Visibility in to the operations in plant floor has historically been a major challenge in most manufacturing firms. Manufacturers lose millions of dollars every year due to lack of visibility which can be brought in through the latest development in technology.
What it does
At high level the solution addresses three main concerns
- Order Time Prediction: Predicts the right order time and quantity of raw materials in line with materials processed in the plant and brings in visibility to the supplier
- Order Validation- Sales visibility: Validate the sales order based on system constraints and suggest go – no go decision based on the cost, availability and existing production orders, assisting the sales team.
- Order Tracking – Operation visibility : Brings in visibility into the operations, which would assist the plant manager in making informed decisions
How I built it
The solution is developed on Predix platform and uses BMS’s WIP API’s extensively to capture the plant floor data.
Order Time Prediction: Part / Material availability is captured in the Predix based application and is monitored for consumption in operations. To capture consumption of Part/Material, the system traverse through all open work orders where the parts are used and captures the exact time at which the part inventory reaches Safety stock level.
Order Validation- Sales visibility: Application uses the WIP API to capture the open work orders to get the earliest availability of work centers for the new work order. Also, application’s Part inventory is validated to check the availability of Parts required for the work order based on current consumption rate. In addition, system also lists any work orders with low priority (based on promised date) to show any options available to swap the work order for earlier delivery.
Order Tracking – Operation visibility: System uses WIP APIs to get all open work orders. A drill-down on specific work order, all Work order related information like, Time of operations, work centers , Part/Material availability (at time of operation) and current projected time of work order completion are captured from system and reported. The solution was originally proposed as three different ideas and on detailed study we understood the combined system offers better synergies.
Challenges I ran into
- Most BMS APIs requires mandatory filters to query data from the system. Also, search on all fields in the object is not supported. This resulted in additional processing after fetching data from APIs.
- Work order priority is not captured in BMS system. Since work order priority was one of the key parameter in this solution, we had to define a workaround based on’ promised date’ to achieve the results
- We expected WIP to manage / make available the material inventory. We had to define a temporary inventory in the application and manage. Also location of Material is not captured in system.
- Adding data to the system was challenging. We lost time to understand the clock-in API to feed data in work orders.
- Objects in different modules are not internally connected. Needs significant to map data fed to the system.
Accomplishments that I'm proud of
- After losing considerable time in exploration of the framework and APIs, the team developed the system from scratch within the short timeframe.
- Quick decisions and workarounds within team to overcome the limitations / unavailability of features which we expected in the APIs
What's next for Order Time prediction - Plant Floor Visibility
- Extend Part / Material location definition to enable augmented reality based system to show material availability in real-time to plant supervisors.
- Add operations defects as parameter to improve the accuracy of order time prediction
- Graphical representation of work center availability and options for early completion time
- Integration of Predictive Maintenance solution (our IP solution) to improve accuracy of order processing and enable flexibility in planning