Flywheel AFE energy storage power supply system for data center

Abstract

The invention relates to the technical field of data center energy storage and power supply, and discloses a flywheel AFE energy storage and power supply system for a data center, which comprises a flywheel energy storage unit, an AFE control module, a hierarchical energy scheduling module, a dynamic balance adjustment module and a feedback optimization module. The flywheel energy storage unit constructs a multi-stage energy buffer topology through a rotating speed monitoring node and a kinetic energy conversion network; the AFE control module realizes steady-state and transient-state control by using a multi-mode switching link of the vector modulation unit and inhibits harmonic waves in combination with the direct-current bus compensation unit; the hierarchical energy scheduling module generates a bus voltage reference strategy through a dynamic coordination model of the master control coordination node, and load power supply priorities are sorted through a dynamic priority distribution unit. The dynamic equalization adjustment module realizes power phase synchronization compensation and modulation frequency correction, and the feedback optimization module corrects a power supply strategy based on a closed loop of a bus harmonic evaluation network. The system improves the power supply stability, efficiency and reliability of the data center.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage power supply for data centers, and particularly to a flywheel AFE energy storage power supply system for data centers. Background Art

[0002] With the rapid development of technologies such as cloud computing, big data, and artificial intelligence, the scale and energy consumption of data centers have increased explosively, posing extremely high requirements for the stability, efficiency, and reliability of power supply systems. Traditional data center power supply systems mainly rely on battery energy storage, but battery energy storage has problems such as low energy density, short cycle life, high maintenance costs, and environmental protection. Especially when dealing with high-frequency and short-time power fluctuations, the response speed and regulation ability of batteries are limited, and it is difficult to meet the strict requirements of data centers for power quality.

[0003] On the other hand, the loads in data centers are diverse and dynamically changing, and different loads have significantly different requirements for power supply priority and reliability. Traditional power supply systems lack flexible energy scheduling and load priority management mechanisms, and cannot dynamically adjust power supply strategies according to real-time load demands, easily leading to insufficient power supply for some critical loads or waste of energy for non-critical loads.

[0004] In addition, during the operation of data centers, transient events such as grid voltage fluctuations and load mutations occur frequently. The transient response ability of traditional power supply systems is insufficient, and it is difficult to quickly suppress harmonics and stabilize the bus voltage, which may cause abnormal operation or even failure of data center equipment. At the same time, in the process of energy conversion and distribution in traditional systems, energy losses are large, and the kinetic energy conversion efficiency needs to be improved, and it is impossible to achieve efficient utilization and optimized allocation of energy.

[0005] To solve the above problems, a new type of energy storage power supply system is needed, which can improve energy storage efficiency, enhance the dynamic response ability and stability of the system, realize dynamic energy scheduling and priority management of loads, and at the same time reduce energy losses and improve power quality. Flywheel energy storage, as a new type of energy storage technology, has the advantages of fast response speed, long cycle life, and low maintenance costs, and can effectively cope with high-frequency power fluctuations in data centers. AFE (Active Front End) control technology has advantages such as high power factor, small harmonic pollution, and bidirectional energy flow, and can improve power quality and energy conversion efficiency. Therefore, combining flywheel energy storage with AFE control technology to construct a new type of energy storage power supply system suitable for data centers has important practical significance. Summary of the Invention

[0006] The purpose of the present invention is to provide a flywheel AFE energy storage power supply system for data centers to solve the problems put forward in the above background art.

[0007] To achieve the above object, the present invention provides the following technical solution: a flywheel AFE energy storage power supply system for a data center, the system comprising: a flywheel energy storage unit, an AFE control module, and a hierarchical energy scheduling module; The flywheel energy storage unit includes a rotational speed monitoring node and a kinetic energy conversion network; the AFE control module includes a vector modulation unit and a DC bus compensation unit; the hierarchical energy scheduling module includes a main control coordination node and a dynamic priority allocation unit; the rotational speed monitoring node is configured to collect in real time the angular velocity data of the flywheel rotor and the energy storage state parameters; according to the flywheel kinetic energy conversion efficiency and the bus voltage fluctuation range, a multi-level energy buffer topology is constructed through the kinetic energy conversion network; the kinetic energy conversion network is composed of K energy storage subnets based on rotational speed thresholds connected in parallel; based on the real-time load response characteristics of the K energy storage subnets, a bus voltage reference strategy is generated through the main control coordination node, and combined with the harmonic suppression signal output by the DC bus compensation unit, the dynamic priority allocation unit performs dynamic power supply priority sorting on the data center load.

[0008] Preferably, the system further includes a dynamic balance adjustment module and a feedback optimization module; the dynamic balance adjustment module includes a load balance unit and a frequency compensation unit; According to the generated bus voltage reference strategy and the dynamic power supply priority sorting result, the load balance unit performs phase synchronization compensation on the output power of the flywheel energy storage unit, and the frequency compensation unit dynamically corrects the modulation frequency of the AFE control module; at the same time, the DC bus voltage distortion rate is analyzed in real time through the bus harmonic evaluation network constructed by the feedback optimization module, and the analysis result is fed back to the main control coordination node and the dynamic priority allocation unit to perform closed-loop correction on the power supply strategy.

