Optimizing Smelter Uptime Through Digital Asset Management

Abstract

In most smelting operations, there is a strong incentive to improve safety, reduce downtime and extend the campaigns of furnaces and related equipment. An asset management strategy that is based on information, rather than simple throughput metrics and intuition, can be developed to improve uptime and unlock value currently unrealized at existing smelters. Advances in digital technology , including data analysis, modelling and monitoring for hot pyro-metallurgical vessels, make it possible to assess the current condition of an asset, as well as to make more informed projections about its future condition, using on-line and historical data. This would allow practices around asset management to be more proactive and less reactive, with the end-result of optimized uptime. ‘Digital Asset Management ’ describes a concept in which digital technology is used to support an asset management strategy and accelerate decision-making by providing information on what, why and how an asset should be designed, operated or maintained. With the overall goals of effective resource utilization , optimized operations , capital efficiency and social acceptance, Smelter 4.0 is a program of holistic improvement and is underpinned by the effective use of data to drive decisions. This paper focuses on Asset Reliability, which is one of the improvement pillars of Smelter 4.0 , and introduces a methodology that is aimed at closing the gap between recorded data and evidence-based decision-making through Digital Asset Management .

Introduction

The profitability of a smelter is heavily dependent on uptime (or availability or on-line time), utilization (how closely a smelter operates to its design capacity) and product quality. Based on a Hatch data collected from 65 smelter sites in a variety of industries, many smelters operate with an uptime of 94% or less, as shown in Fig. 1. The uptime of a typical smelter represents a significant opportunity for improvement, with a 1% improvement resulting in approximately US $3 M per year increase in profit margin, per site. This opportunity can be captured by leveraging two known enablers within the industry—asset management and digital tools.

../images/468727_1_En_40_Chapter/468727_1_En_40_Fig1_HTML.gif — Fig. 1
Significant uptime improvement opportunity with each 1% improvement resulting in an estimated average of US $3 M annual margin increase per site across industries

The recommended approach to manage critical assets is quantitative and risk-based, driven by real-time information rather than “rules of thumb” as the most economically efficient method to optimize uptime and balance total cost of ownership [1]. A risk-based asset management strategy, however, also places a higher demand on data compared to reactive, time-based or purely condition-based strategies [1]. Digital tools, including advanced sensing and modelling technologies, which have been made more available by technological advancements and cost reductions in computing, can provide continuous insight into current asset condition, as well as the rate of wear or deterioration of an asset. This information can be used to estimate the life of the asset based on current operating data, rather than solely based on past campaigns, as well as to inform equipment design changes that are required to meet the future needs of the process. Such information would accelerate the decision making process and be highly valuable in developing an effective and efficient asset management strategy. This combination of digital tools and a risk-based asset management program is referred to as ‘Digital Asset Management ’.

This paper describes the key components of Digital Asset Management , driving towards zero unplanned downtime. It describes a risk-based approach to asset management planning, as well as the digital tools that are needed to make decisions and facilitate risk-controlling actions. Such tools include those that are used in the design of equipment, technologies used for operations and process control , and predictive maintenance technologies. Lastly, this paper describes how Digital Assessment Management fits into the Smelter 4.0 program, in which Asset Reliability is one of the key improvement pillars of the program.

A Risk-Based Approach to Asset Management Planning

With the goal of increasing uptime, smelter operations are looking to extend time between major rebuilds and to correctly balance planned downtime with resultant uptime and total cost of ownership. This is depicted in Fig. 2, which shows the successive increase in campaign life for a particular asset. Every asset, as designed to serve a specific requirement, has an inherent risk of not meeting its uptime and campaign life objective (Point A, Fig. 2), which can be continually reduced through design improvements and advancements in equipment technology. As the asset ages, and as the operating conditions change (e.g. feed characteristics), along with emergence of gaps in operating and maintenance practices, the risk increases. Eventually the risk exceeds a tolerable level (Point B, Fig. 2) leading to a rebuild of that asset or a premature failure event.

../images/468727_1_En_40_Chapter/468727_1_En_40_Fig2_HTML.gif — Fig. 2
Risk-based asset management planning involves continually identifying failure mechanisms and operational risks, and implementing changes that cover all aspects of the asset’s design, operation, maintenance and operating environment . (dots indicate rebuild)

A risk-based management plan is an approach that can be used to better control the rate of risk increase (i.e. the slope of the line in Fig. 2). A risk-based asset management plan proactively manages an asset by addressing the root-cause contributors to failure, while promoting buy-in from stakeholders to encourage the implementation of new technologies and strategies [2–4]. This approach is in alignment with published frameworks for asset management being adopted to manage mining assets, and is already widely adopted in the infrastructure and oil and gas sectors [1].

