Method, device and equipment for UPS power zero switching inverter control and storage medium

Abstract

This application discloses a UPS power supply zero-switching inverter control method, apparatus, device, and storage medium. The method includes: acquiring multi-dimensional electrical signals from the load side of the UPS output terminal; dynamically characterizing the behavior of the multi-dimensional electrical signals to generate characteristic quantities characterizing the load operating state; identifying the load type based on the characteristic quantities and determining the target output voltage amplitude, phase, and instantaneous power of the inverter during the switching process; constructing a non-singular fast terminal sliding mode switching surface based on the target output voltage amplitude, phase, and instantaneous power; driving the inverter output to converge towards the sliding mode switching surface to achieve coordinated adjustment of the output voltage amplitude, phase, and instantaneous power, wherein the gain function of the sliding mode control law adopts a form of dynamic adjustment with the deviation of the sliding mode surface; and performing power supply path switching when the output voltage and the target output state meet a preset continuity condition. This application improves the reliability and efficiency of zero-switching control.

Description

Technical Field

[0001] This application relates to the field of uninterruptible power supply technology, and in particular to a UPS power supply zero-switching inverter control method, device, equipment and storage medium. Background Technology

[0002] Uninterruptible power supplies (UPS) need to quickly and seamlessly switch power from the bypass to the inverter output when the mains power fails, achieving zero or near-zero switching time to ensure that critical loads (such as servers, medical equipment, and industrial control systems) operate without interruption. Current technologies commonly rely on methods such as fast phase-locked loops (PLLs) combined with feedforward compensation, virtual synchronous machines (VSGs) for transition control, ordinary sliding mode control (SMC), or model predictive control (MPC). While these solutions achieve smooth switching to some extent, they still have significant shortcomings.

[0003] When faced with nonlinear, unbalanced, or impulsive loads, PLLs have limited tracking accuracy, which can easily lead to phase jumps and voltage dips. VSG parameters are sensitive, and the response speed is difficult to meet the requirements of extremely short switching windows. Traditional sliding mode control often uses linear sliding surfaces, which only address voltage or voltage-current deviations and cannot simultaneously constrain phase and instantaneous power, resulting in power fluctuations, increased THD, or severe chattering during the switching process. Advanced methods such as MPC are computationally complex and have high real-time requirements for hardware, making them difficult to implement in small and medium power UPS systems.

[0004] Therefore, existing technologies still cannot fully meet the requirements of seamless power supply for high-reliability UPS in terms of complex load adaptability, transient response speed, steady-state output quality and robustness, and there is an urgent need for a more effective zero-switching inverter control method. Summary of the Invention

[0005] This application provides a UPS power supply zero-switching inverter control method, device, equipment, and storage medium, which improves the reliability and efficiency of zero-switching control.

[0006] This application provides the following solution:

[0007] According to a first aspect, a zero-switching inverter control method for a UPS power supply is provided. The method includes: acquiring multi-dimensional electrical signals from the load side of the UPS output terminal; dynamically characterizing the behavior of the multi-dimensional electrical signals to generate characteristic quantities characterizing the load operating state; identifying the load type based on the characteristic quantities and determining the target output voltage amplitude, phase, and instantaneous power of the inverter during the switching process; constructing a non-singular fast terminal sliding mode switching surface based on the target output voltage amplitude, phase, and instantaneous power, wherein the sliding mode switching surface includes at least nonlinear coupling terms of voltage deviation, phase deviation, and power deviation; driving the inverter output to converge toward the sliding mode switching surface according to the sliding mode control law, and sliding on the sliding mode switching surface to achieve coordinated adjustment of the output voltage amplitude, phase, and instantaneous power, wherein the gain function of the sliding mode control law adopts a form of dynamic adjustment with the deviation of the sliding mode surface; and performing power supply path switching when the output voltage and the target output state meet a preset continuity condition.

[0008] According to one achievable method in an embodiment of this application, identifying the load type includes: constructing a load feature vector based on the feature quantity, and classifying the load into at least one of resistive load, inductive load, capacitive load, nonlinear load, or impulsive load through pattern matching.

[0009] According to one achievable method in this application embodiment, determining the target output voltage amplitude, phase, and instantaneous power of the inverter during the switching process includes: constructing a load state evolution model based on the characteristic quantities; predicting the change trajectory of the load-side voltage, current, and power flow within a preset time window using the load state evolution model to obtain the corresponding target output state trajectory; and extracting the target voltage amplitude target interval, phase alignment relationship, and power flow continuity constraint corresponding to the inverter output based on the target output state trajectory, as the target output voltage amplitude, phase, and instantaneous power during the switching process.

[0010] According to one achievable method in an embodiment of this application, the construction of a non-singular fast terminal sliding mode switching surface based on the target output voltage amplitude, phase, and instantaneous power includes: mapping the deviations between the current output voltage amplitude, phase, and instantaneous power of the inverter and the target output voltage amplitude, phase, and instantaneous power into voltage deviation terms, phase deviation terms, and power deviation terms, respectively; constructing a non-singular fast terminal sliding mode function containing nonlinear power function terms and sign function terms based on the voltage deviation terms, phase deviation terms, and power deviation terms, so that each deviation term converges within a finite time; constructing dynamic weighting coefficients according to the rate of change of each deviation term, and adaptively weighting and fusing the voltage deviation terms, phase deviation terms, and power deviation terms to form the multi-state coupled sliding mode switching surface.

[0011] According to one achievable method in an embodiment of this application, the step of constructing dynamic weighting coefficients based on the rate of change of each deviation term includes: normalizing the rate of change of each deviation term to obtain a corresponding rate of change index; constructing a weight adjustment function based on the rate of change index, so that the weight of each deviation term is nonlinearly adjusted according to the relative magnitude of its rate of change; and introducing a hysteresis adjustment factor into the weight adjustment function to suppress frequent weight oscillations caused by instantaneous fluctuations in the rate of change.

[0012] According to one achievable method in the embodiments of this application, the sliding mode control law uses a continuous approximation function to replace the sign function, so as to reduce chattering in the control process while ensuring the convergence of the system state; the gain function of the sliding mode control law is jointly adjusted according to the deviation of the sliding surface and its rate of change, and gradually decreases when the deviation approaches zero, so as to reduce system oscillation.

