We often talk to battery manufacturers who want to better control quality. This is true for manufacturing nearly anything, but for batteries it’s particularly important. Why is this so critical for them?
Battery manufacturers want there to be as little variation as possible between batteries because the impact of variable quality can be severe. Even when an energy storage system has redundancy built in, replacing a battery can be both costly and time consuming, especially if it’s at a remote location.
Safety is an even bigger issue. While a malfunction in the lithium-ion battery in mobile phone can be a safety risk, a malfunction in a battery energy storage system (BESS) can be catastrophic. In 2022, there were at least 12 BESS failure events worldwide.1
One way to mitigate such risks is to track all source materials and components, so that if some batteries start failing, the culprit they have in common is easily found. If a bad batch of anodes is starting to impact the performance of certain batteries, all other batteries made from that batch can be identified and replaced before they impact system performance.
Tracking all source materials and components is key. Isn’t a company already tracking this data through their procurement and manufacturing processes?
In many cases, yes, companies are tracking at least some of this data such as battery cycler data, part numbers, model numbers, frame plate lot numbers, and container ID’s to name a few. The problem, however, is everything is being tracked in different ways and in different systems. For example, we’ve worked with companies where paper checklists are used during the manufacturing process. It’s very difficult (as you can imagine) for this information to be later referenced or correlated. It’s also error prone, with no validations. Estimates of manufacturers relying on paper-based tracking are as high as 62%2.
The same situation applies to processes that capture data in spreadsheets, tied to someone’s computer or a shared network folder. In other cases, outdated software may be used to capture and store data in a proprietary, un-shareable format. It can be very difficult to later extract this data for analysis with data from other systems.
This sounds like a big challenge. How can a battery or component company go about solving this?
To better track source materials and components used in battery manufacturing, you can first create a system that records the origin, supplier, date of manufacture, and specifications of each material or component used in the manufacturing process. Depending on the size of your operation, you may also want to consider implementing a tracking system that records each material and component’s unique identifier, allowing you to keep track of each item throughout the production process. At Peaxy, we always thread assets down to the individual serial number for all of our customers. This is the only way to ensure that they can be tracked for their entire life cycle.
You may want to consider automating part of the tracking process to streamline data collection. Examples might include:
- Pulling the dry and wet weight measurements from measuring scales
- Pulling welding information directly from a PLC connected to a welder
- Pulling QC results directly from test equipment used on the manufacturing line
This step includes creating data pipelines that feed data from proprietary systems into a single, normalized format for easy reference and analysis. Once completed, it becomes possible to query and associate downstream data – including battery life cycle data – with parameters of the materials that went into that battery in a fully automated fashion.
The effort of creating a single, serialized data stream can be challenging, and often exceeds the resources, skill sets or time available internally. Orchestration of data across multiple systems including time series data requires a coordinated effort across multiple disciplines within a company, often with competing priorities and budgets. In those cases, it’s best to partner with a provider that specializes in this often overlooked part of what eventually becomes the basis for predictive analytics capabilities.
How have you implemented such a solution in the past for a customer? How does it work?
We’ve used our “digital traveler” module for this very purpose for our customers. This is a “manufacturing traveler” tool that allows every step of the manufacturing process to be digitized, with data available to the entire production line, and with clear hand-offs between steps. As the battery build process begins, at each stage of the process, it records the individual lots of the components being used for assembly. For a battery manufacturer, this might include the anode and cathode plates and frames. For component manufacturers, this might include cathode and anode materials, electrolytes, foils and binders.
At each stage of the assembly process, the operator electronically signs off on the traveler stage data before proceeding. This kind of “check and balance” approach, fully digitized, is absolutely crucial in ensuring that batteries are manufactured according to their specifications.
In order for the traveler solution to work effectively, there is a stage prior to that, referenced above, around making sure your data is collected and optimized. It’s important to first identify the data sources you will be using and decide at what level you want to collect the data, such as the examples in the diagram below.
This diagram illustrates some of the possible data sources manufacturers may need to identify and collect during battery assembly. (Source)
How do you ensure that manufacturing data is referenceable with downstream data later in the asset life cycle?
In addition to data captured in the digital traveler, there will be a huge volume of data – up to 600x more for a battery block vs. a wind turbine – generated downstream as part of the battery’s life cycle. It’s important to first identify the data sources you will be using and decide at what level you want to collect the data. As any reader of this newsletter will agree, handling the high volume of battery data can be a challenge. (See our previous issue for a more in-depth look at the ideal amount of battery data to collect and store.)
Aside from the importance of acquiring, cleansing, transforming, and integrating the data, it’s perhaps even more critical to have the necessary infrastructure to store the data, preferably in a scalable and secure cloud environment. Finally, you may need to set up metrics for managing and monitoring the quality of the data to triage any issues around collection and normalization.
As a battery or battery component producer, ensuring safety and performance for the final product is rooted in how well you manage your data. Without insights that only careful data collection and analysis can provide, you’re only guessing.