DESPRIMA: Demand-Side- und Produktions-Management für Getränkeabfüllprozesse

Problem Formulation

The increasing consumption of electricity on one side, and the global warming and the limitations of energy resources, e.g. fossil fuels and gas, on the other side, challenge electricity production in the close future. One solution is to use renewable energy resources as a substitution for the classic methods. However, due to the nature of this kind of energy resources, it important that the consuming loads provide a degree of flexibility for the renewable-based power distribution grid. The flexibility potentials are not only found in the beverage industry, but also in various production plants. The systems that are developed in DESPRIMA should therefore be transferable to various production plants. In this way, industrial companies can actively cooperate with grid operators and promote the grid integration of renewable energies. Their advantage: controllable loads are remunerated by the grid operator. Power demand and power availability are thus optimally matched through cooperation between industry and grid operators. In this context, the tools of control technology come into play, linking the production management system and the demand-side management system. An essential prerequisite for this is to digitize the production lines and create communication interfaces between the production plants and the grid operator.

DESPRIMA architecture

A decisive contribution to DSM is to be made by making the production chain more flexible: If production planning is included, individual plants can run at lower load as soon as power bottlenecks occur. This means that the grid operator has to provide less control energy, which prevents rising grid costs. If, on the other hand, a lot of electricity is available, the plant can ramp up all production steps to full capacity. The beverage bottler can offer the adjustment of electricity consumption to the network operator as a so-called system service. In both cases, electricity demand and availability are balanced and network stability is guaranteed.

As shown in the following, two kinds of contracts are considered for a beverage factory; one is the contract between the factory and the grid side operator (“e-contract”), and the other one is the contract between the beverage enterprise and the customer (“p-contract”). In the first case, the producing enterprise is charged for the consumption of the electrical energy by the machines, which is varying with time, either on a scale of hours in a day or on the scale of seasons of a year. The grid operator provides a reference power consumption profile for the factory one day ahead. Tracking this reference power profile is important because the grid operator computes them based on stability and economy analysis of the whole electrical grid. In the second case, the enterprise is responsible for the delivery of a predefined number of various beverages (“production”). Typically, weekly production plans are devised based on customer-specific short, middle and long-term contracts (“p-contracts”).

A description from an industrial demand-side management perspective can be found here.

Solution Approach

To run the machines in the beverage factory, MPC is used here as the centralized controller to provide the machines internal controllers’ setpoints. This DR-based optimal control strategy is proposed not only to react to the electrical grid demands, but also to satisfy the bottle production requirements. The algorithm provides flexibility for the grid by modifying the power consumption of the line that should ideally track a day-ahead profile. The ancillary service requests from the electrical power network contains different time scales which are referred as primary, secondary and tertiary services. The primary service helps the grid to keep the frequency stability and in order to respond to it, it is necessary to have the models with fast dynamics in order of seconds in the MPC. However, the secondary and tertiary services do not need to react as fast as the primary one. Therefore, a more generic model called abstract model is introduced as an internal model for MPC.  Having the one-day ahead power profile, the production planning of lines, and the intra-day demand requested by the energy reseller, an optimization problem is solved to find the set-points of the machines and the bottle inflow rate of the lines.

The proposed algorithms adjust the speeds ωref of the production machines, including the conveyor belts to optimally react to the demands originating from the grid side. At the same time, it is vital to take care of the production goals arising from the customer contracts and production planning bref. Two kinds of contracts for a beverage factory are considered: one is the contract between the factory and the grid side operator and the other one is the contract between the beverage enterprise and the customer. The grid operator provides a day-ahead reference power consumption profile pref,1d. The proposed scheme for DR is depicted in the left figure. During a day, typically every 15 minutes, there might arise requests from the grid side to deviate from the day- ahead profile. This “DR service request” may arise from the market or grid situation, claiming flexibility in energy consumption from the enterprises which amounts to a modified power profile. This invokes a feedback path “system service” in the above figure which is balanced by financial incentives. Such a balance in conjunction with the adaptation of the production plans is managed by predictive control algorithms (MPC).
 

Project Goals

  • Increasing the profit of the production line respecting the production plan and the time varying one day-ahead power profile provided by the grid
  • Implementation of the intra-day ancillary service requested by the grid
  • Designing a generic MPC controller to find the best way of production regarding these two controversial goals
  • Guaranteeing the stability of the grid for the future power distribution networks
  • Providing a library of machine models for different time scales
  • Providing the process flexibility for participation in energy spot market
     

Solutions

Keywords

  • Demand side management
  • Model predictive control (MPC)
  • Time delay system
  • Production planning
     

Funding

Time span

July 2019 - March 2023

 

Project Partners

SWW Wunsiedel GmbH 
TU Kaiserslautern 
Fraunhofer ITWM 
Dresden Elektronik Ingenieurtechnik GmbH 
Software AG 
University of Bremen 
 

Contact

Prof. Dr.-Ing. Naim Bajcinca
Gottlieb-Daimler-Str. 42
67663, Kaiserslautern
+49 (0)631/205-3230
naim.bajcinca(at)mv.uni-kl.de