Interpreting Industrial Data

Information and Understanding

Industrial machines and systems now provide access to a much greater amount of data than in the past. Temperatures, pressures, humidity levels, electrical power consumption, speeds, flow rates, machine states and many other parameters can be acquired, recorded and compared over time. This availability makes it possible to observe what happens during the production process more continuously and to retain information that can later be used to analyse machine behaviour.

The ability to measure an increasing number of variables, however, requires an equally developed ability to interpret them. A value may be correctly acquired and easily accessible, but its meaning depends on the conditions under which it was recorded, the stage of the process, the machine’s previous behaviour and, in some cases, the behaviour of other parameters at the same time. Data availability therefore represents the starting point of a process that includes contextualisation, interpretation and the use of that information within a technical or operational assessment.

This becomes particularly evident when information relating to different phenomena is collected within the same process.

Data and Objectives

In some activities carried out in the food industry, processes were observed in which numerous signals were available relating both to product processing conditions and to machine operation. Operators focused mainly on a limited set of parameters identified as critical for the product and covered in greater depth during training.

Temperature and humidity provide a significant example. In the processing of certain food products, such as pasta dough, these variables can influence product behaviour during the process and therefore require continuous monitoring. Their relevance derives directly from the characteristics of the process and from the conditions that need to be maintained for production to remain within the expected parameters.

At the same time, the machine provides information about its own operation, including electrical power consumption, component states and other operating conditions. These variables describe aspects that differ from those directly connected with product characteristics and become particularly useful when it is necessary to understand how the machine is operating under a specific condition.

Different types of data therefore coexist within the same process and respond to different needs. Some support control of product processing conditions, while others help observe machine behaviour. Their relevance depends on the technical question being addressed and on the aspect of the system that needs to be analysed.

Context in Industrial Data

Electrical power consumption helps illustrate this point. Its instantaneous value may provide limited information when considered in isolation, particularly during a production phase in which the operator’s attention is focused on product conditions. The same parameter can become more meaningful when compared with the machine’s historical behaviour, the stage of the process in which it was recorded or other variables available at the same time.

A deviation from usual behaviour may suggest the need for further investigation to determine whether the machine has experienced a different operating condition or a period of increased load. In this case, the parameter contributes to the analysis without providing, on its own, an explanation of the cause. Its informational value increases when it is compared with other known conditions.

The same principle can be applied to other variables. Interpreting a measurement may require knowing which product was being processed, which stage of the process the machine was in and whether other parameters were changing at the same time. Comparison with historical data also adds a time dimension, helping to understand whether a particular condition is occasional or has already occurred under similar circumstances.

Context therefore makes it possible to assign the data a meaning that is consistent with the phenomenon being analysed. This allows a measurement to move from simple availability towards becoming usable information.

Organising Data

Once the role of context has been recognised, it becomes necessary to consider how information is organised and made available to the people using the machine. The ability to acquire many parameters does not mean that all of them need to receive the same level of attention at the same time.

During normal operation, the operator must be able to identify clearly the variables that are relevant to process control. In the food-processing cases considered, attention was focused mainly on the parameters that were better known and recognised as priorities for the product. Other information remained available and could become more relevant at different times, for example during the analysis of an anomaly or when comparing different operating conditions.

It is therefore useful to distinguish between data that needs to be immediately readable during production and data that can be consulted when a more detailed analysis is required. This organisation makes it possible to keep potentially significant information available without unnecessarily increasing the information load during normal machine operation.

The role of the person consulting the data also influences how that information is used. An operator, a maintenance technician and a designer may access the same information with different objectives: the operator to monitor the progress of the process, the maintenance technician to investigate a machine condition, and the designer to understand system behaviour in relation to the intended functions. The information structure should support these different perspectives while maintaining a common reference to the conditions actually recorded.

Operational Experience

Data interpretation also interacts with the knowledge developed by those who use the machine every day. Continuous exposure to the process allows experienced technicians and operators to recognise behaviours and variations involving several elements at the same time, which may not always be immediately attributable to a single parameter.

