Fernando J. Marianno

Machine learning based multi-physical-model blending for enhancing renewable energy forecast - improvement via situation dependent error correction

With increasing penetration of solar and wind energy to the total energy supply mix, the pressing need for accurate energy forecasting has become well-recognized. Here we report the development of a machine-learning based model blending approach for statistically combining multiple meteorological models for improving the accuracy of solar/wind power forecast. Importantly, we demonstrate that in addition to parameters to be predicted (such as solar irradiance and power), including additional atmospheric state parameters which collectively define weather situations as machine learning input provides further enhanced accuracy for the blended result. Functional analysis of variance shows that the error of individual model has substantial dependence on the weather situation. The machine-learning approach effectively reduces such situation dependent error thus produces more accurate results compared to conventional multi-model ensemble approaches based on simplistic equally or unequally weighted model averaging. Validation over an extended period of time results show over 30% improvement in solar irradiance/power forecast accuracy compared to forecasts based on the best individual model.

PAIRS: A scalable geo-spatial data analytics platform

Geospatial data volume exceeds hundreds of Petabytes and is increasing exponentially mainly driven by images/videos/data generated by mobile devices and high resolution imaging systems. Fast data discovery on historical archives and/or real time datasets is currently limited by various data formats that have different projections and spatial resolution, requiring extensive data processing before analytics can be carried out. A new platform called Physical Analytics Integrated Repository and Services (PAIRS) is presented that enables rapid data discovery by automatically updating, joining, and homogenizing data layers in space and time. Built on top of open source big data software, PAIRS manages automatic data download, data curation, and scalable storage while being simultaneously a computational platform for running physical and statistical models on the curated datasets. By addressing data curation before data being uploaded to the platform, multi-layer queries and filtering can be performed in real time. In addition, PAIRS offers a foundation for developing custom analytics. Towards that end we present two examples with models which are running operationally: (1) high resolution evapo-transpiration and vegetation monitoring for agriculture and (2) hyperlocal weather forecasting driven by machine learning for renewable energy forecasting.

Wireless Sensor Platform for Cultural Heritage Monitoring and Modeling System

Results from three years of continuous monitoring of environmental conditions using a wireless sensor platform installed at The Cloisters, the medieval branch of the New York Metropolitan Museum of Art, are presented. The platform comprises more than 200 sensors that were distributed in five galleries to assess temperature and air flow and to quantify microclimate changes using physics-based and statistical models. The wireless sensor network data shows a very stable environment within the galleries, while the dense monitoring enables localized monitoring of subtle changes in air quality trends and impact of visitors on the microclimate conditions. The high spatial and temporal resolution data serves as a baseline study to understand the impact of visitors and building operations on the long-term preservation of art objects.

Toward large-scale crop production forecasts for global food security

Predicting crop production plays a critical role in food price forecasting and mitigating potential food shortages. Crop models may require parameters from, for example, weather, crop genotype, farm management, and soil. Sources for these data are often found in very different places. Researchers spend a significant amount of time to collect and curate them. In addition, in order to scale yield forecasts from the single-farm level up to the continental scale, crop models have to be coupled with a geospatial big data platform to provide the required data inputs. In a proof-of-concept case study, we investigate the coupling of a scalable geospatial big data platform, Physical Analytics Integrated Repository and Services (PAIRS), to the Decision Support System for Agrotechnology Transfer (DSSAT) crop model. We envision running this system on a global scale. For geospatial analytics, PAIRS provides curation of heterogeneous data sources to simulate crop models using hundreds of terabytes of data.

Energy Efficiency and Air Quality Considerations in Airside Economized Data Centers

Many data center operators are considering the option to convert from mechanical to free air cooling to improve energy efficiency. The main advantage of free air cooling is the elimination of chiller and Air Conditioning Unit operation when outdoor temperature falls below the data center temperature setpoint. Accidental introduction of gaseous pollutants in the data center along the fresh air and potential latency in response of control infrastructure to extreme events are some of the main concerns for adopting outside air cooling in data centers. Recent developments of ultra-high sensitivity corrosion sensors enable the real time monitoring of air quality and thus allow a better understanding of how airflow, relative humidity, and temperature fluctuations affect corrosion rates. Both the sensitivity of sensors and wireless networks ability to detect and react rapidly to any contamination event make them reliable tools to prevent corrosion related failures. A feasibility study is presented for eight legacy data centers that are evaluated to implement free air cooling.Copyright

Discrete Control of Air Conditioning Units in Mission Critical Facilities

For redundancy, almost all mission-critical facilities such as data centers are fitted with more air condition units than required. These units are most of the time heavily underutilized, where the fans within the units are still consuming energy circulating air without actually providing cooling. In more modern facilities such fans are equipped with variable frequency drives, which can reduce substantially the energy consumption if proper controls are implemented. While there have several solutions for controlling and optimizing such variable frequency drive operated air conditioning units, control systems without variable frequency drives (discrete on/off ACU controls) have not been addressed thoroughly. In this paper, we present a practical, distributed and automatic control method for such discrete air conditioning units. The technique includes several safety features and is based on dense environmental sensing and events like hotspots or device failures. We discuss this approach by way of example of a case study.Copyright

Optimal and distributed automatic discrete control of air conditioning units in data centers

In order to prevent potential failures due to lack of cooling systems, data centers are fitted with redundant air conditioning units. Even though such extra elements provide safety and could be turned off, they are typically on and underutilized, thus wasting energy. We propose a distributed and automatic control of air conditioning units based on dense environmental sensing and data center events handling system that prevent hot spots formation or server/device failures. The control is optimal in the sense of determining and controlling the operation of a set with the minimum number of active air conditioning units while preserving the specified data center cooling capacity. In addition, we present an implementation of such system for a discrete state (on/off) control of air conditioning units in a medium sized data center.

Condensation risk in a partially air side economized data center

Air side economization minimizes the usage of Air Conditioning Units (ACU) and water chiller when outside temperature falls below the operating set point of data centers. In many data centers, the relative humidity is less controlled compared to temperature; humidification or de-humidification are used only when upper (80%) or lower limits (10%) thresholds are exceeded. We present a case study of a data center where high humidity combined with abrupt cooling leads to condensation. Since servers racks cannot respond to instant decrease in temperature, the thermal lag time to reach equilibrium can create a transient condition when condensation can appear in locations where the temperature of the servers falls below the dew point within the data center. A new control algorithm is discussed based on dew point control to improve Information Technology (IT) equipment safety by minimizing corrosion and/or reducing condensation risks.

IBM PAIRS curated big data service for accelerated geospatial data analytics and discovery

IBMs Physical Analytics Integrated Data Repository and Services (PAIRS) is a geospatial Big Data service. PAIRS contains a massive amount of curated geospatial (or more precisely spatio-temporal) data from a large number of public and private data resources, and also supports user contributed data layers. PAIRS offers an easy-to-use platform for both rapid assembly and retrieval of geospatial datasets or performing complex analytics, lowering time-to-discovery significantly by reducing the data curation and management burden. In this paper, we review recent progress with PAIRS and showcase a few exemplary analytical applications which the authors are able to build with relative ease leveraging this technology.