This section outlines the key features of the clicSAND software in its initial version, clicSAND 1.0, and highlights the enhancements introduced in its third iteration, clicSAND 3.0. A brief mention will also be made of the macOS version of the software, clicSAND 2.0.
2.1. Key Features of clicSAND 1.0 Software
A basic overview of the clicSAND 1.0 software and its main functionalities is provided in
Figure 1, while the primary advantages and limitations are summarised in
Table 1. It utilises the SAND Excel interface as the sole data entry point for the user. A significant usability advantage is that no interaction with the command line is necessary during the modelling process. Furthermore, the software operates entirely offline, which is especially beneficial for regions with low-bandwidth connections.
A one-click installation package containing all the necessary components of the software is freely available for download on GitHub [
13]. Instructions on correctly installing the clicSAND 1.0 software and the two free solvers can be found on Zenodo [
14]. Additionally, a step-by-step online certified course hosted on the Open Learn Create (OLC) Platform is available as a public good [
15]. This course combines theoretical lessons with practical exercises, teaching users the theory behind energy modelling for investment planning and how to build an energy model from scratch using the clicSAND 1.0 software for OSeMOSYS. The authors recommend that first-time users of the clicSAND 1.0 software enrol in this online course and carefully follow the instructions provided in practical exercises 1, 2, and 3:
Hands-on 1: Download and installation of the clicSAND 1.0 software and the solvers (GLPK and CBC).
Hands-on 2: Best practices for inputting data into the Excel SAND Interface and practical examples for one technology.
Hands-on 3: Instructions on how to save, run, and visualise results using the Microsoft Access database and the Excel template provided with the software.
Another asset of the clicSAND 1.0 software lies in its compatibility with a related project aimed at creating Starter Data Kits (SDKs) [
16,
17]. These SDKs consist of national datasets and simple energy system models, currently available for 69 countries in Africa, East Asia, and South America (
Appendix A). Notably, the models in the Starter Data Kits were constructed using the Excel SAND Interface, and the results were obtained by running the clicSAND 1.0 software. The scientific significance of the SDK lies in its facilitation of access to national datasets and energy system models, thereby streamlining research processes and enhancing the efficiency of energy analysis. By providing a comprehensive set of predefined parameters within the SAND file, the SDK eliminates the arduous task of manual data entry, allowing researchers to focus more on analysis and interpretation.
One key requirement of this first version of the software is the need for at least 8 GB of RAM for smooth operation, along with a Windows computer with Microsoft Access and Excel installed, both of which are commercial software. However, to broaden the accessibility of this public good and to overcome the need for a Windows operating system, a second version of the clicSAND software (clicSAND 2.0), which is compatible with macOS and does not require Microsoft Access, was recently released. Details regarding this are provided in
Section 4.1.1 of this paper. Similarly, to address the 8 GB RAM requirement, a third version of the software, clicSAND 3.0, was developed. This version offers the flexibility to execute a model on an online cloud platform, thereby eliminating the need for dependence on the computer’s memory.
The Excel SAND interface is specifically designed for beginner and intermediate users, allowing them to work on middle-sized models with basic regional and temporal resolution. As explained in
Section 2.3, the Excel SAND Interface serves as the platform where users can intuitively input data for up to 200 technologies, 50 commodities, five types of emissions, one region, and 96 time slices.
It is important to note that when initially using the Excel Interface to input data, users need to be cautious when pasting data while a filter is applied, as Excel will copy the data to the subsequent row and not to the next filtered row. Nevertheless, to assist users in avoiding mistakes during model compilation, step-by-step video guidance has been released on YouTube, demonstrating best practices for effectively inputting data [
18].
Another weakness of the first version of the clicSAND software for Windows was its inability to detect errors if the model failed during the run. In fact, the software platform depicted in
Figure 1 does not specify the exact error that might occur; it simply indicates that the model is not functioning. This issue also arose when running OSeMOSYS models with other interfaces. To address this limitation and enhance the troubleshooting experience, clicSAND 3.0 has been equipped with functionality to identify in which step of the process the model is failing, as detailed in
Section 2.2 of this paper.
All versions of the clicSAND software are fully compatible with the existing OSeMOSYS architecture, meaning that the constraints, variables, and parameters are all represented. Additionally, two widely used solvers to find a model’s optimal solution, GLPK (GNU Linear Programming Kit) [
19] and CBC (COIN-OR branch and cut) [
20], are included as part of the open-source and freely available clicSAND software.
Moreover, the results produced by the software are fully compatible with other tools such as CLEWS (Climate, Land, Food, Energy, and Water systems approach) [
21], the IRENA FlexTool for analysing system flexibility, MUSE (Modular energy system Simulation Environment) [
22], and other OSeMOSYS platforms, including the OSeMOSYS Cloud platform [
23] (which provides online solvers through a cloud service). Further details on this compatibility are provided in
Section 2.4.2 of this paper.