[0013] According to one achievable method in an embodiment of this application, the step of determining that the output voltage and the target output state meet the preset continuity condition includes: constructing voltage deviation, phase deviation, and power deviation judgment indicators based on the inverter output voltage amplitude, phase, and instantaneous power, respectively; when the voltage deviation, phase deviation, and power deviation are all lower than the corresponding thresholds and remain continuously satisfied within a preset time window, it is determined that the output state meets the preset continuity condition; wherein, the thresholds of each judgment indicator are adaptively adjusted according to the load type or operating mode.

[0014] According to a second aspect, a zero-switching inverter control device for a UPS power supply is provided. The device includes: a feature quantity generation unit configured to acquire multi-dimensional electrical signals from the load side of the UPS output terminal, perform dynamic behavior characterization on the multi-dimensional electrical signals, and generate feature quantities characterizing the load operating state; a target parameter determination unit configured to identify the load type based on the feature quantities and determine the target output voltage amplitude, phase, and instantaneous power of the inverter during the switching process; a sliding mode switching surface construction unit configured to construct a non-singular fast-terminal sliding mode switching surface based on the target output voltage amplitude, phase, and instantaneous power, wherein the sliding mode switching surface includes at least nonlinear coupling terms of voltage deviation, phase deviation, and power deviation; a sliding mode coordinated adjustment unit configured to drive the inverter output to converge toward the sliding mode switching surface according to a sliding mode control law, and slide on the sliding mode switching surface to achieve coordinated adjustment of the output voltage amplitude, phase, and instantaneous power, wherein the gain function of the sliding mode control law adopts a form of dynamic adjustment with the deviation of the sliding mode surface; and a path switching unit configured to perform power supply path switching when the output voltage and the target output state meet a preset continuity condition.

[0015] According to a third aspect, an electronic device is provided, comprising: one or more processors; and a memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of the method described in any one of the first aspects.

[0016] According to a fourth aspect, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any one of the first aspects.

[0017] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0018] This application dynamically characterizes the multidimensional electrical signals on the load side and identifies the load type, predicting and determining the target voltage amplitude, phase, and instantaneous power during the switching process. It then constructs a non-singular fast-terminal sliding mode switching surface containing nonlinear coupling terms for voltage, phase, and power deviations, achieving convergence of each deviation within a finite time. Simultaneously, it employs a gain function dynamically adjusted according to the sliding mode surface deviation, enabling rapid recovery from the transient phase and significant suppression of chattering in the steady-state phase. Finally, it performs power path switching under conditions of multi-index continuity. This method effectively solves the problems of large phase jumps, deep voltage drops, instantaneous power interruptions, and high steady-state THD in traditional solutions under complex loads. It achieves true zero-switching with continuous voltage, no significant phase jumps, and no power interruptions in an extremely short time, exhibiting strong adaptability and robustness to resistive, inductive, capacitive, and nonlinear loads, significantly improving the UPS's seamless power supply reliability for critical loads.

[0019] Of course, any product implementing this application does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart of a zero-switching inverter control method for a UPS power supply provided in an embodiment of this application;

[0022] Figure 2 A structural block diagram of a zero-switching inverter control device for a UPS power supply provided in an embodiment of this application;

[0023] Figure 3A schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0025] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0026] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0027] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0028] Figure 1 A flowchart illustrating the zero-switching inverter control method for a UPS power supply provided in this application embodiment. Figure 1 As shown, the method may include the following steps:

[0029] Step 101: Obtain the multi-dimensional electrical signal on the load side of the UPS output terminal, perform dynamic behavior characterization on the multi-dimensional electrical signal, and generate characteristic quantities that characterize the load operating state.

[0030] Step 102: Based on the aforementioned characteristic quantities, identify the load type and determine the target output voltage amplitude, phase, and instantaneous power of the inverter during the switching process.

[0031] Step 103: When the switching conditions are met, a non-singular fast terminal sliding mode switching surface is constructed based on the target output voltage amplitude, phase and instantaneous power. The sliding mode switching surface includes at least nonlinear coupling terms of voltage deviation, phase deviation and power deviation.

[0032] Step 104: According to the sliding mode control law, drive the inverter output to converge toward the sliding mode switching surface and slide on the sliding mode switching surface to achieve coordinated adjustment of the output voltage amplitude, phase and instantaneous power. The gain function of the sliding mode control law is dynamically adjusted according to the deviation of the sliding mode surface.

[0033] Step 105: When the output voltage and the target output state meet the preset continuity conditions, perform power supply path switching.

[0034] As can be seen from the above process, this application dynamically characterizes the multi-dimensional electrical signals on the load side and identifies the load type, predicting and determining the target voltage amplitude, phase, and instantaneous power during the switching process. It then constructs a non-singular fast terminal sliding mode switching surface containing nonlinear coupling terms for voltage, phase, and power deviations, achieving convergence of each deviation within a finite time. Simultaneously, it employs a gain function dynamically adjusted according to the deviation of the sliding mode surface, enabling rapid recovery from the transient phase and significant suppression of chattering in the steady-state phase. Finally, it performs power path switching under conditions of multi-index continuity. This method effectively solves the problems of large phase jumps, deep voltage drops, instantaneous power interruptions, and high steady-state THD in traditional solutions under complex loads. It can achieve true zero-switching with continuous voltage, no significant phase jumps, and no power interruptions in a very short time. It exhibits strong adaptability and robustness to resistive, inductive, capacitive, and nonlinear loads, significantly improving the UPS's seamless power supply reliability for critical loads.

[0035] The following describes in detail each step of the above process and the effects that can be further produced, with reference to the embodiments. First, with reference to the embodiments, step 101, namely "acquiring the multi-dimensional electrical signal on the load side of the UPS output terminal, performing dynamic behavior characterization on the multi-dimensional electrical signal, and generating feature quantities characterizing the load operating state", will be described in detail.

[0036] First, multi-dimensional electrical signals from the load side are directly acquired from the UPS output. These signals typically include three-phase voltage, three-phase current, instantaneous active power, and instantaneous reactive power, and sometimes extend to multiple dimensions such as voltage harmonic components, current distortion rate, and power factor change rate. These quantities are not acquired in isolation, but are obtained synchronously at a high sampling rate, forming a high-dimensional time series, thereby reflecting the comprehensive behavior of the load in terms of voltage, current, and power.