Controlling the rate of risk increase starts with understanding the condition of the asset today. This understanding may come from equipment inspections, condition monitoring, including lining thickness measurement by NDT , and/or analysis of operating data [5]. This information is translated into a quantified risk score $( {\text{Risk}}\,{\text{score}} = {\text{Severity}} \times {\text{Likelihood}} \times {\text{Detectability)}}$ for each identified risk. This is best done collaboratively with input from a cross-functional team and in accordance to the corporate risk standards (Fig. 4a).

The next step is to ensure the risks identified are monitored and that an action plan is generated to control the risks. Developing or improving an asset condition monitoring and management program can be aided through a digital maturity evaluation, in which the existing monitoring and data use practices (acquisition, analysis, predictions and actions) are assessed against a standardized digital maturity matrix (Fig. 4b). This process can quickly reveal gaps between the existing operation and the industry standard, as well as internal gaps (for example, a smelter may have a centralized data system, but some information may be kept separately as logs rather than being stored centrally). The use of the appropriate, cascading key performance indicators (KPIs), targeted at the correct level of the organization then becomes critical to sustain a continuous improvement cycle. To maintain the improved trajectory requires ongoing monitoring and feedback to manage the risk profile and adjust, as required. This describes the ability to change the slope of the line in Fig. 2.

When answering the question, ‘how can risk be controlled’, it is important to have a focus on defect elimination, eliminating root causes of failure rather than mitigating loss of asset function [6]. It is also important to involve aspects of design, operating and maintaining practices as part of the solution [6]. This is extremely important as studies have repeatedly shown that poor asset performance is a combination of flaws in design and supply, operating and maintaining practices, rather than any single source (Fig. 3). This approach can yield short and medium-term plans in which the risk-controlling actions are prioritized based on their potential to reduce the risk and ease of implementation (Fig. 4c).

../images/468727_1_En_40_Chapter/468727_1_En_40_Fig3_HTML.gif — Fig. 3
Root causes of failure, industry-wide

../images/468727_1_En_40_Chapter/468727_1_En_40_Fig4_HTML.gif — Fig. 4
(a) Quantifying risks to operation based on understanding of asset health (b) identifying any gaps in data collected and in its use against a digital maturity matrix (c) a prioritized list of actions that include design, operating and maintaining practices to improve asset reliability

Some strategies for mitigating in each of the three key areas—design, operation and maintenance —are described in the following sections.

Design

To optimize smelter downtime , the design of an asset must suit the process, operation, and consider the practical limits of maintenance , such as access and exposure to hazards. Experience, combined with monitoring data available to drive design changes in the right direction, makes it possible to optimize uptime with improved design.

Our experience, gained from a long history of successful smelter design, rebuilds, repairs and improvement retrofits, as well as involvement in multiple failure investigations and their subsequent rebuilds for smelters of varying design, provides insight into causes of failure and the effect of implemented changes. Additionally, actual measurements and data from the failure investigations provide a crucial source of information for the determination of current condition.

On several occasions, rebuilds or repairs were performed when potential failure or unplanned downtime was foreseeable through observation of monitored trends that deviate from expectations. A number of examples have been previously published [7–10] and the interested reader is referred there for more details.

Implementations of new equipment technology on existing and new builds continue to increase equipment robustness and the information available to understand current conditions and improve safety [11–13].

The future is expected to bring higher variability of feed materials which must be considered in design to accommodate the corresponding wider range of operating conditions. We anticipate therefore that the process equipment design will continue to evolve, but this will be done mainly through capital efficient, incremental retrofits. Each rebuild or repair is an opportunity to retrofit the flexibility required to monitor and adapt to these future changes, such as new sensing and actuation technologies.

Operate

Operating within a set of defined limits is critical to maintaining equipment integrity, achieving the target production/quality and complying with environmental regulations. Historically, mitigating process upsets has been made easier by the supply of a constant feed material to the smelter, allowing smelters to operate using a consistent set of rules that have been developed through experience over many years of operation. The increasing variability in feed materials has made such rules-of-thumb less effective. These factors emphasize the need for fully integrated, data-backed, feed-forward process control systems that are more capable of handling change, with key objectives of maintaining equipment integrity and ensuring safe work conditions, while maximizing production and product quality.

