Accelerating the Adoption and Benefits of Model-based Control in PEMD Applications

控制算法是运行许多系统的软件的核心，无论是洗衣机还是电动汽车。大多数应用程序都有相对基本的控制，不会从任何复杂的东西中获得太多好处。然而，对于更复杂、多方面的系统，如电动汽车，更先进的控制算法可以提供显著的好处，例如更大的航程或更长的使用寿命组件，只需具有针对系统进行优化的控制动作。例如，电动汽车电池的寿命在很大程度上受到驾驶过程中充电或放电速度的影响，而复杂的控制算法可以积极地将这些周期降至最低，并最终延长电池寿命。同样，电动汽车的续航里程可以通过一种算法来最大化，该算法可以在汽车通过交通工具、十字路口或起伏的道路时最大限度地减少刹车和加速。在许多这样的情况下，操作的某些方面可以通过增强控制来改进。最优控制算法基于一些定义的性能度量提供“最佳”控制动作，并且使用内部数学模型来表示系统行为和优化例程来实现这一点。由于对内部模型的依赖，这些控制器通常被称为基于模型的控制器。这种先进的控制策略在炼油厂得到了广泛的应用，在这些炼油厂中，系统的缓慢性质允许复杂的算法有足够的时间进行计算。现在，这种基于模型的控制在汽车和其他独立机器上运行的软件中是可行的，这些软件运行在功能越来越强大的微处理器上。然而，这种控制算法的设计和部署确实需要巨大的工程能力，因此除了拥有大型研发团队的最大公司外，所有公司的采用都存在相当大的障碍。这项工作的目的是开发以应用为重点的培训材料，消除英国PEMD公司采用这些障碍的障碍，使他们能够了解和部署复杂的基于模型的控制方法，并为其产品提供竞争优势。

Control algorithms are at the core of the software that operates many systems, whether it is a washing machine or an EV car. Most applications have relative basic control and will not have much benefit from anything sophisticated. However, for more complex, multi-faceted systems, like an EV, a more advanced control algorithm can offer significant benefits, such as greater range or longer life components simply by having control actions that are optimised to the system. For example, the life of an EV battery is greatly influenced by how rapidly it is charged or discharged during driving, and a sophisticated control algorithm can actively minimise those cycles and ultimately extend battery life. Equally EV range can be maximised by an algorithm that minimises braking and accelerating as a car travels through traffic, junctions or undulating roads. There are many such situations where some aspect of operation can be improved by enhanced control.An optimal control algorithm provides the "best" control action based on some defined performance metrics, and it does this using an internal math model to represent the system behaviour and an optimisation routine. These are often referred to as model-based controllers due to the reliance on the internal model.Such advanced control strategies have been used widely in oil-refineries where the slow nature of the system allowed the complex algorithms plenty of time for their calculations. Such model-based control is now feasible within the software running on ever powerful microprocessors found in cars and other standalone machines. However, such control algorithms do require a significant engineering capability to design and deploy, and so there is a sizable barrier to adoption for all but the biggest of companies with large R&D teams.The aim of this work is to develop application focussed training materials that remove these barriers to adoption for UK PEMD companies to allow them to understand and deploy sophisticated model-based control methods and provide a competitive edge for their products.