Use of Oklahoma Mesonet and National Weather Service Data in the Development of Control Strategies for Winter-time Bridge Heating

This paper describes the procedures, tools, and the weather data products related to control strategies for winter-time bridge heating in a long-term "Geothermal Smart Bridge" (GSB) project at Oklahoma State University. Among various task elements of the GSB project, integration of weather inputs with control strategies has been presented in this paper.

A control system module written in LabView programming has been integrated with the HVACSIM+ (Clark and May, 1985), a Fortran-based finite-difference bridge deck model. The feed forward control system module (near real-time future forecast), which was designed to predict the arrival and movement of cold weather fronts at specific locations of interest, has been tested using weather data from Oklahoma Mesonet and the National Weather Service (NWS). This control system module has been designed to provide some conservative estimates of winter temperature using the historical extreme winter temperatures. The threshold ambient temperature for turning on the bridge deck heating was taken as 0.83 ºC (33.5 ºF) in the model.

Preliminary model results using Oklahoma Mesonet data at 15-min intervals during the winter of 1997-98 indicated that the bridge can be turned on for heating using the ground source heat pumps at least 45 minutes ahead of freezing conditions. This model has been also found to provide a reasonable operating time (simulated time) for heating as against manual operation based on actual freezing weather conditions.

Currently, the efforts are underway to analyze the applicability of various products issued by the NWS such as: Revised Digital Forecast (RDF), Rapid Update Cycle (RUC), and NEXRAD (Next Generation Radar) WSR-88D (Weather Surveillance Radar 1988-Doppler) precipitation maps. These NWS products will be used as input to the feed forward control module. Algorithms or procedures have been/will be developed to incorporate NWS products into the LabView based control module.

One of the major goals of this project will be comparison of model results obtained using weather data from Oklahoma Mesonet (dense network of 114 stations distributed around the State of Oklahoma with an average distance of 32 km between stations) and the NWS network (13-14 stations spread over the State). This comparison will quantify levels of uncertainty associated with using weather data from a very dense network (Mesonet) versus coarsely distributed NWS data.