A key foundation of any good smelting operation and process control is the fundamental understanding of the process metallurgy that governs the process. This knowledge is used to develop mass and energy balances for the operation in the form of flowsheets and advanced process models that recommend operating setpoints based on process data inputs and measured operating conditions. More detailed knowledge of what is happening within and to the vessel can be derived from tools such as Computational Fluid Dynamics (CFD) methods and Discrete Element Method (DEM), as illustrated in Fig. 5. For more information please refer to [14–16]. Unlike unit-based mass and energy balances around the vessel, these tools provide detailed spatial and temporal information that is necessary to explain localized behavior. In many applications, this level of detail is necessary to understand the fundamental process drivers. These tools are calibrated to the actual process and can be used to predict the impact of expected process changes or to facilitate enhancements aimed at maximizing smelting efficiency and/or production without exceeding equipment design or environmental limits.

../images/468727_1_En_40_Chapter/468727_1_En_40_Fig5_HTML.gif — Fig. 5
CFD and DEM modelling of the previous burner and reaction shaft of the BHP Olympic Dam Direct-To-Blister Flash Smelting Furnace ; (a) velocity vectors and (b) temperature isotherms generated using CFD and (c) Discrete Element Method (DEM) modelling used to improve the feed distribution through the design of a new burner [14–16]

Inside the smelter, tools that are typically implemented include calibrated unit-based process models such as the On-line Heat Balance and Furnace Supervisory Control Modules that provide automatic power-feed control [18, 19] and feed distribution control . These systems use on-line process data (material flows, stream temperatures and assays) and measured crucible heat losses to provide an automatic, on-line furnace mass and energy balance and operating setpoints that are matched to the actual conditions. These tools are further enhanced with other sensing technologies (e.g. automatic level measurement , online radar level measurements [17, 20]), with the net result of a correct balance of power to feed and optimized feed distribution helping to maximize energy efficiency and prevent unnecessary downtime from high crucible temperatures.

Knowing how to operate within a set of defined limits can be approached using Machine Learning. Very powerful data-mining techniques, with a sound statistical basis, are used to establish relationships between the key variables that drive the system, and most importantly, to understand the conditions that lead to operation outside those defined limits. An example of such an analysis is shown below in Fig. 6, where these techniques have been used to identify the process variables responsible for the process deviation. These tools are being used to identify relationships between process parameters and measured equipment conditions, particularly those that are too complex to be predicted theoretically or by direct correlation.

../images/468727_1_En_40_Chapter/468727_1_En_40_Fig6_HTML.gif — Fig. 6
Determining attributing variables for specific process events; contribution plots (left) and principal component analysis (right)

The next step is to bring these tools together to develop advanced process models that are based on metallurgical process fundamentals and calibrated to actual operating data in real-time. This involves accounting for relationships between process variables that are defined through Machine Learning in combination with the fundamental understanding from mass and energy balances, as well as 2D and 3D modelling tools, such as CFD and DEM.

Operating improvements have the potential to be extended plant-wide through a smelter-wide process controller that incorporates how each unit operation impacts the downstream process. Such capabilities enable the operator to manage: custom feeds; ore blending; operating windows to manage asset health; upstream upsets by adjusting downstream processes to meet product targets and environmental limits (e.g. emissions limits, impurity levels in slag ).

Maintain

Optimized downtime can only be achieved with a maintenance strategy that includes predictive maintenance . To be successful, the development of predictive maintenance strategies demands the capability to infer the current condition of an asset with sufficient confidence. This requires sufficient data (in both quality and quantity), as well as sound engineering judgement for its analysis and interpretation.

A number of sensing technologies, as shown in Fig. 7, are available for the acquisition of sufficient data for current condition monitoring. Predictive capacity is built by combining current and historical data, with thermal, structural and process models.

../images/468727_1_En_40_Chapter/468727_1_En_40_Fig7_HTML.gif — Fig. 7
Furnace integrity technologies

Sensing technologies and advanced modelling have been combined successfully to improve the accuracy of predictions for furnace condition assessments, as illustrated in the blast furnace hearth example shown in Fig. 8. Blast furnace hearth refractory is exposed to harsh conditions, including high temperatures, thermal stress, chemical attack, erosion and oxidation . Understanding the extent of refractory loss and the nature of the protective skull is important to avoid expensive repairs and optimize production while extending campaign life. By combining 3D modeling analysis with the available thermocouple data, a more representative model of refractory lining condition can be built, in comparison with the 1D approach more widely employed. Limitations to modeling , such as the impact of cracks and thermocouple data reliability, that would normally affect lining thickness predictions, are addressed by incorporating data from Acousto Ultrasonic-Echo (AU-E) measurements performed on the operating furnace [21, 22]. Incorporating knowledge of blast furnace operations, the 3D model supports the identification of key process parameters affecting wear.

