The technology landscape has been shifting rapidly in the past few years, and the use of varied control systems, such as robots and computers, is changing the way in which traditional industries operate. As industrial automation begins to take center-stage, end users and industrial-automation suppliers are faced with an important quandary. They must weigh the risk of adopting new technologies that may result in heartburn against the possibility of losing market share to compatriots that are willing to experiment in order to stay ahead of the curve.
As a result, quality-assurance methods need to constantly evolve and keep pace with the demands of modern-day industrial-automation systems. However, testing programs are often an afterthought in most industrial entities or business enterprises, and most existing approaches use physical devices, such as gateways, communication modules and sensors, for validation. To address this important gap in the testing process, I recommend employing a hybrid testing approach that covers both manual and automated testing. Since manual testing is fairly well-established, let’s focus on test automation, along with simulated testing.
Traditional industrial testing approaches primarily involve the use of physical devices, such as cloud platforms, gateways, communication modules and sensors covering three test levels—namely, system testing, site acceptance and factory acceptance testing. The shortcoming of this approach is that component validation of the firmware occurs very late into the system test cycle. It also leads to higher costs due to multiple test setups, while adding greater lead time during regression cycles.
Requirements of a Hybrid Test Approach
A robust test-automation framework is the first and foremost requirement of a hybrid test approach. Some important considerations of the framework include:
- Support for various industrial protocols at each stage of a deployment ecosystem
- CLI, GUI and API test automation
- Development of software protocol data simulators
- Support for PID and business logic validations
- Integration with test hardware and external vendor tools
- Security and performance test-automation support
- Extent of need for technical and domain knowledge for test-case scripting
- Detailed business-driven test-execution reports
Test-Framework Development
The proposed hybrid test automation needs to have both a framework and simulated testing. This needs to be designed in such a way as to cover the device layer (physical hardware), programming control layer (PLC or gateway) and control layer (HMI or SCADA) and simulated testing integration. There are different approaches from which one can choose:
Test-Driven Development (TDD): a test-first development technique by which the test cases are written first to ensure the goal of the development steps, and implemented code satisfies the requirement.
Behavior-Driven Development (BDD): an evolution of TDD by which tests are defined based on a user story written in a specific format, and on system behavior.
Data-Driven Testing: complex interaction sequences that can be implemented as repeatable automated test cases, by which test data is separated from implementation.
Key-Driven Testing: a set of keywords and data tables by which functionalities are captured and translated into steps, irrespective of dependence on the automation tool.
The core components of this framework would be configuration utilities, protocol data simulators, open-source libraries, custom-built libraries for IA protocols, sensor simulations, keyword libraries, CLI/API/GIU support, libraries for data parsers, cloud connectivity and integration with a CI/CD pipeline.
Protocol and Device Simulation Is Equally Important
In the second part of the hybrid testing approach, protocol and device simulation play a key role in accelerating a product’s time to market, and in solving the after-sales support problem. Sensor simulators with GUI based on PyQt or similar languages can be designed to leverage the open-source technological support currently available. PyQt provides an easy way to design user interfaces with inbuilt designer tools that provide support integration with the test-automation framework.
Sensor simulators will be a dedicated development activity to achieve higher test-automation coverage at every stage, and to reduce capital expenditures (capex) of the required hardware. This will facilitate better software-development cycles, increasing test-automation coverage significantly, along with corner error handling test cases. As such, this approach directly results in capex savings by eliminating the need for complete test setups for each developer and test engineer. In addition, it provides extensive coverage, cycle-time reduction, reusable test-automation scenarios for multiple products, and convenient migration or integration into existing test tools.
With the technology evolution that is being driven by Industry 4.0, there are great demands being made on the further development of OT and IT systems. Therefore, new or enhanced product-development processes will need to be created. A hybrid testing approach is key to supporting this product transformation, as it significantly improves quality-assurance levels through advanced test automation, while bringing about cost optimizations through simulation-based testing. Ultimately, the success or failure of product architectures in the industrial-automation industry are determined by their ability to meet the requirements of a thriving and constantly evolving Industry 4.0 ecosystem.
Jason Chandralal is the general manager of product-engineering services at Happiest Minds Technologies. He is responsible for defining and leading test-engineering solutions in the area of industrial and embedded systems.