Estimating Hull Coating Thickness Distributions Using the EM Algorithm
Abstract
The underwater hull coating system on surface ships is comprised anti-corrosive (AC) and anti-fouling (AF) paint The AF layers are designed to wear away, continuously leaching cuprous oxide to inhibit marine growth. The thickness of the AF paint layers determines the expected service life of a coating system. Thus, it is important to assess the thickness of the AF layers to determine if the current hull coating system is sufficient. The Naval Ship Technical Manual (NSTM) provides specific guidelines as to how much AF paint should be applied. Unfortunately, the AF layers cannot be measured directly. The distribution of total paint thickness measurements is currently used as a proxy for the distribution of the thickness of the AF paint layers when determining if the existing coating system meets the hull coating requirements. We propose a remedy for this situation. A non-parametric maximum likelihood estimator for the cumulative distribution function of the AF layers, based on the EM algorithm, has been developed. Monte Carlo simulation is used to study the properties of this statistical approach for estimating the AF thickness. This model can be used to help decide if sufficient AF paint is on the underwater hull of a ship.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2000
- Accession Number
- ADA386413
Entities
People
- Michael A. Corriere
Organizations
- Naval Postgraduate School