Near edge zero-power sensor system for asset condition monitoring

Abstract

A system for monitoring structural conditions includes a sensor network comprising a plurality of sensor assemblies. Each sensor assembly includes at least one a sensor device comprising a sacrificial layer and a reference layer, wherein the sacrificial layer comprises a first plurality of sensor elements, each of the first plurality of sensor elements comprising a strip of nano-engineered material, wherein the reference layer comprises a second plurality of sensor elements encased in a protective coating, each of the second plurality of sensor elements comprising a strip of nano 10 engineered material.

Description

NEAR EDGE ZERO-POWER SENSOR SYSTEM FOR ASSET CONDITION MONITORINGPRIORITY

[0001] This application claims the benefit of priority to U.S. Provisional Patent Application No. 63 / 725,219, filed November 26, 2024, which is incorporated by reference herein in its entirety.TECHNICAL FIELD

[0002] The present disclosure generally relates to sensor systems, and more specifically, to passive or low-power sensor systems for asset monitoring.BACKGROUND

[0003] Infrastructure condition monitoring provides for inspection of certain structures, such as wind turbines, offshore platforms, pipelines, grid infrastructure, energy assets and industrial facilities, which are often subjected to harsh environmental conditions that can lead to gradual degradation over time. Traditional monitoring approaches face challenges in certain environments. Manual inspections and data acquisition methods can be costly, time-consuming, and may require skilled technicians to perform. As infrastructure continues to age and environmental conditions become more challenging, there is an ongoing need to develop more effective and efficient monitoring solutions.BRIEF DESCRIPTION OF THE DRAWINGS

[0004] In the drawings, which are not necessarily drawn to scale, like numerals may describe similar components in different views. Like numerals having different letter suffixes may represent different instances of similar components. The drawings illustrate generally, by way of example, but not by way of limitation, various embodiments discussed in the present document. Various ones of the appended drawings merely illustrate example embodiments of the present inventive subject matter and cannot be considered as limiting its scope.

[0005] FIG. 1 is a block diagram illustrating a sensor system, in accordance with some examples.

[0006] FIG. 2 illustrates an example structure that can be monitored via the sensing techniques described herein, in accordance with some examples.

[0007] FIG. 3A is a perspective view illustrating example sensor assemblies that include certain geometries useful in sensing modalities that include non-planar surfaces, in accordance with some examples.

[0008] FIG. 3B is another perspective view illustrating example sensor assemblies that include certain geometries useful in sensing modalities that include non-planar surfaces, in accordance with some examples.

[0009] FIG. 4 is a top view of an example of a sensor device illustrating a placement of a stagnation line about a middle of the sensor device, in accordance with some examples.

[0010] FIG. 5A illustrates a perspective view of an example wind turbine blade equipped with a sensor assembly, in accordance with some examples.

[0011] FIG. 5B illustrates a top view of an example wind turbine blade equipped with a sensor assembly, in accordance with some examples.

[0012] FIG. 5C illustrates a side view of a wind turbine blade with an integrated sensor assembly, in accordance with some examples.

[0013] FIG. 6A is a perspective top view of an example of the sensor assembly having a pre-engineered crack, in accordance with some examples.

[0014] FIG. 6B illustrates a bottom perspective view of the sensor assembly of FIG. 6A, illustrating the pre-engineered crack, in accordance with some examples.

[0015] FIG. 7A illustrates a side view of an example monitoring of a structure (e.g., pipe) utilizing two sensor assemblies attached to a structure via an adhesive layer, in accordance with some examples.

[0016] FIG. 7B illustrates sensor assemblies that include sensor elements suitable for use as strain gauges, in accordance with some examples.

[0017] FIG. 8A is a sectional side view illustrating a sensor assembly disposed onto a structure, in accordance with some examples.

[0018] FIG. 8B is a sectional side view illustrating the sensor assembly of FIG. 8A but with a cover removed, in accordance with some examples.

[0019] FIG. 9A is a block diagram of a sensor assembly having a preengineered crack illustrating a 4-point bend setup, according to some examples.