[0009] Preferably, the load balance unit is composed of a power distribution circuit corresponding to the energy storage subnets connected in parallel in the flywheel energy storage unit; the steps for constructing the multi-level energy buffer topology include: Collect the angular velocity data of the flywheel rotor and the bus voltage fluctuation signal, input them into the sliding mode observer configured for each level of energy storage subnet, generate a rotational speed deviation compensation coefficient and store it in the corresponding energy storage subnet; Based on the rotational speed deviation compensation coefficients stored in each level of energy storage subnet, combined with the instantaneous change rate of the DC bus voltage, the flywheel kinetic energy output power is distributed to the K energy storage subnets through an adaptive weight algorithm to form a dynamic power distribution topology.

[0010] Preferably, the conditions for constructing the energy storage subnet configured by the rotational speed monitoring node include: the angular velocity of the flywheel rotor is within a preset rotational speed range, the kinetic energy conversion efficiency corresponding to the energy storage subnet is greater than a threshold, and the DC bus voltage fluctuation amplitude is less than the allowable deviation range; The vector modulation unit is also configured with a multimodal switching link; the multimodal switching link includes a steady-state operation sub-link and a transient compensation sub-link; the steady-state operation sub-link is generated based on the matching degree between the PWM modulation signal of the AFE control module and the bus voltage reference strategy; the transient compensation sub-link is constructed based on the correlation between the instantaneous power deficit of the flywheel energy storage unit and the load mutation event.

[0011] Preferably, the master control coordination node is configured with a dynamic coordination model; the dynamic coordination model includes a speed-power mapping table corresponding to the flywheel energy storage unit and a harmonic suppression parameter library of the AFE control module; the steps for generating the bus voltage reference strategy include: According to the real-time angular velocity data of the flywheel energy storage unit and the load priority sorting result, calculate the theoretical output power upper limit of each energy storage subnet in the steady-state operation sub-link; Input the theoretical output power upper limit and the transient load demand into the dynamic coordination model to generate a bus voltage reference waveform, and correct the distortion rate of the reference waveform through the transient compensation sub-link to form a bus voltage reference strategy.

[0012] Preferably, the parameter update steps of the dynamic coordination model include: When it is detected that the angular velocity of the flywheel rotor exceeds the preset speed range or the distortion rate of the DC bus voltage exceeds the threshold, trigger a first transient compensation instruction, that is, start the standby energy storage unit and adjust the modulation depth of the AFE control module; If the bus voltage still does not return to the allowable range after the execution of the first transient compensation instruction, trigger a second transient compensation instruction, that is, switch to the transient compensation sub-link through the multimodal switching link, and reallocate the output power weight of the flywheel energy storage unit based on the duration of the load mutation event.

[0013] Preferably, the parameter update steps of the dynamic coordination model further include: When the kinetic energy conversion efficiency of a certain energy storage subnet in the flywheel energy storage unit is lower than the threshold, trigger a third transient compensation instruction, that is, block the power output channel of this subnet, and transfer the corresponding load to other energy storage subnets through the load balancing unit; During the execution of the third transient compensation instruction, if it is detected that other energy storage subnets are overloaded, trigger a fourth transient compensation instruction, that is, call the feedback optimization module to degrade the output strategy of the dynamic priority allocation unit and limit the power supply demand of non-critical loads.

[0014] Preferably, the degradation processing logic of the feedback optimization module includes: Construct a load power supply safety level table according to the distortion rate data output by the bus harmonic evaluation network and the flywheel speed stability index; When the fourth transient compensation instruction is triggered, the power supply priority of the load is dynamically degraded according to the safety level table, and the modulation frequency of the AFE control module is synchronously adjusted to match the degraded power supply strategy.

[0015] Preferably, the present invention further includes a flywheel AFE energy storage power supply control method for a data center, which is implemented based on the above-mentioned flywheel AFE energy storage power supply system for a data center. The method includes: Real-time collect the angular velocity data of the flywheel rotor, the bus voltage fluctuation signal and the load priority parameters, and construct a multi-level energy buffer topology through a sliding mode observer and an adaptive weight algorithm; Generate a bus voltage reference strategy and a dynamic power supply priority ranking based on the output power distribution result of the multi-level energy buffer topology and the load mutation event; Perform multi-modal switching on the modulation parameters of the AFE control module according to the execution status of the transient compensation instruction and the bus voltage distortion rate correction requirement; Perform closed-loop correction on the power supply strategy through a feedback optimization module, and update the load power supply safety level table in real time.

[0016] Preferably, the present invention further includes a computer-readable storage medium, on which computer instructions are stored, and when the computer instructions run, the above-mentioned flywheel AFE energy storage power supply control method for a data center is executed.

[0017] Compared with the prior art, the beneficial effects of the present invention are: The flywheel energy storage unit real-time collects the angular velocity data of the flywheel rotor and the energy storage state parameters through the speed monitoring node. Based on the flywheel kinetic energy conversion efficiency and the bus voltage fluctuation range, a multi-level energy buffer topology is constructed by using a kinetic energy conversion network. This topology is composed of K energy storage subnets based on speed thresholds connected in parallel. This structure can dynamically adjust the working state of the energy storage subnets according to the real-time speed of the flywheel and the bus voltage fluctuation, realize the efficient utilization of the flywheel kinetic energy and the multi-level buffering of energy, improve the energy conversion efficiency of the system and the response ability to power fluctuations. When the angular velocity of the flywheel rotor is within the preset speed range, and the kinetic energy conversion efficiency corresponding to the energy storage subnet is greater than the threshold, and the DC bus voltage fluctuation amplitude is less than the allowable deviation range, the energy storage subnet will be put into operation, ensuring that each energy storage subnet can operate in an efficient and stable state, and avoiding energy loss and system risks caused by inefficient or unstable operation.

