Search

In this work, we aim to enhance the reliability of health information technology (HIT) systems by detection of plausible HIT hazards in clinical order transactions. In the absence of well-defined event logs in corporate data warehouses, our proposed approach identifies relevant timestamped data fields that could indicate transactions in the clinical order life cycle generating raw event sequences. Subsequently, we adopt state transitions of the OASIS Human Task standard to map the raw event sequences and simplify the complex process that clinical radiology orders go through. We describe how the current approach provides the potential to investigate areas of improvement and potential hazards in HIT systems using process mining. The discussion concludes with a use case and opportunities for future applications.

The recent increase of smart meters in the residential sector has led to large available datasets. The electricity consumption of individual households/devices can be accessed in close to real time, and allows both the demand and supply side to extract valuable information for efficient energy management. Predicting electricity consumption should help utilities improve planning generation and demand side management, however this is not a trivial task as consumption at the individual household level varies with occupant behavior. In residential buildings, many loads have some power flexibility. One of them is water heater, which accounts for up to 20% of home daily electricity use. Conventional methods for water heater power prediction, which heavily rely on physical principles, have limited applicability as their performance is subject to many physical assumptions. Recently, black-box models have gained huge interest due to their flexibility in model development and the rich availability of data in modern buildings. Black-box modelling methods can be further categorized into two types, i.e., statistical methods and supervised machine learning (ML). While statistical methods are relatively easy to implement, they can only capture linear relationships among building variables. Since building operations are typically complicated and nonlinear, the resulting accuracy can be poor. Many ML-based, data-driven approaches have the ability to characterize and forecast total energy consumption of commercial data. However, a paucity of research applying data-driven methods have been tested on the hour ahead energy consumption forecasts for typical detached residential houses in the US. With the advances in smart metering, sub meter usage forecasts at the household-level is also gaining popularity for smart building control and demand response programs. This led us to develop a hybrid model to address the problem of residential hour and day ahead load forecasting through the integration of data-driven techniques. The developed forecasting models are built using three common ML algorithms, support vector machines(SVM), Gaussian Naire Bayes, and Random Forest. Performance comparison among these ML methods was carried out. The results suggest that all models were able to correctly predict a greater proportion of the actual power consumption with prediction accuracy yields between 94% ~ 96%. The SVM model performs the best, while the RF works the worst.

The residential building sector accounts for approximately 37% of total U.S. electricity consumption. Within the residential building sector, heating and cooling is the main target for peak load shifting/reduction since it is the largest contributor to peak demand. In fact, the flexibility of residential HVAC loads can provide continuous variation of demand to provide grid services. Such load can be taken as virtual energy storage (VES) resources by varying their demand over a baseline so that they appear to be providing the same service as electrical battery energy storage to the grid. The performance of HVAC load control to provide VES relies heavily on the accuracy of indoor temperature or cooling/heating demand predictions and therefore the quality of building model. Most building models can be broken down into two categories: gray-box models, i.e. Resistance-Capacitance (RC) model and data-mining based models. The RC model, built based on a combination of physical principles and measured data, is constituted with electrical analogue pattern with resistance (R) and capacitance (C). Different RC models with different structures and orders have been widely applied in prediction of building indoor temperature or cooling/heating demand. But, in general, the RC models require long periods of data to train model coefficients as well as considerable training time and computation burden. Recently, data-mining based models have gained increasing interest due to their capability in analyzing large-scale data and flexibility in practical applications. But, the data-mining based models tend to have invisible model structures which poses a problem when trying to use the model for optimal control or model predictive control of the HVAC system. Hence, for this control-oriented modeling scope, there is a continuing need for efficient online system identification techniques, which can provide explicit parameters for the model. Traditional regression models fit well for this specific purpose. This paper presents a rolling horizon building thermal model for identifying indoor temperatures in separate floors of a typical detached residential house. The developed model includes online adaptive correction component. Since this adaptive algorithm needs to be implemented online, a less computation-demanding polynomial fitting algorithm is adopted. This kind of fitting problem can be cast as linear regression problem with multiple variables, parameters of which can be efficiently obtained by well-known gradient descent method. The validation is realized by comparing the predicted results with the measured data from a real typical detached two-floor house. The results show that the developed model has satisfactory performance in predicting the building indoor temperature in 1st and 2nd floors.