[0020] FIG. 9B is a perspective view of the now calibrated sensor assembly 902, according to some examples.

[0021] FIG. 10 illustrates a sensor erosion detection system having a leading edge protection (LEP) layer that is disposed on two example sensor assemblies, in accordance with some examples.

[0022] FIG. 11 illustrates a flowchart depicting a process for applying the techniques described herein, according to some examples.

[0023] FIG. 12 is a block diagram depicting a machine suitable for executing instructions via one or more processors, according to some embodiments.

[0024] FIG. 13 machine learning engine for wind turbine blade leading edge material erosion rate prediction and cumulative damage index, in accordance with some embodiments.

[0025] FIG. 14 illustrates a multi-agent agent system (MAS) for infrastructure monitoring in accordance with some embodiments.DETAILED DESCRIPTION

[0026] The described examples relate to condition monitoring sensor systems for infrastructure, particularly focusing on power-passive sensor technologies that address challenges in monitoring large-scale structures in harsh or remote environments. These systems utilize zero power or low power sensor elements engineered to detect and measure various forms of structural degradation, including erosion, fatigue, cracks, displacement, leaks and corrosion. The sensors operate without the need for continuous electrical power, overcoming limitations associated with traditional powered sensors that undergo battery replacements or use electrical wiring systems for long-term asset monitoring.

[0027] In certain examples, the techniques described herein include a (near edge zero-power) NEZP sensor element constructed using an electrically conductive material (such as a silver or a chromium alloy). This sensor element is engineered to monitor surface erosion by detecting variations in electrical parameters such as resistance, inductance, and capacitance. As sensor material erodes, these electrical properties change, providing a measurable indication of the extent of the change, e.g., through erosion.

[0028] In some examples, a sensor assembly comprises multiple individual sensors mounted onto a single base. One or more NEZP sensor elements are positioned near an edge to serve as reference sensors, remaining shielded from loading or environmental exposure, for compensation and / or calibration of remaining NEZP sensor element measurements. In contrast, one or more NEZP sensor elements are positioned near the edge and are exposed to the loading or environment, allowing them to measure changes in certain physical or chemical property that they are monitoring, such as material erosion, corrosion, leak, material fatigue, crack and / or displacement.

[0029] In some examples, a NEZP sensors includes a sacrificial or witness layer that mimics the behavior of physical degradation processes on the monitored asset. This layer can be designed to respond to various types of degradation, such as erosion, cracks, gaps, separation, fatigue, chemical exposure, and so on. The sacrificial layer may be metallic or non-metallic and can be manufactured through machining, molding, and / or additive manufacturing techniques. That is, NEZP sensors may include a sacrificial or witness layer that is engineered to mimic or “twin” the behavior of asset degradation (e.g., mechanical - crack or displacement, chemical - corrosion or erosion, and the like) and changed properties (e.g., electrical - resistance, inductance, chemical - color) permanently without using power.

[0030] In examples of the sensor system that include an electrical system, the electrical system provides for power generation as well as data acquisition from the one or more NEZP sensors. A power system for data acquisition can be wired, energy harvested, and / or use wireless power. Data collection from the NEZP sensors can be triggered digitally (based on aschedule or specific events) or physically (based on a level of physical degradation). Communication of the collected data can be achieved through wireless (long or short distance) or wired methods.

[0031] The sensor system is capable of detecting and capturing multiple loading conditions or environmental modalities, including erosion, corrosion, cracks, fatigue, elongation, breaks, gas presence, gaps, icing, temperature, and / or pressure. This multi-modal sensing capability allows for comprehensive monitoring of various aspects of infrastructure health. Data analysis is then used to determine the extent of erosion, corrosion, cracking, fatigue, and so on. Data acquisition can operate intermittently, powered only when measurements are taken, which conserves energy and reduces operational costs. Indeed, data capture can operate continuously without power based on permanent changes to NEZP sensor element. That is, continuous monitoring can occur without power due to permanent changes in the NEZP sensor element. Data upload or acquisition to edge electronics can operate intermittently, powered only to record change in NEZP sensor clement, which conserves energy, extend system life and reduce operational cost.