../images/468727_1_En_40_Chapter/468727_1_En_40_Fig8_HTML.gif — Fig. 8
Blast furnace hearth profile developed with combination of thermocouple data and AU-E

Capability for real-time condition monitoring and continuous indications of asset health form the basis for actions to extend life through proactive maintenance . These capabilities will be built on advanced sensing and modelling technologies, with physical inspections performed as required for validation and calibration of digital models.

Looking Toward the Smelter of the Future

For over 60 years, Hatch has made continuous improvements to the design of smelting equipment and associated process and control technologies. However, the time is right for a step change in improvement driven by advances in data analysis and computing. Hatch foresees a Smelter of the Future in which digital platforms host industrial applications that are used to provide real-time condition monitoring visualized as a furnace ‘digital twin’, advanced analytics and process control , and predictive maintenance tools. These applications would be built upon the subject matter expertise and furnace design experience that have so far provided the foundation for Hatch Furnace Technologies. Hatch ’s vision of a Smelter of the Future is being pursued through a holistic program of improvement, referred to as Smelter 4.0 , that is underpinned by the effective use of data to drive decisions. The philosophy that governs this program is represented in Fig. 9.

../images/468727_1_En_40_Chapter/468727_1_En_40_Fig9_HTML.gif — Fig. 9
Smelter 4.0 foundation, improvement pillars and overall goals. Initial focus on asset reliability

Smelter 4.0 aims to meet the objectives and challenges faced by smelting operations today; including Optimizing Operations to counter higher labour and energy costs, Effective Resource Utilization by processing a greater variety of feed materials and valorising previously discarded byproducts, and lifting the level of Social Acceptance by improving safety and environmental performance, while maximizing Capital Efficiency (return on initial and sustaining investment). These objectives comprise the goals of Smelter 4.0 , as illustrated in Fig. 9.

Smelter 4.0 is built upon three foundational areas; Work Processes, Technology and People, all of which must be advanced simultaneously to achieve sustainable results and the overall goals of the program. Work Processes refer to any prescriptive or structured work process, including operating and maintenance practices, planning processes and the definition of key performance indicators that promote alignment throughout an organization. Technology refers to the actual equipment or assets and related technology that allow an asset to function, as well as the digital technology that makes data available and accessible to drive decisions. Lastly, the program also depends on the People who are responsible for the assets, and recognizes that change management will be critical to the success of any improved or new work processes, including the adoption of new technology .

The optimization of smelter uptime through improvements to Asset Reliability is a pillar of the program and is the initial focus, as described in this paper. However, Smelter 4.0 is designed to advance and consolidate improvements in four key areas, all working toward a fully optimized operation, which include:

Zero Harm—Programs and practices to ensure zero harm to people, equipment and the environment
Ore -to-Metal Optimization —Plant-wide prescriptive process models to optimize operations across the entire value chain with rapid response capability for varying feed inputs and prediction of impact on equipment integrity in real time
Asset Reliability—Digital asset management with a focus on enhancing design, operating and maintaining practices to achieve optimized uptime
Project Delivery—Excellence in project delivery and supply chain management

Advancements in digital technology are occurring rapidly. The Smelter 4.0 program recognizes the need to address all the areas described above. Combining each of the elements (work processes, technology, people) in the correct way is needed to ensure that the results of the program are successful and sustainable, and that the advantages available from digital technology can be fully realized.

