Mechanical Engineering Department
BS in Mechanical Engineering
The purpose of this document is to demonstrate the design, production, and testing of optimum pump sequencing logic for variable frequency drive (VFD) pumps in commercial heating ventilating and air conditioning (HVAC) systems. The product will be developed by Pump Efficiency Solutions (PES), which is a group of Cal Poly students working on their senior project. The product is being developed for Trane, an industry leader in large scale HVAC systems. MATLAB™, Microsoft Excel, Pump System Improvement Modeling Tool™ (PSIM), and fluid mechanics hand calculations were used during software development. Next, we converted the program into Trane’s graphical programming language (TGP) and applied these algorithms to a MP580 Controller.
There were three separate phases to the project. The phases, each lasting one quarter, can be designated as conceptualizing, modeling, and programming. During the design conceptualizing PES gathered information on commercial HVAC systems and VFD pumps which can be implemented into these systems. We then modeled a simple system using PSIM and developed a series of optimization algorithms within excel to optimize the quantity of the pumps for a given system load (flow-rate) and pressure requirements. The completion of the model represented the end of the conceptualizing phase and beginning of the modeling phase.
For the second phase of the project, we took the analysis from Excel and used it to develop a general algorithm which can be applied to a variety of different systems. Essentially, it incorporates pump curves for the specific pumps selected for a system and calculates the theoretical best number of pumps to run for maximum system energy efficiency at a given load. This algorithm can be adaptable to any system and will be incorporated into Trane’s pre-existing system control software. The program will have permanent inputs based on the components and geometry of the system in addition to dynamic inputs that define the current operating conditions due to a transient load. The outputs of the program will be number of pumps that should run in order to match the load and minimize power consumption.
In the programming phase, we translated our MATLAB™ program into a series TGP files to make the algorithms compatible with the controller. To simulate changes in system head and flow rate, we connected two potentiometers to the controller. Both MATLAB™ code and MP580 controller simulation successfully calculated the optimum number of pumps to operate given any system inputs within a standard operating range.