Jun-2021

Energy network monitoring and optimisation

A software program for optimisation of an energy network supports decisions on equipment use in refineries that produce and import power.

ELIF GÜL GÖÇER, ELIF MELEK ÖZTÜRK, GÜLSEN SAHIN ANDAS, YAHYA AKTAS and ELIF METE
Tüpras

Article Summary

Optimisation of energy costs and use is among the most important factors affecting operating expenditure and competitiveness in the refining sector. In order to reduce energy costs, there are many options to consider such as benchmarking internal and external best practices and optimising the use of existing capacity.

We think we can do this through a new conceptual framework, energy optimisation, which means using — and not using — energy in the refinery to maximise the reliability of processes for both safety and economic performance. For this reason, TüpraÅŸ refineries operate their own power plants to meet energy demand.

In these power plants, energy is produced by conventional and waste heat boilers, and gas and steam turbines with different capacities and efficiencies which have been commissioned at different times. The refineries are also connected to the local grid. Thus, electricity traded between the refineries and the electricity grid is an option. In refineries, there is energy consuming equipment including furnaces, boilers, exchangers, pumps, and compressors. Different types of fuel can be fired in the furnaces and boilers.  The energy demand of pumps and compressors is supplied by turbines or electrically driven motors. When electricity and steam demand and production costs are considered, the need for a decision support tool which ensures effective management of the utility system arises. Hence the Energy Network Monitoring and Optimization Program has been developed.

Previously, energy optimisation was performed via commercial software. However, the software did not exactly match refinery configurations, and software updates could not be adapted by the end user. Eventually, the expected results could not be achieved by the software. The most important objective of the program developed here is to provide a decision support system that reflects the refinery in a realistic way, and determines the load at which equipment should operate with which type of fuel. In addition, the program is able to produce, buy, or sell output for electricity trade on an hourly basis by taking into account the electricity/steam balance in the refinery and current electricity purchase/sell unit prices.

Methodology
Within the scope of the Energy Network Monitoring and Optimi-zation Program, an interface has been developed wherein 25 different items of equipment can be selected from the palette by means of drag and drop. Power plant and process units are defined.

When modelling process units, energy conservation, mass conservation, and data collected from the field are used. Gas turbine modelling is described in this section. A gas turbine compresses air for the combustion process. High pressure and high velocity gas expand through the turbine to produce work. During the compression of air, the ratio of compression is calculated by:¹

rp  = Pr2/Pr1

where, rp is the compression ratio, and Pr1 and Pr2 are the inlet and outlet pressure of the compressor. The isentropic efficiency is calculated by: 1
nc = (T2s-T1)/(T2-T1)

where T2s is the isentropic outlet temperature of the compressor, and T1 and T2 are the inlet and outlet temperature of the compressor.

The work done by the compressor is calculated using:

where Cpa is the specific heat of air, γa is a constant(1.4), nm and nc are the mechanical and isentropic efficiency of the compressor. Energy balance around the combustion chamber is expressed using:¹

where ma is the mass flow of air, mf is the mass flow of fuel, Cpf is the specific heat of the fuel, HVf is the heating value of the fuel, Tf is the fuel temperature, and TIT is the inlet temperature of the turbine. The shaft work produced in the turbine is calculated by:¹

where nt is the efficiency of the turbine and γg is a constant value of 1.33. The network can be calculated from the difference between Wt and Wc.

Steps in optimisation
The optimisation program consists of three main components:
- The database
- Modelling and optimisation
- Interface and reporting

The database
At this stage of the project, adapters that are independent of each other and which collect data from different data sources were developed. The process data measurements are impaired due to errors in measurement, processing, and transmission. These errors can be classified as gross errors and random errors. Gross errors include major errors caused by sensor calibration shifts, sensor surface contamination, and sensor failure. Random errors are minor errors caused by noise and external conditions. These errors lead to inconsistency in measurement and refinery applications, such as simulation and optimisation, and decision support systems that use these measurements produce false results.

The detection of gross errors is performed using various filters for these data. The data is filtered according to upper and lower limits and measurement quality. The effect of an error is reduced by using the last good value or a safe value instead of the detected error. In addition, data received in different units is adapted to the SI system to ensure integrity between data.

The equipment in the energy network is defined in the diagram developed within the program. The identified equipment, rings, and parameters are stored in the SQL database. Refinery, process, and equipment ontologies are designed in order to make collected data meaningful. In this way, how much utility each equipment produces or consumes and the total energy demand of a refinery for a given time are determined. The data warehouse had been designed for the storage of collected and semantically related data. The average hourly data from process is stored in the data warehouse. This archive provides the opportunity to conduct studies related to past data.