Waste heat recovery method and system of phase change heat storage data center

Abstract

The invention provides a waste heat recovery method and system for a phase change heat storage data center, and the method comprises the steps: dynamically adjusting the heat storage capability of a heat storage system of the phase change heat storage data center according to a correlation mapping relation between load fluctuation and heat generation, and analyzing the phase change temperature and latent heat value of a material; determining the heat absorption rate range of the material under different loading conditions; aiming at a heat absorption rate range, combining a nonlinear change rule of a heat storage material to construct a prediction model of a heat release rate, acquiring a dynamic adjustment parameter of heat release from the prediction model as model output, and judging whether the heat release rate meets a heat demand of a user side or not; and according to the optimized heat release rate data, the constant temperature maintaining time of the heat storage material in the solid-liquid conversion process is monitored in real time, and the stability index of the temperature platform is determined.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a waste heat recovery method and system for a phase change thermal storage data center. Background Art

[0002] Against the backdrop of the rapid development of modern information technology, energy management and waste heat recovery in data centers have become key areas for energy conservation, emission reduction, and sustainable development. Their importance is self-evident. As high-energy-consuming facilities, data centers generate waste heat that, if effectively utilized, not only reduces operating costs but also provides a stable heat supply for surrounding users. However, current technologies often struggle to adapt to complex dynamic environments when managing data center load fluctuations. In particular, they lack the flexibility and continuity of heat storage and release. Existing methods focus primarily on thermal management under static conditions, ignoring the transient response requirements brought about by sudden load fluctuations. This results in low waste heat recovery efficiency and even the risk of heat supply interruption. The core challenge in this area lies in ensuring the stability of heat storage and release during drastic load fluctuations. In particular, when server loads fluctuate suddenly, the rate of heat absorption and release exhibits nonlinear variations. This irregularity makes stable temperature control extremely difficult. Furthermore, this nonlinear response directly affects the isothermal properties of the heat storage material during the solid-liquid transition process. The duration of the constant temperature can be shortened or prolonged by load fluctuations, posing a threat to the continuity of waste heat recovery. Balancing heat absorption rate and temperature stability in a dynamic environment has become a pressing technical challenge. Therefore, optimizing the dynamic response characteristics of heat storage materials to ensure temperature stability, maintain continuity of waste heat recovery, and reliable heat supply to end users in the face of sudden changes in data center loads has become a key issue. Summary of the Invention

[0003] The present invention provides a waste heat recovery method for a phase change thermal storage data center, which mainly includes:

[0004] By monitoring the load fluctuation characteristics of the phase-change thermal storage data center in real time, the heat output data of the phase-change thermal storage data center computing server under operating conditions is obtained. Combined with the preset load variation model, correlation analysis is performed to obtain the correlation mapping relationship between load fluctuation and heat generation;

[0005] Dynamically adjust the heat storage capacity of the phase change thermal storage data center thermal storage system based on the correlation mapping relationship between load fluctuations and heat generation. By analyzing the phase change temperature and latent heat value of the material, the heat absorption rate range of the material under different load conditions is determined.

[0006] Based on the range of heat absorption rates and the nonlinear variation of heat storage materials, a heat release rate prediction model is constructed. Dynamic adjustment parameters for heat release are obtained from the prediction model as model output to determine whether the heat release rate meets the user's thermal requirements.

[0007] If the heat release rate is lower than the user-end heat demand, the matching degree of heat absorption and release in the phase change thermal storage data center heat storage system is increased to obtain optimized heat release rate data. If the heat release rate is higher than the user-end heat demand, the phase change rate of the thermal storage material is reduced to avoid excessive heat release and energy waste.

[0008] Based on the optimized heat release rate data, the constant temperature maintenance time of the heat storage material during the solid-liquid transition process is monitored in real time to determine the stability index of the temperature platform;

[0009] If the stability index is lower than the preset threshold, the phase change material composition ratio parameters of the heat storage material are adjusted to enhance the adaptability of the constant temperature maintenance time, obtain the adjusted temperature control data, analyze the overall efficiency of heat storage and release in the thermal storage system, and obtain the continuity parameter of waste heat recovery;

[0010] Combined with the real-time feedback of user-side heat demand, the results of the updated heat storage supply reliability strategy are used to determine whether the heat storage supply meets the continuity requirements. If the heat supply continuity requirements are not met, the heat output mode is adjusted to obtain the heat storage supply reliability data of the phase change thermal storage data center.

[0011] The present invention provides a waste heat recovery system for a phase change thermal storage data center, which mainly includes:

[0012] The load monitoring and correlation analysis module is used to monitor the load fluctuation characteristics of the phase change thermal storage data center in real time, obtain the heat output data of the phase change thermal storage data center computing server under the operating state, and perform correlation analysis based on the preset load variation model to obtain the correlation mapping relationship between load fluctuation and heat generation;

[0013] A dynamic heat storage capacity adjustment module is used to dynamically adjust the heat storage capacity of the phase change thermal storage data center thermal storage system based on the correlation mapping relationship between load fluctuations and heat generation. By analyzing the phase change temperature and latent heat value of the material, the heat absorption rate range of the material under different load conditions is determined;

[0014] The heat release prediction model construction module is used to build a heat release rate prediction model based on the heat absorption rate range and the nonlinear change law of the heat storage material. The dynamic adjustment parameters of heat release are obtained from the prediction model as the model output to determine whether the heat release rate meets the user's thermal requirements.