[0037] Next, the dynamic behavior of these multidimensional electrical signals is characterized. This step goes beyond simple amplitude, RMS, or average value calculations, employing advanced analytical methods capable of revealing the inherent nonlinearity, chaos, periodic abrupt changes, or trend evolution of the signal. Short-time Fourier transform can be used to capture local spectral changes, empirical mode decomposition to extract intrinsic mode functions, wavelet transform to analyze multi-scale transient features, permutation entropy or sample entropy to quantify the complexity and degree of loss of regularity of the signal, and phase space reconstruction methods to extend one-dimensional time series into high-dimensional attractor trajectories through delayed embedding. These methods collectively uncover dynamic details that are difficult to detect with traditional mean or RMS methods, such as trajectory bifurcation when the load suddenly enters the impact phase from steady state, entropy jumps when harmonic content increases sharply, or periodic distortion patterns caused by nonlinear rectified loads.

[0038] For example, dynamic behavior representation methods employ phase space reconstruction based on delayed coordinate embedding, often combined with complexity metrics such as permutation entropy or sample entropy to generate feature quantities. This approach is particularly suitable for handling the nonlinear, chaotic, or abrupt characteristics of UPS load-side voltage or current time series.

[0039] First, select a key signal on the load side as the observation sequence, such as the single-phase instantaneous value of the output current, to form a time series of length N. Then determine two key parameters: time delay. and embedding dimension Time delay Embedding dimension is typically selected using mutual information functions or average displacement methods to minimize the autocorrelation of the corresponding sequence (typically between 5 and 20 sampling points, depending on the sampling rate). The calculation can be gradually increased using the False Nearest Neighbors method or the Cao method until the addition of dimensions no longer significantly reduces the proportion of false neighbors.

[0040] Next, we will reconstruct the phase space: construct... dimensional vector ,in From 1 to These vectors form orbital points in the reconstructed phase space. Theoretically, these orbits are topologically equivalent to the attractors of the original system and can preserve the core invariants of the load dynamics, such as periodicity, degree of chaos, or bifurcation behavior.

[0041] Finally, feature quantities are extracted from the reconstructed phase space. The most direct approach is to calculate orbital complexity metrics, such as permutation entropy: for each The elements in the dimensional vector are sorted and encoded into patterns. The probability distribution of each pattern is statistically analyzed, and then the Shannon entropy value is calculated. A higher permutation entropy value indicates a more random and disordered sequence (such as distortion caused by a nonlinear rectifier load); a lower value indicates a more regular sequence (such as a steady-state resistive load). Additionally, sample entropy can be combined to quantify the self-similarity and complexity changes of the sequence by comparing the similarity of adjacent vector segments.

[0042] Through the aforementioned dynamic behavior representation, a set of feature quantities characterizing the load's operating state is ultimately generated. These feature quantities typically constitute a low-dimensional but information-condensed vector, including, for example, the main spectrum energy ratio, the maximum Lyapunov exponent estimate, permutation entropy, sample entropy, attractor dimension approximation, short-term prediction residual trend, and power fluctuation severity index. They are no longer simple statistics of the original signal, but rather a high-level summary of the load's current "operating mode" and "imminent change trend," providing crucial input for subsequent load type identification and target output state prediction.

[0043] The design philosophy behind this feature is that UPS load behavior is highly complex and variable, and traditional judgments based on a single voltage or current amplitude can no longer meet the millisecond-level accurate prediction requirements of zero switching. Only through the dynamic and nonlinear characterization of multi-dimensional signals can the "precursor" information of the load be captured in advance, allowing the inverter to prepare the most matched target voltage, phase, and power command the instant the mains power is abnormal, thus achieving true voltage continuity, phase transition-free operation, and uninterrupted power supply.

[0044] The following describes in detail step 102, namely, "based on the aforementioned characteristic quantities, identifying the load type and determining the target output voltage amplitude, phase, and instantaneous power of the inverter during the switching process," with reference to an embodiment.

[0045] First, load type identification is performed using the previously generated features. These features highly condense the dynamic characteristics of the load; for example, permutation entropy reflects the randomness of the sequence, sample entropy quantifies changes in self-similarity, attractor dimension approximates track complexity, and short-term trend indicators capture the direction of power fluctuations. A multi-dimensional feature vector is constructed using these features. Then, pattern matching, cluster analysis, or preset decision rules can be used to classify the load in real time into resistive, inductive, capacitive, nonlinear rectifier loads, impulsive loads, or a mixture thereof. The identification result is not a static label but a dynamically updated operating mode judgment, providing a basis for the subsequent personalized setting of target states. For example, nonlinear rectifier loads are usually accompanied by high harmonics and periodic current spikes, while inductive loads exhibit significant phase hysteresis and transient overshoot. These differences directly affect the setting strategies for target voltage and phase.

[0046] After performing load type identification, the system constructs a load state evolution model based on the current feature values ​​and the identified load pattern. This model can be a simple autoregressive form, a multinomial trend fitting, a state-space observer, or a lightweight grey prediction model, used to describe the possible behavioral changes of the load over the next few milliseconds to tens of milliseconds. The model input includes the current feature vector and historical short-term sequences, and the output is the predicted trajectory of the load-side voltage, current, and power flow within a preset time window. These trajectories include not only amplitude changes but also phase evolution and power fluctuation trends.

[0047] Based on the predicted load state trajectory, the system further extracts and calculates the target parameters that the inverter should immediately output at the moment of switching. Specifically, the target output voltage amplitude is set to a range, usually determined by the center value of the predicted voltage RMS value ± a certain margin, to ensure coverage of possible small fluctuations in the load; the target phase is required to be aligned with the predicted voltage phase as much as possible, usually controlled within a few degrees, to avoid significant phase jumps after switching; the instantaneous power target emphasizes the continuity constraints of active and reactive power, for example, requiring that the instantaneous active power change rate after switching does not exceed a certain percentage, and that the reactive power direction is consistent with the load demand. These target parameters are not fixed values, but dynamically given and adaptively adjusted according to the load type: for impulsive loads, the amplitude margin is slightly larger to prevent overshoot; for nonlinear loads, phase alignment is more stringent to reduce harmonic injection; for inductive loads, reactive power constraints focus more on lag compensation.

