Pulping Process Analysis with Dynamic Delay Compensation

Size: px
Start display at page:

Download "Pulping Process Analysis with Dynamic Delay Compensation"

Transcription

1 Pulping Process Analysis with Dynamic Delay Compensation Mika Suojärvi: Product Manager, Savcor Oy, FI Mikkeli, Finland, Rodrigo Prado: Project Manager, Savcor Forest, São José dos Campos SP Jarmo Kahala: Vice President, Process Analysis & Biorefineries, Savcor Oy, Abstract This article summarizes the implementation of an online process management, analysis and optimization at pulp and paper companies, within the PIMS and MES philosophy, for process indicator implementation, understanding of variables, and their cause-effect relationship in the process, as well as the achievement of operational benchmarking. The process analysis, by means of statistical tools, allied to mathematical modeling algorithms used in Wedge, makes it possible to identify the relevant information for process and quality improvement with regard to efficiency, equipment maintenance, productivity, and uniformity of the end product quality, reduction of unwanted shutdowns, and quicker resolution of process problems. Keywords: Optimization, Process indicators, Benchmarking, Continuous improvement, On-line monitoring Introduction The profile of the present economic scenario, associated with the continuous pursuit of cost reduction in the pulp and paper companies, result in the need for a better understanding of the mechanism according which the process variables related to the increase in productivity behave. The latter should have easy access and manipulation for the whole technical and managerial body of the companies. Different existing methods of information management and operational follow-up are based on specific software for database construction and process monitoring. In parallel therewith, there is the need for statistical tool application, aiming to better understand pertinent transformation to the dynamics involved in the processes, with greater interest in those of online follow-up. A tool that in addition to information management makes it also possible for the technical and operational body of the company to better understand the process variables, establishing their cause-effect relationship in real time, has been up to now a step to be concluded. The case presented herein demonstrates the application of such a tool. Savcor Wedge process analysis software was developed to detect and analyze disturbances in any industrial process, by using advanced mathematical and statistical methods of online process analysis and follow-up. The analyses make it possible to better understand the process and quality variables, as well as to put into practice improvement measures aiming at efficiency, electric maintenance, productivity, and uniformity of the end product quality, reduction of unwanted shutdowns, and faster resolution of process problems. WEDGE methods of analysis are divided into two categories: - Those finding variations from the database of process variables. - Those finding the cause-effect correlations, among which: SPC, PCA, MAR, and FFT, based on multivariable analysis. The case described in the present article demonstrates direct Wedge application to Kappa and Brightness investigation as well as the results obtained by applying the delay tool as online monitoring technology.

2 Purpose By means of the online process management tool, to identify and quantify the process variables, identify disturbances and generators of anomalies occurred in the addressed areas, providing a feedback for benchmark and process optimization. Development The communication was established through an interface between Wedge server and the mill database, by means of an OPC-HDA server for online process and laboratory follow-up. The diagram below shows the information management: Long process delays (lags) of the pulping process considerably complicate the process analysis and troubleshooting. A typical example is when a process engineer wants to find the root cause for some pulp quality problem like suddenly happened pulp brightness drop, but the process delays hide the relevant correlations between the measurements. In this kind of study the first problem is nowadays caused by the huge amounts of data in the mill s databases. An efficient tool to acquire all relevant data from the mill s databases for process analysis should be available. The next problem is that the data typically have some outlier values. These bad values should be removed before any further analysis; otherwise the analysis results will be useless. After these tasks the real process data analysis can be performed.

3 This is a good starting point but still long process delays will cause errors in the analysis results. The process delay in a pulp mill s fiber line can easily be hours or even longer. If some measurements are selected from the beginning of the process and compared to some end product quality measurement the real cause-effect relations can be hidden by the process delay. To tackle the process delay problem a dynamic process delay compensation tool has been developed for Wedge process diagnostics system. With the delay tool a process delay model can easily be created and utilized in troubleshooting and process diagnostics. The delay compensation is done dynamically meaning that the compensation varies based on the production speed and the tower levels. After delay compensation all measurements are aligned and can be assumed to have been measured from the same point of the process from the process analysis point of view. The correct root cause of the problem can be found more easily. Case study: Quality problem Brightness drop of the ready pulp! What is causing the drop? Long process delays complicates the analysis After process delay compensation it is possible to compare all process measurements to the quality measurement Easier to find the root cause of the quality problem Kappa variations Original measurements in a trend window Blow off kappa & Oxygen stage kappa Process delay can be noticed Figure 1.: Kappa and Oxigen feed normal process trend Correlation between the original measurements Blow off kappa Oxygen stage kappa

4 Figure 2.: Kappa and Oxigen feed normal process correlation grade Delay compensation Delay model is configured to Savcor Wedge Figure 3.: Fiber Line process diagram with process delay configuration Process delays are compensated The same kappa measurements in a trend window again

5 Figure 4.: Kappa and Oxigen feed process delay trend Process delays are compensated More reliable results Better correlation between the measurements Easier to find the root cause of a problem Figure 5.: Kappa and Oxigen feed process delay correlation

6 Case study: Effect of poor chip quality and Energy optmization Incoming chips can be tracked through the process For example bad quality chips or raw material change can be marked with background color Figure 6.: Chips feeding to digester (1 Pine, 0 Eucaliptus) Wedge can calculate when the marked chips are in each process stage The effect of poor raw material quality can be taken into account in control actions Figure 7.: Delay process trend compensation showing time diference between the stages

7 The process is followed up by process measurements with automatic updating according to the mill database. The links provide direct access to the variable trends, allowing the operators to follow up any disturbances in the process, either by dashboards or by alarms made available on the follow-up screen itself. The well-known process oscillations (different grades, shutdowns), which make it difficult to conduct a precise analysis of the cause-effect relationship of the process, can be easily dealt with through the software, which separates the desired process range to be analyzed, as per function of the focus of the analysis. A quite appropriate resource to analyze the cause of the process oscillation can be obtained through the best correlation tool, where, after filters are applied, the system instantaneously correlates all variables by a matrix-based method, reporting to the user the best way of optimizing the process in question. Other tools are also available, such as histograms, same shape wave identification, correlation matrix, statistics, and the possibility of creating sensors calculated on the basis of MATLAB language. Figure 8.: Process dashboards regarding on-line process dignostics Figure 9.: Best correlation tool in order to find instantaneous correlation grade beetween variables.

8 Conclusions The online process management software showed to be a rather useful process follow-up tool. By means of continuous use of Wedge and changes in the process behavior, corrective actions could be identified and quickly performed, allowing the mill to act in real time on the disturbances. It represents an applicable tool: 1. At the research center, for product and process improvement; 2. At the office, for process monitoring, optimization, and problems solution; 3. In the control room, for process monitoring and optimization. Statistical and problem tracing tools are extremely important to identify sources of problem in process. The use of ranking the major disturbance generators allows the mill to evaluate changes for process optimization and to act more dynamically to solve such disturbances. Benefits for proper process delay compensation in process analysis Long process delays cause problems in process analysis Wedge has easy-to-use and intuitive tools to compensate the delays After the compensations more reliable results can be achieved from analysis Quicker troubleshooting Poor raw material can be followed through the process Effects to the quality can be minimized Channeling cases can be detected Improved process efficiency