Not long ago, industry, whether process or manufacturing, was operated in silos. All functions including engineering, operation & maintenance functioned in isolation. But thanks to advancements in technology, today, most of the systems are connected & there is widespread availability of data everywhere. The reasons for this huge deluge of data are multifold; one is advanced sensing, which has made it possible to get data, which was always relevant for the business, but sensors on a techno economic scale were not developed.
Secondly, the advent of data storage technology and also storage techniques has increased the storage period enormously with the result that data at very fine time resolution is available for wishful periods and third, the advanced connectivity makes the data ubiquitous.
For instance, ten years back, only a few signals of a power station related to generation was brought to the head office. Today, practically all critical data of each generating unit is brought to the head office. All this translates to the term we call ‘Big Data’. I will use the most prevalent definition of Big Data, i.e. data so big that organisation finds it difficult to extract useful information for its business processes. Its similar to the situation in ‘The Rime of the Ancient Mariner’ by Coleridge–water, water everywhere/not any drop to drink. In order to ‘drink’ the water i.e. to make the data useful, it needs to be leveraged into business triggers to improve operational efficiency; to use unknown correlations and hidden signatures or patterns to improve business processes, ultimately for increasing profits. The difference between having data and acting on it can be a big game changer. This concept is depicted pictorially in Figure1.
Some worldwide efforts by governments and corporates in the field of information driven enterprise include Industry 4.0, Industrial Internet, Connected Manufacturing, Smart Manufacturing Leadership Coalition. Mainly related to the manufacturing sector, all aim to extend access and visibility throughout the value chain to improve adaptability to dynamic markets, for increasing profitability and customer satisfaction. The availability of data at all corners of the organisation pitchesfor real-time performance monitoring in order to bridge the gap between vision & realisation. The Steven Covey model for the four disciplines of Execution is presented in Figure 2. A corporate governance model can be based on this concept with KPIs percolating into each functional area of the organisation. For instance, in power plant operation, the lead measure can be heat rate. If the compelling score card indicates a deterioration and if it is due to a spray valve passing, then accountability is fixed on mechanical maintenance. If it is due to improper tuning, it is C&I maintenance. So, the cadence of accountability is there. Identification of the relevant lead measures in other functional areas will also generate pertinent and actionable information.
Power sector challenges
Having described the relevance of the concept of Information Driven Enterprise in a connected world to an organisation like NTPC, let us examine some of the challenges of the power sector spanning all stakeholders in the power sector value chain, i.e, generators, transmission companies, distribution companies, consumer and regulator. Figure 3 depicts the basic challenge generators in the power sector are facing today. Fuel uncertainty is one major challenge. The coal source cannot be pre determined on regular basis & due to this, seldom we get to fire the coal of the desired quality (i.e. the quality on which the boiler is designed). Land & water are also major issues, due to which we need to optimise land & water usage. Further, CERC regulations change over time, especially in the tariff fixation, leading to change in operating regime of the power generating business.
The latest challenge confronting the power sector is stabilisation of grid due to the huge solar power capacity planned to be added. This is posing tough demands to the existing utilities, especially thermal power plants for keeping spinning reserve for quick start up/shut down, a situation unheard of in earlier days.
Overcoming the above challenges requires optimising the resources at hand & here is where technology comes to the rescue. Some technologies which can improve resource utilisation & drive efficiency are ultra supercritical technology, renewables, high voltage transmission and smart grid/smart metering. Smart grid is likely to generate the Big Data in the power sector and many business analytics models can be developed on this Big Data.
NTPC follows a three phase business process as described in Figure 4. At the engineering phase, many software tools have been developed in house. DREAMs package was conceived way back in 2003 and has now become the backbone of engineering. About 2,00,000 drawings are getting reviewed & approved using this package. ICONS & IDEAL are also two tools developed in-house by our C&I engineers & used in C&I detailed engg.
C&I system Factory Acceptance Test was another area where automation using technology has saved many man hours & streamlined the entire testing process. The software testing could be separated from the hardware assembly, staging, integration & hardware tests, creating a win-win situation for both suppliers & NTPC. The Construction & O&M phase has many automation areas, which are described in subsequent sections of advanced analytics, advanced sensing and advanced connectivity.
First, we start with analytics. Real-time data in our generating units has lot of stories in-built in it and we need to harness every bit of it, not by the operators sitting in the control room, but a centralised expert team catering to all the plants. This vision was created as early as 2008 in NTPC & from this vision was created Antariksh, NTPC’s fleet monitoring centre at the Noida office with data brought from each site (PI server) through MPLS line. Data of each unit is analysed & expert advisories are provided to stations. A centralised expertise pool ensures uniform operation and maintenance practices across the organisation. You can take it as an equivalent of cloud computing in NTPC. Also, last but not the least, the experience is captured; the knowledge gained is institutionalised. This is so important for an organisation like NTPC where expert manpower is due to retire in a few years.
