Simulating Inhibitor Application in Industrial Environments -
A Critical Review
G. Banerjee
Department of Chemical Engineering
University of Notre Dame
Notre Dame, IN 46556
K. L. Vasanth
Electrochemistry Branch
Naval Surface Warfare Center - Carderock Division
Silver Spring, MD 20903
Abstract
Rapid developments in computer technology have changed the traditional approach to research in areas of science and engineering, and corrosion is no exception to this. Corrosion affects us all and an enormous amount of savings can be achieved with proper control of corrosion in terms of financial, human and natural resources. It is only natural for researchers and engineers dealing with corrosion to adopt the new developments in computer hardware and software.. In the area of corrosion inhibitors, computers have been used by various researchers to simulate different industrial environments for determining the proper inhibitor treatments. This paper presents a critical review of such simulation studies as have been reported so far.
A pilot-scale field test unit was developed by Aronson et al[2] to simulate operation of a full-scale power plant recirculating cooling water system. Test results indicated that there was no scale formation during operation at 20 cycles of concentration. Treatment of a combined stream of raw water and recirculating water in a conventional soda-lime softener can produce the effluent with low concentrations of Ca, Mg and SiO2. The process computer model accurately simulates the recirculating water system but not the softening system.
1. The treatment chemical is fed at a rate of x ppm prior to feeding the boiler.
2. The treatment chemical concentrates C times in the boiler. (This concentration may be different than the mechanical operating cycles based on boiler water solids.)
3. The treatment chemical has a volatility or vapor-liquid(V/L) distribution ratio D at the boiler pressure.
4. R% of the condensate is reused as boiler feedwater.
5. Yn, n=1,2,..., is the treatment chemical concentration after n returns.
Phase separation occurs only in the boiler.
An equation for calculating the concentration of a treatment chemical can be stated as follows:
Therefore, the concentration of the treatment chemical after n returns, Yn, can be computed from a known amine feed rate, the V/L distribution ratio, the percentage condensate return and the cycles of concentration. An equilibrium condition is established when n is large enough. This basic model was then expanded into multiple condensations, considering real condensate system configurations. This model was found to be an effective tool for optimizing chemical treatment for corrosion protection of condensate system. The authors described case histories to prove that the modeling analysis resulted in improved equipment protection against condensate corrosion. One of these case histories is described below.
A refinery, operating steam generators at 450 and 1250 psi(3100 and 8620 kPa), uses feedwater pretreated with hot lime and sodium zeolite softeners, with no condensate return. The condensate system was not treated chemically. Corrosion occurring in the network was reducing equipment life, plugging the condensate piping with corrosion products, and contaminating the condensate with iron. The computer model first determined the energy savings available by returning the condensate-an estimated $80,000 per year from recovering condensate heat value. The returned condensate would decrease feedwater solids, thus allowing cycles of concentration to increase, making possible a reduction in blowdown and improving energy efficiency. To ensure that corrosion product contamination would be kept at a minimum, the model was used to determine the CO2 content in the condensate. Measured values were 11 to 16 mg/L CO2, compared to 14 mg/L predicted by the model, indicating a good correlation. Finally, the model was used to determine the dosage of a neutralizing amine blend necessary to maintain condensate pH between 8.2 and 9.2. The dosage determined by the model was 15 mg/L. The actual dosage was 14 mg/L. Again, the model showed excellent correlation. The model was used in this case to illustrate the energy benefits of returned condensate and to select the chemical program necessary to ensure its quality as a feedwater source. Modeling analysis helped minimize dissolved CO2 levels, increase condensate pH and improve corrosion protection.
Hepp et al[5] illustrated the value of a computer based tool for analyzing corrosion protection of steam and condensate systems using amines by its application to optimizing feedwater treatment for a copper-based system. To arrive at the optimum treatment, a system specific computer program was developed to calculate amine requirements for various alternatives. Key inputs to the program are the mass balances around the circuit and the amine chemistry. For proper determination of mass balances, the program takes into account total feedwater flow, flow through the lime/soda hot-process tank, condensate returns, boiler blowdown, steam recovery from the blowdown flash tanks and steam requirements for the deaerators and the hot-press tank. Factoring in each major flow and component, this system specific approach was found to be more useful than a generalized method that must force fit the myriad possible permutations and combinations of steam-system equipment within the constraints of the program. The amine chemistry stipulation in the program input was done in terms of the distribution ratio - a measure of the relative volatility - and the neutralizing capacity of the amine. The program output predicted the amine concentration in the surface condenser condensate, amine consumption prorated against makeup, feedwater and steam production and projected daily amine use. The program further evaluated different chemical and physical scenarios. one of the important variables examined was alkalinity, the main source of carbon dioxide. Once the plant water chemistry was optimized, the program could be used to show the effects of boiler cycles of concentration, extraction or export steam and other changes in mass balance around the cycle.
Petersen[6] developed a computer model for providing practical guidelines on controlling acid corrosion of overhead systems in refineries and chemical plants. This model was based on an extensive study into the chemistry and properties of neutralizing amines which were used as inhibitors in laboratory experiments. Results of research into the use of neutralizing amines have been made accessible in a set of computer programs, based on mathematical models developed and later confirmed in the laboratory. The programs have been used to improve corrosion control, optimize neutralizer usage and to develop new products. The outcome of these factors has been the improvement in corrosion control performance and economics. The case of a refinery running a mixed slate of sour crudes can be explained to illustrate the application of the Program. The program is routinely used to compare observed pH and neutralizer demand with pH changes. A significant difference between theoretical and observed pH alerts the refiner to the presence of unidentified acids or bases. Routine water analysis consists of the measurement of the accumulator pH and the amounts of chloride, hydrogen sulfide and ammonia and these are used for the program data input. Total and soluble iron are also measured as a means of detecting potential corrosion rate increase and to help confirm corrosion rates measured with resistance probes. In general, the theoretical and observed pH values agree within 0.2 units. Considering the accuracy of all measurements such as crude feed rates, steam rates and overhead naptha rates, this is exceptionally good.
When the refinery ran a certain imported crude, the demand for neutralizing amine increased dramatically, sometimes requiring seven to nine times as much neutralizer as was normally used. The first time that this happened, the pH in the accumulator dropped to about 3. There were great concerns that the neutralizer supply was not reaching the overhead transfer line, that the neutralizer pump had failed and that the neutralizer was not as active as it should be, etc. Ion chromatographic analysis of overhead waters showed that there were large amounts of low molecular weight organic acids and increased amounts of anions from sulfur dioxide acids present in the accumulator water. When these acidic species were taken into account in the program, the predicted pH and neutralizer demand were very close to the observed values, confirming that there were no problems caused by the chemical or injection system. Further, the prediction of the increased usage of neutralizer amine helped make an important economic decision on whether to run this imported crude or not.
A proprietary computer model and its use to assist in the formulation, application and trouble shooting of condensate treatments has been evaluated by Burgmayer et al[7]. According to the authors, application of this diagnostic tool would reduce the need for costly manpower-intensive empirical testing procedures. The condensate modeling system provided accurate system analysis, increased trouble shooting capability and improved condensate corrosion inhibitor performance.