Artificial Intelligence powers complete process optimisation: Huaxiang slashes scrap by over 57% with DISA’s Monitizer® suite

Scrap levels on test castings at Shanxi Huaxiang Group Co., Ltd. (“Huaxiang”) have fallen dramatically after the foundry deployed the Monitizer® digital suite, supplied and supported by DISA, a Norican technology.

After a digital transformation project, Huaxiang is now collecting, analysing and monitoring live process data with Monitizer® | DISCOVER, providing a complete digital view of its process in one system. Huaxiang is also optimising casting quality using Monitizer® | PRESCRIBE’s Artificial-Intelligence-driven analytics. During recent testing, Huaxiang reported excellent scrap reduction figures for several automotive castings: 86.1%, 57.3%, 46.6% and 30.3%.

“Monitizer│PRESCRIBE’s first results from four different patterns showed an average scrap rate reduction of 57.2%. We are truly amazed by these results,” says Mr. Shuang Zhankui, Director of Informatization and Digitalization at Huaxiang.

Huaxiang picks digital to boost competitive

Huaxiang is China’s fifth largest private foundry and produces over 250,000 tons of iron castings annually. Originally specialising in compressor castings, the company has grown and diversified into new markets like automotive.

Aiming for a 10 billion Yuan annual turnover, Huaxiang wants to increase competitiveness through better quality, lower costs, higher sustainability and faster product delivery. But a Lean manufacturing project it began three years ago identified significant waste in its process, not just scrap, but unplanned downtime and excess metal consumption too.

To improve, Huaxiang knew digital foundry technology was the best option. Digitalization is a key element of its development strategy and Huaxiang had already worked with China Mobile to deploy a facility-wide 5G network to support equipment networking as well as automated melting and pouring. It had also started to collect and display data from some equipment but each system only served individual machines and their operators.

Huaxiang needed a solution that could deliver a real-time, data-driven view across each of its lines. And once it identified problems, the foundry wanted to solve them quickly, completely and permanently.

“We needed a smart system that could spot upcoming errors and adjust process settings to correct them,” says Mr Yuan Wang, President at Huaxiang. “That would make our process more stable and produce higher quality components. The right digital management tool would support our company’s continued growth, improve overall performance, control costs and boost profitability, as well as making us more sustainable and ensuring a safe production environment.”

DISA and Monitizer: the proven digital foundry solution

Huaxiang found its smart solution: the Monitizer suite. Monitizer | DISCOVER is a cloud-based IIoT platform designed specifically for foundries and proven in many applications. It delivers robust, secure data collection, timestamping, transfer and storage, with user-friendly reporting and visualisation tools that presents live and historical data in easy-to-understand, customised dashboards. DISCOVER supports remote monitoring and alerting across multiple lines and sites, makes performance comparisons simple and speeding up fault-finding.

Powered by Norican’s strategic Artificial Intelligence (AI) partner DataProphet, Monitizer | PRESCRIBE’s sophisticated AI-driven prescriptive analytics optimise an entire production line in real time and recommend changes to process parameters that prevent defects occurring. It is runs on the same IIoT platform as DISCOVER and, like DISCOVER, is a Software-as-a-Service (SaaS) tool which helps foundries deploy optimisation rapidly, effectively and with low project risk. Both applications incorporate the deep process knowledge and expert foundry support of DISA and the other Norican technology brands.

Since 2001, DISA had been a trusted Huaxiang supplier and seven out of the foundry's 11 casting lines already included DISA equipment. Monitizer | DISCOVER would initially connect three test lines, featuring two DISAMATIC 131 moulding machines and one DISAMATIC D3.

Huaxiang and DISA started implementing Monitizer in January 2021, along with DISA’s Mould Accuracy Controller (MAC) to identify any mould string problems in real time before pouring. First came DISA’s Monitizer | CIM and NoriGate IIoT gateways to digitally enable Huaxiang’s equipment. This hardware would collect live data from Huaxiang’s existing machines and systems, and connect them to the Monitizer | DISCOVER cloud application.