[0032] The NEZP sensor elements generate accumulated condition data that is aggregated into a Cumulative Measurement Index (CMI) specific to the monitored sensing parameters. In certain embodiments, the NEZP sensor elements undergo permanent changes due to erosive loading caused by wind turbine operation in rainy and humid environments. The cumulative effects of such loading are periodically captured and aggregated into the CMI, representing the Damage Equivalent Loading (DEL) experienced by the wind turbine blade system

[0033] In some examples, the data analysis incorporates the concept of CMI and DEL, or a combination of both. Through simulation, laboratory tests and / or field observations, a relationship is established between the environmental (e.g, erosive) damage equivalent load measured by the sensor and the damage equivalent load experienced by the surface or coating being monitored. This relationship allows for the conversion of NEZP sensor measurements into an assessment of the current damage equivalent erosionload on the material surface of interest. The relationship between sensor element DEL and asset surface DEL can be established through empirical testing and curve fitting using mathematical relationships and / or machine learning algorithms. Physics-based simulations and finite element analysis of impacts, fatigue damage, and wear may also be employed to determine this relationship. Additional data such as weather conditions (e.g., rainfall rate, wind, temperature) and impact speed of droplets can be included in the analysis to account for non-linearity that may exist between different material types.

[0034] In some embodiments, the condition monitoring system (e.g., edge condition monitoring system) can include a combination of NEZP sensor elements and general-purpose (GP) sensors, such as accelerometers, gyroscopes, and / or LiDAR. This integration enables a robust, multi-faceted approach to condition monitoring, allowing for the collection of complementary data streams. The data fusion between NEZP sensor elements, which are more optimized for specific measurements like erosion and surface degradation, and GP sensor elements, which can capture broader dynamic parameters like vibration and structural motion, results in a more comprehensive assessment of the turbine’s health. By combining these diverse data sources, the system can optimize the calculation of the Cumulative Measurement Index (CMI), thereby enhancing the accuracy and reliability of asset condition monitoring. This fusion of data not only improves the detection of potential failure modes but also enables more precise predictions of the remaining useful life (RUL) of the monitored assets.

[0035] The analysis of sensor data and DEL relationships can be performed on a data processing unit in situ and / or in post-processing, for example, in a cloud-hosted data repository and processing system. This flexibility in data processing allows for scalable and adaptable monitoring solutions. In some examples, the sensor system leverages a multi-agent system (MAS) architecture to optimize the analysis of sensor data and Cumulative Measurement Index (CMI) and / or Damage Equivalent Loading (DEL) relationships. The MAS framework is composed of multiple independent agents that can operate collaboratively across various system levels,including on-site edge devices, centralized hubs, and cloud-hosted data repositories. Each agent is responsible for processing specific subsets of data, making decisions, and communicating results to other agents, thereby enabling a distributed and more efficient approach to condition monitoring and predictive maintenance. This decentralized system architecture improves responsiveness and improves fault tolerance, particularly in industrial environments where network connectivity may be intermittent, such as offshore wind farms.

[0036] The analysis of sensor data and DEL relationships (and / or CMI) can be performed at multiple layers of the system. At the edge level, NEZP sensor nodes equipped with local or edge processing capabilities can perform preliminary data analysis, allowing for immediate detection of critical conditions and localized damage assessment. Advanced algorithms including edge Machine Learning (ML) can also compress the dataset for wireless transmission efficiency and saving power on the edge. At the hub level, data from multiple sensor nodes on a single asset (such as a wind turbine blade) or a group of assets (such as multiple blades in one or more turbines) are aggregated and further analyzed to provide a more comprehensive view of asset condition. Advanced algorithm including Machine Learning (ML) can be applied to draw additional intelligence through combination of these sensor data. Finally, at the cloud level, large- scale data aggregation and advanced analytics, including machine learning models, can be applied to derive long-term insights, optimize maintenance schedules, and refine predictive algorithms. The ability to shift analysis between the edge, hub, and cloud layers allows for scalable, adaptable, and resource-efficient monitoring solutions.