[0015] The heat release rate optimization module is used to increase the matching degree of heat absorption and release in the phase change thermal storage data center thermal storage system if the heat release rate is lower than the user-end heat demand, and obtain optimized heat release rate data. If the heat release rate is higher than the user-end heat demand, the phase change rate of the thermal storage material is reduced to avoid excessive heat release and energy waste;

[0016] The constant temperature stability monitoring module is used to monitor the constant temperature maintenance time of the heat storage material during the solid-liquid transition process in real time based on the optimized heat release rate data, and determine the stability index of the temperature platform;

[0017] The phase change material ratio adjustment module is used to adjust the phase change material composition ratio parameters of the heat storage material to enhance the adaptability of the constant temperature maintenance time if the stability index is lower than the preset threshold, obtain the adjusted temperature control data, analyze the overall efficiency of heat storage and release in the thermal storage system, and obtain the continuity parameter of the waste heat recovery;

[0018] The heat storage supply continuity judgment module is used to combine the real-time feedback of user-side heat demand and determine whether the heat storage supply meets the continuity requirements from the results of the updated heat storage supply reliability strategy. If the heat supply continuity requirements are not met, the heat output mode is adjusted to obtain the heat storage supply reliability data of the phase change thermal storage data center.

[0019] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0020] The present invention discloses a waste heat recovery method and system for a phase-change thermal storage data center. By real-time monitoring of the load fluctuation characteristics of the data center, heat output data is obtained and a correlation mapping relationship between load and heat is established. The heat storage capacity of the thermal storage system is dynamically adjusted according to this relationship, and the solid-liquid transition characteristics of the heat storage material are used to determine the range of heat absorption rates under different loads. A heat release rate prediction model is constructed to obtain dynamic adjustment parameters to determine whether the user's thermal needs are met. The heat release rate is optimized by adjusting the dynamic response characteristics of the material. The temperature stability of the phase change process is monitored in real time, and the material composition ratio is adjusted to enhance the constant temperature maintenance time. Combined with the waste heat recovery efficiency evaluation model, the overall performance is analyzed and the supply reliability strategy is updated. The present invention realizes the precise control of heat storage and release in the phase-change thermal storage data center, thereby improving energy utilization efficiency and heating reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 The present invention is a flow chart of a waste heat recovery method for a phase change thermal storage data center.

[0022] Figure 2 This is a structural schematic diagram of a waste heat recovery system for a phase change thermal storage data center according to the present invention. DETAILED DESCRIPTION

[0023] To further understand the content of the present invention, the present invention is described in detail with reference to the accompanying drawings and examples. The present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the invention are shown in the accompanying drawings.

[0024] like Figure 1-2 In this embodiment, a waste heat recovery method and system for a phase change thermal storage data center may specifically include:

[0025] Step S101: By real-time monitoring of the load fluctuation characteristics of the phase change thermal storage data center, the heat output data of the computing server in the phase change thermal storage data center under the operating state is obtained, and a correlation analysis is performed in combination with a preset load change model to obtain a correlation mapping relationship between load fluctuation and heat generation.

[0026] By deploying temperature sensor arrays and power monitoring devices within data center cabinets, real-time CPU utilization, memory occupancy, and temperature data at corresponding locations are collected from each computing server. The temperature and phase state of the phase-change material in the phase-change thermal storage device are also recorded. Temperature sensors are placed at the air inlets of the computer room to collect ambient temperature, generating a time series dataset of server load fluctuations and heat output. Based on this collected time series dataset, a bidirectional time series of load and heat output is constructed. A sliding window approach is used to calculate the load difference between adjacent moments divided by the time interval, based on a preset time interval, to obtain the load change rate. The heat change rate is also calculated by dividing the temperature difference between adjacent moments by the time interval. The Pearson correlation coefficient is used to calculate the load-heat correlation under different time delays, and the time lag parameter from load change to heat output is determined. Based on the determined time lag parameter, a multivariate linear regression method is used to establish a mapping function from load indicators to heat output. CPU utilization and memory occupancy are used as independent variables, and the difference between the temperature sensor measurement and the ambient temperature is used as the dependent variable. The regression coefficients are determined using the least squares method, resulting in a quantitative correlation between load fluctuation and heat generation.

[0027] In one possible implementation, a phase-change thermal storage data center achieves comprehensive thermal monitoring through a sophisticated sensor network. Temperature sensor arrays are arranged in the top, middle, and bottom layers of each cabinet, with at least four sensors per layer, forming a three-dimensional monitoring network. This arrangement captures temperature gradients within the cabinet, as heat dissipation at different heights varies significantly when servers are operating at high loads. Power monitoring devices are directly connected to the server's power supply circuits, recording power consumption changes in real time and simultaneously obtaining CPU utilization and memory usage data through the server management interface.

[0028] It's important to note that the phase-change material in a phase-change thermal storage device undergoes a solid-liquid phase transition at specific temperatures, absorbing or releasing a significant amount of latent heat. When a data center's load suddenly increases, the heat generated by the servers is absorbed and stored by the phase-change material, preventing a sharp rise in the room's temperature. By deploying temperature probes and phase-state monitoring sensors within the phase-change material container, the thermal storage device's operating status can be monitored in real time.

[0029] Specifically, the sliding window method involves setting a fixed-length time window, such as 5 minutes, and then sliding the window forward in 1-minute increments. Within each window, the load difference between the start and end times is calculated and divided by the window duration to obtain the average rate of change for that period. This method smooths out short-term fluctuations and extracts the main trends in load variation. The heat rate of change is calculated using the same method, dividing the temperature difference by the time interval.

[0030] In one embodiment, the calculation of the Pearson correlation coefficient takes into account time delay. Due to the inertia of heat transfer, there is a time lag between load changes and temperature responses. By setting different delay times, from 0 to 300 seconds, at 10-second intervals, the correlation coefficient between the load series and the delayed temperature series is calculated. When the correlation coefficient reaches its maximum value, the corresponding delay time is the system's thermal response time lag parameter.

[0031] Preferably, the least squares method is used to construct the multivariate linear regression model. CPU utilization and memory occupancy are used as independent variables x1 and x2, and the temperature rise value is used as the dependent variable y. By fitting a large amount of historical data, the regression equation y = a × x1 + b × x2 + c is obtained. The coefficients a and b reflect the contribution of the CPU load and memory load to heat generation, respectively. The establishment of this quantitative relationship enables the system to predict future heat output based on the current load state, providing an accurate basis for the control strategy of the phase change thermal storage device and achieving optimized management of data center energy consumption.

