Real-time mechanistic analysis of electrochemical noise data is essential for rapid identification of localized corrosion by plant operators. Recent developments towards this goal include: Intelligent Noise Data Reduction techniques to eliminate uninformative data; neural nets which learn how to categorize corrosion mechanisms from data patterns; multivariate analysis which allows the identification of combinations of plant process parameters that cause damage. These techniques can be combined to facilitate pro-active management of the corrosion problem, including consideration of corrosion mechanisms within the plant optimization process.
Plant performance monitoring and process optimization has historically been driven by economic factors that have been considered separately from corrosion monitoring data. Corrosion monitoring tends to be performed as a procedure in response to emergencies, or indeed not at all as part of normal operations.
The corrosion monitoring solution chosen is, of course, constrained by the value of the plant. However, when corrosion problems are considered outside of other plant conditions that effect plant profitability then this can prove to be a false economy. It is perceived that there is a need to define sets of operational regimes in which corrosion is considered as a strong factor biasing the choice of mode of operation depending upon market conditions for the plant?s product.
There is often strong pressure when the process parameters are optimized for production throughput, that conditions such as chemical injection for inhibition are maintained regardless of market demand. With certain plant, when product demand, and hence need for high output drops, the focus tends to stay with maintaining a low unit cost of production and existing (or degraded) protection. In fact, the value of the plant and the importance in protecting its integrity should demand more attention so that the plant is available for high output upon demand. Of course, when market demand is high a premium can be charged for the product. Corrosion data can have less impact at this time, as long as production integrity is not compromised, but profit should be retained for remediation work caused by high throughput. It is often not.
