by Ron Chebra, DNV KEMA
Recently I had some interesting discussions with a few power utilities about their advanced metering infrastructure (AMI) systems and how they will use the 15-minute consumption interval data they’ve collected.
Some utilities are gathering a mound of information, but they are uncertain about who, when and how this data will be used to optimize operations and what information will be used for analysis, as well as billing.
Before I go down that path, let me convey a brief anecdote. About 17 years ago, my brother-in-law, Walt, owned and operated a dry cleaning operation. He invested time, energy and money in an innovation he believed would provide great value to him and his customers. He installed a bar-coded system that tracked each garment through all cleaning stages. Each operator had a pen-shaped, battery-operated bar code reader and was instructed to wand every item through the process: as an item was sorted, pre-spotted, placed in the cleaner, put into the dryer, passed through inspection, handed over to pressing, sorted, bagged and finally put on the rack. Within a few weeks, Walt realized a few things:
- All this data collection took extra time and added complexity to the existing process.
- He began to accumulate mounds of data.
- Processing (486 processors) and storage (20 MB) were a bit pricey. (This was 1995.)
- There were gaps in the data. (Some operators forgot to wand garments at their stations.)
- He had a lot of data, but he needed valuable information.
- Customers didn’t care what stage of the process their garments were in; they wanted to know if their clothes were ready.
The data could be used for time-motion studies and process optimization, but Walt needed more time and resources. So he stopped collecting intermediate data that didn’t provide value and focused on what was important. Is there a lesson to be learned? I am a strong advocate of collecting interval data, but I ask: What, when and how will this collected data be used?
- It is useful for customer service representatives to have ready access to time-stamped, interval data because it enables better customer service.
- Conducting load research anywhere without having to deploy additional resources also is valuable.
- Using time-synchronized, consumption data for transformer load profiles can help optimize assets.
- Leveraging information to balance production with use has great potential.
- Gathering granular consumption information in the middle of the night can help identify other problems such as water leaks and losses.
Many applications are emerging, and we can imagine hundreds of others once we sort the possible data analytics that extract value from AMI systems. But being a bit of a pragmatist, I have some tough questions:
- Are we collecting data just because we can, not because we need it?
- Is there a value target defined to justify the cost of the collection investment (bandwidth, processing, etc.)?
- Who will be responsible for extracting the value that AMI provides?
The next wave of innovation in AMI might not be technology-based; more than likely it will be founded in meaningful analytics.
By the way, back when my cell phone bill was based on call time and duration, I was one of those people who looked through 18 pages of bill details to pick out the high-cost calls to figure out who I called and why it cost so much. Will we find most utility customers doing this?
I want to thank Walt, the dry cleaner, for the practical lesson in data acquisition and data mining. His bar-coded system determined garment sorting needed improvement and he could offer more accurate dry cleaning services.
Ron Chebra is vice president of management and operations consulting at DNV KEMA Energy & Sustainability.
View Power Generation Articles on PennEnergy.com