Font Size: a A A

Research On Soft-sensor Methods Of Biomass In Fermentation Process

Posted on:2008-06-03Degree:MasterType:Thesis
Country:ChinaCandidate:R L DuFull Text:PDF
GTID:2178360218952875Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Fermentation process is strong nonlinear and time varying and its inherence mechanism is very complex. It's difficultly to measure some important process variables online, which made the modeling and controlling more complexity and difficultly. To implement optimization and automatic control performance of fermentation system and further enhance the quality and yield of the output, the research on soft sensor methods of biomass in fermentation process and the implementation are discussed and studied primarily.Based on analysis the present status of biomass soft sensor methods, the biomass soft sensor model is established by studying the combination particle swarm optimization algorithm, neural network and the penicillium fermentation process. On the base of the model, the performances of optimization neural network are researched by particle swarm optimization algorithm, quantum-behaved particle swarm optimization algorithm, multi-velocity particle swarm optimization algorithm and BP algorithm. The experiment results indicates that this model features small training errors, high learning speed, well generalization ability, high estimation precision by simulation experiment of making use of penicillium ferment process data set. The Recurrent Neural Network model is established by taking model output as model input of next time to achieve the multi-step pre-estimate of fermentation process. The experiment results indicates that the model of multi-step pre-estimate feature high precise. On base of the multi-step pre-estimate model, both the control trajectories of pH and the temperature of ferment process are optimized with QPSO algorithm. The experiment indicated that optimized temperature and pH control trajectories enhanced the output of penicillium fermentation.The biomass soft-sensor method is applied in engineering project to complete the computer control system of 50L fermentation tank in biological laboratory. The system hardware is made of Industrial computer and PCI multi-function data access card etc. The software partially is developed by Borland C++ builder 6. Each item of the system meets the design request. The graphically handling interface is flexibly and conveniently. This system has high performance price ratio. The system provide the conditions to further study the Optimized control of fermentation process.
Keywords/Search Tags:fermentation process, biomass, soft sensor, modeling, optimization
PDF Full Text Request
Related items