[0037] By leveraging a multi-agent system approach, the monitoring solution gains several advantages. First, it allows for scalable deployment, enabling improved integration of additional sensor nodes and turbines into the monitoring network without significant changes to the existing infrastructure. Second, the MAS approach supports adaptive intelligence, where agents can autonomously adjust their behavior based on the operational context, ensuring optimal performance under varying environmental conditions. Third, by distributing computational tasks acrossmultiple layers, the system minimizes latency and optimizes bandwidth usage, which is particularly beneficial in remote offshore environments where connectivity can be limited. The MAS architecture also enhances fault tolerance and system resilience, as each agent can operate independently, improving continued functionality even if a portion of the network experiences failures.

[0038] The sensor system can be used in various applications where erosion or wear needs to be monitored. For instance, the sensor can be applied to measure environmental degradation (e.g., erosion, corrosion, cracks, fatigue, elongation, breaks, gas presence, gaps, icing, temperature, and / or pressure) in turbine blades, pipelines, marine environments, and / or other settings where material degradation is a concern. The sensor's design allows for easy customization to meet the specific needs of different industries. Alternative embodiments of the sensor system may use different electrically conductive materials to construct the sensor, chosen to be more sensitive to specific erosive environments. In some examples, a thickness of the sacrificial layer of the NEZP sensor to be eroded can be adjusted so that the sensor lifetime is similar to or greater than that of the coating or surface being monitored. Another variation involves the use of multiple conductive layers to build the NEZP sensor. In this configuration, the erosive rate of the material changes with the amount of erosion, allowing the damage equivalent erosion loading sensitivity to change as a function of the sensor’s lifetime.

[0039] In some examples, the integration of data from NEZP sensor elements with general-purpose (GP) sensors such as accelerometers can significantly enhance the accuracy and depth of condition monitoring. For instance, an NEZP sensor element designed to measure crack propagation due to fatigue loading on the surface of a wind turbine blade can be combined with an accelerometer that captures the blade’s vibration response in real-time. In this scenario, the NEZP sensor continuously monitors changes in electrical resistance, which are indicative of micro-crack formation and growth on the blade surface. Simultaneously, the accelerometer measures vibration patterns and dynamic stresses experienced by the blade during operation.

[0040]

[0041] By fusing these two data streams, the system can correlate the rate of crack propagation detected by the NEZP sensor with specific vibration frequencies and amplitudes recorded by the accelerometer. This correlation provides deeper insights into how dynamic loading conditions, such as gusts of wind or turbulent airflow, are contributing to fatigue damage. For example, an increase in vibration amplitude in conjunction with accelerated crack growth could indicate that specific operational conditions are exacerbating the damage.

[0042]

[0043] The fusion of NEZP and accelerometer data allows the system to adjust the Cumulative Measurement Index (CMI) in a more nuanced manner, reflecting both the microscopic structural changes and the macroscopic dynamic responses of the blade. This holistic approach not only improves the accuracy of damage assessment but also enables predictive maintenance by identifying conditions that are likely to accelerate fatigue failure, thereby optimizing maintenance schedules and extending the asset’s lifespan.

[0044] In another example, data fusion can be utilized to monitor blade deflection and shape using accelerometers placed at two locations along the blade. By comparing the vibration and acceleration signals from these sensors, the system can estimate the blade’s deflection profile under varying wind loads. This method captures changes in blade curvature and bending behavior, providing insights into structural deformation. When combined with data from NEZP sensors measuring erosion or surface damage, the system can enhance the Cumulative Measurement Index (CMI) for more accurate condition monitoring, allowing for early detection of potential fatigue issues and optimizing maintenance schedules.

[0045] The described examples also include a wireless sensor network that enables continuous lifecycle data collection and automated field connectivity. This network integrates with the various NEZP sensors to create a more comprehensive monitoring system. In some examples, the data collected by the sensor network feeds into a hybrid Al analytics system. This system incorporates online training with field data and digital twins foroperations and maintenance, as further described below. The Al analytics component processes the continuous stream of data from the sensors to deliver actionable insights for predictive maintenance and asset management.

