Computer Room Air Research Articles

A determination method on the supply airflow rate with the impact of leakage and bypass airflow of computer room air conditioning

Raised-floor air supply (RFAS) system is widely used in data centers (DCs) for its superior thermal environment and air distribution. However, the occurrence of leakage and bypass airflow (LBA) results in a poor thermal environment and ineffective air cooling in a computer room. Most existing DCs either failed to consider the above airflow effects incurring hot spots, or considered these effects using over-cooling which resulted in high energy consumption. In this paper, we propose a determination method of supply airflow rate for computer room air conditioning (CRAC). With the two-dimensional network diagram of airflow and heat transfer in a DC, the relationship of airflow rates between server demand and CRAC supply is deduced based on the conservation of energy. Then a CFD model was built according to a practical DC and its accuracy was verified by field tests, and the room thermal environment is simulated under three different supply airflow rates obtained by traditional method, proposed method and empirical method. The thermal environment performance is evaluated and compared using SHI. Besides, the energy consumption of these three cases is also investigated. Compared with traditional method, the proportion of racks with SHI>0.2 is reduced from 39.5% to 24.2% and 17.2% for proposed method and empirical method, respectively. The corresponding incremental energy consumption of the empirical method is 4.8 times higher than that of the proposed method. The comparison results show that the proposed determination method of supply airflow rate is a proper option that can reduce hot spots with an energy saving of 406.24 kW of the cooling system.

Design and Thermal Environment Analysis of a Decentralized Cooling System with Surface-Mount Heat Pipe Exchangers on Servers in Data Centers

This paper proposes a decentralized cooling system combined with a traditional computer room air conditioning unit and server-level heat pipe exchangers for thermal environment optimization in a data center. Two cooling strategies, with heat exchangers installed above and below the servers respectively, are proposed and compared with the original CRACs system in terms of thermal environment. The simulation results of the original data center model are in good agreement with the on-site measurement results, and thus its reliability can be validated. The results show that a decentralized cooling system can effectively improve the thermal environment in data centers. To obtain the highest cooling efficiency, altogether 18 cases, where heat pipe exchangers were installed at different locations and heights, are analyzed and compared. The results show that the thermal environment is optimal when heat pipe exchangers are installed 0.01 m below each server. The local hotspot temperature is reduced by 6.8 °C, and the temperature distribution of the rack is the most uniform, which can effectively reduce the heat accumulation in data centers.

Experimental analysis of airflow uniformity and energy consumption in data centers

Temperature and airflow uniformity are experimentally investigated in this research study along with energy consumption analysis on the laboratory container size small data center. A raised floor plenum (RFP) data center system of size 6 (L) × 3.1 (W) × 2.3 (H) m with one Computer Room Air Handler (CRAH), two rows and each row contains five racks, five 8U simulator cabinets in each rack was constructed to test partial and closed cold aisle containment (CAC). Seven different cases were investigated with variables of perforated tiles (32%, 50 % & 70%), CRAH fan speed (60 Hz, 42 Hz), chiller return water temperature (CRWT) (15 °C & 17 °C) and heating load (full 50 kW, 75% load 37.5 kW). These cases were further analyzed using Rack Cooling Index (RCI) and Supply Heat Index (SHI) along with power consumption scenarios in terms of Coefficient of Performance (COP) and Power Usage Effectiveness (PUE). The uniqueness of this study is experimental investigations of system COP and PUE analysis for a wide range of important variables for partial and closed CAC. It was found that 50% porous tile provided the best airflow and temperature uniformity for the case of partial containment with full and 25% reduced heat load along with full and 30% reduced CRAH speed respectively. Slight recirculation was observed during the reduced CRAH fan speed case at the corner edge of data racks. However, due to additional supply airflow usage, these recirculation’s were within the acceptable SHI. Reduction of CRWT reduces the total Heating Ventilation and Air Conditioning (HVAC) power consumption and improves the overall system COP.

