Challenge

Pumping large amounts of water over large distances with multiple pumping stations and intermediary tanks results in high energy consumption and requires balancing the workload and switches of pumps to optimize their lifespan and decrease the number of unplanned maintenance events.

Operators controlling the pumping stations must continuously make the right decisions as to which pumps to turn on and off to achieve the following objectives and constraints:

  • Meet variable plant demand for water.

  • Minimize energy consumption (specific energy per cubic metre of water).

  • Minimize pump wear (decrease the number of pump switches).

  • Ensure that tank levels don't go below a minimum level or overflow.

  • Balance the load across pumps and monitor their performance for better maintenance planning.

Operators are not able to continuously predict future system performance and need to perform other tasks, resulting in different performance across shifts and sub-optimal operations. Using normal cascade control logic is not a good solution either as it looks at upstream tank levels individually rather than at the entire pipeline level and is not effective at optimizing all these objectives.

Achievable benefits

A solution ensuring that pipeline pumping operators are provided with the best set points to continuously reach the objectives and constraints allows to:

  • Optimize the way pumps are utilized 24/7 independently of the operator.

  • Decrease specific energy costs.

  • Decrease pump failures and associated costs.

  • Identify and act on sub-performing pumps.

  • Ensure environmental compliance and plant needs.

Our solution

IntelliSense.io's Pipeline Pumping Optimization application provides operators with the information they need to efficiently monitor pumping operations and provides them with an optimized pumping schedule constantly adapting to changing conditions to optimally run operations.

Figure 1: General solution concept

Live data from sensors such as Level Indicator Transmitters, Flow Indicator Transmitters, and temperature and vibration sensors are integrated to provide digital twins of the pumping stations and tanks/ponds.

The solution is configured for each site to answer the needs of operators and provide a transparent view of the system and KPIs at any time. The following image shows an example of the general live overview of a site's pumping stations and pond levels.

Figure 2: Example of a live overview of pumping stations and pond levels

IntelliSense.io's proprietary pump schedule optimization algorithm uses machine learning to generate an optimized schedule for the pipeline control in real-time to reach the target objectives.

This schedule gets automatically refreshed as the situation evolves, for example, when:

  • Water demand at the plant increases or decreases.

  • A pump becomes unavailable due to failure or planned maintenance.

Operators are provided with a simple operator screen allowing them to implement the recommended optimized schedule.

Figure 3: Pipeline Pumping operator screen

The Operator screen is composed of three main sections:

  • Matrix for the tracking of pumps and recommended actions in the top left corner.

  • Graph of past, current, and predicted tank levels in the top right corner.

  • Matrix to toggle the status of pumps at the bottom.

  1. Recommended pump actions

Figure 4: Recommended pump actions

This matrix is refreshed every minute and represents:

  • The stations as columns (5151, 5152, etc.).

  • The pumps at each station as rows.

  • The current status of these pumps.

    • Red if currently running.

    • Green if in standby.

    • Yellow if unavailable (failure, maintenance, interlock).

  • Where a number is provided: in how many minutes the status of these pumps needs to be changed (ON to OFF or OFF to ON).

It is effectively a way to represent the optimized schedule in a format that is focused on the action to be taken to optimise energy efficiency and pump switches.

2. Graph of tank levels

Figure 5: Graph of tank levels

This graph enables the operator to track and predict the evolution of tanks’ levels in an easy-to-understand format.

It represents:

  • Past performance on the left, with inflection points in the curves representing a change of pumping regime at the station

    • e.g. on the blue TK51 curve turning on an additional pump filling this tank or turning off a pump that was emptying it

  • Current tank levels in the middle

  • Predicted future state if the schedule recommended by the application is followed

    • This schedule adapts automatically to any change of demand or unexpected failure of pumps.

3. Matrix of pump status

Figure 6: Matrix of pump status

This matrix is used by operators to manually toggle the status of a pump, in case of failure or planned maintenance. Pumps flagged as in maintenance or failure are taken out of the schedule and optimization is done on the remaining pumps.

Analyze the efficiency of your pumps

Figure 7: Example of a chart comparing specific costs for pumps at multiple stations

Virtual Sensors such as the Specific Energy Cost can be used to compare pumps and the evolution of their efficiency over time. Observations and alerts related to the temperature and vibration of the pump and motor drives are used by operators to prioritize maintenance.

Value

Using IntelliSense.io's Pipeline Pumping application, sites benefit from the optimal operation of pumps 24/7, irrespective of the operator.

By providing optimization at the entire pipeline level and across all pumps, the adaptive and safe optimized schedule results in decreases in energy usage and improvements to pump life with significantly fewer pump switching events (observed decrease of 60-80% compared to manual or loop control).

Figure 8: Example comparison of pump switches and specific energy

For more information, please do reach out to us via Instant Messaging. We're happy to help!

Did this answer your question?