ML Work Plan Template Selection for Maintenance Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Customer service representatives face challenges in accurately creating work orders for millions of assets across thousands of locations, often relying on their own knowledge and consulting experts, which can lead to errors.
Innovation Solution
A database system trains a machine learning model to select work plan templates and steps based on user inputs, using previous data to generate comprehensive work orders, reducing human error by automating the creation of work plans and steps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a user manually creates work orders based on their own knowledge and expert consultation, then the work order creation process relies on human judgment, but errors increase particularly when handling millions of assets across thousands of locations
Solution Approach 1:
The system enables automated self-service work order creation by training a machine learning model to autonomously select work plan templates and generate work steps based on asset data, eliminating the need for manual expert consultation and reducing human error in work order creation
Solution Approach 2:
The patent replaces the mechanical human decision-making process with an automated machine learning system that processes asset data, selects appropriate work plans, and generates work steps algorithmically, thereby improving consistency and reliability across millions of assets
2Productivity
If a user manually enters work plan details for each work order, then comprehensive work plans can be created, but time consumption increases significantly
Solution Approach 1:
The system performs preliminary action by pre-training a machine learning model on comprehensive work plan templates and expert knowledge before actual work order creation, enabling rapid automated generation of accurate work plans without manual input during execution
Solution Approach 2:
The patent uses copying by training the machine learning model on existing comprehensive work plan templates and expert knowledge articles, allowing the system to replicate proven work plans automatically for similar assets, thereby speeding up work order creation while maintaining comprehensiveness
3Reliability
If work orders are created without automated assistance, then users have full control over work plan details, but mistakes are easily made particularly when dealing with large volumes of assets
Solution Approach 1:
The system implements feedback by training the machine learning model on historical work order data, completion notes, and expert corrections, allowing the model to learn from past mistakes and continuously improve the accuracy of generated work steps and plan selections
Data Source
AI summary
Generating work plans which include work steps when creating new work orders is described. A database system trains a machine learning model to use inputs for creating work orders to select work plan templates, which include sets of work steps, from a training set of work plan templates, in response to receiving the inputs for creating the work orders. The database system receives an input for creating a work order and identifies work plan criteria based on the input for creating the work order. The trained machine learning model uses the work plan criteria to select at least one work plan template, which includes work steps, from work plan templates. The database system creates a work order that includes work plan(s) corresponding to the selected work plan template(s) and includes at least part of the input for creating the work order. The database system outputs the created work order.