DISCOVER delivers data-driven quality insights

The Covid pandemic caused many lockdowns during the project but, as Monitizer is purpose-built for remote working and collaboration, DISA’s engineers were usually able to support Huaxiang remotely. As the data started to flow in from the three connected lines, it quickly highlighted areas for improvement. Comparing data from different lines helped the project team to trace – and fix – some fundamental causes of poor casting quality

For example, comparing data for six similar castings produced across multiple lines revealed high scrap rates for two of them. Checking that line’s moulding machine operating data against DISA benchmarks showed that the sand shot pressures were low. DISA’s engineers investigated further and discovered unexpected, serious damage to the machine body caused by fitting non-genuine, locally manufactured wear plates.

“The quality of the fake parts was so bad that air and sand were escaping from the chamber during the sand shot,” says Mr. Shuang. “All that loose sand had then worn out the moulding machine body prematurely. We fixed it using new genuine DISA parts and the scrap rates immediately dropped dramatically.”

While one of the DISAMATIC 131 lines produced complex castings with excellent quality, the other had longstanding problems that the foundry had been unable to solve. Monitizer | DISCOVER’s data helped to locate the root cause – a non-DISA sand plant. Although the settings on the two DISAMATIC 131 lines were identical, the non-DISA plant’s sand mix would drift out of tolerance during production while the DISA TM mixer on the other line stayed consistent.

“DISCOVER helped Huaxiang’s managers identify and understand these process problems and find out what was really causing them,” says Michael Hu, APAC Sales Manager and Digital Leader for DISA. “Now they can spot a variance in the data between different lines, find out why that is happening and quickly find its root cause. Once they have implemented them, they can check the data again to see if the effect has been positive.”

Monitizer produces lower cost, higher quality castings

A further important learning was that the higher cost of smaller castings was due to the greater number of pours required which meant using extra energy to maintain the melt temperature in the holding furnace. Moving the lighter castings’ production slot to the night shift when electricity is cheaper instantly reduced their production cost.

Huaxiang now stores all its machine data in DISCOVER, showing when wear parts were changed along with detailed data on subsequent machine performance. “This helps us understand how long different parts last and the best time to replace them,” says Mr Shuang. “We can schedule maintenance more accurately and order the parts so they are on site, ready to fit.”

With all three lines now stable and running well, Huaxiang had sufficient data to start the optimisation phase of the project with Monitizer | PRESCRIBE. In early April 2022, the DataProphet team started to build the first predictive AI models for each production line.

PRESCRIBE’s AI-driven prescriptive modelling shows which combinations of machine, material and other settings produce high quality, stable production. PRESCRIBE’s built-in Expert Execution System then recommends the best parameter settings for every pattern and process situation before defects occur.

The prescriptions were arriving daily by June 2022 and, by August, early testing for 4 patterns delivered an average scrap rate reduction of 57.4%. By November, more than 50 tests had been run and results have continued to improve.

Early quality wins help maintain project momentum

“Patterns with the highest scrap rates have already seen big reductions, for example, from 7% down to 4%,” says Michael Hu. “Scrap rates that are already low are more challenging of course but the overall trend is improvement. Huaxiang’s technical team are sometimes surprised by PRESCRIBE’s advice, but they can see that results improve if they comply. I think it is helping them learn more about how to control each line and what works and what doesn’t.”

“We definitely want our employees to learn how to solve process problems by using the PRESCRIBE system,” says Mr. Shaung. “We will continue to use the prescriptions ourselves to reproduce the same good results from the test and cannot wait to implement prescriptions for the rest of our patterns and see further scrap reductions.”

Huaxiang is now also installing DISA’s innovative Trace And Guidance (TAG) casting tracking system. TAG adds a unique ID to each mould so that each individual casting can be linked directly to its process and quality data stored in Monitizer | DISCOVER. This enhances optimisation effectiveness, with easy-to-use hand scanners employed to quickly classify each casting’s quality following shakeout.

With Monitizer, Huaxiang will keep learning and improving

With rising success in scrap reduction, Huaxiang’s current focus is on expanding DISCOVER adoption throughout the foundry and encouraging more senior managers to make use of its reporting and real-time monitoring capabilities.

“Huaxiang is the first Chinese foundry to implement the full Monitizer platform and we’re delighted they have already benefited so much from it,” says Ulla Tønnesen, President at DISA. “Once again, shows how our engineers’ expertise combined with Monitizer’s tools deliver Industry 4.0 infrastructure and effective analytics very quickly. That means foundries can see significant benefits in higher yields, reduced costs and increased sustainability with rapid, low risk digital projects.”