Heat Rate Optimization through Machine Learning Implementation and Its Impact on Sustainability
It is fair to say that power, and hence electricity generation has been one of the main driving forces of the economy. The growth in electricity generation would, most definitely, spur the growth in the economy. In 2018, Indonesia had total generated electricity of 267.3 terawatt-hours (TWh); a 4.9% increment from the previous year. The interesting part was that 58.5% of it was fueled by coal. Furthermore, the emission that is created from the electricity generation process would also increase with respect to the generation process itself. In the same year, Indonesia recorded 543.0 million ton of carbon dioxide emissions from power-related industry; an increase of 5.2% from 2017 and 1.6% of world-wide emission contribution.
These concerns lead us into a constructive question of how we can reduce the coal consumption as well as the emission. One of them is to implement Heat Rate Optimization on coal-fueled power plants.
A glimpse of coal-fueled closed-loop Power Plant
In order to generate electricity, there are several steps to do before gaining the result. Firstly, water will be pumped from the reservoir to the boiler. In the boiler, coal is burnt (combusted) to create high temperature and pressure. In there, water will be heated until it changes its state to vapor (or steam) and reaches high pressure and temperature – around 8.0 Mega Pascal (MPa) and 480°C, respectively, in a power plant with medium-to-large capacity. This high pressure of steam will be directed to rotate the turbine which is coupled to generator. The electricity is generated with respect to the rotation of the turbine-coupled generator. Coming out from the turbine, the low-pressure vapor will be cooled in the condenser before being re-pumped (re-used) into the boiler. Figure 1 shows the visualization of the whole closed-loop cycle.
Figure 1 Schematic of a Closed-loop Power Plant.
As a means of measuring the efficiency of a power plant, we should take into account of how much heat that we create in the boiler versus how much power that we are able to generate. This term is commonly called heat rate. The lower the heat rate is, the more efficient is the power plant.
How could we optimize Heat Rate?
Heat rate can be improved by doing physical and operational optimization. Physical optimization can be performed by introducing substantial change (or replacement) in the power plant. For instance, by changing the layout of the piping system, installing a new insulation system, etc. This option, however, is less preferential due to the massive capital expenditure that has to be invested with a low yield and long return on the investment. Not mentioning the high risk of engineering complexity in introducing the new installation.
The less painful approach can be taken by implying operational optimization through Machine Learning implementation. At first, all parameters that might influence the value of heat rate are selected. The list may include value of pressure, flow, and temperature of the circulating fluid in different states. The Machine Learning algorithm (decision tree and supervised random forest are the most suitable ones for this particular case) is then used to determine the contribution of each parameter to the value of heat rate. The importance of parameters can then be ranked after obtaining the result from the model. At that instant, we can put a threshold into the parameters and, consequently, make an adjustment on the operating profile (e.g. limiting the pressure to XX MPa, managing the flow between XX to YY ton/minute, etc.). These initiatives would result to an improved value of heat rate.
How big is the reward?
Recent studies show that the machine-learning-based operational optimization initiatives would result to a total potential saving of 3.8% in annual coal consumption of a small-sized power plant. If applied to all coal-based power plants in Indonesia, that number would result to an increase of 5,94 Terawatt-hours (TWh) of annual generated electricity. It is equivalent to additional number of 1.45 million electrified families in Indonesia. Combined with government’s masterplan of electricity distribution, it will be very beneficial for those who live in an area which infrastructure is rather under-developed (e.g. families in Papua whose electrification is only 60% as of 2017).
In addition to that, the same initiatives might result to 0.6% to 0.8% of decrease of annual carbon dioxide emission. This number would certainly have a significant impact on the sustainability in a purpose to preserve our nature.
References:
BP Statistical Review of World Energy (2019)
PwC Power in Indonesia: Investment and Taxation Guide (2018)