Ilkka Korhonen

Researcher

Lappeenranta U niversity of

Technology

ilkka.korhonen@lut.fi

Finland

The operational environment of a production plant can be described in several ways and it can be modelled as a multi-level structure. From the process data generation, utilization and reporting perspective, it can be modelled as concentric circles (Figure 1).

The physical and chemical phenomena creating the process events and the process itself are in the centre of the model. The process is harnessed to produce the final products and the physical and chemical conditions must be controlled and managed in order to run the process and produce the desired quality and quantity of final products.

In general, the processes are well identified and known. However this is not always the case and sometimes different process phenomena can confuse, disturb and even block out each other and in addition, some of them cannot even be directly measured. As a consequence, these conditions must be defined and measured by indirect methods by calculating the value of the unknown parameter from other process information.

The component being measured can be strong or almost undetectable regardless of the mass flows in the process. Typical highvolume processes are the combustion process in energy production, paper making, oil refining and concentration processes in mining industry. In such processes the differences between the main flows and component to be measured are huge. For example in a combustion process the concentrations of trace metals and other small contaminants are measured at ppm (parts-per-million) or ppb (parts-per-billion) levels in flue gas flows of millions of cubic meters.

Sensors

The chemical, physical and other process conditions are measured with different types of sensors for process control and reporting purposes. In a mid-sized process the number of measuring points, with direct or indirect sensors, can be several thousands (Figure 2). The life cycle costs of sensors, including sensor selections, purchasing, installation, calibrations and recycling is between 3–5 % of the budgetary and operational costs of a process plant.

Click Image to Enlarge

Figure 1. In the focus of measurement technologies, thera are physical and chemical phenomena.

Figure 2. In the big factory or power plant, there can be thousands of sensors.

Figure 3. Slagging and dust can cover boiler walls with tight and thick layers (fig. E. Vakkilainen).

Click Image to Enlarge

Figure 4. Data from sensors is needed for many purposes.

Measuring can be very difficult because of the challenges of the basic measurement and the often very harsh process conditions. For example the temperature in a power plant’s combustion chamber can be over +1000 °C and in blast furnaces it is even higher. Aggressive chemical processes inside combustion chambers cause corrosion, grinding sand and fuel flows, especially in circulating fluidizing boilers (CFB), cause erosion to the boiler itself and to the sensors. The sensors must be designed to withstand erosion and the sensor structure must not disturb the mass flows inside CFB boilers in order to avoid extra erosion.

Slagging (Figure 3) on boiler surfaces can totally inhibit the operation of the sensors and any layers of slag and soot must be removed to ensure reliablemeasurements and to avoid sensor damage.

Sensors produce valuable information of the process conditions and the information-users can be divided basically in three categories shown in figure 4: ONLINE NEEDS, INTERNAL REPORTING and EXTERNAL REPORTING.

Online Needs

Online needs means “real-time” usage of the sensor information. The cycle time to read the sensor data varies from 10 to 100 milliseconds in automatic process control functions, and from some tens of seconds to some minutes in manual operations. The measurement information can be used directly from the sensing element or it can be processed and converted to a standardized signal with a transmitter. In the following text the term sensor means a sensor or a combination of a sensor and a transmitter.

Most of the measurement information is used as input to the distributed process control system (DCS) either as standard analog signal, typically 4–20 mA, or digital signal using standardized or proprietary fieldbus technologies. The analog mA signal has been in use for many decades and it is still the most common way of transmitting information from sensors to DCS and from DCS to field actuators

In critical processes with a potential safety risk to personnel, environment or assets, measurement information is needed for the Safety Instrumented Systems (SIS) which are responsible for the safe control of the process if dangerous conditions occur as a result of operation mistake, process equipment or process control system failures. Dedicated sensors must be used for the measurements to ensure fully independent operation of the safety systems during failures and the use of specific digital field busses has recently been approved in these applications.

Condition monitoring and diagnostics of sensors, DCS and process equipment is done by separate stand-alone software applications, which are most preferably running in dedicated computers. The updating intervals for condition monitoring and diagnostic purposes can be longer and the amount of sensor data can be less than for process stand control purposes. Condition monitoring and diagnostic information from the sensors is generally transmitted as a digital HART signal over the same wires as the analog 4–20 mA measurement signal.

Wireless HART has recently been introduced as a much more flexible and cost effective technology than the conventional wired HART. Condition monitoring and diagnostic system require a multiplexing method to read the HART data from one sensor at a time. Wired HART with multiplexer systems are quite complicated, require a lot of space from the cross connection rooms and are quite expensive to implement (figure 5). Wireless HART devices act inherently also as multiplexers and it is obvious that this functionality will be the key driver of Wireless HART’s penetration to the market in the near future.