[0018] The vector modulation unit in the AFE control module is configured with a multi-modal switching link, including a steady-state operation sub-link and a transient compensation sub-link. The steady-state operation sub-link is generated based on the matching degree between the PWM modulation signal of the AFE control module and the bus voltage reference strategy, and can achieve efficient and stable power conversion and power quality control under normal operating conditions; the transient compensation sub-link is constructed based on the correlation between the instantaneous power deficit of the flywheel energy storage unit and the load mutation event, and can quickly respond when encountering transient events, suppress harmonics, stabilize the bus voltage, and improve the transient response ability and anti-interference ability of the system. The DC bus compensation unit outputs a harmonic suppression signal, which, combined with the dynamic priority allocation unit, can perform dynamic power supply priority sorting on the data center load, ensure the priority power supply of critical loads, and improve the reliability and flexibility of power supply.

[0019] The master control coordination node in the hierarchical energy scheduling module is configured with a dynamic coordination model, which includes a speed-power mapping table corresponding to the flywheel energy storage unit and a harmonic suppression parameter library of the AFE control module. Through this model, it is possible to calculate the theoretical output power upper limit of each energy storage subnet in the steady-state operation sub-link according to the real-time angular velocity data of the flywheel energy storage unit and the load priority sorting result, generate a bus voltage reference waveform in combination with the transient load demand, and then correct the distortion rate of the reference waveform through the transient compensation sub-link to form an accurate bus voltage reference strategy. This dynamic coordination mechanism realizes the coordinated control of the flywheel energy storage unit and the AFE control module, and improves the accuracy and rationality of energy scheduling.

[0020] The dynamic equalization adjustment module includes a load balancing unit and a frequency compensation unit. The load balancing unit is composed of a power distribution circuit corresponding to the energy storage subnets connected in parallel in the flywheel energy storage unit, and can perform phase synchronization compensation on the output power of the flywheel energy storage unit according to the bus voltage reference strategy and the dynamic power supply priority sorting result, realizing the balanced distribution of power among the energy storage subnets, avoiding the situation of overloading or underloading of individual subnets, and improving the stability and reliability of the system. The frequency compensation unit dynamically corrects the modulation frequency of the AFE control module, ensuring that the AFE control module can maintain the best modulation performance under different working conditions, and further improving the power quality and energy conversion efficiency.

[0021] The bus harmonic evaluation network constructed by the feedback optimization module can analyze the distortion rate of the DC bus voltage in real time and feed the analysis results back to the main control coordination node and the dynamic priority allocation unit to perform closed-loop correction on the power supply strategy. This closed-loop feedback mechanism realizes real-time monitoring and optimization of the system operation state, can timely detect and solve problems existing in the system, and improves the adaptive ability and stability of the system. At the same time, the degradation processing logic of the feedback optimization module can construct a load power supply safety level table according to the distortion rate data output by the bus harmonic evaluation network and the flywheel speed stability index. When triggering the transient compensation instruction, it dynamically degrades the load power supply priority according to the safety level table and synchronously adjusts the modulation frequency of the AFE control module to match the degraded power supply strategy, ensuring that the system can still maintain the basic power supply capacity in extreme cases and improving the fault tolerance and reliability of the system.

[0022] Through the collaborative work of multiple modules such as the flywheel energy storage unit, AFE control module, hierarchical energy scheduling module, dynamic equalization adjustment module, and feedback optimization module, the present invention realizes dynamic energy scheduling and priority management of the data center load, improves the energy conversion efficiency, dynamic response ability and stability of the system, reduces energy loss, improves power quality, and provides a reliable power supply guarantee for the safe and efficient operation of the data center. Brief Description of the Drawings

[0023] Figure 1 is the working principle diagram of the flywheel AFE energy storage power supply system for the data center described in the present invention; Figure 2 is the working principle diagram of dynamic equalization adjustment and feedback optimization; Figure 3 is the flow chart of bus voltage reference strategy generation; Figure 4 is the working principle diagram of the first and second transient compensation instructions for updating the dynamic coordination model parameters; Figure 5 is the flow chart of the flywheel AFE energy storage power supply control method for the data center. Detailed Embodiments

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] Please refer to Figures 1-5, the flywheel AFE energy storage power supply system for the data center of the present invention includes a flywheel energy storage unit, an AFE control module, and a hierarchical energy scheduling module. The specific steps are as follows: Step S1: The flywheel energy storage unit includes a rotational speed monitoring node and a kinetic energy conversion network. The rotational speed monitoring node collects the angular velocity data and energy storage state parameters of the flywheel rotor in real time, and transmits the data to the kinetic energy conversion network. The kinetic energy conversion network constructs a multi-level energy buffer topology formed by the parallel connection of K energy storage subnets based on rotational speed thresholds according to the flywheel kinetic energy conversion efficiency and the DC bus voltage fluctuation range. Each energy storage subnet corresponds to a different rotational speed interval, and its access conditions include that the angular velocity of the flywheel rotor is within a preset interval, the kinetic energy conversion efficiency is greater than a threshold, and the fluctuation amplitude of the DC bus voltage is less than the allowable deviation range.

[0026] Step S2: The AFE control module includes a vector modulation unit and a DC bus compensation unit. The vector modulation unit receives the bus voltage reference strategy from the hierarchical energy scheduling module and generates a PWM modulation signal to control the output of the AFE inverter. The DC bus compensation unit monitors the harmonic components of the bus voltage in real time, generates a harmonic suppression signal, and transmits it to the hierarchical energy scheduling module.

