EXASAGE: The first data center operational data analysis assistant

After years of informatization construction, the State Grid Corporation’s data centers have reached a very large scale in the headquarters and provincial company’s network companies, providing good support for the deployment of various business systems. Compared with the rapid development of the data center infrastructure construction, the data center-related operation and maintenance management work concept is relatively lagging behind, the data center generates a large amount of operation and maintenance data every day, and relevant analysis and processing work of the data was ignored. In response to the overall requirements of the State Grid Corporation to build a ubiquitous power Internet of Things, the operation and maintenance of the computer room has been upgraded to a new level, and the operation and maintenance work will be transformed from the traditional lag problem solving problem to the forward-looking work of predicting the operation and maintenance risk in advance. Through the analysis and processing of operation and maintenance data, this paper establishes the functional relationship between operation and maintenance indicator data and time, so as to predict the events that may cause system downtime in the future.

Information gathering and assimilation is normally performed by data mining tools and Online analytic processing (OLAP) operating on historic data stored in a data warehouse. Data mining and OLAP queries are very complex, access a significant fraction of a database and require significant time and resources to be executed. Therefore, it has been impossible to draw the data analysis benefits in operational data environments. When it comes to analysis of operational (dynamic) data, running complex queries on frequently changing data is next to impossible. The complexity of active data integration increases dramatically in distributed applications which are very common in automated or e-commerce applications. We suggest a remote data analysis approach to find hidden patterns and relationships in distributed operational data, which does not adversely affect routine transaction processing. Distributed data integration on frequently updated data has been performed by analysing SQL commands coming to the distributed databases and aggregating data centrally to produce a real-time view of fast changing data. This approach has been successfully evaluated on data sources for over 30 data sources for hotel properties. This paper presents the performance results of the method, and its comparative study of the state-of-the art data integration techniques. The remote approach to data integration and analysis has been built into a scalable data monitoring system. It demonstrates the ease of application and performance results of operational data integration.

Attempts to leverage operational time-series data in Condition Based Maintenance (CBM) approaches to optimize the life cycle management and Reliability, Availability, and Maintainability (RAM) of military vehicles have encountered several obstacles over decades of data collection. These obstacles have beset similar approaches on civilian ground vehicles, as well as on aircraft and other complex systems. Analysis of operational data is critical because it represents a continuous recording of the state of the system. Applying rudimentary data analytics to operational data can provide insights like fuel usage patterns or observed reliability of one vehicle or even a fleet. Monitoring trends and analyzing patterns in this data over time, however, can provide insight into the health of a vehicle, a complex system, or a fleet, predicting mean time to failure or compiling logistic or life cycle needs. Such High-Performance Data Analytics (HPDA) on operational time-series datasets has been historically difficult due to the large amount of data gathered from vehicle sensors, the lack of association between clusters observed in the data and failures or unscheduled maintenance events, and the deficiency of unsupervised learning techniques for time-series data. We present an HPDA environment and a method of discovering patterns in vehicle operational data that determines models for predicting the likelihood of imminent failure, referred to as Parameter-Based Indicators (PBIs). Our method is a data-driven approach that uses both time-series and relational maintenance data. This hybrid approach combines both supervised and unsupervised machine learning and data analytic techniques to correlate labeled, relational maintenance event data with unlabeled operational time-series data utilizing the DoD High Performance Computing (HPC) capabilities at the U.S. Army Engineer Research and Development Center. In leveraging both time-series and relational data, we demonstrate a means of fast, purely data-driven model creation that is more broadly applicable and requires less a priori information than physics informed, data-driven models. By blending these approaches, this system will be able to relate some lifecycle management goals through the workflow to generate specific PBIs that will predict failures or highlight appropriate areas of concern in individual or collective vehicle histories.