[0032] Step S102 , dynamically adjusting the heat storage capacity of the phase change thermal storage data center thermal storage system based on the correlation mapping relationship between load fluctuation and heat generation, and determining the heat absorption rate range of the material under different load conditions by analyzing the phase change temperature and latent heat value of the material.

[0033] Based on the obtained mapping relationship between load fluctuations and heat generation, the heat storage material's phase transition onset temperature, phase transition completion temperature, and latent heat per unit mass were retrieved from a pre-established phase change material property table containing the physical properties of paraffin, fatty acids, and hydrated salts. This data set of solid-liquid transition characteristics for different types of phase change materials was then obtained. The heat absorption capacity per unit mass of each material was calculated by multiplying the phase transition temperature range by the latent heat value. Based on the comparison of the current data center load level with a preset load threshold, if the load exceeds the threshold, a material combination with a higher heat absorption capacity per unit mass was selected to determine the appropriate phase change material configuration. Based on the determined phase change material configuration, the time from solid-state phase transition to liquid-state phase transition completion was measured using differential scanning calorimetry as the phase transition duration. The heat absorption rate of the material under different load conditions was calculated by dividing the latent heat value by the phase transition duration, resulting in a heat absorption rate range corresponding to each load range.

[0034] In one possible implementation, the construction of a phase change material property table involves systematic testing and classification of multiple phase change materials. Paraffin wax materials typically have a phase change temperature range of 18°C ​​to 65°C, making them suitable for data centers operating in various ambient temperatures. Fatty acid materials, such as lauric acid and myristic acid, have phase change temperatures of approximately 44°C and 58°C, respectively, with latent heat values ​​of 170 to 200 kilojoules per kilogram. Hydrated salt materials, such as sodium sulfate decahydrate, have a phase change temperature of approximately 32°C and a latent heat value of up to 241 kilojoules per kilogram, demonstrating a high heat storage density.

[0035] It's important to note that the mapping relationship between load fluctuations and heat generation is the foundation for dynamic adjustments. If server load suddenly increases from 30% to 80%, the previously established mapping relationship predicts that heat output will increase accordingly after a specific time delay. This predictive capability enables the thermal storage system to respond proactively, rather than passively waiting for temperature increases before making adjustments.

[0036] Specifically, heat absorption capacity is calculated by multiplying the phase transition temperature range by the latent heat value. For example, if a paraffin wax material has a phase transition starting temperature of 25°C and a completion temperature of 28°C, a temperature range of 3°C, and a latent heat value of 180 kilojoules per kilogram, then the heat absorption capacity per unit mass is 540 kilojoules per kilogram. This value reflects the total amount of heat that the material can absorb during the complete phase transition process.

[0037] In one embodiment, the preset load threshold is based on historical data center operating data. When the real-time monitored load exceeds 70%, the system identifies a high load state and selects a hydrated salt material or combination thereof with a higher latent heat value. When the load is below 50%, paraffin wax materials with a lower phase transition temperature can meet the cooling requirements, avoiding resource waste caused by over-configuration.

[0038] Preferably, the application of a differential scanning calorimeter provides an accurate determination of the time of the phase change process. The instrument records the curve of the change in the heat absorption power of the material with temperature by controlling the heating rate. The time points when the phase change starts and ends can be accurately identified from the curve. For example, the phase change process of a certain fatty acid material lasts for 15 minutes at a heating rate of 2°C per minute. Based on the measured duration of the phase change process, the calculation of the heat absorption rate becomes intuitive and clear. If the latent heat value of the material is 190 kilojoules per kilogram and the phase change time is 900 seconds, the heat absorption rate is 0.21 kilojoules per kilogram per second. Under different load conditions, by adjusting the amount and type combination of phase change materials, it is possible to achieve an accurate match with the heat generation rate of the data center, thereby achieving the goal of dynamically adjusting the storage capacity of the thermal storage system and improving the energy utilization efficiency of the data center.

[0039] In step S103, a heat release rate prediction model is constructed based on the heat absorption rate range and the nonlinear change law of the heat storage material. The dynamic adjustment parameters of heat release are obtained from the prediction model as the model output to determine whether the heat release rate meets the user's thermal demand.

[0040] Based on the heat absorption rate range, the temperature and time curve data of the phase change material during the liquid-to-solid transition are collected. The ratio of the temperature difference between adjacent time points to the time interval is calculated as the instantaneous slope. When the slope shows a characteristic change over time, first fast and then slow, it is identified as a nonlinear law of material heat release, and the measured heat release rate values ​​within different temperature ranges are obtained. Based on the measured heat release rate values ​​and the solid-liquid phase ratio calculated by mass ratio at the corresponding time, the support vector regression method is used. The current material temperature, phase ratio, and the difference between the material temperature and the ambient temperature are used as input parameters. A prediction model is trained to output the predicted heat release rate as a dynamic adjustment parameter. By obtaining the real-time heat load demand value from the user-side heating monitoring device, the predicted heat release rate is compared with the demand value. If the release rate is lower than the demand value, it is determined that the current release rate does not meet the user's heat demand.

[0041] In one possible implementation, the transition from liquid to solid in a phase-change material exhibits distinct nonlinear characteristics. When the material begins to solidify, the initial release rate is slow due to the energy barrier required to form nuclei. As the number of nuclei increases and the crystals grow, the release rate reaches a peak. At the end of solidification, as the remaining liquid decreases, the release rate gradually decreases. This slow-to-fast-then-slow-against-a-time process is a typical example of nonlinear heat release.

[0042] It should be noted that the calculation method of the instantaneous slope directly reflects the speed of temperature change. For example, if the temperature drops from 35°C to 34.5°C within a 5-second time interval, the instantaneous slope is 0.1°C per second. By continuously recording the slope values ​​at multiple time points, a slope-time curve can be plotted. When the slope value is found to be 0.05°C per second in the early stage, reaches 0.15°C per second in the middle stage, and drops to 0.03°C per second in the late stage, this change pattern clearly shows nonlinear characteristics.