[0048] Through this step, the system transforms abstract dynamic characteristics into specific control targets that can be directly used to construct the sliding mode switching surface. This proactive determination method avoids the drawbacks of the traditional approach of "passively following" the mains power or load. Instead, it allows the inverter to "predict" the electrical demands that the load will face in advance, thereby achieving coordinated regulation of continuous voltage amplitude, no significant phase mismatch, and uninterrupted power flow within a millisecond-level switching window, ultimately supporting a truly zero-switching effect.

[0049] The following describes in detail step 103, namely, "constructing a non-singular fast terminal sliding mode switching surface based on the target output voltage amplitude, phase, and instantaneous power, wherein the sliding mode switching surface includes at least nonlinear coupling terms of voltage deviation, phase deviation, and power deviation," with reference to the embodiments.

[0050] This step is the core control design part of the entire zero-switching control method. Its fundamental purpose is to construct a nonlinear switching surface that can simultaneously constrain the three-dimensional deviations of voltage amplitude, phase and instantaneous power, so that the system state can quickly converge to the surface within a finite time and slide along the surface to the equilibrium point, thereby achieving voltage continuity, no significant phase jump and no power interruption during the switching process.

[0051] First, the system compares the inverter's current output voltage amplitude, phase, and instantaneous power with the previously predicted target values, calculating voltage deviation, phase deviation, and power deviation terms respectively. These deviation terms are not simple linear differences but directly reflect the degree of mismatch between the inverter and load demand at the moment of switching. Traditional sliding mode control often only focuses on voltage deviation or the voltage plus current loop, while this method explicitly incorporates phase deviation and instantaneous power deviation into the same sliding mode surface, forming a multi-state coupled structure. The key to this coupled design is that voltage amplitude deviation mainly affects the load's voltage regulation performance, phase deviation directly causes instantaneous power oscillation or reverse flow after switching, and instantaneous power deviation reflects the transient balance of active and reactive power. Only when all three are constrained simultaneously can the power continuity and voltage waveform seamlessness during power supply path switching be truly guaranteed.

[0052] Next, the constructed sliding mode switching surface adopts the form of non-singular fast termination sliding mode, which is a special nonlinear sliding mode surface design. Unlike traditional linear sliding mode surfaces, non-singular fast termination sliding mode surfaces exhibit stronger attraction and faster convergence speed when the deviation is large, while smoothly transitioning when the deviation is close to zero, avoiding control failure caused by singularities. In terms of specific construction, this sliding mode surface includes at least nonlinear power function terms and sign function terms for each deviation term. For example, a nonlinear term with a power exponent less than one is introduced for each deviation term, making the system state exhibit finite-time convergence characteristics. At the same time, the ability to suppress steady-state errors is further enhanced through integral terms or cross-coupling terms. The design of nonlinear coupling terms ensures that different deviations influence each other: when the phase deviation is large, the weight of the power deviation is automatically strengthened; when the power fluctuation is severe, the convergence of the voltage deviation is accelerated, thereby achieving a dynamic balance of coordinated regulation among the three.

[0053] This method of constructing the sliding mode switching surface is significantly superior to existing linear or simple nonlinear sliding modes. The convergence time of a linear sliding mode surface is proportional to the initial deviation and cannot guarantee finite-time stability. While ordinary terminal sliding modes have finite-time convergence, they are prone to singularities that cause control input divergence. This method, through a non-singular fast termination structure, retains the advantages of finite-time convergence while eliminating the risk of singularities, enabling the switching process to perform state recovery within milliseconds, while maintaining the continuity and feasibility of the control signal. In UPS zero-switching scenarios, this characteristic directly determines whether the inverter output can be pulled to a state highly matched with the load prediction trajectory within an extremely short window, avoiding voltage dips, phase mismatches, or power interruptions common in traditional methods.

[0054] Specifically, sliding mode switching surface Defined as: .

[0055] in, Voltage amplitude deviation (the difference between the current inverter output voltage amplitude and the target voltage amplitude); Phase deviation (the difference between the current output phase and the target phase, usually expressed in radians); Instantaneous power deviation (instantaneous active power or the combined deviation of active and reactive power, which can be defined according to requirements) (instantaneous value difference); Nonlinear term coefficients (typically ranging from 0.5 to 2.0, tuned according to system dynamic requirements); Power exponent (recommended 0.4–0.8, ensuring non-singularity and fast convergence; commonly used) or ); Integral term coefficient (to enhance steady-state error suppression, commonly 0.5 to 5.0); : Integral weight (can be set to 1, or adjusted according to the importance of each deviation); These are dynamic weighting coefficients, calculated in real time based on the rate of change of each deviation term. For example:

[0056] .

[0057] in, Represents the corresponding deviation term. Its rate of change, Use small positive numbers to prevent the denominator from being zero.

[0058] Preferably, the step of constructing dynamic weighting coefficients based on the rate of change of each deviation term includes: normalizing the rate of change of each deviation term to obtain a corresponding rate of change index; constructing a weight adjustment function based on the rate of change index, so that the weight of each deviation term is nonlinearly adjusted according to the relative magnitude of its rate of change; and introducing a hysteresis adjustment factor into the weight adjustment function to suppress frequent weight oscillations caused by instantaneous fluctuations in the rate of change.

[0059] Specifically, firstly, the rate of change of each deviation term. Normalization is performed: Then, based on Construct a nonlinear weight adjustment function: Finally, a hysteresis adjustment factor is introduced for smoothing: ;in .

[0060] This parameter combination can rapidly increase the power deviation weight to above 0.6 under impact loads, while the three weights tend to be balanced in steady state, avoiding violent weight oscillations caused by noise or instantaneous changes.

[0061] This approach achieves nonlinear sensitivity to the relative magnitude of the rate of change, and effectively suppresses control chattering caused by frequent weight switching through hysteresis filtering. It demonstrates good stability and adaptability in actual UPS switching experiments.

[0062] In summary, this technology, by introducing a non-singular fast terminal sliding mode switching surface with multi-state nonlinear coupling, unifies voltage, phase, and instantaneous power deviations under a unified dynamic framework for coordinated control. This construction approach not only improves the system's transient response speed and robustness but also significantly reduces transient shocks and steady-state chattering during switching, providing a solid theoretical foundation and engineering feasibility for achieving truly seamless zero-switching.

[0063] The following describes in detail step 104, namely, "according to the sliding mode control law, driving the inverter output to converge toward the sliding mode switching surface and sliding on the sliding mode switching surface to achieve coordinated adjustment of the output voltage amplitude, phase and instantaneous power, wherein the gain function of the sliding mode control law adopts a form of dynamic adjustment with the deviation of the sliding mode surface", with reference to the embodiment.

