Edward Masato Ishiyama

Heat exchanger fouling is a persistent problem contributing to process economics, plant capacity, environmental concerns, and safety. Advances in research on hydrocarbon fouling have changed the impact of fouling in crude refinery operations, through the use of rigorous performance monitoring and process optimization methods, use of fouling predictive models, use of fouling mitigation technologies (e.g., use of tube inserts), etc. This work presents case studies from two TOTAL refineries preheat trains where the exchangers are subject to fouling. TOTAL has implemented performance monitoring and predictive maintenance software (HTRI SmartPM™) throughout their refineries to improve existing fouling management practice. The software performs advanced data reconciliation, including simulation of detailed exchanger operational data using HTRI shell-and-tube heat exchanger calculation methods and inferring fouling resistance for individual shells for shells-in-series based on dynamic fouling behavior. Dynamic fouling models are used to assess the impact of fouling on the overall network performance. Several heat exchangers in the network have tube inserts (Turbotal® and Spirelf^® inserts from Petroval), and their performances are monitored and predicted. The case studies demonstrate successful implementation of SmartPM software in TOTAL Normandy and Grandpuits refineries, enabling economically feasible, technically viable, and environmentally desirable decision-making for refinery operation.

Heat exchanger fouling has been studied for some time in the petroleum industry. As understanding of fouling dynamics and mitigation methods improves, refinery fouling mitigation strategies are changing. The implications of deposit aging in refinery units have not been addressed in detail: aging refers to where the deposit undergoes physical and chemical conversion over time. In the 2009 Heat Exchanger Fouling and Cleaning conference, Wilson et al. [Ageing: Looking back and looking forward] presented a simple framework illustrating how deposit aging impacts heat exchanger thermal and hydraulic performance. This paper presents insights into deposit aging gained from analysis of refinery monitoring data. Two case studies are presented: (i) one from the Preem refinery in Sweden where stream temperature, flow and gauge pressure measurements indicated a higher deposit thermal conductivity in exchangers located in the hotter section of the preheat train. (ii) US refinery stream temperature, flow and plant cleaning log data, showing an increased resistance to cleaning when deposits are exposed to high temperature for a prolonged period. The use of deposit aging analysis to improve exchanger operation is discussed.

Parallel branches are commonly observed in industrial heat exchanger networks (HENs). Despite the important relationship between flow distribution and network efficiency, not all parallel branches comprise of flow controllers or not least, flow measurements. When the network is subject to fouling, uncontrolled flow branches can introduce undesired phenomenon such as thermo-hydraulic channeling (THC) [presented at the 2007 HEFC conference; Ishiyama et al., Effect of fouling on heat transfer, pressure drop and throughput in refinery preheat trains]. Recent analysis of crude preheat train heat exchangers has shown the need to use THC models, in particular, for situations where there is insufficient flow measurement data, especially in nonsymmetric branches. This paper revisits the THC model and highlight practical importance of the THC phenomenon through analysis of plant data. The hydraulic aspect of the analysis is strongly linked to the knowledge of deposit thermal conductivity. A case study of a section of a crude refinery HEN is used to illustrate the use of thermo-hydraulic models in data reconciliation to understand flow imbalances caused due to differences in operating conditions and fouling of heat exchangers in each branch of a parallel network.

Visbreakers and other thermal cracking units are thermal process units in crude oil refineries that upgrade heavy petroleum, usually residual oils produced from atmospheric or vacuum distillation of crude oil. The associated process streams of these units consist of heavy hydrocarbons with very high viscosities and impurities, resulting in fouling of the heat exchangers used to cool or heat these streams. This paper presents a practical fouling analysis for thermal cracking units in a refinery in Germany. Fouling management at this refinery was initiated as part of the refinery energy-saving program. Following similar analysis of the refinery's crude preheat trains, heat exchanger networks associated in the thermal cracking units were modeled by entering the plant monitoring data, network topology, and heat exchanger geometries into a commercial heat exchanger network simulator, SmartPM. Fouling behaviors of vacuum residue streams and thermal cracker residue streams were identified and quantified. Both chemical reaction fouling and particulate fouling mechanisms were identified to be responsible for the fouling in these streams. Dynamic fouling models were fitted and used to predict fouling of these heavy petroleum streams, which fouled on both the shell and tube sides of the shell-and-tube heat exchangers.

Ebert and Panchal introduced the “threshold fouling” approach for describing the initial rate of crude oil chemical reaction fouling at the meeting in this series of conferences held in San Luis Obispo, CA, in 1995. This paper summarizes reviews of developments in the threshold modeling approach over the last 10 years, following the review by Wilson et al. at the 2005 meeting. Three areas are considered: (i) The development of quantitative models, which has seen little activity but a switch toward using the threshold models to describe fouling dynamics. One of the reasons for the stagnation in development is the need to incorporate chemical understanding. (ii) The types and range of data sets that have been processed with these models, and an evaluation of the parameters. (iii) Applications where the models are used to predict fouling, or the likelihood of fouling. This is the area that has seen greatest activity, linked to the use of threshold models to describe fouling dynamics. Topics for future research and development are discussed.

Oil refinery preheat trains can exhibit unwanted two-phase flow behavior. An example is boiling of crude oil inside heat exchangers, when the local pressure is not high enough to keep crude in a liquid state. This often arises when the pump is undersized. Understanding the two-phase behavior and assessing the boiling heat transfer coefficients would result in a better prediction and estimation of exchanger fouling. Where single-phase modeling is used under boiling conditions, the anomalous behavior leads to unrealistic estimates of fouling resistance, and can severely underpredict the increased pressure drop and consequent loss of crude throughput. There is little public information on fouling in two-phase flows as laboratory experiments are very costly, despite the importance of this in refinery heat exchangers and furnaces. Indeed, the importance of crude boiling is likely to increase as lighter crudes such as shale oils are processed. These lighter crudes are often blended with heavier crudes to maintain an appropriate refining average density. This paper consists of two sections. The first section uses industrial monitoring data to illustrate fouling behavior for a heat exchanger that undergo both boiling and fouling. The second section discusses simulations to evaluate thermohydraulic behavior when the crude undergoes boiling. The analysis requires coupled heat transfer, and hydraulic and surface fouling aspects; a commercial preheat train network simulator, SmartPM, was used for this study.

Process units subject to fouling often require regular cleaning, giving rise to repeated cycles of fouling and cleaning. The initial stages of fouling are strongly influenced by the effectiveness of the most recent cleaning step and, similarly, the effectiveness and rate of cleaning are determined by the extent and nature of the deposit layer present on the surface. The optimal operating cycle will therefore be determined by fouling–cleaning interactions. Deposit aging is an important factor in this, as an aged deposit is usually more difficult to clean. Aging therefore introduces an element of choice into fouling–cleaning operating cycles, between in situ “chemical” methods and ex situ “mechanical” methods, with associated differences in effectiveness, time, and cost. This paper reports a reformulation of the cleaning scheduling problem to consider the choice of cleaning method, as well as the timing of cleaning. A case study based on a shell-and-tube heat exchanger processing crude oil is used to illustrate the concepts and scope of application of this approach. A novel and more general formulation of the problem, linking design, fouling, and cost aspects via dimensionless groups, is then presented and illustrated with a second case study based on a simpler exchanger model.