[0043] Specifically, the calculation of the solid-liquid phase ratio is based on the principle of conservation of mass. In a phase change container, the total material mass can be determined by weighing or volumetric methods. Taking advantage of density differences, when the material partially solidifies, the solid phase sinks to the bottom and the liquid phase rises to the top. By measuring the interface height and combining it with the container's cross-sectional area, the proportion of the solid phase to the total mass can be calculated. For example, when the interface height is 40% of the total container height, considering that the solid phase density is typically about 5% greater than the liquid phase, the solid phase mass ratio can be calculated to be approximately 42%.

[0044] In one embodiment, the support vector regression model training process requires extensive historical data. The collected data includes material temperature, ambient temperature, phase ratio, and the corresponding measured release rate at different times. For example, when the material temperature is 45°C, the ambient temperature is 20°C, and the solid phase ratio is 30%, the measured release rate is 0.8 kilowatts per square meter. By collecting hundreds of sets of such data, the model learns that the release rate increases with increasing temperature differences and higher liquid phase ratios.

[0045] Preferably, the user-side heating monitoring device is usually installed at the inlet of the heating pipeline and is equipped with a flow meter and a temperature sensor. The heat load demand value is calculated by measuring the flow rate and temperature difference of the heating medium. For example, when the hot water flow rate is 10 cubic meters per hour and the inlet and outlet temperature difference is required to reach 15°C, the heat load demand value is 174 kilowatts. This real-time value is compared with the predicted release rate to accurately determine whether the heat storage system meets the heating requirements. When the predicted release rate is only 150 kilowatts, it is determined that the demand cannot be met, and it is necessary to start auxiliary heating or increase the heat release area of ​​the phase change material.

[0046] In step S104, if the heat release rate is lower than the user's heat demand, the matching degree of heat absorption and release in the phase change thermal storage data center thermal storage system is increased to obtain optimized heat release rate data. If the heat release rate is higher than the user's heat demand, the phase change rate of the thermal storage material is reduced to avoid excessive heat release and energy waste.

[0047] If the heat release rate is lower than the user's heat demand, the dynamic response characteristics of the heat storage material are adjusted by increasing the contact area between the phase change material and the heating pipe or adding heat-conducting fins in the material container. The heat transfer per unit time before and after the adjustment is measured, and the ratio of the transfer amount after adjustment to the transfer amount before adjustment is calculated as the matching improvement factor. Based on the obtained matching improvement factor, the heat release rate of the phase change material is recalculated, and the optimized heat release rate data is obtained by multiplying the original release rate by the matching improvement factor. If the heat release rate is higher than the user's heat demand, the flow rate of the circulating medium in the heating pipe is reduced or the thickness of the insulation layer outside the phase change material container is increased to reduce the temperature difference change rate between the inside and outside of the material, thereby reducing the phase change rate of the heat storage material. The adjusted release rate is measured as the optimized heat release rate data.

[0048] In one possible implementation, when the heat release rate is insufficient, increasing the contact area is the most direct way to enhance heat transfer. Phase change materials are typically encapsulated in rectangular or cylindrical containers. Original designs may only have the bottom in contact with the heating pipe. By transforming the container into a multi-tube interpenetrating structure, allowing the heating pipe to pass through the phase change material in a serpentine or spiral shape, the contact area can be increased by 5 to 10 times. This modification allows all parts of the phase change material to transfer heat to the pipe nearby, significantly shortening the heat transfer path.

[0049] It should be noted that the principle of installing thermal fins is based on expanding the heat transfer area and improving the uniformity of temperature distribution. The fins are usually made of aluminum or copper and are arranged radially or in a parallel plate pattern inside the container. When the phase change material begins to solidify and release heat, the material near the container wall solidifies first, forming a solid layer that blocks the internal heat from escaping. The presence of the fins is equivalent to establishing multiple heat conduction channels within the material, allowing internal heat to be quickly transferred to the container wall.

[0050] Specifically, the matching improvement factor is measured by comparing heat flux densities. Before the adjustment, a heat flux meter measured a heat flux density of 2 kilowatts per square meter. After the contact area was expanded and the fins were installed, the heat flux increased to 3.5 kilowatts per square meter under the same temperature difference. The ratio of the two, 1.75, is the matching improvement factor. This factor directly reflects the actual effectiveness of the heat transfer enhancement measures.

[0051] In one embodiment, the flow rate of the circulating medium within the heating pipeline is regulated using a variable-frequency water pump. When excessive heat release is detected, the pump speed is reduced, lowering the flow rate from 20 cubic meters per hour to 12 cubic meters. This reduced flow rate causes the medium in the pipeline to rise in temperature more rapidly, reducing the temperature difference with the phase change material and thus reducing the driving force for heat transfer. This regulation method responds quickly, taking effect within minutes.

[0052] Preferably, the increase in the thickness of the insulation layer adopts a modular design. The basic insulation layer is 50 mm thick polyurethane foam. When it is necessary to further reduce the heat release rate, a 25 mm thick insulation module can be superimposed on the outside. With each additional layer of module, the thermal resistance increases by about 0.5 square meters Kelvin per watt, which reduces the heat loss accordingly. In this way, gradient regulation of the heat release rate can be achieved. The process of measuring the release rate after adjustment requires waiting for the system to reach a new steady state. Usually 30 minutes after the adjustment measures are implemented, the temperature field is redistributed and stabilized. At this time, the actual heat release power is calculated by the flow rate of the heating medium and the inlet and outlet temperature difference. For example, when the flow rate is 0.003 cubic meters per second and the temperature difference is 8°C, the heat release power is about 100 kilowatts. This optimized data becomes the benchmark for subsequent control, achieving accurate matching of heat supply and demand.

[0053] Step S105 , based on the optimized heat release rate data, the constant temperature maintenance time of the heat storage material during the solid-liquid transition process is monitored in real time to determine the stability index of the temperature platform.