In the era of big data, data has become a new factor of production. During the operation and maintenance of the data center, rich operation data have been accumulated. How to effectively manage, analyze and mine the massive data accumulated in the operation and maintenance process, solve the problems that cannot be solved by automatic operation and maintenance, improve the fine management level of data center, and enable data center to better provide services for oilfield enterprises and users will become the challenge faced by oilfield data center in the era of big data. Taking "design and implementation of operation and maintenance log analysis system" as an example, this paper introduces the use of big data technology for further analysis and mining of operation and maintenance data, which is helpful to discover new value points, improve operation and maintenance initiative, and provide reference for fine operation and maintenance management of data center.

This paper focuses on extended operation testing and data analysis of free-piston Stirling convertors at the NASA Glenn Research Center (GRC). Extended operation testing is essential to the development of radioisotope power systems and their potential use for long duration missions. To document the reliability of the convertors, regular monitoring and analysis of the extended operation data is particularly valuable; allowing us to better understand and quantity the long life characteristics of the convertors. Further, investigation and comparison of the extended operation data to baseline performance data provides us an opportunity for understanding system behavior should any off-nominal performance occur. GRC currently has 14 Stirling convertors under 24-hour unattended extended operation testing, including two operating the Advanced Stirling Radioisotope Generator Engineering Unit (ASRG-EU). 10 of the 14 Stirling convertors at GRC are the Advanced Stirling Convertors (ASC) developed by Sunpower, Incorporated. These are highly efficient (up to > 33.5% conversion efficiency), low mass convertors that have evolved through technologically progressive convertor builds. The remaining four convertors at GRC are Technology Demonstration Convertors (TDC) from Infinia Corporation. They have achieved> 27% conversion efficiency and have accumulated over 178,000 of the total 250,622 hours of extended operation currently at GRC. A synopsis of the Stirling convertor extended operation testing and data analysis at NASA GRC is presented in this paper, as well as how this testing has contributed to the Stirling convertor's progression toward flight.

This paper presents the development and application of a super performance dew point cooling technology for data centres. The novel super performance dew point cooler showed considerably improved energy saving and carbon reduction for data centre cooling. The innovations of this technology are built upon a series of technological breakthroughs including, a novel hybrid flat/corrugated heat and mass exchanging sheets, an innovative highly water absorptive and diffusive wet-material for the sheets which enable an intermittent water supply with well-tuned water pressure and flow rate, and the optimised fan configurations. Following a list of fundamental research including theoretical, numerical and lab experimental testing of a small scale prototype system, a specialist 100 kW rated data centre dew point cooling system was dedicated designed, constructed, installed and real life tested in an operational live data centre environment, i.e., Maritime Data Centre at Hull (UK) to investigate its dynamic performance, suitability and stability for application in operational data centre environment conditions.During the testing period, the system showed its reliability and capability to remove a tremendous amount of heat dissipated from the IT equipment and maintain an adequate space temperature in the operational live data centre. The dynamic data collection and analysis during the continuous testing and monitoring period showed the average COP of 29.7 with the maximum COP of 48.3. Compared to the existing traditional vapor compression air conditioning system in the data centre, the energy saving using the super performance dew point cooling system is around 90 %. The work presented in this paper include detailed innovation aspects of the technology and the system operation, as well as the established bridging knowledges, methodology and technical procedure for bringing this new technology into real life operation which involve in data centre survey, optimum design and modularization of the specialist cooling system for data centre application, proven system installation, operating method and cooling air management for data centre as well as the assurance of the continuous sufficient cooling supply to the data centre.

