Business Intelligence - from measurement data to smart actions
In recent times companies have shown a growing interest in generating and using measurement data to streamline and develop their operations. Business intelligence has, in fact, become a key success factor in this regard. In order to maintain competitiveness and gain competitive advantages, data use has to be integrated into all organisational levels. Organisations that are able to collect, analyse and process key performance-related data into information that supports smart decision making, can gain competitive advantages.
Rapid developments in measurement and data transfer technologies have enabled the processing of large masses of data. With the help of the right tools, data analyses can be conducted quickly and cost efficiently. This paves the way for continuous monitoring of operations, rapid problem solving, and the prediction of actions required.
In some cases, large amounts of measurement data is already available but is only partially exploited due to a lack of resources. In such cases, well planned and implemented data analyses can produce excellent results. The measurement data can also be deficient or fragmented in several different systems, in which case its wholesale utilisation is demanding.
Different production problems often provide the trigger to start benefitting from data analysis. These include, for example, fluctuations in production costs, or efficiency per production type/shift, or for some other unknown reason.
Digital transformation is driving change
Digital transformation in our societies is taking place and is accelerating continuously (see Figure 1). Large changes are also looming in business structures and familiar operational models. It is time to take advantage of the possibilities offered by digitisation. If we don’t, our competitiveness will weaken as our competitors bring efficiencies to their operations.
In essence, competitiveness and a competitive edge can be achieved by targeting three areas: customer experience, business models and operational processes. These three areas guide and integrate the digital capability to do things better, which includes measurement and data analysis.
By combining management skills with business intelligence, we can improve decision making in continuous process and business development. At the same time, we allow customers and users to see and experience the effects of decision making better.
The understanding of the above-mentioned factors starts with data collection, but analysis and transforming information into solutions are at the core of maintaining competitiveness. Diagram 1 presents this process in its simplest form.
Developing data collection and processing data into information
The data analysis process in companies can be divided into the following three phases:
- Measurement and monitoring development (planning)
- Data collection development and system infrastructure
- Data processing and analysis
In the measurement and monitoring development phase, it is important to gain an understanding of what data is already available, how reliable is it, and what kind of measurement data would be useful for analysis. This development phase is very practical and includes, for example, going through and updating measurement instrumentation lists.
In addition, new measurement devices are dimensioned and decisions are taken regarding the device locations, as well as the formats in which the measurement data is transferred to systems or databases. All procurements are evaluated in relation to needs and goals in order to minimise the investment cost.
In the data collection development and system infrastructure phase, the systems and databases that host data are selected. Such systems may already be in place, or they can be set up if required. A key factor here is that the data is collected and stored in places where it is easily accessible for analysis. This can be done, for example, with a centralised database that collects information for analyses from several different systems (see Figure 2). These different systems include, inter alia, building automation process automation systems, as well as utility control systems (e.g. compressed air systems, cooling systems etc.).
The data processing and analysis phase is the most important of the entire process. It is here where essential information from the large data mass is summarised and processed into usable formats. In practice, data has to be handled in many different ways, such as sorting, filtering, calculating refinements and taking time delays into account. In this phase, it is crucial to analyse the interdependence of measurement variables at different operating points.
Root cause analysis can also be used to produce useful information for process development and prediction needs. These kinds of analyses are often very difficult to fully automate. It is important to generate concrete measures that can be implemented. This means that the analyst also requires know-how about the process and its operation.
Benefitting from results and implementation
A company’s management group is responsible for making use of data analysis results and implementing the recommended measures. It is, therefore, important that recommendations based on analysis results contain sufficiently detailed supporting reasoning so that management can justify trying new operational models or make investments.
Once measures have been implemented, it is important that results can be authenticated with data. Thus, different measurement periods that are viable for comparison need to be identified. In these measurement periods, for example, the time of year, production quantity and production type can be similar.
Root-cause analysis may lead to observations that can help prediction of the cost-efficiency and maintenance needs of production processes. These predictions can be used to ease the definition of different instruction and alarm limits for process control systems.
Significant savings can be achieved
Data analysis can identify significant savings measures. Recently, at a chemical plant in Finland, for example, data analysis was used to identify measures to improve energy efficiency. The plant’s energy consumption data was analysed alongside measurement data from production and peripheral processes. The analysis identified optimisation targets in process cooling and heating systems, as well as the potential to benefit from heat recovery. The recommended measures are currently being implemented and the plant is on the verge of reducing its energy consumption by about 10%, resulting in annual savings of approximately 120,000 euros.
It needs to be noted that the implementation of data analysis does not necessarily require the procurement of new analysis applications. Existing applications can, in some cases, be utilised. If suitable applications are not available, the use of a remote connection is an alternative. There are several operating models in this approach, for example, analyses and reports that are conducted periodically, or checks and analyses that are conducted when a given indicator is exceeded.
This kind of approach lends itself to cost-efficient implementation and enables quick access to experts when required. The whole data analysis system can also be integrated with management systems such as ISO 14001 or ISO 50001, thereby infusing it into the company’s daily operations.
With the help of data analysis, process measurement information can be effectively used to streamline and improve business operations. Data processing identifies problem areas in production processes and can be used to develop production efficiency. The results can be used to boost production process models and enhance the predictability of the process from maintenance and cost efficiency perspectives.
Now is the time to take advantage of the possibilities offered by the digital age. Failure to do so could hand hard-earned competitive advantages to competitors that have sensed these opportunities better.
Authors: Teemu Turunen and Jussi Jääskeläinen