[0054] The temperature sampling frequency is set based on the optimized heat release rate data. A high-precision temperature sensor array is arranged within the phase change material container. Temperature data of the material during the solid-liquid transition process is collected at preset time intervals. When the temperature fluctuations of multiple consecutive sampling points are within the preset threshold range, it is identified as the starting point of the temperature platform interval. Timing begins at the starting point of the temperature platform interval and continuously monitors the material temperature until the temperature begins to rise or fall significantly beyond the preset threshold range. The time difference between this moment and the starting point is recorded as the constant temperature maintenance time. At the same time, the sum of the squares of the differences between all temperature data and the average temperature during this period is calculated and divided by the number of data points, and the square root is taken to obtain the temperature standard deviation. Based on the obtained constant temperature maintenance time and temperature standard deviation, the time proportion is calculated by dividing the constant temperature maintenance time by the total duration from the start to the end of the phase change. The stability index of the temperature platform is determined by multiplying the time proportion by a preset weight coefficient and adding the inverse of the temperature standard deviation multiplied by another preset weight coefficient.

[0055] In one possible implementation, there's a direct correlation between optimized heat release rate data and the temperature sampling frequency. When the release rate is high, temperature fluctuations are relatively dramatic, necessitating a higher sampling frequency to capture subtle changes. For example, when the release rate reaches 200 watts per second, the sampling interval is set to 2 seconds; when the release rate drops to 50 watts per second, the sampling interval can be extended to 10 seconds. This dynamic adjustment ensures data accuracy while avoiding resource waste caused by oversampling.

[0056] It's important to note that the formation of a temperature plateau arises from the physical properties of phase change materials. During the phase change process, the material maintains a relatively constant temperature as it absorbs or releases latent heat, forming a gentle temperature plateau. The key to identifying this plateau lies in setting a reasonable temperature fluctuation threshold, typically ±0.5°C. When the temperature fluctuations at five or more consecutive sampling points fall within this range, the temperature plateau is identified.

[0057] Specifically, accurate measurement of the constant temperature maintenance time relies on precise determination of the start and end points of the temperature platform. Once the starting point is identified, the system begins timing and continuously monitors temperature changes. When the temperature value at a sampling point exceeds a preset threshold of 0.5°C compared to the previous sampling point, and the temperature at three subsequent sampling points shows the same trend, the platform interval is determined to have ended. This multi-point confirmation mechanism avoids misjudgments caused by occasional interference.

[0058] In one embodiment, the temperature standard deviation calculation process reflects the degree of data dispersion. First, the arithmetic mean of all temperature data within the platform interval is calculated. For example, in a monitoring session, 100 data points are obtained, and the average temperature is 35.2°C. Then, the difference between each data point and the mean is calculated. The squares of the differences are summed and divided by the total number of data points. Finally, the square root is taken to obtain the standard deviation. The smaller the standard deviation, the more stable the temperature control.

[0059] Preferably, the comprehensive calculation of the stability index adopts a weighted summation method. The time proportion reflects the ability of the phase change material to maintain a constant temperature, and the weight coefficient is usually set to 0.6. The reciprocal of the temperature standard deviation reflects the accuracy of temperature control, and the weight coefficient is set to 0.4. For example, the constant temperature maintenance time of a certain material is 45 minutes, the total phase change time is 60 minutes, and the time proportion is 0.75; the temperature standard deviation is 0.2°C, and its reciprocal is 5. The stability index is calculated as 0.75×0.6+5×0.4×0.1=0.65. Here, the reciprocal of the temperature standard deviation is multiplied by 0.1 to normalize it to the same order of magnitude as the time proportion. This multi-dimensional evaluation method can comprehensively reflect the temperature control performance of the phase change material. The higher the stability index, the better the temperature control of the material during the phase change process, and it is more suitable for data center cooling applications with strict requirements on temperature stability.

[0060] In step S106, if the stability index is lower than the preset threshold, the phase change material composition ratio parameters of the heat storage material are adjusted to enhance the adaptability of the constant temperature maintenance time, and the adjusted temperature control data is obtained. The overall efficiency of heat storage and release in the thermal storage system is analyzed to obtain the continuity parameter of the waste heat recovery.

[0061] For the obtained temperature platform stability index, if the index is lower than the preset threshold, then according to the mass ratio of different melting point components in the existing phase change material, the phase change temperature range is widened by increasing the proportion of high melting point components. After adjusting the ratio, the phase change starting temperature, completion temperature and constant temperature maintenance time of the material are re-measured to obtain the adjusted temperature control data. Based on the constant temperature maintenance time in the adjusted temperature control data, the total heat generated by the data center during this period, the heat absorbed by the phase change material and the heat released to the user end are calculated. The heat conversion efficiency is obtained by dividing the released heat by the absorbed heat as the overall efficiency value. Based on the calculated overall efficiency value, the continuity parameter of the waste heat recovery is obtained by counting the cumulative time that the efficiency value of the phase change material remains above the preset percentage of the initial value during multiple consecutive heat absorption and heat release processes, and dividing it by the total operating time.

[0062] In one possible implementation, the adjustment of the components of the phase change material involves the precise ratio of a variety of substances with different melting points. Common combinations include a mixed system of paraffin and fatty acids, in which the low-melting-point paraffin has a melting point of about 25°C and the high-melting-point paraffin can reach 60°C. When the original ratio is 7:3, the phase change temperature of the material is concentrated in a narrow range of 30-35°C. By adjusting the ratio to 5:5, the phase change temperature range is widened to 28-45°C, significantly extending the constant temperature maintenance time. This widening effect stems from the characteristics of the step-by-step phase change of different components. The low-melting-point component melts first and absorbs heat, and the high-melting-point component melts later, forming a stepped temperature platform.

[0063] It's important to note that obtaining temperature control data requires a complete thermal cycle test after adjusting the mix. During the test, a precision temperature recorder collects temperature data every 30 seconds, while also recording heating power and time. For example, one test showed that the adjusted material maintained a constant temperature near 35°C for 80 minutes, 35 minutes longer than the original mix. This data provides the basis for subsequent performance calculations.