Experimental and Numerical Analysis of Data Center Pressure and Flow Fields Induced by Backward and Forward CRAH Technology

Abstract An increasingly common power saving practice in data center thermal management is to swap out air cooling unit blower fans with electronically commutated plug fans, Although, both are centrifugal blowers. The blade design changes: forward versus backward curved with peak static efficiencies of 60% and 75%, respectively, which results in operation power savings. The side effects of which are not fully understood. Therefore, it has become necessary to develop an overall understanding of backward curved blowers and compare the resulting flow, pressure, and temperature fields with forwarding curved ones in which the induced fields are characterized, compared, and visualized in a reference data center which may aid data center planning and operation when making the decisions of which computer room air handler (CRAH) technology to be used. In this study, experimental and numerical characterization of backward curved blowers is introduced. Then, a physics-based computational fluid dynamics model is built using the 6sigmaroom tool to predict/simulate the measured fields. Five different scenarios were applied at the room level for the experimental characterization of the cooling units and another two scenarios were applied for comparison and illustration of the interaction between different CRAH technologies. Four scenarios were used to characterize a CRAH with backward curved blowers, during which a CRAH with forwarding curved was powered off. An alternate arrangement was examined to quantify the effect of possible flow constraints on the backward curved blower's performance. Then parametric and sensitivity of the baseline modeling are investigated and considered. Different operating conditions are applied at the room level for experimental characterization, comparison, and illustration of the interaction between different CRAH technologies. The measured data is plotted and compared with the computational fluid dynamics (CFD) model assessment to visualize the fields of interest. The results show that the fields are highly dependent on CRAH technology. The tile to CRAH airflow ratios for the flow constraints of scenarios 1, 2, 3, and 4 are 85.5%, 83.9%, 61%, and 59%, respectively. The corresponding leakage ratios are 14.5%, 16%, 38.9%, and 41%, respectively. Furthermore, the validated CFD model was used to investigate and compare the airflow pattern and plenum pressure distribution. Lastly, it is notable that a potential side effect of backward curved technology is the creation of an airflow dead zone.

Comprehensive energy and economic assessment of CRAH bypass method in air-cooled data centers

The computer room air handling (CRAH) bypass (BP) method is a method to reduce the fan energy consumption in data centers. To provide a fraction of the server airflow, CRAH BP fans typically operate across an order of magnitude lower pressure drop across floor tiles compared to that across CRAH units. However, at a higher fraction of BP airflow, CRAH units are subject to low-temperature operation, which poses an optimization problem. Earlier studies indicated the potential to reduce the total cooling energy consumption in air-cooled data centers by using the CRAH bypass method. This study provides the most comprehensive assessment of the CRAH BP method in the literature for air-cooled data centers based on detailed thermodynamic modeling using hour-by-hour annual energy simulations. The parametric modeling and optimization results in this study address energy and economics of induced and forced BP methods in open and enclosed-aisle data centers considering both economized and non-economized operations. Despite the significant energy-saving potential of the CRAH BP method, economics plays a pivotal role in the feasibility of such an investment. The results and following discussions provide design considerations for the successful application of the CRAH BP method in air-cooled data centers.

Adaptive control algorithm with a retraining technique to predict the optimal amount of chilled water in a data center cooling system

We developed control algorithms based on one of three artificial-intelligence-based retraining techniques (sliding window, vector adaptation, and vector augmentation) to provide the optimal indoor temperature and save on the energy expenditure for cooling in data centers. The artificial neural network prediction model predicts the computer room air handler supply air temperature of a central chilled water system and is added to the control algorithm. The proposed algorithm can determine the optimal chilled water flow rate required to cool the server to the set temperature by using the predicted computer room air handler supply air temperature. We developed a control algorithm embedded in an artificial neural network predictive model that includes three retraining techniques. Afterward, we compared the control performance and verified its adaptability by using computer simulation. When using the algorithm with sliding window control, the root mean-squared error between the set temperature and the control temperature was 0.08 °C, the maximum error was 0.81 °C, and the cooling load was 21,026.27 kWh. The accuracy, stability, and energy-saving ability of the sliding window control algorithm were higher those of the other two algorithms, and its superior adaptability and scalability under changing environmental conditions were demonstrated.

In-situ sensor correction method for data center cooling systems using Bayesian Inference coupling with autoencoder

• In-situ sensor correction method based on Bayesian inference and autoencoder. • Impacts of the system model accuracy on the correction performance. • A series of case studies were designed to fully demonstrate the performance. • Result shows that deviation rates of sensor measurements are decreased to 7.67%. The quality of measurements from working sensors has a considerable influence on the system operation and energy usage in data center cooling systems. However, due to the malfunction, aging, and installation positions, it is very common phenomenon that the measurements are far away from their true values. Virtual sensor correction is a promising solution to correct erroneous measurements. This study proposed a novel sensor correction method for data center cooling systems using the Bayesian Inference coupling with autoencoder (AE) to eliminate the sensor errors. The correction performance of the proposed method was comprehensively investigated under a series of single/multiple sensor error scenarios in a typical computer room air handler (CRAH) unit. Furthermore, the impact of the system model accuracy on the correction performance was also quantified. The results show the excellent correction accuracies of the proposed method in the data center cooling systems: following correction, the sensor deviation rates are decreased to 7.67% and 3.00% for the single and multiple simultaneous sensor error scenarios, respectively. Additionally, the sensor correction error increases to 18% when the deviation rate of the heat transfer coefficient is set to 10%.