Refrigeration systems in supermarkets and convenience stores operate continuously to maintain proper product storage conditions, and these refrigeration systems can account for 40% or more of the electrical energy consumption of a supermarket or convenience store. Peak refrigeration system electricity demand typically occurs during the afternoon and summer months, corresponding to when the general demand and price for electricity are the highest. In order to reduce peak demand and its associated costs, it may be possible to shift refrigeration system electrical energy use by utilizing the thermal mass of the stored product. Refrigeration system capacities vary from 30-60KW in small store to over 400KW for large stores with annual energy consumption varying from 1-1.5 million kWh. During early morning hours, when electricity demand and price are lower and refrigeration system efficiency is greater, the temperature set point of the refrigeration system can be reduced in order to pre-cool the stored products to below their normal storage temperature. Subsequently, refrigeration system operation can be reduced during the mid to late afternoon when electricity demand and price are high, and product temperature may be allowed to drift upwards. This operating strategy is particularly feasible with products that are not adversely affected by variations in temperature, such as water and canned or bottled beverages. The key to utilizing the thermal storage is to understand the storage potential and associated time constants of the display cases. To determine the feasibility of reducing peak demand by shifting the refrigeration load to off-peak times, experimental and analytical analyses were performed. Simulated product, consisting of one-pint containers filled with a 50% ethylene glycol and 50% water solution, were stored in a medium-temperature vertical open refrigerated display case. Product temperature rise as a function of time was determined by turning off the refrigeration to the display case, while product temperature pull-down time was subsequently determined by turning on the refrigeration to the display case. It was found that the thermal mass of the product in the display case was such that during a 2.5 hour period with no refrigeration, the average product temperature increased by 10°F. In addition, it took approximately 3.5 hours for the product to recover to its initial temperature after the refrigeration was turned on. Using transient heat conduction analyses for one-dimensional objects (such as infinite slabs, infinite cylinders, or spheres), cooling or heating times of various objects may be estimated. For example, analytical methods predict that heating a cylindrical product by 10°F (from 30°F to 40°F) would take approximately 2.2 hours, which is in good agreement with the experimental results obtained in this study. From the analysis, it appears that the thermal mass of the stored product in refrigerated display cases is sufficient to allow product temperatures to safely drift for a significant time under reduced refrigeration system operation. Thus, strategies for shifting refrigeration system electrical demand can be developed. The use of an advanced refrigeration system controller that can respond to utility signals can enable demand shifting with minimal impact.

The primary energy consumer of smart buildings are Heating, Ventilation, and Air-Conditioning (HVAC) systems, approximately 30% of the building energy use, which usually operate on a fixed schedule. Currently, most modern buildings still condition rooms with a set-point assuming maximum occupancy rather than actual usage. As a result, rooms are often over-conditioned needlessly. Occupancy-based controls can achieve significant energy savings by temporally matching the building energy consumption and building usage, conservative user behavior can save a third of expended energy.  In this paper, we present a simple yet effective algorithm to automatically assign reference temperature set-points based on the occupancy information. Both the binary and detailed occupancy estimation cases are considered. In the first case study, we assume the schedule involves only binary states (occupied or not occupied), i.e. the room is invariant. With long-term observations occupancy levels can be estimated using statistical tools. In the second case study, three techniques are introduced. Firstly, we propose an identification-based approaches. More precisely, we identify the models via Expectation Maximization (EM) approach. The statistical state space model is built in linear form for the mapping between the occupancy measurements and real occupancy states with noise considered. Secondly, we propose a method based on uncertain basis functions for modeling and prediction purposes. In literature, basis functions (e.g., radial basis functions, wavelets) are fixed; instead, we assume that the basis functions are random. We consider basis functions with three different distributions, which are Gaussian, Laplace and Uniform, respectively. Finally, we introduce a novel finite state automata (FSA) which is successfully reconstructed by general systems problem solver (GSPS). As far as we know, no studies have used the finite state machine or general system theory to estimate occupancy in buildings. All above estimates can be used to adaptively update the temperature set-points for HVAC control strategy.  To demonstrate effectiveness of proposed approach, a simulation-based experimental analysis is carried out using occupancy data. We define the estimation accuracy as the total number of correct estimations divided by the total number of estimations, and both Root Mean Squared Error (RMSE) and estimation accuracy analysis are provided. All the proposed estimation techniques could achieve at least 70% accuracy rate. Generally, accuracy for binary states estimation is much higher than that of detailed occupancy. For GSPS model, more training data improves performance of estimation. It should be remarked that although some mismatch exist for non-zero jumps, estimation performance tracks the zero base line (non-occupied status) perfectly. Therefore, the estimation techniques are effective for binary estimation with over 90% accuracy. Finally, the estimated occupancy is applied into temperature set algorithm to generate reference temperature curve. By adjusting temperature set curve, we can achieve significant energy without sacrificing customer’s comfort.  In this paper, we propose three real-time occupancy estimation methods that can be incorporated into HVAC controls . We have shown the effectiveness of all the proposed approaches by simulation examples. We have seen great potential of energy saving by integrating the proposed technique into real HVAC control system.    Â

We describe a supervisory control strategy for limiting peak power demand by small and medium commercial buildings while still meeting the business needs of the occupants. The objective of the supervisory control is to operate no more than