[0064] Specifically, the calculation of heat conversion efficiency is based on the principle of conservation of energy. Within 80 minutes of constant temperature maintenance in the data center, assuming that the server continuously outputs a thermal power of 10 kilowatts, the total heat generated is 48,000 kilojoules. The heat absorbed by the phase change material during this period is calculated by multiplying the mass and the latent heat value. For example, 200 kilograms of material and a latent heat value of 180 kilojoules per kilogram absorb 36,000 kilojoules of heat. The heat released to the user end during the same period is measured by the flow rate and temperature difference of the heating pipe. For example, if the flow rate is 0.005 cubic meters per second and the temperature difference is 10°C, the heat released in 80 minutes is 33,600 kilojoules. The heat conversion efficiency is 33,600 divided by 36,000, which is approximately 0.93.

[0065] In one embodiment, the statistics of the continuity parameter require support from long-term monitoring data. The performance value of each heat absorption and heat release cycle is recorded. When the performance value falls below 90% of the initial value, the cycle is considered to have failed to meet the standard. For example, within a 1000-hour operation period, a total of 500 cycles are completed, of which the performance value remains above 0.837 for 450 cycles, which is 90% of the initial value of 0.93. The cumulative time to meet the standard is 900 hours, and the continuity parameter is calculated as 900 divided by 1000, which is 0.9.

[0066] Ideally, this multi-dimensional evaluation system comprehensively reflects system performance. While the addition of high-melting-point components may slightly reduce the heat storage density per unit mass, it generally improves the reliability of waste heat recovery by extending the hold-temperature period and enhancing system stability. A continuity parameter of 0.9 means the system maintains good heat recovery 90% of the time, which is crucial for applications requiring continuous and stable heating.

[0067] In step S107, based on real-time feedback from user-side heat demand, the results of the updated heat storage supply reliability strategy are used to determine whether the heat storage supply meets the continuity requirements. If the heat supply continuity requirements are not met, the heat output mode is adjusted and the heat storage supply reliability data of the phase change thermal storage data center is obtained.

[0068] Based on the waste heat recovery continuity parameter, the real-time heat demand data and demand change frequency data transmitted by the user-side heat load monitoring device are received. The ratio of the continuity parameter to the demand change frequency is calculated as the supply and demand matching degree. When the supply and demand matching degree is multiplied by the preset weight coefficient, the updated heat storage supply reliability value is obtained. Based on the updated heat storage supply reliability value, if the value is lower than the preset continuity threshold, it is judged that the heat supply continuity requirement is not met. At this time, the time interval from the current heat storage material receiving heat to the start of heat release is extracted from the monitoring data as the temperature response time, and the heat release rate data is obtained. Based on the temperature response time and heat release rate, the heat exchange intensity is changed by increasing the number of heat exchange pipes or adjusting the circulating medium flow rate to match the temperature response time with the user demand response time. At the same time, the heat release ratio in different time periods is adjusted to obtain the heat storage supply reliability data of the phase change thermal storage data center.

[0069] In one possible implementation, matching the waste heat recovery continuity parameter with the frequency of user-side demand fluctuations is key to stable system operation. As mentioned previously, a continuity parameter of 0.9 indicates that the system maintains good performance 90% of the time. The frequency of user-side demand fluctuations is determined through statistical analysis. For example, the heating demand of an office building exhibits a clear periodicity during weekdays: it begins to rise at 7:00 AM, peaks at 9:00 AM, decreases slightly at noon, rises again at 2:00 PM, and gradually decreases after 6:00 PM. This frequency of fluctuation can be quantified as 0.2 major fluctuations per hour.

[0070] It should be noted that the calculation of the supply-demand matching degree reflects the degree of adaptability of the system's responsiveness to changes in demand. When the continuity parameter 0.9 is divided by the demand change frequency 0.2, the basic matching degree is 4.5. This value indicates that the system's stable operation time far exceeds the demand change cycle and has good adaptability. The setting of the weight coefficient takes into account the importance of different application scenarios. The weight coefficient of key facilities such as hospitals can be set to 0.95, and that of ordinary office buildings to 0.85. The thermal storage supply reliability value is 4.5 multiplied by 0.85, which is approximately 3.8.

[0071] Specifically, extracting the temperature response time involves precise time recording. Timing begins when a sudden increase in server load in the data center causes an increase in heat output, and ends when the phase-change material temperature begins to noticeably change. For example, one monitoring session showed a sudden increase in server load at 2:30:00 PM, and the phase-change material temperature began to rise at 2:32:30 PM, resulting in a temperature response time of 150 seconds. This time reflects the transfer process from heat generation to absorption by the material.

[0072] In one embodiment, heat exchange intensity can be adjusted in a variety of ways. Increasing the number of heat exchange pipes is the most direct method. The original design may have only four 50mm diameter pipes passing through the phase change material container. By increasing the number to six, the heat exchange area increases by 50%. Adjusting the circulating medium flow rate is more flexible. Adjusting the flow rate from 100 liters per minute to 150 liters per minute using a variable frequency pump can increase the heat transfer coefficient by approximately 30%. This adjustment shortens the temperature response time from 150 seconds to 105 seconds, better matching the user's 100-second demand response requirement.

[0073] The heat release ratio is optimally adjusted over time based on user demand curves. During peak demand periods, such as 9:00 AM to 11:00 AM, the release ratio is set at 40% of the total stored heat; during the midday trough, it is reduced to 15%; during the afternoon peak, it is increased to 30%; and at night, a baseline release ratio of 15% is maintained. This refined time-based control is achieved by adjusting the circulation pump speed and valve opening. Reliability data shows that the adjusted system achieved a heat supply satisfaction rate of 98.5% during a 72-hour test, significantly improving the reliability and cost-effectiveness of waste heat utilization in the data center.

