant is calculated based on the buffer coefficient of the main circuit. This time constant is determined by the pipe volume, refrigerant density, and flow resistance coefficient. A larger migration time constant extends the duration of the compensation channel, while a smaller migration time constant shortens it. Then, the opening sequence of the pulse valve assembly is dynamically adjusted according to the pressure difference between the main circuit and the compensation channel. A larger pressure difference increases the opening to enhance the compensation effect, while a smaller pressure difference decreases the opening to avoid overcompensation. Subsequently, real-time monitoring... The system monitors the COP value change trend and adjusts the flow distribution ratio of the compensation channel using an adaptive PID controller. When the COP value decreases at a rate exceeding the preset COP value, the compensation flow rate is increased; when the COP value begins to rise, the compensation flow rate is gradually decreased. Simultaneously, the response characteristics of the compensation channel are dynamically optimized based on the flow state characteristics of the refrigerant in the main loop. When the refrigerant is in the two-phase region, a high-frequency compensation component is added to suppress pressure fluctuations caused by boiling; when the refrigerant is in the single-phase region, a low-frequency compensation component is added to smooth pressure fluctuations caused by flow inertia. Finally, a criterion for exiting the compensation channel is established based on the system's instantaneous energy efficiency ratio. When the fluctuation amplitude of the COP value is less than the set range and the change trend is stable over multiple consecutive sampling periods, the compensation channel is gradually closed, allowing the system to return to normal operating conditions.

[0056] The specific process for determining the pressure threshold is as follows: First, statistical analysis is performed on historical operating data of the system during steady-state operation. Pressure fluctuation data samples are collected for several consecutive working cycles, and their standard deviation σ and mean μ are calculated. μ±nσ is defined as the pressure fluctuation range during normal system operation, where n is the fluctuation multiple coefficient. Then, the pressure fluctuation characteristics under different operating condition switching scenarios are classified and studied, including typical operating conditions such as cooling mode switching, defrosting mode switching, and load changes. The peak value and duration of pressure fluctuation under each switching scenario are recorded. Next, based on the system's safety margin requirements, the pressure fluctuation threshold is initially set as a multiple of the upper limit of the normal operating pressure fluctuation range, i.e., μ+kσ, where k is the safety margin coefficient. Finally, the set threshold is verified through system trial operation, and the impact of the threshold setting on the system response characteristics is evaluated to ensure that the threshold can trigger compensation control in a timely manner without causing frequent system switching, thus forming the final threshold setting scheme.

[0057] The specific process for setting the COP preset value is as follows: First, obtain the standard COP value of the system under steady-state conditions. Collect COP change data under different loads and environmental conditions through continuous operation to establish the COP fluctuation characteristic curve. Then, analyze the dynamic response characteristics of COP during operating condition switching, extract the instantaneous rate of change data of COP, and use wavelet transform to decompose the COP rate of change into multiple scales to obtain its main frequency components and energy distribution. Next, based on the actual operating data of the system in multiple operating condition switching cycles, statistically analyze the decay trend of the COP value, extract the probability distribution characteristics of the COP rate of change, calculate its mean and standard deviation, and determine the fluctuation range of the COP rate of change. Subsequently, based on the system's thermal inertia and refrigerant migration time constant, evaluate the system's self-recovery capability to COP fluctuations and obtain the COP recovery limit of the system under uncompensated conditions. Finally, use the fluctuation characteristics and self-recovery limit to comprehensively determine the preset value of the COP reduction rate, so that it can respond promptly to abnormal degradation of system performance without overcompensating for normal system fluctuations.

[0058] In summary, the hybrid heat pump energy efficiency adaptive optimization method based on the embodiments of the present invention is explained. It obtains the circulation cycle parameters of the refrigerant in the main loop and the bypass pressure parameters of the refrigerant in the bypass loop of the hybrid heat pump system. Based on the circulation cycle parameters, it calculates the residence time of the refrigerant in the main loop and accurately predicts the critical point of operating condition switching by constructing a dual buffer matrix. Before detecting the critical point, the opening sequence of the pulse valve group in the bypass loop is pre-adjusted to form a reverse compensation channel, thereby effectively suppressing the inertial migration of refrigerant in the main loop and reducing instantaneous pressure fluctuations during operating condition switching. This improves the energy efficiency stability and adjustment response accuracy of the hybrid heat pump system under dynamic operating conditions, helps to achieve rapid stabilization of the system's COP, thereby improving the overall operating efficiency of the system and reducing operating energy consumption.

[0059] Here, those skilled in the art will understand that the specific operations of each step in the above-described hybrid heat pump energy efficiency adaptive optimization method have been referenced above. Figure 1 and Figure 2 The description of the adaptive optimization method for energy efficiency of hybrid heat pumps is detailed here, and therefore, its repeated description will be omitted.

[0060] In summary, the hybrid heat pump energy efficiency adaptive optimization method based on the embodiments of the present invention is explained. It obtains the circulation cycle parameters of the refrigerant in the main loop and the bypass pressure parameters of the refrigerant in the bypass loop of the hybrid heat pump system. Based on the circulation cycle parameters, it calculates the residence time of the refrigerant in the main loop and accurately predicts the critical point of operating condition switching by constructing a dual buffer matrix. Before detecting the critical point, the opening sequence of the pulse valve group in the bypass loop is pre-adjusted to form a reverse compensation channel, thereby effectively suppressing the inertial migration of refrigerant in the main loop and reducing instantaneous pressure fluctuations during operating condition switching. This improves the energy efficiency stability and adjustment response accuracy of the hybrid heat pump system under dynamic operating conditions, helps to achieve rapid stabilization of the system's COP, thereby improving the overall operating efficiency of the system and reducing operating energy consumption.