[0064] The basic principle of sliding mode control law is to decompose the control input into two parts: the equivalent control part and the switching control part. The equivalent control part is responsible for maintaining the force required for the system to continue sliding along the sliding surface once the system state reaches it. It is usually derived by differentiating the system state with respect to the sliding surface and setting the derivative to zero. The switching control part ensures that the system state rapidly approaches the sliding surface from any initial position and resists external disturbances and parameter uncertainties.

[0065] As an implementable approach, the sliding mode control law uses a continuous approximation function instead of a sign function to reduce chattering during the control process while ensuring system state convergence.

[0066] In traditional sliding mode control laws, the switching control term is usually represented by a sign function. (Right now ), in the form of ,in This refers to the sliding surface deviation. When the system state crosses the sliding surface... Will instantly from Jump to Conversely, this can lead to high-frequency switching actions at the control input. Such high-frequency switching in actual inverter systems can cause significant chattering, manifesting as high-frequency, small-amplitude oscillations in the output voltage, increased current harmonics, increased switching losses in power devices, and exacerbated electromagnetic interference. It may even induce mechanical resonance or filter resonance in the system. These chattering problems are particularly critical in UPS zero-switching scenarios because they not only degrade steady-state output power quality (increased THD) but can also amplify transient impacts during the switching transient phase, affecting voltage continuity and the load's ability to perceive the change.

[0067] To address this problem, this invention replaces the traditional sign function with a continuous approximation function in the design of the sliding mode control law. Continuous approximation functions refer to a class of smooth, differentiable, or approximately differentiable functions that... There will be no sudden jumps in the vicinity, but rather the approach will be continuous. This characteristic allows for a significant reduction in the high-frequency switching behavior of the control input while maintaining the robustness and convergence of sliding mode control.

[0068] After using a continuous approximation function, the switching control term becomes ,in For example, when using the hyperbolic tangent function: .in, To adjust the parameters and control the steepness of the approximation ( The larger it is, the closer it gets. The faster it is, the weaker the jitter suppression effect. When it is large, near or ,and Consistent behavior ensures rapid convergence; when When approaching zero, The smooth transition avoids abrupt changes in the sign function, thus significantly reducing chattering amplitude.

[0069] This replacement method effectively suppresses chattering while ensuring system state convergence. Specifically, it results in: rapid return of deviations to the sliding mode surface during the transient switching phase (finite-time convergence characteristics are largely preserved); smoother output voltage waveforms, significantly reduced harmonic content, reduced switching stress on power devices, and improved electromagnetic compatibility during the steady-state phase. This is particularly important in UPS zero-switching applications, as it allows the inverter to maintain low THD during long-term grid-connected operation after power path switching, without sacrificing robustness to load disturbances.

[0070] Of particular note is the key optimization of the sliding mode control law gain function in this method: the gain is no longer a fixed value, but rather dynamically adjusted according to the deviation of the sliding surface. The deviation of the sliding surface is typically defined as the absolute value of the sliding function s or some norm thereof. When the system state is far from the sliding surface, i.e., the deviation is large, the gain function automatically takes a larger value, strengthening the switching control term and generating a strong attraction effect, accelerating the state's approach to the sliding surface. Once the system state approaches or reaches the sliding surface, i.e., the deviation decreases, the gain function automatically decreases to a lower level, significantly weakening the amplitude of the switching term and effectively suppressing the chattering phenomenon common in traditional sliding mode control. This dynamic adjustment mechanism cleverly resolves the contradiction between transient response speed and steady-state control quality: ensuring rapid convergence during the extremely short transient phase of switching, and maintaining high-quality output with low chattering and low harmonics during the long-term operation after switching.

[0071] Specifically, the gain function of the sliding mode control law The method of dynamic adjustment based on the deviation of the sliding surface is adopted, and the specific expression is as follows: .in, This refers to the non-singular fast terminal sliding mode switching surface; This is the maximum gain during the approach phase (typically 20-100, adjusted according to switching frequency and load power). This is the steady-state minimum gain (typically 1 to 10, to suppress chattering). This is the adjustment factor (typically 5 to 20, controlling the steepness of the gain change with deviation).

[0072] This form ensures that when When the gain is large, it is close to To achieve rapid convergence; when Gain approaches zero This effectively reduces control chattering and output voltage harmonic distortion.

[0073] Preferably, the gain function of the sliding mode control law is jointly adjusted according to the deviation of the sliding surface and its rate of change, and gradually decreases when the deviation approaches zero, so as to reduce system oscillation.

[0074] Specifically, when the deviation of the sliding surface |s| is large and the rate of change is large... Pointing away from the sliding surface (i.e.) When the deviation is large, it indicates that the system is moving away from the target surface. At this time, the gain function automatically takes a large value, providing a strong attraction and accelerating the approach of the state to the sliding surface; when the deviation is large... Larger but with a higher rate of change Pointing to the sliding surface ( When the deviation is high, the gain can be appropriately reduced to avoid overshoot; when the deviation is low... Approaching zero and rate of change When the value is relatively small, it indicates that the system has entered the vicinity of the sliding surface or the steady-state sliding phase. At this time, the gain function gradually decreases to a low level, or even approaches a minimum value. This gradual decrease mechanism effectively weakens the strength of the switching control term, avoiding continuous high-frequency oscillations caused by small fluctuations in the sign function or the continuously approximating function near the sliding surface. This significantly reduces system oscillation phenomena, including high-frequency ripple in the output voltage, current spikes, power fluctuations, and potential mechanical resonance risks.

[0075] Through this dynamic gain sliding mode control law, the system can achieve true coordinated regulation of output voltage amplitude, phase, and instantaneous power. Rapid elimination of voltage amplitude deviation ensures that the load voltage does not experience significant drops or overshoots; synchronous convergence of phase deviation avoids instantaneous power oscillations or reverse flow after switching; and constraint of instantaneous power deviation maintains the continuity of active and reactive power, preventing power outages or transient overcurrents. This synergistic effect is difficult to achieve with traditional fixed-gain sliding mode or linear control, especially when facing impulsive loads, nonlinear loads, or severe grid disturbances. The adaptability of the dynamic gain makes the control process more robust and smoother.