Numerical analysis of layout of air conditioning in data center considering seasonal factors

In this paper, a better layout of the computer room air conditioning (CRAC) units in the data center (DC) is obtained by comparatively analyzing the indices about velocity and temperature in the room. First, Computational Fluid Dynamics (CFD) is utilized to model the computer room of DC, and two layout schemes of air conditioners are presented. Then, to determine the better layout, the distribution of average air velocity and temperature field of the two layout schemes are compared. Besides, a uniformity index is proposed to conduct the quantitative analysis of the temperature uniformity at the inlet of racks under two layout modes. Considering the influences of seasonal factors on the cooling efficiency, the temperature distribution analysis with different outdoor temperature under two layouts is carried out contrastively.

Synergy optimization analysis of heat transfer performance and energy consumption in heat transfer process and its application in data centers

Most industrial heat exchange processes are carried out between two fluids. Usually the thermal fluids are driven by pumps or fans. The power consumption of pumps and fans is quite high in the transport systems, especially in data centers. Nearly 60% of the computer room air conditioning (CRAC) system’s energy consumption is consumed by pumps and fans in winter. A lot of literatures discussed the heat transfer enhancement in heat transfer process but few focused on the energy consumption. In this paper, a new synergy optimization method is proposed to save energy in heat transfer process. The heat transfer constraint equations of the heat transfer process are established. Taking the energy efficiency as the objective function, the synergistic relationship between the power consumption and temperature difference is derived analytically by using variation principle. A synergy operation factor is defined to guide the practical operation optimization of heat transfer process. The closer the synergy operation factor to 1, the higher the system energy efficiency is. An experiment of a separated heat pipe system is carried out to verify the accuracy of the synergy optimization analysis. The synergy optimization analysis is implemented in a real data center, and the cooling system can save energy by 30.9% under test conditions.

Tunicate swarm algorithm-trained multi-layered perceptron for data centre energy demand forecasting and relative percentage contribution analysis of input parameters

PurposeEnergy management is critical to data centres (DCs) majorly because they are high energy-consuming facilities and demand for their services continue to rise due to rapidly increasing global demand for cloud services and other technological services. This projected sectoral growth is expected to translate into increased energy demand from the sector, which is already considered a major energy consumer unless innovative steps are used to drive effective energy management systems. The purpose of this study is to provide insights into the expected energy demand of the DC and the impact each measured parameter has on the building's energy demand profile. This serves as a basis for the design of an effective energy management system.Design/methodology/approachThis study proposes novel tunicate swarm algorithm (TSA) for training an artificial neural network model used for predicting the energy demand of a DC. The objective is to find the optimal weights and biases of the model while avoiding commonly faced challenges when using the backpropagation algorithm. The model implementation is based on historical energy consumption data of an anonymous DC operator in Cape Town, South Africa. The data set provided consists of variables such as ambient temperature, ambient relative humidity, chiller output temperature and computer room air conditioning air supply temperature, which serve as inputs to the neural network that is designed to predict the DC’s hourly energy consumption for July 2020. Upon preprocessing of the data set, total sample number for each represented variable was 464. The 80:20 splitting ratio was used to divide the data set into training and testing set respectively, making 452 samples for the training set and 112 samples for the testing set. A weights-based approach has also been used to analyze the relative impact of the model’s input parameters on the DC’s energy demand pattern.FindingsThe performance of the proposed model has been compared with those of neural network models trained using state of the art algorithms such as moth flame optimization, whale optimization algorithm and ant lion optimizer. From analysis, it was found that the proposed TSA outperformed the other methods in training the model based on their mean squared error, root mean squared error, mean absolute error, mean absolute percentage error and prediction accuracy. Analyzing the relative percentage contribution of the model's input parameters based on the weights of the neural network also shows that the ambient temperature of the DC has the highest impact on the building’s energy demand pattern.Research limitations/implicationsThe proposed novel model can be applied to solving other complex engineering problems such as regression and classification. The methodology for optimizing the multi-layered perceptron neural network can also be further applied to other forms of neural networks for improved performance.Practical implicationsBased on the forecasted energy demand of the DC and an understanding of how the input parameters impact the building's energy demand pattern, neural networks can be deployed to optimize the cooling systems of the DC for reduced energy cost.Originality/valueThe use of TSA for optimizing the weights and biases of a neural network is a novel study. The application context of this study which is DCs is quite untapped in the literature, leaving many gaps for further research. The proposed prediction model can be further applied to other regression tasks and classification tasks. Another contribution of this study is the analysis of the neural network's input parameters, which provides insight into the level to which each parameter influences the DC’s energy demand profile.

