Departamento de Ingeniería Informática, Universidad de Santiago de Chile (USACH), Av. Ecuador, 3659, Santiago, Chile,
Bioprocess Biosyst Eng. 2014 Jan;37(1):27-36. doi: 10.1007/s00449-013-0925-3. Epub 2013 Feb 22.
The lack of sensors for some relevant state variables in fermentation processes can be coped by developing appropriate software sensors. In this work, NARX-ANN, NARMAX-ANN, NARX-SVM and NARMAX-SVM models are compared when acting as software sensors of biomass concentration for a solid substrate cultivation (SSC) process. Results show that NARMAX-SVM outperforms the other models with an SMAPE index under 9 for a 20 % amplitude noise. In addition, NARMAX models perform better than NARX models under the same noise conditions because of their better predictive capabilities as they include prediction errors as inputs. In the case of perturbation of initial conditions of the autoregressive variable, NARX models exhibited better convergence capabilities. This work also confirms that a difficult to measure variable, like biomass concentration, can be estimated on-line from easy to measure variables like CO₂ and O₂ using an adequate software sensor based on computational intelligence techniques.
发酵过程中某些相关状态变量缺乏传感器，可以通过开发适当的软件传感器来解决。在这项工作中，当作为固体基质培养（SSC）过程中生物量浓度的软件传感器时，比较了 NARX-ANN、NARMAX-ANN、NARX-SVM 和 NARMAX-SVM 模型。结果表明，在 20%幅度噪声下，NARMAX-SVM 的 SMAPE 指数低于 9，优于其他模型。此外，在相同噪声条件下，NARMAX 模型比 NARX 模型表现更好，因为它们具有更好的预测能力，因为它们将预测误差作为输入。在自回归变量初始条件扰动的情况下，NARX 模型表现出更好的收敛能力。这项工作还证实，像生物量浓度这样难以测量的变量可以使用基于计算智能技术的适当软件传感器，从像 CO₂ 和 O₂ 这样易于测量的变量在线估计。