[0076] The following describes step 105, namely "when the output voltage and the target output state meet the preset continuity conditions, the power supply path is switched", in detail with reference to the embodiments.

[0077] The preset continuity condition is a quantitative criterion for judging the safety of switching. It doesn't simply look at whether the deviation at a certain moment is less than a threshold, but comprehensively examines whether the deviations of three key indicators—output voltage amplitude, phase, and instantaneous power—are simultaneously below their respective allowable ranges, and this satisfied state must remain stable within a continuous time window. Specifically, the system calculates voltage deviation, phase deviation, and power deviation indicators in real time. For example, voltage deviation can be defined as the relative error between the current inverter output voltage amplitude and the target voltage amplitude; phase deviation is the absolute value of the phase difference between the two; and power deviation is the difference or rate of change between instantaneous active or reactive power and the target value. When all these indicators are less than the preset threshold, switching does not occur immediately. Instead, a time window monitoring mechanism is initiated, requiring all indicators to remain satisfied for a continuous period of 3 to 10 milliseconds. Only through this strict criterion of "continuous satisfaction" can it be confirmed that the inverter output has stably tracked the predicted target state of the load, avoiding erroneous switching caused by instantaneous noise, calculation delays, or transient disturbances.

[0078] The threshold settings are adaptive, dynamically adjusting based on the identified load type or operating mode. For example, for resistive loads, the voltage deviation threshold can be more lenient while the phase deviation threshold is more stringent; for inductive loads, the phase deviation threshold needs to be tightened to prevent reactive power from flowing backward; for nonlinear rectifier loads, the power deviation threshold and rate of change limits are more stringent to suppress transient fluctuations caused by harmonics. This adaptive threshold design ensures applicability and reliability under continuous conditions, avoiding the inconsistency of fixed thresholds under different loads.

[0079] Once all continuity conditions are met, the system immediately performs a power supply path switch. Specific operations typically include: shutting down the bypass thyristor or contactor while simultaneously opening the output channel on the inverter side, or executing the corresponding module's hot-switching logic in a modular UPS. This action occurs within a millisecond window. Because the inverter's output voltage amplitude, phase, and instantaneous power are already highly consistent with the load side at this time, the switching process will not cause significant voltage drops, overshoots, phase jumps, or power interruptions. The transition from bypass power to inverter power is virtually imperceptible to the load equipment, achieving the goal of zero or near-zero switching.

[0080] To further illustrate the technical effects of this application, a specific implementation method and the test results of this implementation method are given below.

[0081] A three-phase online UPS system with a rated capacity of 20kVA is adopted. The inverter topology is a three-phase full-bridge structure, the switching device is IGBT, the switching frequency is 10kHz, and the output filter is LC type. , The control platform uses a TI TMS320F28379D digital signal processor with a sampling frequency of 20kHz. Multi-dimensional electrical signal acquisition on the load side includes three-phase voltage, three-phase current, and calculated instantaneous active and reactive power.

[0082] When a mains voltage drop (e.g., from rated 220V to 176V, a drop of 20%) triggers the switching condition, the system first acquires the load-side signal with a 2ms sampling window and then uses a delayed coordinate embedding method for phase space reconstruction: time delay A sampling point (0.4ms) with an embedding dimension m=4 is used to generate a reconstructed phase space trajectory. Subsequently, the permutation entropy and sample entropy are calculated as feature quantities, and the load type is identified as "hybrid inductive load + nonlinear rectifier load" through a preset pattern matching rule. Based on the AR(2) autoregressive model, the load current trajectory within the next 6ms is predicted, and the target output voltage amplitude is set to ±3% of the rated value (214V~226V). The target phase is aligned with the predicted voltage phase (the deviation is controlled within 3°), and the instantaneous power continuity constraint is that the active power change rate does not exceed .

[0083] Subsequently, a non-singular fast terminal sliding mode switching surface is constructed, in the following form: .in, , , These are the voltage amplitude deviation, phase deviation, and instantaneous active power deviation, calculated after unit normalization.

[0084] The sliding mode control law uses a continuous approximation form, and the switching term uses a hyperbolic tangent function. Substitution sign function. Gain function. A combined regulation approach is adopted: .

[0085] The gain is in the deviation When the rate of change is large and points away from the surface, it approaches 60. When the value is close to zero and the rate of change is small, it gradually decreases to around 8. The control output is then... The transformed signal generates a PWM modulation signal to drive the inverter.

[0086] The continuity conditions are set as follows: voltage deviation <4%, phase deviation <4°, instantaneous power deviation <10%, and maintained continuously for more than 5ms. The threshold is adaptively adjusted according to the load type (in this example, the inductive component is high, so the phase threshold is tightened to 3.5°).

[0087] The test was conducted under laboratory conditions, using a programmable AC power supply to simulate mains power slump. The load was a 10kW resistive-inductive load (power factor 0.8) connected in parallel with a 5kW nonlinear rectified load (THD approximately 35%). The comparison schemes were the traditional zero-switching method based on PLL + feedforward compensation and the fixed-gain linear sliding mode method.

[0088] Test results show that the maximum voltage drop depth of the proposed solution during a mains power slump is 3.8% (compared to 8.2% for the traditional PLL solution and 6.5% for the fixed-gain linear sliding mode), the maximum phase jump is 2.9° (compared to 7.1° for the traditional PLL and 5.4° for the fixed-gain linear sliding mode), and the switching recovery time (deviation convergence to within 5%) is 1.4ms (compared to 3.8ms for the traditional PLL and 2.7ms for the fixed-gain linear sliding mode). The steady-state output voltage THD after switching is 1.9% (compared to 4.3% for the traditional PLL and 3.6% for the fixed-gain linear sliding mode), and the peak power fluctuation does not exceed 7% of the rated power. In a 100% load surge / relief test, the proposed solution exhibits a voltage transient overshoot of <5% and a recovery time of <15ms, demonstrating significantly better robustness and output power quality than the comparative solutions.

[0089] The above-described embodiments and test results show that the present invention can achieve zero switching with continuous voltage, no significant phase jump, and no power interruption under complex load conditions, significantly improving the UPS's ability to provide seamless power supply to critical loads and has good engineering practical value.