Coupled Calculations of Data Center Cooling and Power Distribution Systems

Abstract Physics-based modeling aids in designing efficient data center power and cooling systems. These systems have traditionally been modeled independently under the assumption that the inherent coupling of effects between the systems has negligible impact. This study tests the assumption through uncertainty quantification of models for a typical 300 kW data center supplied through either an alternating current (AC)-based or direct current (DC)-based power distribution system. A novel calculation scheme is introduced that couples the calculations of these two systems to estimate the resultant impact on predicted power usage effectiveness (PUE), computer room air conditioning (CRAC) return temperature, total system power requirement, and system power loss values. A two-sample z-test for comparing means is used to test for statistical significance with 95% confidence. The power distribution component efficiencies are calibrated to available published and experimental data. The predictions for a typical data center with an AC-based system suggest that the coupling of system calculations results in statistically significant differences for the cooling system PUE, the overall PUE, the CRAC return air temperature, and total electrical losses. However, none of the tested metrics are statistically significant for a DC-based system. The predictions also suggest that a DC-based system provides statistically significant lower overall PUE and electrical losses compared to the AC-based system, but only when coupled calculations are used. These results indicate that the coupled calculations impact predicted general energy efficiency metrics and enable statistically significant conclusions when comparing different data center cooling and power distribution strategies.

Evaluation of the Thermal Performance and Energy Efficiency of CRAC Equipment through Mathematical Modeling Using a New Index COP WEUED

As the world data traffic increasingly grows, the need for computer room air conditioning (CRAC)-type equipment grows proportionally. The air conditioning equipment is responsible for approximately 38% of the energy consumption of data centers. The energy efficiency of these pieces of equipment is compared according to the Energy Standard ASHRAE 90.1-2019, using the index Net Sensible Coefficient Of Performance (NetSCOP). This method benefits fixed-speed compressor equipment with a constant inlet temperature air-cooled condenser (35 °C). A new method, COP WEUED (COP–world energy usage effectiveness design), is proposed based on the IPLV (integrated part load value) methodology. The IPLV is an index focused on partial thermal loads and outdoor temperature data variation for air intake in the condenser. It is based on the average temperatures of the USA’s 29 major cities. The new method is based on the 29 largest cities worldwide and with data-center-specific indoor temperature conditions. For the same inverter compressor, efficiencies of 4.03 and 4.92 kW/kW were obtained, using ASHRAE 90.1-2019 and the proposed method, respectively. This difference of almost 20% between methods is justified because, during less than 5% of the annual hours, the inlet air temperature in the condenser is close to the NetSCOP indication.

Data centre day-ahead energy demand prediction and energy dispatch with solar PV integration

This paper presents a novel Marine Predators Algorithm for both training an Artificial Neural Network model used for predicting the energy demand and for solving a dynamic combined economic and emission dispatch of a data centre. The MPA is proposed for first finding the optimal weights and biases of the neural network based on a Mean Squared Error and Mean Absolute Error minimization objective function. Real life dataset obtained from an anonymous data centre operator in Cape Town, South Africa was used for the model implementation. The dataset was made up of a total of 564 samples and was split into training and testing set using an 80:20 ratio. The input variables contained in the dataset are the data centre’s ambient temperature, ambient relative humidity, chiller output temperature and Computer Room Air Conditioning supply temperature while the energy demand is the target variable. The optimal weights of the neural network model were analysed using a weights-based approach to determine the level of influence each input parameter of the model has on the data centre’s energy demand. Then based on the predicted energy demand of the data centre, a dynamic economic and emission dispatch problem is solved for the building while considering thermal and solar photovoltaic generations. A spinning reserve is also incorporated in the energy dispatch model to cater for any shortfall that may exist between the predicted and actual energy demand of the data centre due to possible inaccuracies in the energy demand prediction model. Results for the energy demand prediction task showed that the proposed method outperformed the state-of-the-art by producing the least Mean Squared Error, Root Mean Squared Error, Mean Absolute Error, Mean Absolute Percentage Error and highest prediction accuracy for the training and testing sets. Further analyses also highlighted that the data centre’s ambient temperature has the highest influence of about 37.63% on its energy demand pattern. For the energy dispatch task, the proposed method also identified solar photovoltaic as the preferred energy source over conventional thermal generators in fulfilling the objective function, depending on its availability. Overall, the findings presented in this study emphasize the robustness of the proposed method in solving the problems considered and its potential application towards solving even more complex problems.