Measurement information is used increasingly also in remote control and monitoring applications where the sensor data must be transmitted to remote operating points which may be thousands of kilometers apart. In these cases the data is converted to some suitable standard format for a wired or wireless modem (transceiver), wired connections can be based on wired telephone lines or LAN / WAN networks, wireless systems typically rely on cellular networks (GSM, GPRS, 3G, etc.). Also, dedicated wireless systems are used which are typically operating at license free frequency bands at about 433 MHz and 868 MHz in Europe, 915 MHz in USA or 2,4 GHz worldwide.

Figure 5. HART multiplexer for 32 channels (Fig. H. Leppänen).

The sensor data can also be “recycled”, which means the use of the sensor output as an input to so called soft sensors. These are software applications using mathematical methods to calculate new “sensor level information” combining data from various sensors, DCS and other plant information. The mathematical functions can be anything from a simple operation to advanced modelling and simulation methods. Soft sensor solutions have greatly improved the visibility of process behaviour and the controllability of plant operations. Soft sensor solutions can be easily implemented and taken in to use as stand-alone software packages running either in the DCS or in dedicated PC’s with LAN connection to the DCS.

The data communication between online sensors and systems is taken care of by special automation busses or LAN networks which are isolated from the factory’s office networks. The basic LAN type for process usage is typically either standard or strengthened industrial Ethernet. If different LAN applications are not compatible with each other, gateway units are needed to adapt the protocols and physical signal properties.

Internal Reporting

Compressed and formulated sensor data is used for internal reporting. The first task is to produce reports to people who are controlling the processes and are responsible for the plant operations. These people need daily, weekly, monthly, production period, annual etc. reports which are created from the data from DCS and condition monitoring systems and indirectly from history databases and log files. Logistic operations are responsible for the material transportation and storage and they are also steered by reports to optimize the size of transportations, inventories and warehouse personnel.

Plant quality control follows the qualitative and quantitative parameters of the plant production to be able to give suggestions for qualitative improvements in all levels of plant operations. Improving the overall quality has a major impact on the future of the plant, quality being the key factor to keep the existing customers satisfied and getting new ones.

Quality control needs to make more sample based measurements and analyses in addition to the normal process measurements. The samples taken from the process for quality control analyses are more and more often prepared and measured by robots and automatic analyzing systems but manual methods are still part of the daily routines in many process plants. Improving the quality and product features requires a lot of research and development work (R&D) based on process modelling. This requires a lot of raw process data, difficult special measurements and a deep understanding of process behaviour.

Many people working in the factories do not think of the fact that additional production bonuses need basic process level measurements. In order to determine these bonuses, exact data of production volumes and quality per shift, per day, per month or per year is needed.

Last but not least, the plant’s financial management needs different kinds of reports of the processes and the total operation of the plant for both short and long term planning of the plant operations.

Factory networks, which are isolated from the outside world by firewalls and other security methods, are taking care of the communication between the process systems, internal users, ERP (Enterprise Resource Planning) and other automatic data management solutions used to store and utilize all the internal data.

External Reporting

The term External Reporting includes all compulsory reporting about the values, strategies, business and everyday operations of the process plant, company or corporation to the community. One of the most important connections between process measurements and external reporting are the local requirements for environmental measurements covering everything from flue gas emissions and VOC (Volume of Organic Compounds) to waste water emissions and solid waste dumping to landfill. The authorities must have direct access to the measured emission volumes, measurement procedures and measuring equipment to guarantee the credibility of the measurement results. Emission measurements are also utilized in the process plant as they offer information of the used raw materials and process operations and give valuable feedback for process control purposes.

Measurements have an effect at all levels of process plants, companies and corporations. Although the physical or chemical phenomena and process behaviours are normally very complicated, they still must be measured with adequate resolution and accuracy – they are the foundation for safe and optimal process operation.

»»Who is the author? Ilkka Korhonen has done studies in many phases: ››1978 Engineer in Measurement and Control techniques. ››1995 B.Sc. in Telecommunication Systems ››2006 M.Sc. in Computer Sciences Ilkka has about 30 years work experience in I ndustrial Automation and Telecommunication segments in engineering, project managing, management of engineer team and business development. Nowadays he is working for Lappeenranta U niversity of Technology (www.lut.fi) in Varkaus as researcher.