[0027] Step S3: The hierarchical energy scheduling module includes a master coordination node and a dynamic priority allocation unit. The master coordination node generates a bus voltage reference strategy based on the real-time angular velocity data of the flywheel energy storage unit and the load power supply priority ranking result generated by the dynamic priority allocation unit, combined with the harmonic suppression signal output by the DC bus compensation unit. The dynamic priority allocation unit dynamically adjusts the power supply priority according to the real-time demand characteristics of the data center load to ensure the power supply continuity of critical loads.

[0028] The present invention will be further described below in conjunction with Embodiments 1 to 5: Embodiment 1: This embodiment relates to the expansion module and dynamic adjustment mechanism of the system. The system further includes a dynamic balance adjustment module and a feedback optimization module. The dynamic balance adjustment module includes a load balance unit and a frequency compensation unit.

[0029] During the operation of the system, after the master control coordination node generates the bus voltage reference strategy and the dynamic power supply priority ranking result, the load balancing unit starts to perform phase synchronization compensation on the output power of the flywheel energy storage unit. Specifically, the load balancing unit monitors the output power phase of each parallel energy storage subnet in the flywheel energy storage unit in real time. By analyzing the deviation between the phase of the output power of each subnet and the standard phase required by the bus voltage reference strategy, it adjusts the power distribution circuit parameters corresponding to each subnet, so that the phase of the output power of each subnet gradually approaches the standard phase to achieve phase synchronization compensation. This process avoids problems such as power oscillation and circulating current loss caused by inconsistent output power phases of each subnet through continuous real-time monitoring and parameter adjustment, ensuring the stability and efficiency of the output power of the flywheel energy storage unit.

[0030] The frequency compensation unit dynamically corrects the modulation frequency of the AFE control module. This unit collects the frequency signal of the DC bus in real time, compares it with the standard frequency set by the system (such as 50Hz or 60Hz), and calculates the frequency deviation value. According to the magnitude and direction of the frequency deviation value, the frequency compensation unit sends a corresponding correction signal to the vector modulation unit of the AFE control module to adjust the PWM modulation frequency of the vector modulation unit. For example, when it is detected that the bus frequency is higher than the standard frequency, the frequency compensation unit issues an instruction to reduce the modulation frequency to reduce the power output and suppress the further increase of the frequency; when the bus frequency is lower than the standard frequency, the modulation frequency is increased to increase the power output and promote the recovery of the frequency. Through this dynamic correction mechanism, it is ensured that the power frequency output by the AFE control module is stable within the standard range, meeting the requirements of the data center load for frequency stability.

[0031] The feedback optimization module is constructed with a bus harmonic evaluation network. This evaluation network collects bus voltage signals in real time through voltage sensors arranged on the DC bus, performs harmonic analysis on the collected voltage signals, and calculates the distortion rate of the bus voltage, that is, the proportion of harmonic content in the total voltage. The bus harmonic evaluation network transmits the distortion rate data calculated in real time and the harmonic component distribution to the main control coordination node and the dynamic priority allocation unit through the feedback link. After receiving the feedback data, the main control coordination node evaluates and adjusts the generated bus voltage reference strategy in combination with the current operating state of the flywheel energy storage unit (such as the angular velocity of the flywheel rotor, the power output of each energy storage subnet, etc.). For example, if it is detected that a certain harmonic component increases significantly, resulting in an increase in the distortion rate, the main control coordination node may adjust the harmonic suppression signal parameters of the DC bus compensation unit to enhance the suppression ability of this harmonic and reduce the distortion rate. The dynamic priority allocation unit re-evaluates the urgency of the power supply requirements of each data center load under harmonic interference according to the feedback distortion rate data. For critical loads sensitive to harmonics (such as precision servers, core network devices, etc.), the dynamic priority allocation unit will increase their power supply priority to ensure that they can still obtain high-quality electric energy preferentially during harmonic interference; for non-critical loads insensitive to harmonics, their power supply priority will be appropriately reduced to optimize the overall allocation of power supply resources.

[0032] Through the coordinated operation of the dynamic balance adjustment module and the feedback optimization module, the system forms a closed-loop correction mechanism. The dynamic balance adjustment module makes real-time adjustments to the system output from the aspects of power phase and frequency to ensure the stability of the basic quality indicators of electric energy (phase, frequency); the feedback optimization module deeply optimizes the power supply strategy from the aspects of harmonic suppression and load priority adjustment to address the high-order problems (harmonic distortion) in power quality. The two cooperate with each other, enabling the system to continuously optimize the power supply strategy according to the real-time operating state and external load changes, improve the power supply stability and power quality, and ensure the reliable operation of the data center.

[0033] Embodiment 2: This embodiment details the composition of the load balancing unit and the construction steps of the multi-level energy buffer topology. The load balancing unit is composed of the power distribution circuits corresponding to the energy storage subnets connected in parallel in the flywheel energy storage unit. Each energy storage subnet is equipped with an independent power distribution circuit for adjusting the output power magnitude and flow direction of this subnet.

[0034] The construction steps of the multi-level energy buffer topology are as follows: First, collect the angular velocity data of the flywheel rotor and the bus voltage fluctuation signal, and input these data into the sliding mode observers configured in each energy storage subnet. By analyzing the input data, the sliding mode observer calculates the rotational speed deviation compensation coefficient, which is used to compensate for the influence of flywheel rotational speed fluctuations on the output power of the energy storage subnet. The calculated rotational speed deviation compensation coefficient is stored in the corresponding energy storage subnet for subsequent power allocation.

