Professor Feng Jiang's team from the School of Environmental Science and Engineering at Sun Yat-sen University has published a cover article titled "SIMPO—A Simplified Intelligent Modelling Platform Online for Code-free and Shareable Wastewater Treatment Process Modelling" in Volume 1, Issue 4 of Water & Ecology.
Water & Ecology Volume 1, Issue 4 Cover
The article introduces SIMPO, a code-free intelligent modeling and sharing platform developed by Professor Feng Jiang's team for the mathematical simulation of wastewater treatment processes. SIMPO is a free, open-source, and shareable SaaS (Software as a Service) cloud platform. It aims to provide everyone with zero-threshold access to a professional-grade wastewater treatment process modeling system. This empowers faculty and students, scientists, and engineers in the field of environmental engineering to more easily engage in learning, research, and improvement of wastewater treatment processes, thereby lowering the barriers to the global dissemination of technology. Following its release, the article has been widely reposted by international science media outlets such as EurekAlert! .
DOI: 10.1016/j.wateco.2025.100021
First Authors: Jun Wang, Hao Li
Corresponding Author: Feng Jiang
Affiliation: School of Environmental Science & Engineering, Sun Yat-sen University
Highlights
- SIMPO is an open, cloud-based, code-free, and shareable platform for wastewater treatment modelling.
- SIMPO features a drag-and-drop graphical user interfaces (GUI), adaptive solvers, and automated error-proofing.
- Advanced algorithms support sensitivity analysis, uncertainty assessment and parameter estimation.
- The platform can be widely applied in scientific research, education, engineering optimization, and peer review.
Article Summary
Wastewater treatment process models, such as the Activated Sludge Model (ASM) series, are crucial tools for research, simulation, and design in environmental engineering. However, as these models become increasingly complex, their development, application, and validation pose significant challenges for researchers, engineers, and even peer reviewers. Traditional software often suffers from steep learning curves, cumbersome operations, a lack of collaborative features, and high costs. This not only hinders the widespread adoption and innovation of these models but also restricts the application and improvement of water treatment technologies in developing countries. Addressing this critical need, the author's team has developed SIMPO— a free, zero-barrier, open, and shareable SaaS cloud platform. The platform aims to significantly lower the threshold for wastewater treatment process modeling through a code-free, drag-and-drop graphical interface, while integrating advanced algorithms to ensure the professionalism, accuracy, and reliability of the models.
Compared to traditional modeling software like AQUASIM, SIMPO offers a simpler, more convenient, and smoother modeling process. SIMPO simplifies complex modeling into intuitive drag-and-drop operations, combined with an AI-driven core for model building, achieving "zero-barrier" modeling:
- AQUASIM: Users must manually input variables, define processes, partition units, set connections, configure time steps, and edit plots within multi-level dialog boxes, resulting in a complex and non-intuitive workflow.
- SIMPO: Users only need to follow three straightforward steps: "Edit Model (BioModel)" → "Drag-and-Drop to Build Process (DataSet)" → "Automatically Generate Project and Calculate (Project)". The platform automatically handles time stepping and result plotting.
Figure 1. Comparison of modelling processes between AQUASIM and SIMPO.
SIMPO's platform is powerful, covering the entire process from input, simulation, to evaluation and sharing. Figures 2 and 3 illustrate SIMPO's core operational interfaces:
- Visual Matrix Editing: Chemical stoichiometric and rate expressions are edited in a matrix format, consistent with ASM series models. The system provides color block highlighting to effectively prevent input errors.
- Drag-and-Drop Flow Chart Design:: Users can rapidly build complex wastewater treatment process flowsheet designs by simply dragging and dropping with the mouse, much like assembling building blocks.
Figure 2. The model editing interface of SIMPO.
Figure 3. The drag-and-drop flowchart editing interface of SIMPO: (A) module addition menu; (B) add process components; (C) connection process flow; (D) fill in the flows in the connections.
To further lower the learning curve for water treatment process modeling, the team has recently developed an AI-assisted modeling function. SIMPO, by leveraging large language models such as Gemini 3, can automatically read text and diagrams from research papers to generate biochemical models (Biomodel).
The reliability of a new tool must be rigorously validated. The research team selected a classic case study to compare SIMPO's simulation results against the industry benchmark software, AQUASIM.
As shown in Figure 4, in batch and continuous reactor simulations of the biological nutrient removal model, SIMPO's simulation curves almost perfectly overlap with those of AQUASIM, achieving a weighted Nash-Sutcliffe efficiency (WNSE) above 0.999. This demonstrates that SIMPO maintains operational simplicity while matching the computational accuracy of professional software.
SIMPO's built-in automatic balance checking and visual result evaluation features help researchers identify subtle errors or deficiencies in published models that might otherwise go unnoticed.
Case 1: Detecting Errors in Stoichiometric Matrices
Figure 5 demonstrates SIMPO's performance when checking an extended ASM3 model. By inputting theoretical oxygen demand and nitrogen balance terms, the platform successfully detected three stoichiometric errors in the original literature and precisely pinpointed the specific reaction processes involved. This provides clear direction for model correction.