Heat exchanger design based on earthen materials

We propose herein the design of a heat exchanger based on clay materials in order to reduce the energy consumed due to data center cooling. Our Thermal Energy Storage (TES) system is to be placed upstream of the CRAC (Computer Room Air Conditioner) unit. A direct contact heat exchanger is thereby built as a vertical packed bed, composed of earthen spheres whose imposed heating/cooling cycles comply with ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers) standards. A raw earthen material has been chosen owing to its hygrothermal characteristics, which are recognized as relevant for indoor thermal regulation, in addition to a low embodied energy required for manufacturing. Results show that the pore Reynolds number and sphere diameter exert significant effects on the energy buffering capacity. A simple relationship is thus proposed to link the mass-specific power of the energy storage system, the internal heat exchanger architecture and the pore Reynolds number for purposes of designing raw earth bricks at an elevated scale.

The impact of the operation mode of CRACs on the temperature distribution of hot and cold aisles for the data center room

Redundancy is generally considered in the computer room air conditioners (CRACs) of the data center room. The position of the CRAC in the on state and the operating air volume have a great influence on the temperature distribution. The temperature distribution of the data center room will affect the failure rate of the servers and the energy consumption of the air-conditioning system. Taking a typical data center room as an example, this paper uses numerical simulation to study the influence of the on/off state and air volume of CRACs on the temperature distribution of the hot and cold aisles and energy consumption. The calculation results show that we should preferably turn off the staggered CRACs on the opposite side and close to the middle position under the partial load. When the IT load is unevenly distributed, it is irrational to use the setting method of even distribution of the air volume of each CRAC. The air volume of each CRAC should be appropriately adjusted based on the distribution characteristic of IT load to reduce the energy consumption.

Prototype energy models for data centers

Data centers in the United States consume about twopercent of the nation’s electricity. Because heat gains from IT equipment drive cooling demand, data centers offer unique opportunities for energy savings. However, no prototype energy model for data centers is available in the suite of existing U.S. Department of Energy’s Commercial Prototype Building Models. This study presented the development of two new data center prototype models and their implementation in OpenStudio and EnergyPlus. The small-size data center model represents a computer room in a building served by computer room air conditioners (CRACs); while the large-sized model represents stand-alone data centers served by computer room air handlers (CRAHs) with a central chiller plant. For each data center model, two levels of IT equipment (ITE) load density were considered, to cover the wide range of IT power density of data centers: 40 and 100 W/ft2 (430 and 1076 W/m2) for the computer room, and 100 and 500 W/ft2 (1076 and 5382 W/m2) for the stand-alone data center. All other assumptions, such as building envelope, lighting, HVAC efficiencies and schedules, were based on the minimal requirements of ASHRAE Standard 90.1 at various vintages. We introduced a novel concept of supply and return air approach temperatures to capture the essential effects of non-uniform airflow and temperature distribution in data centers. The approach temperatures were pre-computed by computational fluid dynamics (CFD) simulations for various configurations of ITE loads and airflow containment management in data centers. A new feature was developed in EnergyPlus to implement the approach temperature method. A case study was conducted to demonstrate the use of the data center models. The two data center models cover all U.S. climate zones and can be used to evaluate energy saving measures for data centers, as well as to support development of data center energy efficiency codes and standards.