[0046] The described examples address several challenges in infrastructure condition monitoring. By providing a zero-power or near-zero-power solution, they minimize or eliminate the need for frequent battery replacements or complex wiring systems, making them suitable for deployment in hazardous zones, harsh conditions, and during extreme events. This capability opens up novel applications such as lifecycle equipment condition tracking for original equipment manufacturers (OEMs) and monitoring of remote areas. The sensors can be installed by facility personnel and do not require battery changeouts, resulting in lower operational expenses and maintenance requirements. This makes large-scale deployment of sensor networks more feasible and affordable.

[0047] The techniques described herein thus have potential applications across various industries, including wind energy, onshore / offshore production and drilling, and refinery operations, among others. In wind asset monitoring, for example, the sensor systems can be used to monitor certain structures, such as turbine blades, turbine towers, fixed or floating foundations, and / or offshore substation structures.

[0048] Reference will now be made in detail to specific example embodiments for carrying out the inventive subject matter. Examples of these specific embodiments are illustrated in the accompanying drawings, and specific details are set forth in the following description in order to provide a thorough understanding of the subject matter. It will be understood that these examples are not intended to limit the scope of the claims to the illustrated embodiments. On the contrary, they are intended to cover such alternatives, modifications, and equivalents as may be included within the scope of the disclosure.

[0049] FIG. 1 is a block diagram of a sensor system 100, in accordance with some examples. In the depicted embodiment, the sensor system 100 includes a sensor network 102 communicatively coupled to a dataacquisition and logging system 120, and to a data analytics system 122 designed to continuously monitor and analyze infrastructure conditions. For example, one or more sensor assemblies 104 included in the sensor network 102 are disposed on turbine blades, pipelines, marine environments, building structures, and / or other settings where material degradation is a concern. The sensor assemblies 104 detect, via a sensor device 106, variations in electrical parameters such as resistance, inductance, and capacitance, which change as certain material in the sensor device 106 changes or degrades. More specifically, the sensor device 106 includes a sacrificial layer 108 that in turn includes one or more sensor elements 110. In some examples, the sensor element 110 are manufactured via sputter silver materials and will change or degrade based on environmental conditions, such as material erosion, particulate buildup, material fatigue, cracks, and / or corrosion.

[0050] Various individual sensor elements 110 are positioned on along the sensor device 106, thus providing coverage at a variety of physical locations. Similarly, multiple sensor devices 106 can be positioned along desired areas to be monitored, for example, in a physical structure such as a wind turbine, a bridge column, a roof, and underwater support structure, and so on.During monitoring, the sensor elements 110 of the sacrificial layer 108 will experience the same environmental conditions as the structure(s) that they are placed in. For example, chemical erosion such as through acid rain, mechanical erosion due through windborne-sand, ultraviolet (UV) erosion, and the like, are experienced by both the sensor elements 110 of the sacrificial layer 108 and by the physical structure. Accordingly, both the sensor elements 110 of the sacrificial layer 108 and by the physical structure generally show the same types and rate of erosion.

[0051] Also shown is a reference layer 112 that includes multiple sensor element 110 shielded from environmental exposure. That is, the reference layer 112 includes a protective layer 114 that minimizes or eliminates erosion. Accordingly, the sensor elements 110 of the reference layer 112 are used to maintain baseline measurements, providing a comparison point for the exposed sensor elements 110 of the sacrificial layer 108. By comparing various sensor readings, such as voltage, capacitance, resistance,temperature, and / or pressure, between sensor elements 110 of the sacrificial layer 108 and sensor elements 110 of the reference layer 112, rates of erosion and / or change are detected.

[0052] As illustrated, the sensor elements 110 do not use power, but instead, their material makeup is designed to report changes, for example, due to erosion, via external techniques. The sensor element 110 is constructed using sputter metal or non-metallic witness materials, including strips of metal or non-metallic material, which are engineered to detect variations in electrical parameters such as resistance, inductance, and capacitance as the material erodes. These changes in electrical properties provide a measurable indication of the extent of erosion or degradation occurring on the sensor element 110. In some examples, the material used for the sensor element 110 includes conductive nanoscale particles, such as carbon nanotubes (CNTs), graphene, or “Mxenes”, embedded in the sensor material enhances its sensitivity and functionality. These nanoparticles can form networks, layers, agglomerates, or fibers within the sensor element 110 material, creating conductive pathways that change as the material erodes. The electrical resistance of the sensor changes under deformation and failure, for example, due to the breaking of contacts between adjacent nanoparticles and changes in the tunneling resistance between them.

