What can AI deliver with regard to power system modernization, and specifically with regard to HVDC and renewables, regarding technical concepts and solutions, regarding project management and regarding implementation time? To shorten the paths towards clean, equally available energy? Can it be generative AI (GAI) to bring things forward faster, more efficient? or better? in what respect? To get an answer, or better, to try to get an understanding about realistic goals we should distinguish between different types of AI, those where sensors are used to add data and provide information through computation, where these data and the information are "scanned" and analyzed with statistical methods to extract valuable hidden information. Sensor type of AI applies certainly in traffic, robotics, maintenance and repair, in all thinkable fields where human thinking and action can be enhanced, substituted, conditioned (potentially dangerous!) through use of sensors. Its application in our field, in running power system equipment and systems, e.g. for collecting and digesting data for use in maintenance and conclusions drawn for design is partly under way. 

Another type does not use sensors, it uses available basic and advanced collected brain information and data from human beings in sciences, art, technologies, medicine, business, daily life, etc. It searches through these big data - stored in libraries, in news, in the "cloud", etc., for designing, building and purely inciting new thinking in new directions.

International Conference on AC/DC Power Transmission, London, 1985.

In trying to find what could be AI implementable parts in the above work we scan the project tasks.

There are items in systems engineering, detailed engineering, civil and mechanical engineering and in project execution which need a lot of overlapping knowledge in many disciplines - a good project manager with sufficient experience might be able to cover this - however, the chance that this is not complete and items forgotten and lost - exists.

The digital twin is certainly some means to reduce the risks. But even when experienced, in such projects like Pacific DC Intertie and Blackwater HVDC the knowledge needed arose often in the course of project execution. A lot of flexibility was needed and money and time to be reserved for unforeseen needs. Such needs could be due to incompleteness of clients specifications, limited a priori knowledge and/or updated requirements in equipment, functions and system integration. This can be early or later in the design stage, depend on test results, at commissioning or after start of commercial operation. Documentation and document control are essential to have a consistent data base.

The digital twin is thought as a documentary tool, though important, but probably to my imagination not the AI tool providing own generative usuable output of value. Why? Since develpments occur to a good deal through an exchange of personal knowledge, in teams, constantly and not always uncovered - in contrast: often secretely and veiled against external access. Some decades ago the analysis of business processes was en vogue using methods and software stemming from manufacturing, trying to adapt it to power systems projects. Own experience: no openess of knowledgeable personnel for revealing essential project relevant factors and to integrate these into the available software.

So the question to the engineers who work in the field of renewables and their integration via HVDC power transmission: What is your opinion regarding the applicability of AI in that field? What should the contents of an AI data base be? How can the data base be used? What is it? Facts? Selecting process of design and project execution steps? Procedural, mathematically analytical, verbally analytical, compository? A connection of all? Are the various subjects described in the Blackwater paper amenable to AI? What are the limits if applicability is given in some fields?

Human intelligence is intuitive. Is there an adequate substitute in AI?

Goals of generative AI in power system design and construction must be defined and its applicability checked by HI. Where is the point where AI can take over? Which parts, to what extent? The Blackwater HVDC system and in addition a powerful long distance transmission, e.g. the Pacific Intertie could be candidates for investigating a single question: can and should renewable resources like wind farms be added to the systems at their terminals and/or along the AC and DC lines using taps? Which modifications are needed and technically possible and financially viable? Can we construct an AI tool to give a viable response?

After these lines were written an article was encountered centering around the question "How can tasks be managed by applying generative AI?"

IBM's game changing approach to AI

This web site desribes verbally engineering tasks in connection with electric power system modernization. Behind much of the text stands as foundations mathematics. Can the verbal description be used to generate real design computations making design possible starting from verbal descriptions by searching for available circuits and data available in the web, in transactions, in publications?