Predictive Ethylene Production

The growing consumption of polyethylene (PE) and polypropylene (PP) demands an increase in the production of their monomer, ethylene and propylene. Refiners are exploring the integration of petrochemical complexes with the refinery, as the petrochemical industry is expected to grow in the near future primarily due to the increased demand for plastics instead of traditional transportation fuel.

A combination of proper simulation models and accurate measurement of feedstock properties, which represent the type and chemical composition of the feedstock, feedback from temperature sensors, flow sensors, and historical data, provides a valuable tool for predicting the ethylene/propylene composition. This can be further utilised to determine the parameters of the fractionator. These form the basis to determine those KPIs that have the most influence on production efficiency.

In practice, predictive ethylene production is about getting dependable information from the plant when operators need it. In continuous industries, laboratory data remains important, but it represents a sample taken at a particular moment. Online analysers add a different capability: they show how the process is changing between laboratory samples and provide the frequency of information needed for control, optimisation and early detection of abnormal conditions.

In industrial AI and process optimisation, the analyser specification is only part of the job. Reliable results also depend on how the measurement is installed and operated. That means checking sample representativeness, pressure and temperature, phase behaviour, stream composition, response time, hazardous-area requirements, calibration method and integration with the plant DCS or control system. Even a good analyser can give misleading results when the sample arrives late, changes phase or is conditioned in a way that alters the property being measured.

Our engineering work brings together industrial AI, process optimisation, real-time data, refinery performance and decision support around the actual process objective. The right arrangement may use direct in-situ measurement, an engineered sample conditioning system, optical or thermal-conductivity sensing, NIR analysis, or integration of several measurements into a wider optimisation strategy. Live analyser data can also improve the quality of information available to APC, digital twins and industrial AI by giving these systems current composition and quality data instead of relying only on inferred values.

For predictive ethylene production, the final configuration should be checked against the project datasheet and the full operating envelope. The aim is straightforward: a measurement that works under real plant conditions, can be maintained without drama and gives operators information they can use.

Modcon Systems Ltd. supports these applications with engineering, analyser integration and lifecycle services.

Predictive is an important part of the discussion in this article, which examines predictive ethylene production from a practical process-industry perspective.

[dt_divider style=”thin” /]

ethyl
Skip to content