[0090] The method provided in this application can be applied to various application scenarios, including but not limited to: The first typical application scenario is data centers and cloud computing facilities. These locations house a large number of servers, storage devices, and network switches. Any millisecond-level voltage interruption or phase jump can lead to data loss, virtual machine migration failure, or service downtime. This method can still control the switching recovery time to within 2ms and the voltage drop depth to less than 4% even under complex nonlinear loads and high harmonic environments, significantly improving system availability. The second application scenario is medical imaging equipment and operating room power supply systems, such as CT, MRI, DSA, or shadowless lamps, which are extremely sensitive to power supply fluctuations. Instantaneous voltage dips can cause image artifacts or equipment restarts. This method, through dynamic load prediction and multi-state nonlinear sliding mode coupling, ensures continuous power flow and a phase deviation of less than 3° during switching, guaranteeing safe and continuous medical operations. The third scenario is industrial automated production lines, especially precision CNC machine tools, robotic arms, or semiconductor manufacturing equipment. These loads often have inductive or impact characteristics, and traditional switching can easily cause overshoot or chattering, leading to decreased processing accuracy or downtime. The adaptive gain and dynamic weighting mechanism of this method can effectively suppress transient shocks, achieve highly robust zero-switching, and maintain production continuity.

[0091] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0092] According to another embodiment, a zero-switching inverter control device for a UPS power supply is provided. Figure 2 A schematic block diagram of a zero-switching inverter control device for a UPS power supply according to one embodiment is shown. Figure 2 As shown, the device 200 includes:

[0093] The feature quantity generation unit 201 is configured to acquire multi-dimensional electrical signals from the load side of the UPS output terminal, perform dynamic behavior characterization on the multi-dimensional electrical signals, and generate feature quantities characterizing the load operating state.

[0094] The target parameter determination unit 202 is configured to identify the load type and determine the target output voltage amplitude, phase and instantaneous power of the inverter during the switching process based on the characteristic quantities.

[0095] The sliding mode switching surface construction unit 203 is configured to construct a non-singular fast terminal sliding mode switching surface based on the target output voltage amplitude, phase and instantaneous power, wherein the sliding mode switching surface includes at least nonlinear coupling terms of voltage deviation, phase deviation and power deviation.

[0096] The sliding mode coordinated adjustment unit 204 is configured to drive the inverter output to converge toward the sliding mode switching surface according to the sliding mode control law, and slide on the sliding mode switching surface to realize coordinated adjustment of the output voltage amplitude, phase and instantaneous power, wherein the gain function of the sliding mode control law adopts a form of dynamic adjustment with the deviation of the sliding mode surface.

[0097] The path switching unit 205 is configured to perform power supply path switching when the output voltage and the target output state meet a preset continuity condition.

[0098] As an implementable approach, the target parameter determination unit 202 can be configured to: construct a load feature vector based on the feature quantity, and classify the load into at least one of resistive load, inductive load, capacitive load, nonlinear load, or impulsive load through pattern matching.

[0099] As an implementable approach, the target parameter determination unit 202 can be configured to: construct a load state evolution model based on the aforementioned characteristic quantities to obtain the current load state and its changing trend; combine the load state evolution model to predict the changing trajectories of the load-side voltage, current, and power flow within a preset time window to obtain the corresponding target output state trajectory; and, based on the target output state trajectory, extract the target voltage amplitude range, phase alignment relationship, and power flow continuity constraint corresponding to the inverter output as the target output voltage amplitude, phase, and instantaneous power during the switching process.

[0100] As an implementable approach, the sliding mode switching surface construction unit 203, when constructing a non-singular fast terminal sliding mode switching surface based on the target output voltage amplitude, phase, and instantaneous power, can be configured to: map the deviations between the current output voltage amplitude, phase, and instantaneous power of the inverter and the target output voltage amplitude, phase, and instantaneous power into voltage deviation terms, phase deviation terms, and power deviation terms, respectively; construct a non-singular fast terminal sliding mode function containing nonlinear power function terms and sign function terms based on the voltage deviation terms, phase deviation terms, and power deviation terms, so that each deviation term converges within a finite time; construct dynamic weighting coefficients according to the rate of change of each deviation term, and adaptively weight and fuse the voltage deviation terms, phase deviation terms, and power deviation terms to form the multi-state coupled sliding mode switching surface.

[0101] As an implementable approach, when constructing dynamic weighting coefficients based on the rate of change of each deviation term, the sliding mode switching surface construction unit 203 can be configured as follows: normalizing the rate of change of each deviation term to obtain the corresponding rate of change index; constructing a weight adjustment function based on the rate of change index, so that the weight of each deviation term is nonlinearly adjusted according to the relative magnitude of its rate of change; and introducing a hysteresis adjustment factor into the weight adjustment function to suppress frequent weight oscillations caused by instantaneous fluctuations in the rate of change.

[0102] As an implementable approach, the sliding mode control law uses a continuous approximation function to replace the sign function, so as to reduce chattering during the control process while ensuring system state convergence; the gain function of the sliding mode control law is jointly adjusted according to the deviation of the sliding surface and its rate of change, and gradually decreases when the deviation approaches zero, so as to reduce system oscillation.

[0103] As an implementable approach, the path switching unit 205, when the output voltage and the target output state meet the preset continuity conditions, includes: constructing voltage deviation, phase deviation, and power deviation judgment indicators based on the inverter output voltage amplitude, phase, and instantaneous power, respectively; when the voltage deviation, phase deviation, and power deviation are all lower than the corresponding thresholds and remain continuously satisfied within a preset time window, it is determined that the output state meets the preset continuity conditions; wherein, the thresholds of each judgment indicator are adaptively adjusted according to the load type or operating mode.

[0104] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0105] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0106] In addition, embodiments of this application also provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method described in any of the foregoing method embodiments.

[0107] And an electronic device, comprising:

[0108] One or more processors; and

[0109] A memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of the method described in any of the foregoing method embodiments.

[0110] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the method described in any of the foregoing method embodiments.

[0111] in, Figure 3 The architecture of an electronic device is illustrated, which may include a processor 310, a video display adapter 311, a disk drive 312, an input / output interface 313, a network interface 314, and a memory 320. The processor 310, video display adapter 311, disk drive 312, input / output interface 313, network interface 314, and memory 320 can communicate with each other via a communication bus 330.

[0112] The processor 310 can be implemented using a general-purpose CPU, microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits to execute relevant programs in order to implement the technical solution provided in this application.