A decade of nitrous oxide (N2O) monitoring in full-scale wastewater treatment processes: A critical review

Traditionally within the energy industry, much emphasis has been placed upon the gathering and management of geoscientific data for easy access, analysis, and decision-making purposes. Typically, however, much less emphasis has been placed on the management of operational data and its integration with geoscientific, economic, and public-domain data sets. As a result, much of the data is never captured or stored, and attempting to access these data is time-consuming and sometimes unsuccessful. In 1995, Halliburton Energy Services launched a major project to design and implement an environment for managing operational and technical data with integrated applications to access, retrieve, and analyze the data. The foundation of this project was Halliburton's technical data model (HTDM), which has been under development since 1991. While the initial goal was to provide operational and technical data to Halliburton's internal users, it quickly became apparent that efficient management of wellsite data from the point of acquisition presents a major opportunity to the industry for more effective and efficient operations. Informed, timely, optimized decision making is enabled by easy access to these data and by the analysis of these data with integrated application software. Halliburton's plan includes development of a globally distributed relational database with application and communications methods that allow acquisition, access, and analysis of data from any location. This paper also presents examples of successful implementation of the system to lower costs and improve efficiency by effective data mangement. Introduction For years, the energy industry has gathered and managed geoscientific data for easy access, detailed analysis, and critical decision-making. Such data have included those derived from seismic, geologic, petrophysical, and well-test surveys. Typically, however, the industry has placed much less emphasis on managing operational data and integrating that data with geoscientific, economic, and public-domain data sets. Although much operational data has been captured, they are usually found over disconnected and frequently incompatible computer databases. Just as often, operational data are found in hardcopy files in widely separated locations. Thus, the difficulty in gathering operational data from these disjointed sources have made attempts at comprehensive analysis of such data expensive, time consuming and often impractical. Today, operators and service companies concentrate on overall reservoir management and seek the most cost-effective methods to achieve and maintain optimal reservoir performance. A high-quality, well-populated, easily accessible operational database can be analyzed to help minimize operational costs and determine the most successful operational procedures. In 1995, Halliburton Energy Services launched a major project to design and implement an environment for managing operational and technical data with integrated applications to access and analyze the data. This project was named TIMS (Technical Information Management Systems). To implement the operational aspects of the TIMS project, project goals were clarified and workflow processes were reviewed. Then, a comprehensive relational data model and an associated workflow scheme were developed, and software and hardware were selected to implement the database and workflow scheme. The database and workflow are currently being phased into field operations. Project Goals The database and workflow scheme that were to be developed through the TIMS project would:–support and improve mainstay workflow processes across all the company's product-service lines (PSLs)–capture and manage job experience to support decision functions

Data centers play an important role in modern business. They require a large amount of electricity and cooling energy simultaneously. In 2007, the percentage of total energy consumed by data centers in total US energy doubled over seven years and it is estimated to be double again by 2012. Currently data centers typically employ separate cooling, heating and power (SCHP) systems. The Combined Cooling Heating and Power (CCHP) system is an efficient, clean, and reliable approach to generating power and thermal energy from a single fuel source simultaneously on site. They could be suitable energy supply systems for data centers since demands from data centers match with the energy generation of the CCHP systems. This paper assesses the energy performance of a CCHP system for the Qualcomm data center in San Diego, California, by means of modeling and operational data analysis. The CCHP system mainly consists of four gas turbines, one exhaust fired absorption chiller, three hot water fired absorption chillers, three electrical chillers and seven cooling towers. System performance models have been developed and validated by experimental data in TRNSYS. The modeling result shows that the CCHP system is capable of meeting the electricity and cooling demands with an overall system efficiency of 46%. As a result, the CCHP system could approximately save 12.9GWh of energy per year compared with SCHP systems. Therefore, the CCHP system is a sustainable and green option for data centers.

<div class="section abstract"><div class="htmlview paragraph">Within the heavy commercial vehicle sector, fleet availability stands as a crucial factor impacting the productivity and competitiveness of companies. Despite this, the core element of maintenance strategies applied in the sector still relies solely on mileage or component usage time. On the other hand, the evolution of the industry, particularly the advancement of Industry 4.0 enabling technologies such as sensorization embedded in components, now provides a vast amount of operational data. The severity levels of application, driving style influence, and vehicle operating conditions can be indicated through the treatment of these data. However, there is still little practical application of using this data for effective decision-making regarding maintenance strategy in the sector, correlating the severity level with component failure possibility. Seeking a disruptive approach to this scenario where data analysis supports decisions related to component maintenance strategy, a literature review was conducted to understand how aspects of Industry 4.0 and data analysis can influence maintenance strategies. As a result of this review, a methodology is proposed for applying structured data analysis based on a robust statistical foundation. A case study of applying this methodology is presented, with the analysis of operational data from a specific component installed in a fleet of heavy commercial vehicles. Through the application of statistical techniques, a variable representing component wear is correlated with variables describing application severity, demonstrating that enhancing maintenance strategies based on data analysis is feasible. With the increased accuracy of component maintenance criteria, a 10% increase in availability is estimated.</div></div>