Rack Temperature Prediction Model Using Machine Learning after Stopping Computer Room Air Conditioner in Server Room

Data centers (DCs) are becoming increasingly important in recent years, and highly efficient and reliable operation and management of DCs is now required. The generated heat density of the rack and information and communication technology (ICT) equipment is predicted to get higher in the future, so it is crucial to maintain the appropriate temperature environment in the server room where high heat is generated in order to ensure continuous service. It is especially important to predict changes of rack intake temperature in the server room when the computer room air conditioner (CRAC) is shut down, which can cause a rapid rise in temperature. However, it is quite difficult to predict the rack temperature accurately, which in turn makes it difficult to determine the impact on service in advance. In this research, we propose a model that predicts the rack intake temperature after the CRAC is shut down. Specifically, we use machine learning to construct a gradient boosting decision tree model with data from the CRAC, ICT equipment, and rack intake temperature. Experimental results demonstrate that the proposed method has a very high prediction accuracy: the coefficient of determination was 0.90 and the root mean square error (RMSE) was 0.54. Our model makes it possible to evaluate the impact on service and determine if action to maintain the temperature environment is required. We also clarify the effect of explanatory variables and training data of the machine learning on the model accuracy.

Numerical and experimental investigations on the thermal performance of a data center

Designing an efficient cooling system for a data center can significantly affect its thermal performance. The present paper deals with two kinds of factors that influence the thermal performance of a data center. First the geometry configurations include plenum height (0.4, 0.5, 0.6, and 0.7 m), perforation percentage (10, 20, 30, and 40%) and computer room air conditioning arrangement. Second, the operating parameters include the temperature and volume of air. A total number of 32 configurations have been designed by adjusting the plenum height perforation percentage to investigate the effect of the cooling system arrangement on the overall thermal performance of the data center. The results present a critical range for plenum height (up to 0.6 m) and perforation percentage (higher than 20%) in which the system's thermal performance can be significantly enhanced by adjusting the geometry configurations. Beyond this range, the performance does not change much. Supply air temperature is more related to the thermal performance of the racks near the farther end of the cold aisle. Increasing the supply air volume can effectively improve the air uniformity along the direction of the cold aisle. Plenum with gradient cross-sections created using inclined partitions significantly reduced the inlet and outlet air temperature by 1.7–2.9 °C. By investigating the effects of geometry configurations and operating parameters on the thermal performance of data centers, the optimal geometry parameters and feasible operating strategy of “high temperature, high air volume” are proposed in this article as the reference and guidance for newly built data centers or retrofitted practices.

Energy-saving effect of integrated cooling unit with rotary booster and compressor for data center

Abstract To reduce the power consumption of the cooling system in the data center effectively, a prototype of the integrated cooling unit with rotary booster and compressor was developed, and the operating characteristics were tested comprehensively according to its cooling requirements throughout the year. A data center model reflecting the characteristics of energy consumption and environmental requirements was established, the energy consumption characteristics and influencing factors of the model data center when the integrated unit was used to meet its yearly cooling requirements in the typical climate zones in China were simulated by DeST software based on the measured characteristics of the integrated cooling unit. The results showed that compared with the traditional computer room air conditioner, the energy-saving ratio of the integrated system can be as high as 39.55% for the whole year, and the energy-saving effect is remarkable. The energy-saving potential of the integrated unit becomes more significant while the heat transfer coefficient of the envelope structure becomes larger, and the load rate of IT equipment and the annually-averaged ambient temperature lower.

Temperature distribution estimation via data-driven model and adaptive Kalman filter in modular data centers

With the rapid development of information and communications technology, increasing number of data centers is required to support the cloud computing, and critical web-based services that run our daily lives. The conventional cloud data centers usually adopt computer room air conditioner or inRow units as the cooling sytem, while the rack mountable cooling unit is a more promising equipment due to the economy, exact controllability, flexibility, and scalability. To ensure the efficiency of control system in rack mountable cooling unit and the security of servers in the data centers, the information of temperature distribution is very essential. Basically, the temperature distribution could be obtained through physical sensors easily. However, considering the cost of whole system and the burden of fault diagnosis in sensor networks, the number of temperature sensors should be kept down to a bare minimum. Therefore, it is necessary to develop an effective and real-time observer to estimate the temperature distribution in the system. Besides, due to the complex air flow and heat transfer in the container, it is quite difficult to construct a physics model. To this end, a novel observer embracing data-driven model and adaptive Kalman filter is proposed in this work. Auto regression exogenous model is adopted as the framework of data-driven model, and the model is identified through a algorithm of partial least square. Moreover, to represent the nonlinear behaviors in the system, fuzzy c-means is applied for data classification and getting multiple local linear models. Finally, adaptive Kalman filter is utilized to estimate the temperature distribution on the basis of proposed data-driven model. The estimation results based on experimental data indicate the performance of proposed approach is remarkable.