[0035] Next, based on the rotational speed deviation compensation coefficients stored in each energy storage subnet and combined with the instantaneous change rate of the DC bus voltage, the flywheel kinetic energy output power is allocated to K energy storage subnets through the adaptive weight algorithm. The adaptive weight algorithm assigns corresponding power weights to each subnet according to the real-time operating status of each subnet (such as rotational speed deviation compensation coefficient, current load size, energy conversion efficiency, etc.). For example, for a subnet with stable rotational speed and high energy conversion efficiency, a higher power weight is assigned to make it bear more load; for a subnet with larger rotational speed fluctuations or lower efficiency, a lower power weight is assigned to avoid overloading it. Through this dynamic power allocation method, a dynamic power allocation topology that can adapt to real-time load changes is formed, ensuring that the output power of the flywheel energy storage unit can efficiently and stably meet the power consumption requirements of the data center.

[0036] During the operation of the system, the angular velocity data of the flywheel rotor is collected in real time, which reflects the rotation state and energy storage level of the flywheel. At the same time, the voltage fluctuation signal of the DC bus is collected, which reflects the influence of load changes on the bus voltage. These two groups of data are respectively input into the sliding mode observers configured in each energy storage subnet. The sliding mode observer is a special state observer with strong robustness, which can effectively handle the uncertainties and disturbances in the system. In this system, the sliding mode observer analyzes the input angular velocity data and voltage fluctuation signal, establishes a mathematical model between the flywheel rotational speed and the bus voltage, and calculates the rotational speed deviation compensation coefficient based on this model. The magnitude of this coefficient is proportional to the amplitude of flywheel rotational speed fluctuations and inversely proportional to the amplitude of bus voltage fluctuations. Its function is to compensate for the output power change caused by flywheel rotational speed fluctuations during the power allocation process, ensuring the stability of the output power of the energy storage subnet.

[0037] The calculated rotational speed deviation compensation coefficient is stored in the corresponding energy storage subnet as an important parameter for subsequent power distribution. When performing power distribution, the system first obtains the rotational speed deviation compensation coefficients of each level of energy storage subnet and calculates the instantaneous change rate of the DC bus voltage, which reflects the severity of the load change. Then, these parameters are input into the adaptive weight algorithm. The adaptive weight algorithm is an intelligent algorithm based on machine learning that can dynamically adjust the output result according to the changes in the input parameters. In this system, the adaptive weight algorithm assigns a power weight to each energy storage subnet based on parameters such as the rotational speed deviation compensation coefficient of each energy storage subnet, the current load size, the energy conversion efficiency, and the instantaneous change rate of the bus voltage. This weight value represents the proportion of the subnet in the total output power.

[0038] For energy storage subnets with stable rotational speed and high energy conversion efficiency, the adaptive weight algorithm will assign a higher power weight, enabling them to bear a larger proportion in the total output power. This is because these subnets can convert the flywheel kinetic energy into electrical energy more efficiently and stably to meet the load demand. On the contrary, for subnets with large rotational speed fluctuations and low energy conversion efficiency, the algorithm will assign a lower power weight to reduce their proportion in the total output power and avoid affecting the power supply quality of the entire system due to the unstable operation of these subnets. By this way of dynamically adjusting the power weight, the system can form a dynamic power distribution topology that adapts to real-time load changes.

[0039] In the case of relatively gentle load changes, the adaptive weight algorithm will assign power weights according to the basic performance indicators of each energy storage subnet (such as rotational speed stability, energy conversion efficiency, etc.) to ensure the optimal efficiency of the system during steady-state operation. When the load changes drastically, such as when a large server is started or some equipment is shut down in the data center, the instantaneous change rate of the bus voltage will increase significantly. At this time, the adaptive weight algorithm will respond quickly and readjust the power weights of each energy storage subnet according to the magnitude and direction of the change rate, enabling the system to quickly adapt to the load change and maintain the stability of the bus voltage.

[0040] Through this dynamic power distribution topology, the system can flexibly adjust the output power of each energy storage subnet according to the real-time load change situation, ensuring that the output power of the flywheel energy storage unit can efficiently and stably meet the power consumption requirements of the data center. This multi-level energy buffer topology structure not only improves the system's response ability to load changes, but also optimizes the energy utilization efficiency of the flywheel energy storage unit, extends the equipment service life, and reduces the system operation cost.

[0041] Embodiment 3: This embodiment details the composition of the load balancing unit and the construction steps of the multi-level energy buffer topology. The load balancing unit consists of the power distribution circuits corresponding to the energy storage subnets connected in parallel in the flywheel energy storage unit. Each energy storage subnet is configured with an independent power distribution circuit to adjust the output power magnitude and flow direction of the subnet, achieving balanced power distribution among the subnets.

[0042] The construction steps of the multi-level energy buffer topology are as follows: First, collect the angular velocity data of the flywheel rotor and the bus voltage fluctuation signal, and input these data into the sliding mode observers configured in each level of energy storage subnet. By analyzing the input data, the sliding mode observer calculates the rotational speed deviation compensation coefficient and stores it in the corresponding energy storage subnet. Specifically, the rotational speed deviation compensation coefficient is calculated based on the real-time angular velocity of the flywheel rotor and the center value of the preset rotational speed interval, and the expression is: where is the preset compensation coefficient adjustment factor (dimensionless constant) for adjusting the compensation intensity; is the real-time collected angular velocity of the flywheel rotor (unit: rad / s); is the center value of the preset rotational speed interval of the corresponding energy storage subnet (unit: rad / s). This coefficient reflects the degree of deviation of the flywheel rotational speed from the ideal state and is used to correct the output power of the subnet during subsequent power distribution.