[0074] The present invention provides a waste heat recovery system for a phase change thermal storage data center, which mainly includes:

[0075] The load monitoring and correlation analysis module is used to monitor the load fluctuation characteristics of the phase change thermal storage data center in real time, obtain the heat output data of the phase change thermal storage data center computing server under the operating state, and perform correlation analysis based on the preset load variation model to obtain the correlation mapping relationship between load fluctuation and heat generation;

[0076] A dynamic heat storage capacity adjustment module is used to dynamically adjust the heat storage capacity of the phase change thermal storage data center thermal storage system based on the correlation mapping relationship between load fluctuations and heat generation. By analyzing the phase change temperature and latent heat value of the material, the heat absorption rate range of the material under different load conditions is determined;

[0077] The heat release prediction model construction module is used to build a heat release rate prediction model based on the heat absorption rate range and the nonlinear change law of the heat storage material. The dynamic adjustment parameters of heat release are obtained from the prediction model as the model output to determine whether the heat release rate meets the user's thermal requirements.

[0078] The heat release rate optimization module is used to increase the matching degree of heat absorption and release in the phase change thermal storage data center thermal storage system if the heat release rate is lower than the user-end heat demand, and obtain optimized heat release rate data. If the heat release rate is higher than the user-end heat demand, the phase change rate of the thermal storage material is reduced to avoid excessive heat release and energy waste;

[0079] The constant temperature stability monitoring module is used to monitor the constant temperature maintenance time of the heat storage material during the solid-liquid transition process in real time based on the optimized heat release rate data, and determine the stability index of the temperature platform;

[0080] The phase change material ratio adjustment module is used to adjust the phase change material composition ratio parameters of the heat storage material to enhance the adaptability of the constant temperature maintenance time if the stability index is lower than the preset threshold, obtain the adjusted temperature control data, analyze the overall efficiency of heat storage and release in the thermal storage system, and obtain the continuity parameter of the waste heat recovery;

[0081] The heat storage supply continuity judgment module is used to combine the real-time feedback of user-side heat demand and determine whether the heat storage supply meets the continuity requirements from the results of the updated heat storage supply reliability strategy. If the heat supply continuity requirements are not met, the heat output mode is adjusted to obtain the heat storage supply reliability data of the phase change thermal storage data center.

[0082] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the concept of this application. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A waste heat recovery method for a phase change thermal storage data center, characterized in that: The method comprises: By monitoring the load fluctuation characteristics of the phase-change thermal storage data center in real time, the heat output data of the phase-change thermal storage data center computing server under operating conditions is obtained. Combined with the preset load variation model, correlation analysis is performed to obtain the correlation mapping relationship between load fluctuation and heat generation; Dynamically adjust the heat storage capacity of the phase change thermal storage data center thermal storage system based on the correlation mapping relationship between load fluctuations and heat generation. By analyzing the phase change temperature and latent heat value of the material, the heat absorption rate range of the material under different load conditions is determined. Based on the range of heat absorption rates and the nonlinear variation of heat storage materials, a heat release rate prediction model is constructed. Dynamic adjustment parameters for heat release are obtained from the prediction model as model output to determine whether the heat release rate meets the user's thermal requirements. If the heat release rate is lower than the user-side heat demand, the matching degree of heat absorption and release in the phase change thermal storage data center heat storage system is increased to obtain optimized heat release rate data; Based on the optimized heat release rate data, the constant temperature maintenance time of the heat storage material during the solid-liquid transition process is monitored in real time to determine the stability index of the temperature platform; If the stability index is lower than the preset threshold, the adaptability of the constant temperature maintenance time is enhanced, the adjusted temperature control data is obtained, the overall efficiency of heat storage and release in the thermal storage system is analyzed, and the continuity parameters of waste heat recovery are obtained; Combined with real-time feedback on user-side heat demand, it is determined whether the heat storage supply meets the continuity requirements. If the heat supply continuity requirements are not met, the heat output mode is adjusted to obtain the heat storage supply reliability data of the phase change thermal storage data center.

2. The waste heat recovery method for a phase change thermal storage data center according to claim 1, characterized in that: The method obtains heat output data of the computing server in the phase change thermal storage data center under the operation state by real-time monitoring of the load fluctuation characteristics of the phase change thermal storage data center, performs correlation analysis based on a preset load variation model, and obtains a correlation mapping relationship between load fluctuation and heat generation, including: Through a temperature sensor array and a power monitoring device, the CPU utilization rate, memory occupancy rate and temperature data of the corresponding position of the computing server are collected, the ambient temperature data of the phase change thermal storage data center are obtained, and a time series data set of load fluctuation and heat output is generated; based on the time series data set, a bidirectional time series of load and heat is constructed, and the ratio of the load difference and the time interval at adjacent moments in the time series is calculated to obtain the load change rate, and the ratio of the temperature difference and the time interval at adjacent moments is calculated to obtain the heat change rate; based on the load change rate and the heat change rate, the Pearson correlation coefficient is used to calculate the correlation under different time delays to determine the time lag parameter from load change to heat response; based on the time lag parameter, a multivariate linear regression method is used, with the CPU utilization rate and the memory occupancy rate as independent variables and the difference between the temperature data and the ambient temperature data as the dependent variable, to determine the regression coefficient and generate a correlation mapping relationship between the load fluctuation and heat generation.

3. The waste heat recovery method for a phase change thermal storage data center according to claim 1, characterized in that: The method dynamically adjusts the heat storage capacity of the phase change thermal storage data center thermal storage system based on the correlation mapping relationship between load fluctuation and heat generation, and determines the heat absorption rate range of the material under different load conditions by analyzing the phase change temperature and latent heat value of the material, including: The phase change starting temperature, phase change completion temperature and latent heat value per unit mass of the heat storage material are read from the phase change material characteristic table to generate a solid-liquid transition characteristic data set; based on the solid-liquid transition characteristic data set, the heat absorption capacity per unit mass of the heat storage material is calculated; based on the correlation mapping relationship between load fluctuation and heat generation, the current load level is compared with the preset load threshold, and an adaptive phase change material configuration scheme is selected; based on the phase change material configuration scheme, the duration of the phase change process from solid to liquid of the heat storage material is measured, and the ratio of the latent heat value to the duration of the phase change process is calculated to generate the heat absorption rate range.

