Effective Key Performance Measurements in Trackless Mines
Most established maintenance operations in South African mines have a formal fleet performance measurement system. Reliability, maintainability, availability and utilization are ubiquitous indicators shared to give a pulse-check on the health of the maintenance infrastructure.
However, the trackless maintenance world is way behind in using maintenance performance indicators (MPIs) systematically to increase uptime and keep the costs in check. Some easy hacks to identify opportunities to significantly improve the maintenance systems with an existing framework are;
1. Identifying the correct "control levers" in complex multivariate maintenance systems
A maintenance strategy document which is reviewed at least annually should give direction on the control and monitoring plan to maintain the fleet. These documents are however not easy to come by and even if they exist in an organization, the day-to-day decision makers hardly get to see them. Management decisions are done through empirical knowledge of experienced and very migrant-prone personnel. However, with the dynamic nature of trackless maintenance, an emergent strategy approach is quite capable of maintaining a robust set of MPIs to keep the maintenance unit aligned with the overall business strategy.
Measuring improves the control and monitoring process, however, one should take care in not measuring an inexhaustible list of things that do not have "control levers".
Understanding leading and lagging indicators
It is important to understand the lead-lag dimension on any measures being implemented. Lagging indicators tell a story of what was done and what was achieved and are therefore output oriented. Typical indicator here would be availability or tons produced. They are usually easy to measure but hard to influence directly as they usually report final results.
Leading indicators in turn, are future-looking or predictive in nature. These indicators measure how we deal with specific maintenance inputs with the plan to improve a specified outcome/s . Examples of these are tracking the amount of equipment inspection reports conducted, the amount of training courses completed to improve the maintenance capacity and service schedule compliance rate.
When planning what KPIs to use, the levers of influence should be identified ahead of time. If a lagging indicator is out of a predetermined range for instance, one should be able to identify the underlying cause and a have a templated a response to it. As an illustration, if the repair times are too long and are causing major disruption in operation, one must have supporting and correlating MPIs that will show that either the travel distances have become too long, the labor turnover is causing a brain drain in the workshop or the logistic team needs to up the game as the spare shortages are causing long downtimes.
Free open software that can do linear regression models and time-series analysis such as Gretl can assist in establishing connections between inputs and historical results and are most useful in identifying the relevant levers.
A maintenance strategy document which is reviewed at least annually should give direction on the control and monitoring plan to maintain the fleet. These documents are however not easy to come by and even if they exist in an organization, the day-to-day decision makers hardly get to see them. Management decisions are done through empirical knowledge of experienced and very migrant-prone personnel. However, with the dynamic nature of trackless maintenance, an emergent strategy approach is quite capable of maintaining a robust set of MPIs to keep the maintenance unit aligned with the overall business strategy.
Measuring improves the control and monitoring process, however, one should take care in not measuring an inexhaustible list of things that do not have "control levers".
Understanding leading and lagging indicators
It is important to understand the lead-lag dimension on any measures being implemented. Lagging indicators tell a story of what was done and what was achieved and are therefore output oriented. Typical indicator here would be availability or tons produced. They are usually easy to measure but hard to influence directly as they usually report final results.
Leading indicators in turn, are future-looking or predictive in nature. These indicators measure how we deal with specific maintenance inputs with the plan to improve a specified outcome/s . Examples of these are tracking the amount of equipment inspection reports conducted, the amount of training courses completed to improve the maintenance capacity and service schedule compliance rate.
When planning what KPIs to use, the levers of influence should be identified ahead of time. If a lagging indicator is out of a predetermined range for instance, one should be able to identify the underlying cause and a have a templated a response to it. As an illustration, if the repair times are too long and are causing major disruption in operation, one must have supporting and correlating MPIs that will show that either the travel distances have become too long, the labor turnover is causing a brain drain in the workshop or the logistic team needs to up the game as the spare shortages are causing long downtimes.
Free open software that can do linear regression models and time-series analysis such as Gretl can assist in establishing connections between inputs and historical results and are most useful in identifying the relevant levers.
2. Develop a dedicated data support infrastructure
Handling a torrential flow of information can be quite overwhelming in an environment where quick decisions are needed to keep operations going. A good approach is to appreciate that the typical South African planning department is not particularly geared for data analysis. Setting up a dedicated department to support maintenance through data can bring a wealth of insight and opportunities that even seasoned maintenance professional will not pick up easily.
Developing dashboards that give early warnings on critical trends will go a long way instead of waiting on ad-hoc information requests. Some mines have control room reports analyzed monthly or at best weekly. This can be now analyzed in real-time from a smart phone. Maintenance information should be the main initiators of maintenance improvement projects.
Not having a dedicated and centralized data analytics role will not on leave your organization blind to emerging threats and opportunities, but it will also delay adoption of the new data driven technologies that are becoming commonplace in other industries. Data driven decision making fuels the adoption of predictive analytics, machine learning.
3. Rethink accountability and motivation
Mines try improve productivity by being generous with monthly performance bonuses. They are usually based on very broad indicators such as fleet availability, tonnage, square meters and safety incidents recorded. It is also common to find operations that have a myriad of maintenance performance indicators being tracked without having a single person being held accountable. One can improve the KPI-apathy by having a targeted approach to the indices. An example would be holding a General Engineering Superintendent to availability, segregating reliability to the planned maintenance teams, maintainability to the breakdown crews and spare parts availability to the logistics team.
Basing performance bonuses on narrow indices within someone's control is a no-brainer. It brings focus and innovation. Teams who have narrow KPIs should be encouraged to identify and report barriers that inhibit their targets in a bottom-up approach. While a breakdown crew may relocate its default waiting stations to minimize travel time to improve repair times, the service squad may initiate a peer-to-peer conversation with operators to bring equipment on time to make full use of precious maintenance time. After all, creating a culture of empowerment is most likely to succeeds than managing a maintenance team through management resolutions only.
No comments:
Post a Comment