In view of the frequent and costly failures of power converters in wind turbines, a large consortium of research institutes and companies has joined forces to investigate the underlying causes and key driving factors of the failures. This paper presents an exploratory statistical analysis of the comprehensive field data provided by the project partners. The evaluated dataset covers converter failures recorded from 2003–2017 during almost 7400 operating years of variable-speed wind turbines of different manufacturers and types, operating at onshore and offshore sites in 23 countries. The results include the distribution of failures within the converter system and the comparison of converter failure rates among turbines with different generator-converter concepts, from different manufacturers as well as from different turbine generations. By means of combined analyses of converter-failure data with operating and climate data, conditions promoting failure are identified. In line with the results of a previous, much smaller study of the authors, the present analysis provides further indications against the wide-spread assumption that thermal-cycling induced fatigue is the lifetime-limiting mechanism in the power converters of wind turbines. Instead, the results suggest that humidity and condensation play an important role in the emergence of converter failures in this application.

The objective of this work is to introduce the application of an artificial neural network (ANN) to assist in the evaporative cooling in data centers. To achieve this task, we employ the neural network algorithms to predict weather conditions outside the data center for direct evaporative cooling (DEC) operations. The predictive analysis helps optimize the cooling control strategy for maximizing the usage of evaporative cooling thereby improving the efficiency of the overall data center cooling system. A typical artificial neural network architecture is dynamic in nature and can perform adaptive learning in minimal computation time. A neural network model of a data center was created using operational historical data collected from a data center cooling control system. The neural network model allows the control of the modular data center (MDC) cooling at optimum configuration in two ways. First way is that the network model minimizes time delay for switching the cooling from one mode to the other. Second way, it improves the reaction behavior of the cooling equipment if an unexpected ambient condition change should come. The data center in consideration is a test bed modular data center that comprises of information Technology (IT) racks, Direct Evaporative cooling (DEC) and Indirect Evaporative Cooling (IEC) modules; the DEC/IEC are used together or in alternative mode to cool the data center room. The facility essentially utilizes outside ambient temperature and humidity conditions that are further conditioned by the DEC and IEC to cool the electronics, a concept know as air-side economization. Various parameters are related to the cooling system operation such as outside air temperature, IT heat load, cold aisle temperature, cold aisle humidity etc. are considered. Some of these parameters are fed into the artificial neural network as inputs and some are set as targets to train the neural network system. After the training the process is completed, certain bucket of data is tested and further used to validate the outputs for various other weather conditions. To make sure the analysis represents real world scenario, the operational data used are from real time data logged on the MDC cooling control unit. Overall, the neural network model is trained and is used to successfully predict the weather conditions and cooling control parameters. The prediction models have been demonstrated for the outputs that are static in nature (Levenberg Marquardt method) as well as the outputs that are dynamic in nature i.e., step-ahead & multistep ahead techniques.

