Digitalization in the operation, monitoring and control of large-scale biogas plants

Biokraft is the owner and operator of the world's largest facility for liquid biogas, located at Skogn in Trøndelag. The Skogn plant started up in summer 2018 and reached planned production capacity within a few months. The basic idea for the project is to utilize newly developed sensor technology, new analysis methods and computer models to improve the operation, monitoring and control of large-scale biogas plants. The following results have been achieved during the project period: - Installation and testing of on-line analysis of volatile fatty acids (VFA) and total organic carbon (TOC) in connection with a bioreactor that processes water from Norske Skog. The aim was to investigate if on-line instrumentation can reduce the risk of overloading the reactors due to operational disturbances in the feed. Online VFA measurement is considered a useful tool for monitoring and managing the ECSB process, but the utility must be assessed against the cost of analyzer maintenance. - The relationship between TOC and chemical oxygen demand (COD) was mapped through a measurement campaign. The results provided valuable information about how the relationship between TOC and COD varies over time, and to what extent an online measurement of TOC can be used for easy and quick monitoring of COD. - Work with an early warning system based on process data and process modeling started with analysis of process data from a 2-week operating period in January/February 2020. Data was analyzed using various statistical methods, including principal component analysis (PCA) to see if one can identify changes in operating patterns at the facility. Preliminary results indicate that PCA can be a useful tool for early phase prediction. Change in feed composition was recorded using PCA analysis significantly faster than the hydraulic residence time in the mixed tank reactor. Access to data from longer time series is necessary to assess the practical applicability of the method. The work on modeling the multifuel reactors has focused on expanding and improving today's state-of-the-art model "Anaerobic Digestion Model No. 1" (ADM1), among other things, to better describe the bacteria's handling of the nitrogen-containing raw material that is processed at Biokraft Skogn. Today's model framework includes the central process sections with substrate tanks, return sludge, hygienisation and the biogas reactors. - A framework for the exchange of sensor data from the plant as input to the improved ADM1 model was established and has been tested on off-line data. The framework will be able to be used in any future use of the ADM1 model as a digital twin for the multifuel reactors. In combination with data from new sensors, a goal for a digital twin will be to be used for production planning and optimization of operations based on information about varying raw material quality. Cooperation with the development environment and supplier of the process data system will be central to the introduction of the methods at facilities in operation. - Development of new sensors has consisted of a literature review of technology for VFA and H2 sensors for anaerobic fermentation. A new "colorimetric" sensor system has been developed for VFA detection. Testing shows that concentration of 3 specific VFAs can be measured with good accuracy. The relevant key components can provide information about the state of the fermenter and be used to optimize operations against available raw materials. Further development of the sensors is necessary both in terms of robustness and industrialization of the measurement methods. Cooperation with suppliers of sensors will naturally be considered in a future development project. - An optical measurement method based on near infrared spectroscopy (NIRS) has been tested for use in estimating the biogas potential (BMP) in raw materials for biogas production. Conventional determination of BMP involves time-consuming (up to 30 days) lab analyses, while NIRS analysis can be carried out within a day. Tests of the NIRS method on relevant substrates for Biokraft show potential for further development of the methods in collaboration with the research environment and equipment suppliers. - Groundbreaking work has been carried out on the analysis of how the microbial environments in the anaerobic fermenters develop over time - so-called "Microbial community management". The work is based on sample material from the multifuel fermenters collected over a 12-month period. DNA analysis of the sample material provides the opportunity to investigate which microbes thrive in the fermenters in different periods and with different substrate compositions. Combined with existing knowledge about the turnover mechanisms of biogas for the various microbes and under which conditions they produce best, this is a useful tool for increased understanding linked to seasonal variations in substrate composition.

The current project has contributed to exploring and identifying promising technologies which are relevant for increased profitability and reduced cost of biogas production. This is considered a valuable and necessary foundation for subsequent R&D projects which will be more focused on specific applications. The current consortium of Norwegian R&D partners can be extended with participants from the vendor industry who can provide the necessary competence for the commercialization of the technology ideas.

Biokraft is the owner and operator of the worlds largest plant for liquefied biogas (LBG), located at Skogn in Trøndelag. The Skogn plant started operations in summer of 2018 and has reached the planned capacity within a few months. Biokraft is currently working on expansions multiplying the plant's output based on byproducts from marine, agricultural and bio-based industries. The underlying idea of the project is to use emerging new sensor technology, new analysis methods and computer models to improve the operation, monitoring, supervision, and control of large biogas plants. This is a daunting task that needs to be addressed stepwise, over time. The following research challenges will be addressed: * Instrumentation and control of Upflow reactors, major feed disturbances will be identified by means of innovative instrumentation, and these disturbances will be counteracted by appropriate control measures. Challenges are related to whether the selected solutions will be fast and accurate enough to yield enough disturbance compensation. * Early warnings - sensor data and models, critical assessment of available process information will lead to the identification of a set of crucial process states that need to be monitored closely. How to monitor these crucial states (i.e. the mixture of physical sensors, soft sensors, and model-based predictions), as well as how to get sufficient information out of the available data, will be a key challenge * Utilization of opportunities brought by novel optical sensors and spectroscopy. The challenges are related to whether these new sensors, which are potentially fast and inexpensive, can be made robust and sensitive enough for successful field testing under real conditions in a biogas reactor. * Microbial community analysis will focus on the slow microbial shifts that will inevitably occur in a biogas reactor over time, and whether these can be predicted, and, ultimately, controlled.