4. The waste heat recovery method for a phase change thermal storage data center according to claim 1, characterized in that: The heat release rate prediction model is constructed based on the heat absorption rate range and the nonlinear change law of the heat storage material, including: The temperature and time curve data of the heat storage material during the liquid-solid transition process are collected, the ratio of the temperature difference at adjacent time points to the time interval is calculated, the instantaneous slope is generated, the nonlinear law of heat release of the heat storage material is identified, and the actual measured value of the heat release rate is obtained; based on the actual measured value of the heat release rate and the solid-liquid phase ratio, the support vector regression method is used, and the temperature, phase ratio and temperature difference of the heat storage material and the ambient temperature are used as input parameters to generate the prediction model, and the output of the prediction model is the dynamic adjustment parameter of heat release.

5. The waste heat recovery method for a phase change thermal storage data center according to claim 1, characterized in that: If the heat release rate is lower than the user-end heat demand, the matching degree of heat absorption and release in the phase change thermal storage data center heat storage system is increased to obtain optimized heat release rate data, including: According to the matching result of the dynamic adjustment parameters of heat release and the heat demand of the user end, the contact area between the heat storage material and the heating pipe or the configuration of the heat-conducting fins is adjusted, and the ratio of the heat transfer amount per unit time before and after the adjustment is calculated to generate a matching improvement factor; based on the matching improvement factor, the optimized heat release rate data is calculated.

6. The waste heat recovery method for a phase change thermal storage data center according to claim 1, characterized in that: The method of monitoring the constant temperature maintenance time of the heat storage material during the solid-liquid transition process in real time based on the optimized heat release rate data and determining the stability index of the temperature platform includes: According to the optimized heat release rate data, the temperature sampling frequency is set, the temperature data of the heat storage material during the solid-liquid transition process is collected, and the temperature platform interval is identified; the duration of the temperature platform interval is calculated to generate the constant temperature maintenance time; according to the standard deviation of the constant temperature maintenance time and temperature data, the time proportion and the inverse of the standard deviation are calculated to generate the stability index of the temperature platform.

7. The waste heat recovery method for a phase change thermal storage data center according to claim 1, characterized in that: If the stability index is lower than the preset threshold, the adaptability of the constant temperature maintenance time is enhanced, the adjusted temperature control data is obtained, the overall efficiency of heat storage and release in the thermal storage system is analyzed, and the continuity parameters of waste heat recovery are obtained, including: According to the stability index of the temperature platform, the mass ratio of the high-melting-point component in the heat storage material is adjusted, and the adjusted phase change starting temperature, completion temperature and constant temperature maintenance time are measured to generate adjusted temperature control data; based on the adjusted temperature control data, the heat conversion efficiency of the heat storage system during the heat absorption and heat release process is calculated to generate an overall efficiency value; based on the overall efficiency value, the ratio of the cumulative time when the efficiency value is above a preset percentage to the total operating time is calculated to generate the continuity parameter of the waste heat recovery.

8. The waste heat recovery method for a phase change thermal storage data center according to claim 1, characterized in that: The method combines the real-time feedback of user-side heat demand to determine whether the heat storage supply meets the continuity requirement. If the heat supply continuity requirement is not met, the heat output mode is adjusted to obtain the heat storage supply reliability data of the phase change thermal storage data center, including: Based on the continuity parameters of the waste heat recovery and the real-time data of user-end heat demand, the supply and demand matching degree is calculated to generate an updated thermal storage supply reliability value; based on the updated thermal storage supply reliability value, the temperature response time and heat release rate data of the heat storage material from receiving heat to releasing heat are extracted; based on the temperature response time and the heat release rate data, the number of heat exchange pipes or the circulating medium flow rate is adjusted to generate the thermal storage supply reliability data of the phase change thermal storage data center.

9. A waste heat recovery system for a phase change thermal storage data center, characterized in that: The system comprises: The load monitoring and correlation analysis module is used to monitor the load fluctuation characteristics of the phase change thermal storage data center in real time, obtain the heat output data of the phase change thermal storage data center computing server under the operating state, and perform correlation analysis based on the preset load variation model to obtain the correlation mapping relationship between load fluctuation and heat generation; A dynamic heat storage capacity adjustment module is used to dynamically adjust the heat storage capacity of the phase change thermal storage data center thermal storage system based on the correlation mapping relationship between load fluctuations and heat generation. By analyzing the phase change temperature and latent heat value of the material, the heat absorption rate range of the material under different load conditions is determined; The heat release prediction model construction module is used to build a heat release rate prediction model based on the heat absorption rate range and the nonlinear change law of the heat storage material. The dynamic adjustment parameters of heat release are obtained from the prediction model as the model output to determine whether the heat release rate meets the user's thermal requirements. The heat release rate optimization module is used to increase the matching degree of heat absorption and release in the phase change thermal storage data center thermal storage system if the heat release rate is lower than the user-end heat demand, and obtain the optimized heat release rate data; The constant temperature stability monitoring module is used to monitor the constant temperature maintenance time of the heat storage material during the solid-liquid transition process in real time based on the optimized heat release rate data, and determine the stability index of the temperature platform; The phase change material ratio adjustment module is used to enhance the adaptability of the constant temperature maintenance time if the stability index is lower than the preset threshold, obtain the adjusted temperature control data, analyze the overall efficiency of heat storage and release in the thermal storage system, and obtain the continuity parameter of waste heat recovery; The heat storage supply continuity judgment module is used to combine the real-time feedback of user-side heat demand to determine whether the heat storage supply meets the continuity requirements. If the heat supply continuity requirements are not met, the heat output mode is adjusted to obtain the heat storage supply reliability data of the phase change thermal storage data center.

Images