Files and links (1)
Free to read and download, Open Access
Thesis
Closed-loop control of multi-frequency oscillatory ventilation waveforms using iterative spectral adjustment
University of Iowa
Master of Science (MS), University of Iowa
Autumn 2018
DOI: 10.17077/etd.005244
Abstract
Recently, a new kind of mechanical ventilation modality has been purposed termed multi-frequency oscillatory ventilation (MFOV). Similar to high frequency oscillatory ventilation (HFOV), this modality ventilates patients at supraphysiologic rates by using low tidal volumes, thus in theory, mitigating the risks of ventilator induced lung injury (VILI). However, for HFOV, this concept has not improved mortality in most clinical applications. A possible rationale for this is realized when considering the mechanical heterogeneity of the lungs. At high rates, flow throughout the lung is non-uniform and frequency-dependent. MFOV offers a solution to this problem by ventilating different lung regions depending on their local mechanical properties thereby reducing the risk of VILI. However, a control system for controlling MFOV waveforms has not been fully developed. We present a novel and unique control scheme, called Iterative Spectral Adjustment (ISA), for controlling MFOV waveforms. The results outlined in this thesis correspond to the four fundamental components of ISA: measure, predict, update, and mean pressure control.
The spectral content of MFOV waveforms is measured using the discrete Fourier transform (DFT) after acquiring an integer number of cycles for each frequency. A processing delay was incorporated into the measurement process to account for the lag in the system between amplitude updates. We provide simulations that a Blackman or Hann window function will ensure accurate amplitude measurements in the presence of harmonic interference. The effects of spectral leakage between frequencies is also examined, revealing a minimum frequency resolution of 2 Hz to ensure safe and stable operation.
The prediction model developed uses a successive over relaxation method for predicting user defined amplitudes. This process permits an updating method where an MFOV waveform packet is continuously generated by calculating a phase shift for each defined frequency. The optimal weighting parameter to use for updating amplitudes was determined to lie between 0.6 and 0.8.
To control mean pressure during MFOV, we have implemented a proportional-derivative (PD) controller. We show that adding an integral component is not necessary for robust performance. In addition, we compare two PD controllers in terms of the tuning method used over varying user defined amplitudes and mean pressures. We have also provided results of varying set mean pressures which show that steady-state convergence is independent from the difference between the initial and desired state.
Finally, we implemented ISA in animals with healthy and injurious lung conditions as well as a mechanical test lung model with varying compliance and resistance. For both sets of experiments, the results revealed a mean 100% target amplitude convergence across all animal lung conditions and lung model parameters. Overall, ISA is a viable method for enabling physicians to define as well as adjust a multitude of parameters of an MFOV waveform while ventilating patients.
Details
- Title: Subtitle
- Closed-loop control of multi-frequency oscillatory ventilation waveforms using iterative spectral adjustment
- Creators
- Bakir Hajdarevic
- Contributors
- David W Kaczka (Advisor)Joseph M Reinhardt (Committee Member)Osama I Saba (Committee Member)Jason H Bates (Committee Member)
- Resource Type
- Thesis
- Degree Awarded
- Master of Science (MS), University of Iowa
- Degree in
- Biomedical Engineering
- Date degree season
- Autumn 2018
- Publisher
- University of Iowa
- DOI
- 10.17077/etd.005244
- Number of pages
- xviii, 100 pages
- Copyright
- Copyright 2018 Bakir Hajdarevic
- Language
- English
- Description illustrations
- color illustrations
- Description bibliographic
- Includes bibliographical references (pages 90-94).
- Academic Unit
- Roy J. Carver Department of Biomedical Engineering
- Record Identifier
- 9983779800102771
Metrics
13 File views/ downloads
69 Record Views