Customer Introduction
Angang Steel Company is the second largest steel maker in Mainland China, and its Bayuquan Branch is an integrated steel plant in the country's northeast. The coking department of the Bayuquan Branch of Angang Steel looks to become one of the leaders in China, particularly when it comes to green operations.
Manual coal blending takes a long time and relies heavily on the experience of individuals, which is inefficient to pass on.
-
Low replicability
Coal blending know-how is passed on from mentor to apprentice in the most traditional way. New staff takes a while to learn the necessary skills and knowledge, and it's become harder to recruit new talent
Coal blending know-how is passed on from mentor to apprentice in the most traditional way. New staff takes a while to learn the necessary skills and knowledge, and it's become harder to recruit new talent
-
High cost of raw coal
Traditional coal blending methods tend to use overly conservative quality metrics, which drives up the cost of coking.
Traditional coal blending methods tend to use overly conservative quality metrics, which drives up the cost of coking.
-
Varying quality
Manual evaluation of coking results based on the way coal is blended is not accurate enough, so quality might vary.
Manual evaluation of coking results based on the way coal is blended is not accurate enough, so quality might vary.
-
Time-consuming coal blending
Coal has complex metrics, so experts spend a lot of time to refine coal blending plans for different types of raw coal.
Coal has complex metrics, so experts spend a lot of time to refine coal blending plans for different types of raw coal.
Powered by Huawei Cloud ModelArts, the coking department of Angang's Bayuquan branch used cloud, AI, and big data to create and optimize coal blending plans as well as enabling early quality warnings.
-
Low replicability
The smart coal blending solution consists of Huawei Cloud ModelArts' intelligent algorithms, cloud storage, and the customer's coal blending system. The solution aggregates data about raw coal, handling processes, coke, and relevant operations, encrypts and anonymizes the data, and then uploads it to the cloud for storage. Using the data, models are developed and trained on ModelArts. Users can access these models via APIs to predict coke quality and optimize coal blending.
The smart coal blending solution consists of Huawei Cloud ModelArts' intelligent algorithms, cloud storage, and the customer's coal blending system. The solution aggregates data about raw coal, handling processes, coke, and relevant operations, encrypts and anonymizes the data, and then uploads it to the cloud for storage. Using the data, models are developed and trained on ModelArts. Users can access these models via APIs to predict coke quality and optimize coal blending.
Customer Benefits
-
Blended coal now costs 5 RMB less per ton by optimizing coal blending parameters, while guaranteeing the quality of coke.
-
By combining AI with expert experience, it now takes 1 to 2 minutes to develop coal blending plans instead of 1 to 2 days. AI also inspires new ideas and helps share specialized know-how.
-
Coke quality predictions are 95% accurate (up from 90%) due to more precise data analysis and modeling.