A prototype of a cheap IoT system for real-time monitoring of river water quality has been developed. The system consists of monitoring stations and appropriate presentation devices (computer, phone, or similar). Each monitoring station has a possibility for real-time measurement of 4 parameters: temperature, pH, turbidity, and dissolved oxygen. They are measured through probes submerged directly in the water. The probes are connected to the Raspberry Pi 3 model B, and through wi-fi communication, the measured data are transferred to the data storage cloud and then transferred to the visualization platform. The collected data in the IoT system is calculated and presented in real-time. Using real-time we can have instant information on the water quality and an alarm for possible contamination. Collected data on the remote server station can be used for deeper water quality analysis.

The main focus point of this research is the analysis of the computational methods used for Water Quality Index (WQI) calculation using a real-time measurement System. For the calculations, a prototype of a cheap IoT system for measuring surface water quality has been used. The monitoring system incorporates sensors for temperature, turbidity (NTU), and pH. These parameters are measured through probes submerged directly into the water sample. The probes are connected to an Arduino Uno microcontroller, and through Bluetooth wireless communication, the measured data is transferred to a mobile device and displayed on an application. Also, the measured data are saved on a personal computer for their further processing. The collected data from the IoT system is calculated and presented in real time. Using real-time monitoring we can get instant information for the possible water contamination.

In this paper, the design process of neural network-based controller application to hydropower plant control system is presented in order to improve the dynamic behavior when only a Proportional-Integral-Derivative controller is in use. Through simulation experiments, the proper function of the neuro-PID control technique has been successfully verified. Ensuring better results rather than using only gain scheduling PID control lies in designing a suitable neuro controller. That includes sizing the right number of neurons in each layer, a number of hidden and output layers, fitting and training the network behind the neuro-controller, and data normalization. Activation function type determination in each layer is also an important parameter that depends on the output response of the system. In this paper, MATLAB® - Deep Learning Toolbox is used, whereas the simulations are prepared in Simulink. The obtained results show that the optimized turbine model gives a slight improvement in the behavior of the hydropower plant.

Cyber-Physical Systems (CPSs) are becoming more relevant, and ever more present in today's world. The integration of cyber/computer systems with the physical world gives improvements on various efficiency and reliability parameters, as well as convenience. Unfortunately, by their very nature CPSs come with various risks and complications that come from the integration of the cyber and physical worlds. One of these risks are cyber and physical attacks on CPSs. Research across various disciplines on these attacks is growing each passing day. In this paper we simulate the possible effects of False Data Injection Attacks (FDIAs) on a hybrid nonlinear model of a hydropower plant (HPP), and then we implement a classical control engineering method - Kalman filtering, as a detection system for these attacks. It is proven that a Kalman filter designed on a linearized continuous system model is a successful detection method in various attack scenarios on the nonlinear hybrid system model.

Air pollution is an extraordinarily big problem in developing countries, as well as some metropolitan cities. Infiltration of the outdoor air inside buildings, together with other factors can significantly lower the indoor air quality. Particulate matter (PM) pollution is recognized as one of the biggest problems with regards to indoor air quality. An increasingly popular method for dealing with indoor PM pollution is the use of mobile air purifiers. In this paper, a mathematical model for the indoor PM pollution, and the effect of the purifier is presented. This model is verified using an experimental setup. Then an automatic control system is developed based on the model. Finally, a Hardware-in-the-loop (HIL) simulation of the control system is performed using a Raspberry Pi 4 B and an ST Nucleo-64 board.

A new approach to efficient, faster, and intelligent hydropower plant (HPP) control, where constituent equipment is described with highly non-linear mathematical models based on the recommendation from the working group of IEEE on prime movers, is represented in this paper. HPP stability and high efficiency are important factors dependent on the dynamic changes in the energy system demands and the starting time of the plant because the obtained energy is very flexible to those changes in the energy system. This paper is shown and analysed the implementation of the artificial neural network-based controller with PID as an auxiliary controller which helped achieve better behaviour, faster plant stabilization, and operation. The benefits of new technologies and possibilities led to improvements in HPP control and faster system operation. This is achieved by using MATLAB® – Deep Learning Toolbox whereas the simulations are prepared in Simulink. Artificial Neural Networks (ANN) as a technique used in the HPP control systems have advantages in getting a stable and faster response but the complexity of the structure behind the neural networks (NN), meaning algorithms, number of hidden layers, training function, activation function can complicate and destabilize the process. In this paper, the focus is put on the mechanical power responses improvement and the advantages of implementing new technologies contrary to the problems that can occur by using them such as plant destabilization by implementing minor changes, fitting parameters, learning, and training processes, number of hidden layers/neurons, number of epochs, etc.