ASU Scholarship Showcase

Previous studies in building energy assessment clearly state that to meet sustainable energy goals, existing buildings, as well as new buildings, will need to improve their energy efficiency. Thus, meeting energy goals relies on retrofitting existing buildings. Most building energy models are bottom-up engineering models, meaning these models calculate energy demand of individual buildings through their physical properties and energy use for specific end uses (e.g., lighting, appliances, and water heating). Researchers then scale up these model results to represent the building stock of the region studied.

Studies reveal that there is a lack of information about the building stock and associated modeling tools and this lack of knowledge affects the assessment of building energy efficiency strategies. Literature suggests that the level of complexity of energy models needs to be limited. Accuracy of these energy models can be elevated by reducing the input parameters, alleviating the need for users to make many assumptions about building construction and occupancy, among other factors. To mitigate the need for assumptions and the resulting model inaccuracies, the authors argue buildings should be described in a regional stock model with a restricted number of input parameters. One commonly-accepted method of identifying critical input parameters is sensitivity analysis, which requires a large number of runs that are both time consuming and may require high processing capacity.

This paper utilizes the Energy, Carbon and Cost Assessment for Buildings Stocks (ECCABS) model, which calculates the net energy demand of buildings and presents aggregated and individual- building-level, demand for specific end uses, e.g., heating, cooling, lighting, hot water and appliances. The model has already been validated using the Swedish, Spanish, and UK building stock data. This paper discusses potential improvements to this model by assessing the feasibility of using stepwise regression to identify the most important input parameters using the data from UK residential sector. The paper presents results of stepwise regression and compares these to sensitivity analysis; finally, the paper documents the advantages and challenges associated with each method.

Pay-for-performance (PFP) is a relatively new approach to agricultural conservation that attaches an incentive payment to quantified reductions in nutrient runoff from a participating farm. Similar to a payment for ecosystem services approach, PFP lends itself to providing incentives for the most beneficial practices at the field level. To date, PFP conservation in the U.S. has only been applied in small pilot programs. Because monitoring conservation performance for each field enrolled in a program would be cost-prohibitive, field-level modeling can provide cost-effective estimates of anticipated improvements in nutrient runoff. We developed a PFP system that uses a unique application of one of the leading agricultural models, the USDA's Soil and Water Assessment Tool, to evaluate the nutrient load reductions of potential farm practice changes based on field-level agronomic and management data. The initial phase of the project focused on simulating individual fields in the River Raisin watershed in southeastern Michigan. Here we present development of the modeling approach and results from the pilot year, 2015-2016. These results stress that (1) there is variability in practice effectiveness both within and between farms, and thus there is not one "best practice" for all farms, (2) conservation decisions are made most effectively at the scale of the farm field rather than the sub-watershed or watershed level, and (3) detailed, field-level management information is needed to accurately model and manage on-farm nutrient loadings.

As the construction continue to be a leading industry in the number of injuries and fatalities annually, several organizations and agencies are working avidly to ensure the number of injuries and fatalities is minimized. The Occupational Safety and Health Administration (OSHA) is one such effort to assure safe and healthful working conditions for working men and women by setting and enforcing standards and by providing training, outreach, education and assistance. Given the large databases of OSHA historical events and reports, a manual analysis of the fatality and catastrophe investigations content is a time consuming and expensive process. This paper aims to evaluate the strength of unsupervised machine learning and Natural Language Processing (NLP) in supporting safety inspections and reorganizing accidents database on a state level. After collecting construction accident reports from the OSHA Arizona office, the methodology consists of preprocessing the accident reports and weighting terms in order to apply a data-driven unsupervised K-Means-based clustering approach. The proposed method classifies the collected reports in four clusters, each reporting a type of accident. The results show the construction accidents in the state of Arizona to be caused by falls (42.9%), struck by objects (34.3%), electrocutions (12.5%), and trenches collapse (10.3%). The findings of this research empower state and local agencies with a customized presentation of the accidents fitting their regulations and weather conditions. What is applicable to one climate might not be suitable for another; therefore, such rearrangement of the accidents database on a state based level is a necessary prerequisite to enhance the local safety applications and standards.