Background In clinical examinations, it is important to maintain quality and safety with limited resources. Blood collection, which is invasive and data-intensive, requires efforts to reduce the incidence of patient re-puncturing (“failure rate”) and shorten blood collection time by analyzing operational data. Akinaga et al. (2017) developed the blood collection matching system that assigns a patient*s blood collection difficulty level and a blood collector*s skill level, and showed that the system was effective in reducing the failure rate and blood collection time. However, as a decade has passed since the system’s implementation, changes in patient backgrounds and accumulated practical knowledge necessitate a review of the factors influencing blood collection difficulty. This study aims to identify factors with minimal impact on determining blood collection difficulty based on the implemented blood collection difficulty decision flow. Methods First, we extracted factors with minimal impact on the failure rate among the factors that determine the current difficulty level of blood collection (blood vessel condition, blood collection site, procedure, etc.). We used logistic regression analysis based on approximately 100,000 operational data points collected at Iizuka Hospital in FY2023 and interviewed blood collection room managers. Next, we confirmed the impact of the extracted factors on the failure rate by testing for differences in population proportions of the failure rate, stratified by the extracted factors and through interviews with managers and clinicians. Finally, we retrospectively verified numerically whether the failure rates were equivalent or lower after removing the extracted factors using the operational data. Results Extraction of factors with a minimal impact on the failure rate: After logistic regression analysis and expert reviews, we identified the following patient characteristics: taking warfarin, using hot packs, and not using a rubber tourniquet. Confirmation of the impact of the extracted factors: As a result of the population proportion test, we found no statistically significant correlation between the extracted factors and a higher failure rate. Furthermore, interviews revealed that the extracted factors, once considered indicators of difficulty, have shifted in interpretation to operational reminders due to changes in work practices since the system*s implementation. We also found that there was no direct effect on the difficulty of puncturing blood vessels from a clinical viewpoint. These statistical and clinical viewpoints suggested that the extracted factors were not factors that increase patients* blood collection difficulty. Verification of the failure rate after removing extracted factors: After numerical validation with the extracted factors removed, we found that failure count decreased from 1,898 to an expected 1,770 and the failure rate from 1.78% to an expected 1.66%, confirming the limited effect of the extracted factors. Conclusion In this study, we were able to remove the factors that had minimal impact on blood collection difficulty from the implemented blood collection difficulty decision flow, and clarify the factors influencing patient re-puncturing. However, a remaining challenge is the reconstruction of the blood collection difficulty decision flow, incorporating factors that have yet to be evaluated.

Despite continuous developments in the field of MBR technology, membrane fouling together with the associated energy demand and related costs issues remain major challenges. The efficiency of the filtration process in an MBR is governed by the activated sludge filterability, which is still limitedly understood and is determined by the interactions between the biomass, the wastewater and the applied process conditions. The purpose of this thesis is to increase understanding of the factors impacting activated sludge filterability during full-scale membrane bioreactor (MBR) operation. The overall research goal was to determine conditions for enhanced and efficient operation of the MBR technology. The research work included both extended on-site measurements and operational data analysis. Filterability of the activated sludge was experimentally determined in full- and pilot-scale MBRs treating both municipal and industrial wastewater. Subsequently, the most influential parameters influencing activated sludge filterability were determined. In addition, the design, operational and performance data were collected from the selected full-scale MBR plants and analysed in respect to plant functioning, i.e., operation, energy efficiency and operational costs. Therefore, the research links activated sludge filterability assessment and full-scale MBR functioning, i.e., design options, operation, performance and energy efficiency. Overall, it can be concluded that good filterability of the activated sludge is indispensable for efficient and optimal operation of an MBR. Operation with poor sludge filterability will be associated with a cost penalty due to sub-optimal filtration conditions. Wastewater composition and temperature are important influencing parameters with respect to filterability. MBR plant layout and membrane configuration influence overall MBR functioning and should be chosen carefully. The energy efficiency of an MBR is driven by the hydraulic utilization of the membranes and can be improved by implementation of flow equalization, new aeration strategies and adjusting operational settings to the incoming flow.

This report presents ways to avoid mistakes that are sometimes made in analysis of operational event data. It then gives guidance on what to do when a model is rejected, a list of standard types of models to consider, and principles for choosing one model over another. For estimating reliability, it gives advice on which failure modes to model, and moment formulas for combinations of failure modes. The issues are illustrated with many examples and case studies.