An operator may observe how the product is being handled, recognise a change in the overall behaviour of the machine, or associate a current condition with situations encountered previously. This knowledge develops through repeated comparison between different operating conditions and contributes to a broader understanding of how the system behaves in practice.

The designer, on the other hand, has knowledge built around the machine’s functions, the logic according to which it was developed and the parameters intended to describe its behaviour. When these two perspectives are brought together, data can provide a common reference. A behaviour identified by the operator can be compared with historical records to determine whether recurring conditions are associated with the phenomenon; in the same way, a variation identified in the data can be interpreted by considering what was directly observed during operation.

This relationship can also contribute to preserving technical knowledge. Observations based solely on individual experience are difficult to transfer and verify over time. When they can be associated with recorded operating conditions and parameters, they become easier to compare and share, allowing more structured knowledge of machine behaviour to be developed.

Starting from the Process

The way data will be used also depends on the decisions made during the design stage. The definition of which parameters should be acquired can begin with the characteristics of the product and the process the machine is expected to perform, progressively identifying the conditions that need to be controlled and the variables required to observe them.

In a food-processing application, for example, it is necessary to understand which conditions may influence product behaviour. When temperature and humidity are critical variables, monitoring them becomes central. These requirements are complemented by those relating to machine operation, which may require monitoring power consumption, component states or other operating variables.

Parameter selection therefore follows a logic that starts with the product, considers process criticalities and the functions required of the machine, and then identifies the variables that make it possible to observe those functions during operation. This approach helps clarify the purpose of the information being collected and establish under which conditions it should be observed or compared.

Defining the relationship between a variable and the phenomenon it is intended to describe also makes it easier to establish which information requires constant visibility during production and which can primarily be used for subsequent analysis. Data design therefore becomes connected with the way the machine will actually be used.

Data and Decisions

When parameters are defined and organised according to this logic, monitoring can support more informed operational and technical assessments. The instantaneous value of a variable represents one part of the information, while comparison with expected conditions, historical data and other parameters provides a better understanding of the situation in which that measurement was acquired.

When processing a product that is sensitive to environmental conditions, a significant variation in temperature or humidity from the expected range may lead to a review of process conditions and an assessment of whether intervention is required. On the machine side, a variation in power consumption compared with historical behaviour may indicate that further technical investigation is appropriate, particularly when it coincides with other variations or with a specific operating phase.

In both cases, the data supports a decision because there is a reference against which it can be interpreted. This reference may come from the parameters defined for the process, the machine’s historical behaviour or the relationship with other variables. Monitoring therefore becomes more useful when it makes these relationships reconstructable and accessible to the people responsible for assessing the condition being observed.

The same approach can guide the structure of interfaces and supervisory systems. Information required during production can be organised differently from information used for maintenance checks or subsequent technical analysis, even when it comes from the same underlying data set. The difference lies in the purpose for which the information is consulted and in the level of detail required.

From Measurement to Understanding

The increasing availability of sensors and supervisory systems expands the amount of industrial data that can be obtained from machines and production processes. To use this availability coherently, however, it is necessary to consider the entire path through which a measurement becomes part of the system’s real operating context.

The characteristics of the product and the process make it possible to identify the conditions that need to be controlled; the functions of the machine help define the variables that need to be observed; the operating context and historical data allow variations to be interpreted; and people’s experience adds further elements for understanding behaviours that emerge during operation. These steps belong to the same process of building information and make it possible to assign data a specific role within technical assessments.

The availability of a parameter therefore makes it possible to observe one part of the system’s operation. Connecting it with the conditions under which it was acquired, comparing it with meaningful references and making it accessible to the people responsible for interpreting it allow that parameter to be used with greater awareness. In this way, collected data can contribute to understanding the behaviour of both the machine and the process and provide concrete support for the operational decisions that follow.