Claims

1. A hybrid heat pump energy efficiency self-adaptive optimization method, characterized in that, The method comprises the following steps: acquiring a circulation period parameter of main loop refrigerant and a bypass pressure parameter of bypass loop refrigerant in a hybrid heat pump system, calculating a residence time of the main loop refrigerant based on the circulation period parameter; constructing a double-buffering matrix according to the residence time and the bypass pressure parameter, and predicting a critical point of working condition switching through the double-buffering matrix; pre-adjusting an opening sequence of a pulse valve group in the bypass loop before detecting the critical point, the opening sequence of the pulse valve group being opposite to a fluctuation trend of the double-buffering matrix to form a reverse compensation channel; when the working condition switching occurs, inhibiting inertial migration of refrigerant in the main loop based on instantaneous pressure fluctuation generated by the reverse compensation channel, and adjusting a duration of the compensation channel according to a main loop buffering coefficient to realize rapid stabilization of system COP.

2. The adaptive optimization method of hybrid heat pump energy efficiency according to claim 1, wherein, The double-buffering matrix comprises a main loop buffering coefficient and a bypass adjustment coefficient; dividing the residence time into M time sequence nodes according to a preset time interval, recording corresponding bypass pressure parameters at each time sequence node to generate a time sequence state matrix; performing exponential smoothing processing on the residence time sequence in the time sequence state matrix to obtain the main loop buffering coefficient, and performing difference operation on the bypass pressure parameter sequence to obtain the bypass adjustment coefficient.

3. The adaptive optimization method of hybrid heat pump energy efficiency according to claim 2, wherein, performing difference operation on the bypass pressure parameter sequence, obtaining positive and negative pressure gradient cumulative values by performing difference operation on the bypass pressure sequence after reducing sampling to 2-second interval, and then obtaining a preliminary coefficient by comparing the positive and negative pressure gradient cumulative values with a reference pressure value, and finally obtaining a final bypass adjustment coefficient sequence by using cubic spline interpolation smoothing processing.

4. The adaptive optimization method of hybrid heat pump energy efficiency according to claim 3, wherein, The method of predicting the critical point of working condition switching through the double-buffering matrix comprises the following steps: calculating a sequence complexity S1 of the main loop buffering coefficient and a sequence complexity S2 of the bypass adjustment coefficient by using sliding entropy; constructing a phase space trajectory based on the sequence complexity S1 and the sequence complexity S2, and calculating a Lyapunov exponent of the phase space trajectory; when the Lyapunov exponent changes from negative to positive and the duration exceeds a preset period, determining that the system is about to have a working condition switching inflection point.

5. The adaptive optimization method of hybrid heat pump energy efficiency according to claim 4, wherein, The calculation of the Lyapunov exponent is shown in the following formula: wherein, is the total Lyapunov exponent of the system, is the total number of movements of the reference point P0 on the trajectory in the phase space, is the number of effective nearest neighbors for the mth reference point to complete the tracking, is the distance constraint function, and are the sequence complexity coordinate values of the ith evolutionary point, respectively, and are the sequence complexity coordinate values of the reference evolutionary point, is the initial distance, is the number of selected nearest neighbors, is the nearest neighbor weight function, is the hyper-sphere radius.

6. The adaptive optimization method of hybrid heat pump energy efficiency according to claim 1, wherein, The opening sequence of the pulse valve group is obtained by converting a predicted fluctuation trend into a compensation sequence with opposite phase, and adjusting according to adaptive step mapping rules and high and low frequency valve matching adjustment to form an actual valve control amount for actively inhibiting system pressure fluctuation.

7. The adaptive optimization method of hybrid heat pump energy efficiency according to claim 6, wherein, The generation of the compensation sequence comprises the following steps: analyzing a predicted fluctuation sequence to separate instantaneous amplitude and instantaneous phase information therefrom, and performing phase inversion; decomposing the predicted fluctuation sequence to obtain sub-sequences of different frequency bands, calculating an energy proportion of each sub-sequence, and constructing an energy density distribution curve; automatically determining a frequency band screening boundary based on an inflection point feature of the curve, and retaining a frequency band component when the energy density of the sub-sequence is greater than a value corresponding to the inflection point; reconstructing the retained frequency band sub-sequences and corresponding inverse phase information, and constructing a segmented continuous gain function according to a characteristic frequency of a frequency response curve; performing adaptive smoothing and amplitude normalization using local statistical properties of the sequence to generate a signal sequence with equal amplitude but opposite phase of the original sequence.

8. The adaptive optimization method of hybrid heat pump energy efficiency according to claim 7, wherein, The method comprises the following steps: The piecewise continuous gain function is obtained by singular value decomposition of the frequency response curve to obtain characteristic frequencies, and then exponential decay, cosine transition and inverse proportion are respectively used to construct the basic function in different frequency intervals; Combined with least square optimization of parameters, smooth transition and normalization processing, a gain control function with continuous derivative characteristics in the whole frequency range is finally formed.

9. The adaptive optimization method of hybrid heat pump energy efficiency according to claim 7, wherein, The instantaneous amplitude and phase information are separated by performing Fourier transform on the original signal to obtain the analytical signal, using polar coordinate transformation to extract the envelope, four quadrant inverse tangent function to extract the phase, and through phase jump correction and adaptive curvature filtering.

10. The adaptive optimization method of hybrid heat pump energy efficiency according to claim 7, wherein, The refrigerant inertia migration in the main circuit is inhibited by detecting the pressure fluctuation through the main circuit temperature and pressure sensor, starting the reverse compensation channel when the pressure threshold is exceeded, dynamically adjusting the pulse valve group opening degree and compensation duration based on the buffer coefficient and the refrigerant state, and optimizing the response characteristics of the compensation channel according to the COP change trend and the refrigerant flow characteristics.

Images