[0043] Next, based on the rotational speed deviation compensation coefficients stored in each level of energy storage subnet, combined with the instantaneous change rate of the DC bus voltage, the flywheel kinetic energy output power is distributed to energy storage subnets through the adaptive weight algorithm to form a dynamic power distribution topology. The adaptive weight algorithm calculates the power distribution weight for each subnet according to the real-time operating state parameters of each subnet (including rotational speed deviation compensation coefficient, current load power , kinetic energy conversion efficiency , etc.). The weight calculation logic needs to comprehensively consider rotational speed stability, energy conversion efficiency, and bus voltage fluctuation response speed. For example, subnets with smaller rotational speed deviations and higher efficiencies will obtain higher weights to preferentially bear the load power. By dynamically adjusting the power distribution ratio of each subnet, the system can balance the flywheel kinetic energy output and load demand in real time, avoid overloading or inefficient operation of a single subnet, and ensure that the multi-level energy buffer topology can achieve efficient energy distribution and buffering under different working conditions.

[0044] Embodiment 4: This embodiment describes the configuration conditions of the rotational speed monitoring node and the multi-modal switching link of the vector modulation unit. The construction conditions of the energy storage subnet configured by the rotational speed monitoring node include: the angular velocity of the flywheel rotor is within the preset rotational speed range, the kinetic energy conversion efficiency corresponding to the energy storage subnet is greater than the threshold, and the fluctuation amplitude of the DC bus voltage is less than the allowable deviation range.

[0045] Specifically, the preset rotational speed range is a reasonable operating range preset according to the mechanical characteristics and energy conversion efficiency characteristics of the flywheel energy storage unit. Only when the real-time angular velocity of the flywheel rotor falls within this range does it indicate that the flywheel is in a stable and efficient operating state and has the conditions to participate in energy conversion. The kinetic energy conversion efficiency threshold corresponding to the energy storage subnet is a key indicator to measure whether the subnet can effectively convert the flywheel kinetic energy into electrical energy. If the efficiency is lower than the threshold, it means that the energy conversion process of the subnet has too much loss and cannot meet the energy efficiency requirements of the system, so it is not activated. The allowable deviation range of the DC bus voltage fluctuation amplitude is set according to the requirements of the data center load for voltage stability. If the fluctuation amplitude is too large, it indicates that the current flywheel operating state or load conditions may lead to a decrease in power supply quality, and at this time, the corresponding energy storage subnet does not have the conditions to access the system. Only when the above three conditions are met simultaneously will the rotational speed monitoring node include the corresponding energy storage subnet in the energy conversion network of the system, enabling it to participate in the conversion of flywheel kinetic energy and the power supply process for the data center load, thereby ensuring that each energized energy storage subnet can work in an efficient and stable state and avoiding affecting the performance of the entire system due to the abnormal operation of a single subnet.

[0046] The multi-modal switching link configured by the vector modulation unit includes a steady-state operation sub-link and a transient compensation sub-link. The steady-state operation sub-link is generated based on the matching degree between the PWM modulation signal of the AFE control module and the bus voltage reference strategy. When the system is in the steady-state operation stage, the bus voltage reference strategy has been generated according to the real-time state of the flywheel energy storage unit and the load demand. The vector modulation unit of the AFE control module needs to generate the corresponding PWM modulation signal according to this reference strategy to control parameters such as the frequency, phase, and amplitude of the electrical energy. The steady-state operation sub-link determines whether the system is in a stable working state by monitoring the matching degree between the PWM modulation signal and the bus voltage reference strategy in real time. When the matching degree of the two meets the preset standard, the steady-state operation sub-link maintains the current modulation parameters and power distribution strategy to ensure that parameters such as the frequency and amplitude of the output voltage are stable within the set range, providing continuous and reliable electrical energy for the data center load.

[0047] The transient compensation sub-link is constructed based on the correlation between the instantaneous power deficit of the flywheel energy storage unit and the load mutation event. During the operation of the data center, load mutation events such as the sudden startup of a server cluster or the start / stop of high-power cooling equipment may occur. These events cause the load's power demand to suddenly increase or decrease, resulting in the instantaneous power output of the flywheel energy storage unit being unable to meet the load demand in a timely manner, generating a power deficit. The transient compensation sub-link determines whether a load mutation event has occurred and the magnitude and direction of the power deficit by continuously monitoring the difference between the output power of the flywheel energy storage unit and the load demand power. When it is detected that the power deficit exceeds a preset threshold, the transient compensation sub-link is triggered, and the system switches to the transient compensation mode. In this mode, the vector modulation unit adjusts the modulation parameters, such as increasing the modulation depth to improve the power output capacity or changing the modulation frequency to quickly respond to load changes, thereby quickly compensating for the power deficit and maintaining the stability of the bus voltage. At the same time, the transient compensation sub-link also adjusts the power distribution weights of each energy storage subnet according to the type and duration of the load mutation event, preferentially allocating energy to the subnet that can quickly respond to load changes, ensuring that the system can quickly return to stability during the transient process and reducing the impact of load mutations on the normal operation of the data center.

[0048] By ensuring the efficient and stable access of the energy storage subnet through the configuration conditions of the rotational speed monitoring node and realizing the flexible control of the steady-state and transient processes through the multi-modal switching link of the vector modulation unit, the system can maintain good power supply performance under different operating conditions, ensuring both the energy efficiency and stability in the steady state and improving the response ability to load mutations in the transient state, thus providing a reliable energy storage power supply solution for the data center.

