Smart Fault Detection in Nanosatellite Electrical Power System
Alireza Rezaee, Niloofar Nobahari, Amin Asgarifar, Farshid Hajati · 2026
Read Smart Fault Detection in Nanosatellite Electrical Power System by Alireza Rezaee, Niloofar Nobahari, Amin Asgarifar, Farshid Hajati (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-00335.txt.
Introduction
Alireza Rezaee works with Niloofar Nobahari, Amin Asgarifar, and Farshid Hajati on catching power faults aboard tiny satellites. A nanosatellite power system has many fragile parts. Launch stress and harsh orbit conditions threaten each one. Solar panels can short or break open. Converter switches can fail in two directions. Ground battery regulators can fault as well.
Their method needs no attitude control subsystem data. A neural network first learns the healthy system. Solar radiation and panel temperature go in as inputs. Current and load come out as predictions. A classifier then reads the mismatch patterns. Fault type and location are diagnosed from those patterns. Other learners join the comparison for classification. Principal components, decision trees, and nearest neighbors all take part. The setup suits small satellites in low Earth orbit. Early fault names can guide a real response.
Catalog record
| CREATOR | Alireza Rezaee, Niloofar Nobahari, Amin Asgarifar, Farshid Hajati |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-00335
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Smart Fault Detection in Nanosatellite Electrical Power System Alireza Rezaee, Niloofar Nobahari, Amin Asgarifar, Farshid Hajati Abstract This paper presents a new detection method of faults at Nanosatellites’ electrical power without an Attitude Determination Control Subsystem (ADCS) at the LEO orbit. Each part of this system is at risk of fault due to pressure tolerance, launcher pressure, and environmental circumstances. Common faults are line to line fault and open circuit for the photovoltaic subsystem, short circuit and open circuit IGBT at DC to DC converter, and regulator fault of the ground battery. The system is simulated without fault based on a neural network using solar radiation and solar panel’s surface temperature as input data and current and load as outputs. Finally, using the neural network classifier, different faults are diagnosed by pattern and type of fault. For fault classification, other machine learning methods are also used, such as PCA classification, decision tree, and KNN. Keywords: fault diagnosis, machine learning, nanosatellite, electrical power system 1 Introduction Nanosatellites, according to their missions, have several subsystems. The power subsystem generates, manages, and distributes power. The power subsystem is essential in all satellites. According to the mission type and requirements, this subsystem includes parts like photovoltaic solar arrays for converting the sun’s energy and generating electrical power, DC to DC Maximum Power Point Tracker (MPPT) converter and regulator for power optimization and regulation. DC charge manager subsystem, loads, and switch mechanism are designed and implemented for a specific purpose [ 1 ] . The power generated by the photovoltaic solar array and the whole system is affected by the orbital condition and satellite circumstances such as orbital high, inclination, radiant intensity, heat flux, or radiation at orbit temperature that should be analyzed [ 2 ] . When the whole system and components work correctly, the satellite is able to carry on the mission. However, several faults may occur due to the launcher’s shock and pressure. If these faults do not be detected at the right time, they may cause the whole system or mission failure. 1.1 Related Work Beyond satellite power systems, related work on pattern recognition, control, and applied machine learning in power, biomedical, and telecommunication domains provides additional methodological support and motivation for our approach [ 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 ] . Research has been conducted on detecting, classifying, and diagnosing electrical equipment faults at the satellite power subsystem. Zhao et al. [ 3 ] proposed a method to detect faults at photovoltaic solar arrays based on Graph-Based Semi-Supervised Leaning (GBSL). By GBSL, they detected faults that an Overcurrent Protection Device (OCPD) could not uncover. Zhao et al. [ 4 ] detected and diagnosed the photovoltaic solar systems’ faults using a decision tree. Mohamed et al. [ 5 ] also detected photovoltaic solar systems’ faults by neural networks and the genetic algorithm. In a converter proposed in [ 6 ] , there was a method for detecting faults, detecting open circuit power switch faults without using additional hardware, and using the correct control loop in the PWM converter. In aerospace projects, fault in Li-Ion battery is applied [ 7 ] because the density of energy is high and does with methods like diagnosis and distinguish fault based on Kalman Sequential Adaptive filter for looking state charge and DE charge of the battery, rate of charge, and de charge, estimate of state charge and DE charge (SOC), and health state (RUH) with use by back up vector machine and collect load information that is connected to it [ 8 , 9 ] . The above-mentioned researches do not consider the fault diagnosis of the whole satellite’s power system. Only two studies have discussed the fault diagnosis in the satellite’s power subsystem. In the first research, the fault diagnosis and tolerance have been discussed based on the PCA method for any voltage, current, and temperature faults at sensors. This paper is only concerned about the accumulative fault on the power system’s sensors [ 10 ] . In [ 11 ] , each fault probabilities in the satellite’s power system with Bayesian network is defined by knowledge of scientist but for using the Bayesian category without the use of scientist knowledge, fault probilities must be determined or can be estimated by Gaussian equation [ 12 ] . In this paper, the faults of the whole power system, including photovoltaic parts, the DC to DC converter, and the battery are diagnosed and resolved. This paper separate faults using a detective method. 2 Materials and Methods Satellite electrical power system consists of analog and digital circuits that include transistors, trustors, diodes, amplifiers, logic elements, switches, and connections. These parts are manufactured with specific reliabilities. The faulting rate in a component ( \lambda ) is defined as \lambda=\frac{N_{\text{faulty}}}{N_{\text{total}}\times t} where t is the outfit time hour unit (h), N_{\text{faulty}} is the number of faulty elements, and N_{\text{total}} is the total number of elements. This rate depends on temperature and the mode of operation. Table 1 lists the faulting rate in various components. Table 1: Fault rate at temperature 40 ∘ C for various components [ 13 , 14 ] . Equipment \lambda ( 10^{-9} /h) Transistor 10–9 Trustor (power switch family) 10–9 Digital integrated circuit 30 Logical elements 30 Analog switches 2000 Amplifier 10–9 Diode 10–9 Battery Li-Ion 10–9 Solar array 10–9 Using the faulting rate, we can diagnose faults at electrical power system components such as IGBTs, MPPT converters and regulators, solar array, and battery. Open circuit and line-to-line faults in solar array, short circuit and open circuit faults in IGBTs, MPPT converters and regulators, and ground fault in battery are feasible in nanosatellite power systems. Figure 1 shows the overall block diagram of the proposed fault diagnosis model. First of all, the system must be recognized by virtual neural network for detecting and classify faults at different systems. At the next step, possible faults intentionally and virtually are applied to the system, then like a normal state, the faulty system is diagnosised. Afterward, for each fault, one model like figure 1 is created that can be compared by system output with output model. Figure 1: Proposed fault diagnosis model. Create remains based on equation (1) then calcify faults: r_{i}=y(t)-\hat{y}_{i}(t)\,, (1) where, according to figure 1, system output y(t) is contrasted to output model system without fault \hat{y}_{0}(t) and output faulty model like \hat{y}_{1}(t) to \hat{y}_{n}(t) , remained amounts ( r_{0} to r_{n} ) according to relation ( 1 ) are created and for pattern recognition is categorized with neural network. In sequence, at first, the electrical power system is simulated with neural networks without fault which is a powerful tool in system recognition [ 15 ] , and then possible faults are injected into the system and then they are simulated and graph is given.