2.2. Enhancements in the clicSAND 1.0 Software: The Release of clicSAND 3.0
The decision to enhance the functionalities of clicSAND 1.0 and release an improved version, clicSAND 3.0, arose from a continuous improvement process. This process involved collecting feedback from users through surveys distributed at the conclusion of various capacity-building events. During these events, inexperienced users were tasked with constructing basic national OSeMOSYS models within a three-week timeframe using clicSAND 1.0. As indicated in
Table A2 (
Appendix B), the feedback received confirmed the limitations that developers had already identified, primarily concerning extended computational times and the laborious process of visualising results through an Access database.
To address these limitations, the authors developed, tested, and implemented a new version of the clicSAND software, known as clicSAND 3.0. Drawing inspiration from the innovations in clicSANDMac (clicSAND 2.0), this version of the software embeds a Python code that effectively reduces the data file size (both matrix and LP-file) by pre-filtering unused elements—available for download on GitHub [
24]. This optimisation significantly accelerates the solving process. Additionally, this feature eliminates the need for an Access database to visualise results, which will be further elaborated upon in
Section 2.5.2. Instructions to install clicSAND 3.0 and run the model on the OSeMOSYS Cloud are available on Zenodo [
25]. Specific teaching material and exercises using clicSAND 3.0 are also freely available as a public good [
26].
As discussed in the preceding section, clicSAND 1.0 had a specific requirement, necessitating a minimum of 8 GB of RAM for smooth model execution. However, clicSAND 3.0 has comprehensively addressed this requirement by introducing the capability to run models not only offline but also online through a cloud-based platform known as OSeMOSYS Cloud. This innovative approach circumvents the need to rely entirely on the computer’s internal memory, as elaborated further in
Section 2.4.2.
To facilitate this novel functionality, significant modifications were made to the user interface of clicSAND 3.0. As shown in
Figure 2, it now includes a button labelled “Generate OSeMOSYS Cloud input”. This button streamlines the process of formatting the .txt file generated by the Excel SAND interface into a compatible format tailored specifically for seamless integration with the OSeMOSYS Cloud platform. This transformative enhancement substantially expands the accessibility and flexibility of clicSAND 3.0, empowering users to tap into cloud resources without being restricted by local hardware limitations.
The integration with the OSeMOSYS Cloud platform greatly enhances the software’s convenience and adaptability, allowing users to access models and results from virtually anywhere, at any time. This development marks a significant step forward in making clicSAND 3.0 a versatile and user-centric tool in the field.
As highlighted in
Table 1, one of the prominent limitations of clicSAND 1.0 revolved around troubleshooting and debugging challenges. In this earlier version, inexperienced users encountered significant difficulty in identifying errors because the user platform did not provide clear indications of where the error occurred, making the debugging process a challenging task.
However, with the introduction of clicSAND 3.0, this challenge has been effectively addressed through two key enhancements. First, when running a model offline, the user platform now displays the specific line where the error occurred, greatly simplifying the process of identifying and rectifying mistakes. Second, when utilising the OSeMOSYS Cloud to run a model, users have the ability to monitor the entire solving process in real-time, allowing them to pinpoint precisely where the process encounters an issue.
Furthermore, the integration with the cloud offers an exceptional feature for debugging: as a clicSAND model runs, OSeMOSYS Cloud generates a visualisation of the Reference Energy System for that model. This empowers users to effortlessly visualise their input data with a single click [
23]. Any missing connections within the model are highlighted in red, providing valuable assistance to users in identifying and rectifying bugs within their models. These advancements collectively enhance the user experience, making troubleshooting and debugging significantly more accessible and efficient in clicSAND 3.0.
As summarised in
Table 2, clicSAND 3.0 maintains the benefits introduced in clicSAND 1.0 while introducing enhanced features. These include embedded Python code, which reduces data file size and accelerates computation (with clicSAND 3.0, a full Starter Data Kit runs in less than 10 min). Users now have the flexibility to work offline or access a cloud-based service, eliminating the need for dependence on local memory. The tool also streamlines result visualisation with an offline Excel template and seamlessly integrates with the OSeMOSYS Cloud, enabling offline result visualisation.
2.3. Software Licenses
The clicSAND software and its components are licensed under an MIT license (open-source and free of charge), which permits the public to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the software. Users can do so by giving credit, linking to the license, and indicating if they have made any changes to the work in all copies or substantial portions of the software. clicSAND is written using the C# language, enabling easy uptake, improvements, and adaptability.