Case 2: Revealing Model Structural Deficiencies
Figure 6 recreates an autotrophic denitrification model. SIMPO's automatic plotting immediately detected two issues with the model:
- Elemental sulfur exhibited negative concentrations (physically unrealistic).
- The simulation of nitrite nitrogen was poor.
By introducing a simple "switch function" to correct the model structure, the negative concentration issue was successfully resolved. Simultaneously, the visual results clearly showed that the original model's poor fitting of nitrite nitrogen was masked due to plotting scale issues. This highlights the importance of automation and interactive visualization in model evaluation.
Figure 6. (A) Simulation results of the original AD model using SIMPO. The measured data points were extracted
directly from the result figure, and the initial concentration of autotrophic denitrification bacteria (ADB) was determined to be 929 mg·L–1 from the parameter estimation. The public Project：
https://dash.simpowater.org/dashboard/projects/public/name/ADwithsulphide/id/454/did/415. (B) The NO
2--N simulation result.(C) The S
0 simulation result.
For advanced analytical needs, SIMPO integrates professional algorithms such as sensitivity analysis, uncertainty assessment, and parameter estimation.
Sensitivity Analysis
In SIMPO, sensitivity analysis is used to identify sensitive and highly correlated parameters, thereby reducing the number of parameters to be evaluated in subsequent steps and improving computational efficiency.
- Through OAT (One at a Time) single-factor analysis, key parameters with the greatest impact on model output can be quickly identified.
- Correlation heatmaps help discover highly correlated parameter groups, thus reducing the number of parameters to be estimated and enhancing the efficiency of subsequent analyses.
Figure 7. (A) The outcomes of the OAT sensitivity analysis in SIMPO; (B) The weighted sum of the integration of OAT curves in SIMPO; (C) The pairwise correlation heat map in SIMPO; (D) The outcomes of the OAT sensitivity analysis in AQUASIM.不确定性评估（Uncertainty Assessment）
Uncertainty Assessment
SIMPO provides professional uncertainty assessment algorithms. Through Latin Hypercube Sampling (LHS) and Monte Carlo (MC) simulations, it can be used to evaluate how parameter uncertainty propagates to model outputs, analyze parameter and model output distributions, identify equifinality, and narrow down parameter ranges. The results of uncertainty assessment are intuitively evaluated from different angles through various visual forms, including Probability Density Function (PDF) plots, Cumulative Distribution Function (CDF) plots, Band plots, Density plots, and Skeleton plots.
Figure 8. (A) The PDF and CDF plots of the uncertainty assessment in SIMPO; (B) The Band plots; (C) The Density plots; (D) The Skeleton plots.Parameter Estimation
SIMPO employs the Genetic Algorithm (GA) for parameter estimation. SIMPO offers various computational settings, allowing for a flexible balance between computational efficiency and solution accuracy. The computational results from several different evaluation methods are presented in Table 1:
- The Large GA yields the highest accuracy but requires significant time due to its large scale, reaching a WNSE of 0.99161 in 267 seconds after approximately 1,394 iterations.
- The Small GA has a smaller computational scale, resulting in slightly lower accuracy but extremely short computation times, achieving a WNSE of 0.99019 in just 0.83 seconds after only 25 iterations.
- Uncertainty also achieves high accuracy but is the most time-consuming, reaching a WNSE of 0.99161 in 4,658 seconds after approximately 880,510 iterations.
Table 1. Comparison of the optimal results.
Compared to traditional and commercial software such as AQUASIM, BioWin, and GPS-X, SIMPO offers significant advantages in terms of usability, editability, shareability, and cost (Table 2). It is not only the first open-source, free, cloud-native SaaS platform for wastewater treatment modeling but also innovatively integrates AI tools to help users quickly understand literature models.
Table 2. Comparison between SIMPO and peer platforms.
SIMPO successfully combines the professional capabilities of wastewater treatment modeling with an ultimate user experience through its intuitive GUI, dynamic adaptive algorithms, automated error prevention mechanisms, and advanced evaluation algorithms. It significantly lowers the modeling threshold, fosters model sharing, validation, and collaboration, and provides a revolutionary tool for research, teaching, and engineering practice in environmental engineering. In the future, the team plans to enhance SIMPO with GPU acceleration, more reactor types and model libraries, and explore API integration with more AI tools, thereby promoting the fusion of mechanistic and data-driven models.
Corresponding Author: Feng Jiang, Professor and Doctoral Supervisor at the School of Environmental Science and Engineering, Sun Yat-sen University; recipient of the National Science Fund for Distinguished Young Scholars. His research primarily focuses on environmental systems modeling and optimization, urban water systems, and the prevention and resource utilization of heavy metal pollution.
First Authors :
- Jun Wang, PhD, graduated from the School of Environmental Science and Engineering, Sun Yat-sen University. His research primarily focuses on algorithm development, modeling and optimization, and data analysis.
- Hao Li, PhD candidate currently pursuing studies at the School of Environmental Science and Engineering, Sun Yat-sen University. His research primarily focuses on wastewater treatment process modeling and optimization.
Add friend, with the note “SIMPO” to join the group chat
Welcome to follow our WeChat official account