References

1.
D Core Drafting Group: Miller, Sadnicki S, Beck M, Rimmer K (2016) Subject specific guidelines—risk assessment and management. The Institute of Asset Management, Bristol, United Kingdom
2.
Wheeler FM, Twigge-Molecey C, McLean LC (1995) Managing the risk of implementing new technologies. In: 36th mechanical working and processing conference, Baltimore, MD, USA, 1995
3.
Blasi M, Hamill P, Duncan I (2014) Operational performance improvement tools. In: 2014 international conference for study of bauxite, alumina and aluminium (ICSOBA) Conference, Zhengzhou, Henan Province, China, 2014
4.
Babiy S, Gunenkov V, Kumar S, Frieslich M, GordonY (2017) Improving operational efficiency and cost competiveness of steel plants—a case study for electric furnace melt shop and hot rolling mill. In: European steel technology and application days, Vienna, Austria, 2017
5.
Sadri A, Mirkhani K, Ying W-L (2012) Utilization of innovative NDT and condition monitoring techniques for improving maintenance strategies for process vessels. In: 9 Encuentro Internacional de Mantenedores de Plantas Mineras, Santiago, Chile, 2012
6.
Nikolovski N, Spinelli F (2011) Asset strategy development using defect elimination principles. In: ICOMS asset management conference, Sydney, Australia, 2011
7.
Jastrzebski M, Koehler T, Wallace K, Novikov N, Novikov NZB, Shevchenko D, Kryzhanovskaya N, Kapran I (2012) Resdesing of furnace no. 1 at the pobuzhsky ferronickel combine. In: Conference of metallurgists 2012, Niagara Falls, ON, CA, 2012
8.
Daenuway A, Dalvi A, Wasmund B, Solar M (1992) Development of electric furnace design and operation at P.T. INCO (Indonesia). In: International symposium on non-ferrous pyrometallurgy, Edmonton, AB, CA, 1992
9.
Southall S, Damasceno R, Zemanek A, Griebeler A, Stevens G, Emery B, Janzen J (2015) Rapid redesign and restart of the vale onca puma nickel smelter. In: Infacon XIV—the fourteenth international ferroalloys congress, Kiev, Ukraine, 2015
10.
Stober F, Voermann N, Snider B, Munsch A, Koehler T, Norval G, Belanger K, Mclachlan J, Piantino E, Carlos Queiroz J, Rodriques A, Rodriques M, Rodriques A (2015) Optimization of capital cost and smelter revenue—integrated team rebuild of Barro Alto Furnaces. In: COM 2015—the conference of metallurgists hosting AMCAA, America’s conference on aluminum alloys, Toronto, Canada, 2015
11.
Gebski P, Sadri A, Nexhip D, Krippner D, Kaur R, Ying WL, George-Kennedy D (2013) Advances in continuous monitoring of water-cooled tapblocks for pyrometallurgical furnaces. CIM J 4(1)
12.
Voermann N, Walker C, Gonzales T, Nitschke S, Veenstra R (2012) Leveraging technology to economically increase furnace and smelter production rates. In: 16th international conference on non-ferrous metals, 2012
13.
Crowe C, Gerritsen T, MacRosty R, Nitschke S (2007) Advances in furnace monitoring; instrumentation. In: COM Cu2007, Toronto, ON, CA, 2007
14.
White M, Haywood R, Ranasinghe DJ, Chen S (2015) The development and application of a CFD model of copper flash smelting. In: Eleventh international conference on CFD in the minerals and process industries, CSIRO, Melbourne, Australia, 2015
15.
White M, Haywood R, Taylor W, Chen S, Ranasinghe J, West R (2015) Improved copper flash smelting at olympic dam. In AUSIMM MetPlant conference, Perth, Australia, 2015
16.
Wesley AT, White MG (2015) A dispersion apparatus. Patent WO/2015/054739, 23 04 2015
17.
Shameli E, Venditti R, Uyeda B, Kepes A, Gerritsen T, Sadri A, Southall S (2011) Enhanced furnace feed control using radar level measurement. In: Proceeding of European metallurgical conference, Duesseldorf, Germany, 2011
18.
Stober F, Gerritsen T, Janzen J, Kepes A (2004) Developments in integrated furnace controls to enhance furnace operation and crucible integrity in shielded-arc laterite smelting. In: TMS conference—international laterite nickel symposium, Charlotte, NC, 2004
19.
Uyeda B, Sadri A, Gerritsen TA, Shameli E, Southall S, Venditti R, Kepes A (2011) Enhanced furnace feed control using radar level measurement. In: European metallurgical conference, Dusseldorf, Germany, 2011
20.
Sadri A, Shameli E, Venditti R, Kepes A, Gerritsen T, Southall S, Uyeda B (2016) Measurement of charge bank level in a metallurgical furnace. United States of America Patent US9417321B2, 16 08 2016
21.
Chomyn K, Clement A, Ghorbani H, Busser J, Cameron I, Hyde B (2017) Blast furnace hearth thermal assessment and identification of wear zones. In: AISTech, Philadelphia, PA, USA, 2017
22.
Chomyn K, Soltys C, Ghorbani H, Busser J, Cameron I, Hyde B (2017) Importance of thermocouple layout in blast furnace hearth refractory wear monitoring. In: European steel technology and application days, Vienna, Austria, 2017