​Problem we are trying to solve

Today, the engineering of the automation system can still be complex, error-prone and time-consuming.  The trouble-shooting of the automation system is even more critical. All these engineering and troubleshooting tasks require engineers to have deep product knowledge. We would like to lower the barriers to product application so that customers can focus on machine innovation and operation optimization.

​

What solution do we seek?

We are looking for an AI-powered solution, which could act as a technical expert to support and assist engineers to efficiently complete the daily tasks from automation system programming, commissioning to troubleshooting and maintenance.

​

Typical use cases

1: OEM engineers who know the application and machine well, but probably are not very familiar with ABB automation products like PLCs, variable speed drives (VSDs), and servo systems. Then, the machine engineering (e.g. PLC programming, drive commissioning, etc) will be a time-consuming task by checking the technical documents, repeating trial and error and verification, asking support from ABB technical team, etc.   

To support engineers in this case, this AI-powered solution could interact with the engineer to understand the target of the task and may also interact with the products to track the status, then give commissioning steps, or even generate the parameter settings which can be executed directly.

 

2: The end user engineers try to do the trouble shooting and fix the problem quickly when the machine encounters issues, but without a deep understanding of the ABB automation products within the machine, it is difficult to do this work just based on the error codes or alarms.

To support engineers in this case, this AI-powered solution could interact with the engineer to understand the issues, interact with the products to get the error and status information, or take in other information (e.g. photos, videos, etc.), then generate the trouble shooting step for onsite checking or even give proper solutions directly.

​​

More details on the challenge 

We expect the AI-powered solution can solve at least one of the use cases or some other typical use cases in the industrial applications.

In addition, the following aspects need to be considered.

​

• It is a solution which can work with different ABB automation products.

• It is a solution which can also integrate (get trained from) customers’ knowledge and experience in addition to ABB’s expertise.

• It is a solution which can integrate (get trained from) the unstructured and fragmented knowledge and experience in addition to the standard manuals and documents.

• It is an alive solution, which keeps learning and becoming smarter and more experienced for both ABB products and customer machines.

• It is a solution which can ensure information security and local deployment.
   

Furthermore, we are also looking for startups’ innovations to make the AI-powered solution more applicable in industrial scenarios.

​

• How to make the AI solution adapt to the characteristics of the industrial application scenarios, like limited data and low-quality data?

• How to effectively train the AI solution to make it reach the expert knowledge level?

• How to minimize the resources required to run the AI solution?

​