[0113] The memory 320 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 320 can store the operating system 321 for controlling the operation of the electronic device 300, and the basic input / output system (BIOS) 322 for controlling the low-level operations of the electronic device 300. Additionally, it can store a web browser 323, a data storage management system 324, and a zero-switching inverter control device 325 for the UPS power supply, etc. The aforementioned zero-switching inverter control device 325 for the UPS power supply can be the application program that specifically implements the aforementioned steps in this embodiment. In summary, when implementing the technical solution provided in this application through software or firmware, the relevant program code is stored in the memory 320 and is called and executed by the processor 310.

[0114] Input / output interface 313 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, speakers, vibrators, indicator lights, etc.

[0115] Network interface 314 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0116] Bus 330 includes a pathway for transmitting information between various components of the device, such as processor 310, video display adapter 311, disk drive 312, input / output interface 313, network interface 314, and memory 320.

[0117] It should be noted that although the above-described device only shows the processor 310, video display adapter 311, disk drive 312, input / output interface 313, network interface 314, memory 320, bus 330, etc., in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the solution of this application, and does not necessarily include all the components shown in the figures.

[0118] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer program product. This computer program product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0119] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A zero-switching inverter control method for a UPS power supply, characterized in that, The method includes: Acquire multi-dimensional electrical signals from the load side of the UPS output terminal, perform dynamic behavior characterization on the multi-dimensional electrical signals, and generate characteristic quantities characterizing the load operating state; Based on the aforementioned characteristics, the load type is identified, and the target output voltage amplitude, phase, and instantaneous power of the inverter during the switching process are determined. A non-singular fast terminal sliding mode switching surface is constructed based on the target output voltage amplitude, phase, and instantaneous power. The sliding mode switching surface includes at least nonlinear coupling terms of voltage deviation, phase deviation, and power deviation. According to the sliding mode control law, the inverter output is driven to converge toward the sliding mode switching surface and slide on the sliding mode switching surface to achieve coordinated regulation of the output voltage amplitude, phase and instantaneous power. The gain function of the sliding mode control law is dynamically adjusted according to the deviation of the sliding mode surface. When the output voltage and the target output state meet the preset continuity conditions, the power supply path is switched.

2. The method according to claim 1, characterized in that, The identification of load type includes: constructing a load feature vector based on the feature quantity, and classifying the load into at least one of resistive load, inductive load, capacitive load, nonlinear load or impulsive load through pattern matching.

3. The method according to claim 1, characterized in that, Determining the target output voltage amplitude, phase, and instantaneous power of the inverter during the switching process includes: A load state evolution model is constructed based on the aforementioned features; Based on the load state evolution model, the change trajectories of load-side voltage, current, and power flow within a preset time window are predicted to obtain the corresponding target output state trajectory. Based on the target output state trajectory, the target voltage amplitude range, phase alignment relationship, and power flow continuity constraint corresponding to the inverter output are extracted and used as the target output voltage amplitude, phase, and instantaneous power during the switching process.

4. The method according to claim 1, characterized in that, The construction of the non-singular fast terminal sliding mode switching surface based on the target output voltage amplitude, phase, and instantaneous power includes: The deviations between the current output voltage amplitude, phase, and instantaneous power of the inverter and the target output voltage amplitude, phase, and instantaneous power are respectively mapped into voltage deviation terms, phase deviation terms, and power deviation terms; Based on the voltage deviation term, phase deviation term, and power deviation term, a non-singular fast terminal sliding mode function containing nonlinear power function term and sign function term is constructed to make each deviation term converge in a finite time. Dynamic weighting coefficients are constructed based on the rate of change of each deviation term, and adaptive weighting and fusion are performed on the voltage deviation term, phase deviation term, and power deviation term to form the multi-state coupled sliding mode switching surface.

5. The method according to claim 4, characterized in that, The construction of dynamic weighting coefficients based on the rate of change of each deviation term includes: The rate of change of each deviation term is normalized to obtain the corresponding rate of change index. A weight adjustment function is constructed based on the rate of change index, so that the weight of each deviation term is nonlinearly adjusted according to the relative magnitude of its rate of change. A hysteresis adjustment factor is introduced into the weight adjustment function to suppress frequent weight oscillations caused by instantaneous fluctuations in the rate of change.

6. The method according to claim 1, characterized in that, The sliding mode control law uses a continuous approximation function instead of a sign function to reduce chattering during the control process while ensuring system state convergence. The gain function of the sliding mode control law is jointly adjusted according to the deviation of the sliding surface and its rate of change, and gradually decreases when the deviation approaches zero in order to reduce system oscillation.

7. The method according to claim 1, characterized in that, The condition that the output voltage and the target output state satisfy a preset continuity condition includes: Based on the inverter output voltage amplitude, phase, and instantaneous power, voltage deviation, phase deviation, and power deviation judgment indicators are constructed respectively. When the voltage deviation, phase deviation, and power deviation are all lower than the corresponding thresholds and remain continuously satisfied within a preset time window, the output state is determined to meet the preset continuity condition. The threshold values ​​for each criterion are adaptively adjusted based on the load type or operating mode.

8. A zero-switching inverter control device for a UPS power supply, characterized in that, The device includes: The feature quantity generation unit is configured to acquire multi-dimensional electrical signals from the load side of the UPS output terminal, perform dynamic behavior characterization on the multi-dimensional electrical signals, and generate feature quantities characterizing the load operating state. The target parameter determination unit is configured to identify the load type and determine the target output voltage amplitude, phase and instantaneous power of the inverter during the switching process based on the characteristic quantities. The sliding mode switching surface construction unit is configured to construct a non-singular fast terminal sliding mode switching surface based on the target output voltage amplitude, phase and instantaneous power, wherein the sliding mode switching surface includes at least nonlinear coupling terms of voltage deviation, phase deviation and power deviation; The sliding mode coordinated adjustment unit is configured to drive the inverter output to converge toward the sliding mode switching surface according to the sliding mode control law, and slide on the sliding mode switching surface to realize coordinated adjustment of the output voltage amplitude, phase and instantaneous power, wherein the gain function of the sliding mode control law adopts a form of dynamic adjustment with the deviation of the sliding mode surface; The path switching unit is configured to perform power supply path switching when the output voltage and the target output state meet a preset continuity condition.

9. An electronic device, characterized in that, include: One or more processors; and a memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 7.

Images