[0053] In the context of the sensor assembly, multiple sensor elements 110 are arranged in a pattern that allows for comprehensive monitoring of the surface area of interest. This arrangement enables the system to detect and measure erosion or degradation across different regions of the monitored structure. The sensor element 110 is designed to mimic the behavior of physical degradation processes, acting as a sacrificial layer that experiences similar environmental conditions as the structure being monitored. This design allows the sensor to provide a representative measure of the erosive or degradative forces acting on the monitored structure. The erosion of the sensor element 110 results in permanent physical changes that can be measured at any time without the need for continuous power, making it suitable for long-term deployment in remote or harsh environments. This feature addresses limitations of traditional powered sensors that use frequent battery replacements or complex wiring systems

[0054] In some examples, the sensor assembly 104 includes a sensor processing and communications system 116 that can be powered externally, such as via magnetic inductance, radio frequency, and so on, and is communicatively coupled to the one or more sensor elements 110 of both the sacrificial layer 108 and the reference layer 112. Accordingly, the sensor processing and communications system 116 can be used to compare certain properties, such as voltages, inductances, capacitances, resistances, and so on, between the sensor elements 110 that are eroding versus the sensor elements 110 that are being used for reference. For example, the sensor processing and communications system 116 can transmit electric current to each of the sensor elements 110, and the sense a response, such as a voltage, an inductance, a capacitance, a resistance and so on.

[0055] In some examples, comparisons between readings of sensor elements 110 are done via a sensor element 110 in the sacrificial layer 108 that have the same column as a sensor element 110 in the reference layer 112. In other examples, all of the sensor readings of the sensor elements 110 of the reference layer 112 can be combined into an average or mean reference reading.

[0056] In some examples, a power supply 118 is provided, such as a solar power supply, a battery power supply, an inductance-based power supply, and so on, used to power the sensor assembly 104. The power supply 118 can operate intermittently, powered only when measurements are taken, to conserve energy and reduce operational costs. In some examples, the sensor processing and communications system 116 and / or the power supply 118 are communicatively and / or operatively coupled to a data acquisition and logging system 120 and a data analytics system 122 which are external to the sensor assembly 104.

[0057] The data acquisition and logging system 120 is a component of the overall sensor system 100 designed to collect and record data from the sensor assemblies. This system interfaces with the sensor network 102, which includes one or more sensor assemblies 104. The data acquisition and logging system 120 is responsible for, collecting data from the sensor elements 110, for example, via the sensor processing and communicationssystem 116. In some examples, the data collected is raw sensor data, which includes the aforementioned sensor readings such as resistance, inductance, and capacitance that occur as the sensor elements 110 erode or degrade. The collected data is stored and then processed for further analysis.

[0058] The data acquisition and logging system 120 additionally manages the timing and frequency of data collection, which can be triggered by digital (schedule-based or event-based) or physical (based on a level of physical degradation) triggers. In some examples, the data acquisition and logging system 120 interfaces with the power supply 118 to provide for more efficient operation, potentially operating intermittently to conserve energy. The data acquisition and logging system 120 can incorporate software- defined data acquisition capabilities, allowing for adaptive sampling rates and data collection strategies based on environmental conditions or specific monitoring requirements. This flexibility enables the sensor system 100 to optimize data collection for different types of infrastructure and varying degradation processes.

[0059] In some examples, the data acquisition and logging system 120 incorporates software-defined data acquisition capabilities, allowing for adaptive sampling rates and data collection strategies based on environmental conditions or specific monitoring requirements. This flexibility enables the system to optimize data collection for different types of infrastructure and varying degradation processes. The data acquisition and logging system 120 can also perform initial data processing tasks, such as filtering noise, normalizing sensor readings, or calculating preliminary damage equivalent load (DEL) computation and values based on the sensor measurements received.