[0049] Embodiment 5: This embodiment illustrates the parameter update steps of the dynamic coordination model and the degradation processing logic of the feedback optimization module. The parameter update steps of the dynamic coordination model include the triggering mechanism of multiple transient compensation instructions. When the system monitoring module detects that the angular velocity of the flywheel rotor exceeds the preset speed range or the distortion rate of the DC bus voltage exceeds the threshold, the first transient compensation instruction is triggered. At this time, the system first starts the standby energy storage unit, which serves as a supplement to the flywheel energy storage unit and can be quickly put into operation when the main energy storage unit operates abnormally, increasing the system's energy reserve; at the same time, the modulation depth of the AFE control module is adjusted, and by changing the duty cycle of the PWM modulation signal, it quickly responds to abnormal speed or voltage distortion and attempts to restore the bus voltage to the allowable range.

[0050] If, after the execution of the first transient compensation instruction, the monitoring module continuously detects that the bus voltage has not yet returned to the allowable range, it indicates that the current abnormal operating condition exceeds the first-level compensation capacity. At this time, the second transient compensation instruction is triggered. The system switches from the steady-state operation sub-link to the transient compensation sub-link through the multi-modal switching link. The transient compensation sub-link optimizes the modulation parameters for the transient process and can respond more flexibly to power fluctuations. At the same time, based on the duration of the load mutation event, the dynamic coordination model reallocates the output power weights of the flywheel energy storage units. For example, for a load mutation with a long duration, the system will allocate more power weights to the energy storage sub-networks with high kinetic energy conversion efficiency and good rotational speed stability to ensure that critical loads can still obtain stable power supply under long-term abnormal operating conditions.

[0051] When the kinetic energy conversion efficiency of a certain energy storage sub-network in the flywheel energy storage unit is lower than the threshold value, it indicates that this sub-network cannot effectively complete the energy conversion and may affect the overall performance of the system. At this time, the third transient compensation instruction is triggered. The system immediately blocks the power output channel of this sub-network to prevent it from continuing to participate in the power supply process to avoid ineffective energy loss. At the same time, through the power distribution circuit of the load balancing unit, the load originally borne by this sub-network is transferred to other normally operating energy storage sub-networks to maintain the power balance of the system. During the execution of the third transient compensation instruction, if it is detected that other energy storage sub-networks are overloaded (i.e., the output power exceeds their rated capacity), the fourth transient compensation instruction is triggered. At this time, the system calls the feedback optimization module to degrade the output strategy of the dynamic priority allocation unit and restricts the power supply demand of non-critical loads.

[0052] The degradation processing logic of the feedback optimization module is as follows: First, based on the distortion rate data real-time output by the bus harmonic evaluation network and the flywheel rotational speed stability index, a load power supply safety level table is constructed. This level table divides the data center loads into multiple levels. For example, critical loads such as core business servers and network devices are set to a high safety level, and non-critical loads such as office area lighting and non-core application servers are set to a low safety level. When the fourth transient compensation instruction is triggered, the system dynamically degrades the load power supply priority according to the safety level table: The power supply priority of high-safety-level loads is maintained or increased to ensure that they continuously obtain sufficient electrical energy; the power supply priority of low-safety-level loads is reduced, and the system reduces the supply current to them through the power distribution circuit and even temporarily cuts off the power supply when necessary. At the same time, the feedback optimization module synchronously adjusts the modulation frequency of the AFE control module to match the degraded power supply strategy. For example, when reducing the power supply to non-critical loads, the modulation frequency is reduced to reduce the overall energy output, improve the energy efficiency of the system in the load degradation state, and avoid the risk of system collapse caused by overload.

[0053] Through the triggering mechanism of the above-mentioned multiple transient compensation instructions and the degradation processing logic of the feedback optimization module, the system forms a multi-level fault response system. From starting the standby unit, switching the modulation mode to dynamically adjusting the load priority, each link collaborates closely to ensure that in different fault scenarios such as abnormal flywheel speed, voltage distortion, sub-network failure or overload, the system can quickly identify problems and execute corresponding compensation strategies, maximizing the continuous and stable power supply to the critical loads in the data center. At the same time, by reasonably allocating energy resources, the reliability and operating efficiency of the system under abnormal working conditions are improved.

[0054] It should be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device.

[0055] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A flywheel AFE energy storage and power supply system for a data center, characterized in that, Including: A flywheel energy storage unit, an AFE control module, and a hierarchical energy scheduling module; The flywheel energy storage unit includes a rotational speed monitoring node and a kinetic energy conversion network; the AFE control module includes a vector modulation unit and a DC bus compensation unit; the hierarchical energy scheduling module includes a main control coordination node and a dynamic priority allocation unit; the rotational speed monitoring node is used to collect the angular velocity data and energy storage state parameters of the flywheel rotor in real time; according to the flywheel kinetic energy conversion efficiency and the bus voltage fluctuation range, a multi-level energy buffer topology is constructed through the kinetic energy conversion network; the kinetic energy conversion network is composed of K energy storage subnets in parallel based on rotational speed thresholds; based on the real-time load response characteristics of the K energy storage subnets, a bus voltage reference strategy is generated through the main control coordination node, and combined with the harmonic suppression signal output by the DC bus compensation unit, the dynamic priority allocation unit performs dynamic power supply priority sorting on the data center load.