The main components of this fleet monitoring centre is analysis of each unit startup & shutdown w.r.t. OEM curves, problem diagnosis (through in-house developed fish bones), vibration analysis through remote client set-up, and specialised alerts using PI data (sms & emails). Startup & shutdown of each unit is analysed w.r.t. to the curves given by the OEM and each deviation is deliberated and resolved with the aim of improving the process both w.r.t. the time taken, consumption of oil and the life consumption of the boiler & turbine. More important, all this analysis is stored.
The experience of the astute operator in diagnosing problems is encapsulated in fish bones. For instance, in case of condenser vacuum problem, if we click on the poor performance indicator, a fish bone diagram opens pointing to the root cause of the condenser vacuum problem. Vibration analysis data, reveals failure signatures hidden in the data, which is used in predictive analytics to get information about impending equipment failures. Similar to vibration analysis on vibration data, advanced pattern recognition tools can detect anomaly on a parameter, which along with drill down diagnostics can point to an impending equipment problem. Addressing this problem will prevent a major forced outage. This type of tool is available commercially in the market and is in the process of procurement.
Using PI tools, specialised alerts as e-mails and sms are sent to key operation personnel at Noida office as well as site, e.g. mill-off and outlet temp greater than 60 degree which is a warning for mill fire in advance.
NePPS is an in-house development of NTPC’s research wing intended to replace the procured software PADO (Performance Analysis, Diagnostics & Optimisation), being provided for on-line performance improvement of coal fired plants.
Another case of business analytics at work at the power station is Merit Order Rating. Given a load schedule for the station or a stage of a station, it gives the best distribution of the station load amongst the units, which will result in lowest cost of generation. This system will be of additional advantage to the shift charge engineer to decide the backing down of the units and keep spinning reserve for catering to the varying demand from load dispatch centre through the day.
Uptil now, cases were presented wherein data was there, rather Big Data was there, but some kind of analytics was needed to extract useful information out of it. But what about the reverse? We know about a technique or algorithm, which can significantly improve the operating efficiency, but we do not have the relevant data. Here is an example of the same. One of the key challenges was the uncertainty in coal source and quality. Many a time, coal is imported and needs to be blended with domestic coal for getting the desired quality. But the coal analysis data based on which we blend the coal is obtained only after two days, due to which we are not able to optimise on the right blending mix. Installing on-line coal analysers (an advanced sensor based on nucleon penetration), which is being specified for all future projects, will give the elemental analysis of coal in real-time and optimise the usage of imported coal. Hence, at times, we need to demand the relevant data for the analytics to work, i.e. in some way the reverse of data mining.
The next significant aspect is advanced connectivity. Today sitting on Earth, you are able to control satellites orbiting Mars. So, why not use advancements in data connectivity to remotely operate power stations? Its not just an application of technology, but it’s a need of the hour. Over the years, it will be difficult to get qualified manpower posted in remote locations. At the same time, with the increased degree of automation, local interventions are likely to decrease.
With land acquisition an issue, there is a need for construction of power stations on minimum land. More units could be constructed in the space earmarked for staff township. The vision for remote operation for hydro plants was created about ten years ago in NTPC. As a result, this was envisaged in the technical specifications of hydro plants. As an important milestone towards achieving this goal, POC (Proof of Concept) has been carried out by operating drives of Koldam Power station situated in Himachal Pradesh from Engg Office Complex of NTPC at Noida through MPLS line.
The holistic picture
So far, we discussed all these cases of advanced analytics and advanced connectivity in the generation area. There will be equal such cases in other areas too. But the information driven enterprise model in a connected world will be truly achieved only when there is connectivity between all the stakeholders in the value chain.
For instance, when demand forecasting for the generating stations uses load consumption data from smart grid, it will enhance the business value delivered by the power sector as a service to consumers. Similarly, for grid stability, it is imperative that data is shared between grid operator and the utilities for frequency support operation.
Similarly, like the online shopping chain for general purpose items; if we have such a portal created by suppliers for spare parts, we do not have to stock the huge quantity of equipment spare parts as we do today. Manufacturers need to adopt the requisite technology to support their customers and optimise their supply chain. Similarly, best practices can also be shared between customers in the power sector.
With the new focus on renewables and smart grid / micro grid, there will be greater opportunities in the power sector to leverage each bit of data for driving business decisions. The ‘Organisation Culture’ element is seen as critical in driving this change. An organisation can grow when its old people embrace new technology and its new people assimilate the experience of its old people.
To sum up, seamless integration of emerging technologies in the entire value chain of power industry will help percolate the benefits to the society at large and will be instrumental in achieving far reaching goal of Indian power sector, i.e. ‘Smart India’.