The Problem Identified: Energy Costs and Machine Maintenance
How to predict optimal operation of the concurrent blower system regulating dissolved oxygen to reduce energy consumption and unnecessary wear and tear on equipment.
Ensuring clean and safe drinking water is an important task that requires a considerable amount of energy. Freya Systems partnered with the Delaware County Regional Authority (DELCORA), a municipal wastewater treatment company, to identify data-driven optimization opportunities within the treatment process and reduce expenses associated with equipment maintenance and energy consumption. By analyzing operations data, the Freya Systems team established a data-driven approach that will save costs and enhance their impact on the environment.
DELCORA uses a fairly common process to treat wastewater. They begin by separating solids from liquids as it moves through the beginning stages of treatment. During this process, wastewater will reach a series of aeration tanks, which serve a critical function in wastewater treatment. In these tanks, the amount of dissolved oxygen (DO) present is carefully measured and maintained by a series of blowers and valves which keeps the living bacteria integral to the process alive and well.
To ensure the successful conversion of organic waste into inorganic byproducts, DO levels must be carefully monitored and adjusted to support the growth of aerobic microorganisms. This delicate balance is critical to efficient and effective wastewater treatment and requires expert knowledge and diligence.
In 2022, DELCORA had a budget of over $28M for costs related to the operation of treatment plants. This was up 5.26% from 2021. Wastewater treatment plants utilize a hefty amount of energy, with the aeration blower system alone consuming up to 70% – an astounding figure highlighting the necessity for enhanced efficiency in such facilities.
Notably, wastewater facilities across the US account for about 56 billion kilowatt hours of electricity demand, adding an estimated 45 million tons of greenhouse gas emissions to the atmosphere annually. Further, reducing the operating hours on these massive blowers and reducing the instance of cycling on and off increases the lifespan of the expensive equipment and further reduces environmental impact.
Building Predictive Algorithms for Automated Systems
DELCORA provided access to almost three years of once-per-minute exhaustible data to gain insights into their treatment process. In addition to measuring the DO in each of the four tanks, the data included measurements of the opening of valves, blower activity, pressure, airflow, and other aspects of the automated systems. DELCORA also provided additional data relevant to the performance, such as outside weather conditions and inlet wastewater pH values.
Freya Systems first completed an extensive effort to cleanse and organize the data. Then the data science team created a DO-level algorithm to predict future events. The focus was on excessive cost events that increased energy consumption and wear on high-cost machinery (blowers).
In close collaboration with DELCORA, the team discovered that the frequent activation of the fourth blower might be redundant, resulting from pre-existing automated processes within their system.
The approach to the problem differed from typical predictive analysis because it required manual intervention to be taken by the client at certain levels. For example, it is not enough to predict when and by what measure DO levels will drop in the future; the algorithm must deliver controlled variables to adjust by particular amounts in the future and predict when the fourth blower will start to compensate for a loss in DO. Additionally, the client needs a 30-minute window to make the appropriate manual intervention in operation to avoid the added energy cost and wear on the very expensive industrial blower.
Leveraging the data at each minute and in the recent past through an advanced random forest algorithm allowed Freya Systems to create models to predict 30 minutes ahead whether all blowers would be triggered to start.
Analysis of the Solution for Performance Improvement
Based on the analysis, concurrent use of all four blowers was rare. Therefore, precision and recall were used to evaluate the algorithm’s performance. The performance was strong, so Freya Systems implemented the algorithm in the customer’s environment and monitored the operation over several months. During this monitoring phase, performance was as follows:
- Among all days in which the algorithm predicted that all blowers would be used concurrently, all blowers were used concurrently two out of every three days.
- The algorithm successfully predicted the occurrence beforehand on all days all blowers were used concurrently.
Following this monitoring period, DELCORA deemed the algorithm a success and is actively working to fully integrate the algorithm’s predictions into the automated treatment process.
DELCORA can now make more informed decisions regarding the treatment processes. In addition, the algorithm provides them with predictive data that can be used to plan maintenance, lower energy costs, and improve system performance. With this enhanced data-driven approach, the client is better equipped to support its community’s water needs cost-effectively. This novel approach will allow DELCORA to deliver reliable service more efficiently and effectively than ever before.
The Resulting Impact to Conserve Energy
Based on Freya System’s testing and analysis, the team demonstrated that the fourth blower comes online due to a significant drop in DO which, if mitigated using a pre-emptive valve and pressure management, would prevent the fourth blower from firing and reduce energy consumption and equipment wear.
Currently, the fourth blower accounts for 12% of the plant’s kW usage per day. Based on our predictions for intervention preventing activation of the fourth blower, the reduction in use is approximately 5% of the overall annual kW usage. DELCORA is an industrial electricity user and billed on a sliding scale; therefore, we need help calculating the cost reduction achieved.
DELCORA did, however, report that Freya Systems and the results delivered by their algorithms ‘exceeded expectations and have been integrated into systems automation at the plant.’
Implementing advanced data analytics, Freya Systems reduced energy consumption for DELCORA. Looking more globally, with the implementation of advanced data tools in wastewater treatment, plants can drastically reduce energy consumption in a way that benefits both the organization at hand and its rate-payers. Taking such action isn’t merely economic—it has an ecological upside.
If you’d like to discover what operation optimization opportunities await you and your business, Freya’s team of data scientists, engineers, and software developers would love to provide data solutions customized to your specific needs. Contact us to learn more!