2. The flywheel AFE energy storage power supply system for a data center according to claim 1, wherein The system further includes a dynamic balance adjustment module and a feedback optimization module; the dynamic balance adjustment module includes a load balance unit and a frequency compensation unit; According to the generated bus voltage reference strategy and the dynamic power supply priority sorting result, the load balance unit performs phase synchronization compensation on the output power of the flywheel energy storage unit, and the frequency compensation unit dynamically corrects the modulation frequency of the AFE control module; at the same time, the bus harmonic evaluation network constructed by the feedback optimization module analyzes the DC bus voltage distortion rate in real time, and feeds the analysis result back to the main control coordination node and the dynamic priority allocation unit to perform closed-loop correction on the power supply strategy.

3. The flywheel AFE energy storage power supply system for a data center according to claim 2, characterized in that, The load balance unit is composed of power distribution circuits corresponding to the energy storage subnets connected in parallel in the flywheel energy storage unit; The construction steps of the multi-level energy buffer topology include: Collect the angular velocity data of the flywheel rotor and the bus voltage fluctuation signal, input them into the sliding mode observer configured for each energy storage subnet, generate a rotational speed deviation compensation coefficient, and store it in the corresponding energy storage subnet; Based on the rotational speed deviation compensation coefficients stored in each energy storage subnet, combined with the instantaneous change rate of the DC bus voltage, the flywheel kinetic energy output power is distributed to the K energy storage subnets through the adaptive weight algorithm to form a dynamic power distribution topology.

4. The flywheel AFE energy storage power supply system for a data center according to claim 3, wherein, The construction conditions of the energy storage subnet configured by the rotational speed monitoring node include: the angular velocity of the flywheel rotor is within a preset rotational speed range, the kinetic energy conversion efficiency corresponding to the energy storage subnet is greater than a threshold, and the DC bus voltage fluctuation amplitude is less than the allowable deviation range; The vector modulation unit is further configured with a multi-modal switching link; the multi-modal switching link includes a steady-state operation sub-link and a transient compensation sub-link; the steady-state operation sub-link is generated based on the matching degree between the PWM modulation signal of the AFE control module and the bus voltage reference strategy; the transient compensation sub-link is constructed based on the correlation between the instantaneous power shortage of the flywheel energy storage unit and the load mutation event.

5. The flywheel AFE energy storage power supply system for a data center according to claim 4, wherein, The main control coordination node is configured with a dynamic coordination model; the dynamic coordination model includes a rotational speed-power mapping table corresponding to the flywheel energy storage unit and a harmonic suppression parameter library of the AFE control module; the steps for generating the bus voltage reference strategy include: Calculate the upper limit of the theoretical output power of each energy storage subnet in the steady-state operation sub-link according to the real-time angular velocity data of the flywheel energy storage unit and the load priority sorting result; Input the upper limit of the theoretical output power and the transient load demand into the dynamic coordination model to generate the reference waveform of the bus voltage, and correct the distortion rate of the reference waveform through the transient compensation sub-link to form the bus voltage reference strategy.

6. The flywheel AFE energy storage power supply system for a data center according to claim 5, wherein The parameter update steps of the dynamic coordination model include: When it is detected that the angular velocity of the flywheel rotor exceeds the preset speed range or the distortion rate of the DC bus voltage exceeds the threshold, trigger the first transient compensation instruction, that is, start the standby energy storage unit and adjust the modulation depth of the AFE control module; If the bus voltage still fails to recover to the allowable range after the execution of the first transient compensation instruction, trigger the second transient compensation instruction, that is, switch to the transient compensation sub-link through the multi-modal switching link, and re-allocate the output power weight of the flywheel energy storage unit based on the duration of the load mutation event.

7. The flywheel AFE energy storage power supply system for a data center according to claim 6, wherein The parameter update steps of the dynamic coordination model also include: When the kinetic energy conversion efficiency of a certain energy storage subnet in the flywheel energy storage unit is lower than the threshold, trigger the third transient compensation instruction, that is, block the power output channel of the subnet, and transfer the corresponding load to other energy storage subnets through the load balancing unit; During the execution of the third transient compensation instruction, if it is detected that other energy storage subnets are overloaded, trigger the fourth transient compensation instruction, that is, call the feedback optimization module to degrade the output strategy of the dynamic priority allocation unit and limit the power supply demand of non-critical loads.

8. The flywheel AFE energy storage power supply system for a data center according to claim 7, wherein, The degradation processing logic of the feedback optimization module includes: Construct a load power supply safety level table according to the distortion rate data output by the bus harmonic evaluation network and the flywheel speed stability index; When the fourth transient compensation instruction is triggered, dynamically degrade the load power supply priority according to the safety level table, and synchronously adjust the modulation frequency of the AFE control module to match the degraded power supply strategy.

9. A flywheel AFE energy storage power supply control method for a data center, which is implemented based on the flywheel AFE energy storage power supply system for a data center described in any one of claims 1-8, and is characterized in that, Include: Collect the real-time angular velocity data of the flywheel rotor, the bus voltage fluctuation signal and the load priority parameters, and construct a multi-level energy buffer topology through a sliding mode observer and an adaptive weight algorithm; Generate the bus voltage reference strategy and the dynamic power supply priority sorting based on the output power distribution result of the multi-level energy buffer topology and the load mutation event; Perform multi-modal switching on the modulation parameters of the AFE control module according to the execution status of the transient compensation instruction and the bus voltage distortion rate correction requirement; Perform closed-loop correction on the power supply strategy through the feedback optimization module, and update the load power supply safety level table in real time.

Images