How to Streamline Plant Management with Automation and AI/ML Tools

In this blog I am going to guide you through the importance of plant automation, the challenges facing implementation, and 2 case studies with step-by-step solutions.

I went deep into this with Amit Patel in the first session of Integ’s Power: Live series on plant automation. 

It was attended by 30+ plant data experts and business unit managers, who walked away with clear steps to improve their data processes.

If you are part of streamlining plant operations or would like to be, then this blog is for you.

What's in the blog

1. Automation saves millions of dollars & 100s of hours of renewable generators.

2. Renewable companies face BIG challenges with BIG Data.

3. ISO, NERC, and upcoming requirements will require more automation.

4. Utilities and Power Generators lack the tools and processes for renewables automation.

5. Start the automation journey with ISO Market and NERC Compliance requirements.

Banner 1

Automation saves millions of dollars & 100s of hours for renewables generators

Even one hour saved per week across multiple plants can end up in $100s of thousands in annual cost savings. 

So why aren’t renewables embracing automation?

1. You need to process huge amounts of data for ISO and other requirements
ISOs need to have accurate location information and make better predictions because each change has a significant impact on the entire grid. As companies add more units and ISOs impose more requirements, the strain on your data system increases. If you have a wind turbines, you need to closely monitor each turbine and then calculate the total output. So instead of just one plant, you are effectively monitoring a hundred plants.

2. The data needs to be accurate, and in real time, but is often locked in PI/SCADA systems
Data is required in lower granularity and often real-time and is not easily accessible. Data needs to be retrieved from PI tags or SCADA, transformed, and combined with other data sets to become meaningful.

3.  Renewables are new and growing quickly
The industry is adding more plants every week. We have clients where plants are getting added every month, especially with solar and other types of plants. This is straining data teams as they work to keep up with added demand. Clients simply have not had time to build proper data systems.

Renewables companies face BIG challenges with BIG data.

Thera are huge gaps that need to be bridged before renewables companies can build a healthy data system.

1. There is a knowledge gap
While renewables have been around for a while, implementation is still fairly new, resulting in a gap in knowledge and best practices.

2. There is a gap between the tools, processes, and new problems
The tools and processes to address automation problems may not be readily available, and buying off-the-shelf solutions may only offer partial solutions.

3. Varying requirements make it impossible to keep up
There are numerous small to medium requirements and ad hoc requirements, such as predicting faults, optimizing processes, and checking equipment and fault status, which traditional IT and businesses may not be fully equipped to handle.

4. PI & SCADA systems are locking up the data
Data being locked up in systems like PI and SCADA limits the ability to find comprehensive solutions, even when visualization tools like Power BI are utilized.

Blue and Orange Spring Sale eCommerce Website Banner

ISO, NERC, and upcoming requirements will require more automation.

There are new NERC requirements coming soon that are a moving target. ISOs are also demanding increasingly detailed reporting. Unless automation is used, clients are finding it impossible to keep up. 

Examples of requirements that will necessitate automation:

  • 2024 Wind and Solar NERC GADS Requirements, especially creation of events

  • Hybrid Unit Implementations (optimizing bidding)

  • ISO requirements for updated hourly forecasted limit

Utilities and Power Generators lack the tools and processes for renewables automation.

  • We need to understand and bridge the IT business gap.  
  • We need to review and understand required data first. This is a business problem and not a technical issue. 
  • Once data is identified (at least in a preliminary manner), then data exploration needs to take place to ensure that the data required exists, can be accessed, and is coming from the proper source.
  • Often just the first two steps stop a lot of projects, especially smaller projects. 
  • The challenging in understanding data is solved by making sure PI/SCADA engineers and relevant business data owners can work together
  • The challenge in exploring data simply requires proper tools to allow users to access data from any source (PI, Internal, External). Simply being able to view and explore the data can solve 50% of the issues that stop projects from beginning.

    So how do you overcome the process gap?
  • Align the data & refine your tools. 

    Here’s how to build the tools

  • Once you review your data quickly build a prototype and test it.
  • Review the data first. If data is in PI/SCADA, involve the proper engineers 
  • Ensure you have a tool that allows you to access and explore data from multiple sources including PI/SCADA, ISO/RTO, Internal, and Vendor data.
  • Ensure you have the tools to extract data from multiple source, transform and combine data from multiple sources.
  • Using an agile process, build a quick protoype and start small and build. Prototypes should take no longer than 2 weeks typically. Create and build more functions over time.
  • Once a solution is ready, test it in parallel production. You may have missed nuances that you’ll find in parallel production.
  • Operationalize when ready

Start the automation journey with ISO Market and NERC Compliance requirements

I work with dozens of plants data teams to build automation solutions and the best place to start is in ISO Market & NERC requirements.

These are the steps I take to develop best-in-class automation for Automating 
Solar Real-Time Outages:

Identify the need for automating outages, considering the challenges of manual entry and the potential errors it can cause. Here we require PI Data as well as a link to the current ISO outages.

2. Gather the necessary data, such as PI or fault data, from the business using specific tools to connect to the data sources. Here we require PI Data on inverter status, switches, and output. We also have a backup source of data on output and business and engineers worked together to identify and document the required tags. For the ISO outages, we were easily able to connect directly to the PJM ISO via their API (automated interface).

3. Establish a rules engine and incorporate machine learning to determine if an outage is reportable or a derating. This was done and created for each ISO so that only outages/derates that fall into the ISO requirements will be reported. We still track all outages and derates but due to ISO rules, not all outages/derates need to be reported. For example, in certain ISO, you only need to report if you fall below 80%.

4. Break the process into two steps:

            a. Detect a new outage and evaluate if it is a new outage for the ISO.

            b. Store all detected outages or derating and create a ticket if necessary.

Next, perform the same steps to detect the end of an outage.

5. Submit the outage information to the internal system that manages ISO data, ensuring compliance with ISO requirements. Alternatively, you can submit directly to the ISO system.

By following these steps, you can automate the detection and reporting of real-time outages, saving time, improving compliance, and minimizing errors in plant operations.

Building the recipe book for automation with a community approach

Everyone in the renewables data community needs an integrated set of tools and recipes on how solve common problems. But nobody is sharing them. 

Building full stack solutions currently takes months or weeks, but our collaborative community members are able to develop solutions in just days.

Here is how we are building the recipe book:

1. Identify use cases

2. Solve collaboratively as a community

3. Create a prototype using existing tools

4. Document solution steps (recipe) and Distribute to participants

So far we have compiled a list of over 50 use cases from 4 members and are consistently adding more.

Stay Compliant & Save Time with PowerSuite