[0060] The data acquisition and logging system 120 is communicatively coupled to the data analytics system 122, and provides the sensor data to the data analytics system 122 for further processing. The data analytics system 122 applies DEL analysis to provide insights into the cumulative impact of environmental factors on the monitored structures, as further described below. Data analysis mode additional or alternative to DEL analysis is also provided, such as multi-fidelity modeling and physics-informed modeling.Multi-fidelity modeling applies various modeling techniques to analyze the sensor data, including physics-based simulations, machine learning algorithms, and hybrid approaches that combine both. Physics-informed modeling provides for models that integrate physical principles with data- driven approaches to improve the accuracy and interpretability of the analysis. In one example, condition prediction is provided via a digital twin of one or more of the sensor elements 110. That is, the data analytics system 122 creates DEL, multi-fidelity, and / or physics-based digital representations of the materials and / or structures being sensed, enabling predictive maintenance and operational optimization.

[0061] Additionally, the data analytics system 122 creates a digital twin to create a virtual representation of physical assets so that the digital twin mimics the behavior and degradation patterns of the physical assets. In some examples, the data analytics system 122 implements digital twinning through:

[0062] Physical- Virtual Mapping: The sensor network creates a digital representation of the physical structure by using nano-engineered sensors that experience and record the same environmental conditions as the monitored structure.

[0063] Data Collection: The data analytics system 122 collects data about the structure's condition through multiple sensor assemblies 104 and / or sensor devices, creating a dynamic digital model that reflects the current state of the physical asset.

[0064] Predictive Modeling: The data analytics system 122 uses the collected data to create physics-informed models and damage equivalent loading (DEL) calculations that predict how the physical structure will behave and degrade over time.

[0065] Active Learning: The digital twin continuously updates and improves its predictions through active learning and uncertainty quantification, allowing it to provide increasingly accurate condition predictions and actionable insights. This digital twin approach enables the system 100 to provide real-time monitoring, predictive maintenance recommendations, and stochastic assessment of the physical asset's condition without requiring constant power or manual inspection.

[0066] The data analytics system 122 additionally provides for active learning and uncertainty quantification. These techniques are employed to continuously improve the data analytics system's predictive capabilities and provide confidence levels for its analyses. Based on its analyses, the data analytics system 122 can generate alerts and provide recommendations for maintenance or operational adjustments. The data analytics system 122 is designed to handle the continuous stream of data from the sensor network 102, including continuous or intermittent data from each element of the sensor network, providing actionable analysis of infrastructure conditions. The analytic capabilities of the data analytics system 122 enable early detection of anomalies, predictive maintenance planning, and optimization of asset performance across various industries such as wind energy, offshore production, subsea systems, refinery operations, and so on. For example, a digital twin (c.g., virtual representation) of a turbine blade can be created, that mimics the degradation of a physical turbine blade that is being monitored by on or more sensor assemblies 104.

[0067] FIG. 2 illustrates a structure 200 that can be monitored via the sensing techniques described herein, in accordance with some examples. More specifically, the structure 200 of the depicted example is wind turbine blade 202 that has been equipped with multiple sensor assemblies 104 for monitoring of erosion and / or corrosion. It is to be understood that while three sensor assemblies 104 are shown, more (or less) sensor assemblies 104 can be disposed to monitor the wind turbine blade 202, thus creating the sensor network 102.

[0068] The wind turbine blade 202 is shown with its tip 204 and a leading edge 206 shown in picture form. These areas are particularly susceptible to erosion and corrosion during operation and show “cratering” that forms during use. The sensor assemblies 104 are strategically placed at various locations 208 along the blade's surface. These locations 208 are chosen to monitor certain areas prone to degradation, such as leading edge 206. As mentioned earlier, the sensor assemblies 104 are designed to detect and measure environmental conditions or loading that result in changes in the blade’s surface condition over time.

[0069] The figure includes two inset images providing detailed views of the erosion and corrosion effects. An upper inset image shows a close-up view of the leading edge 206, revealing pitting and material loss characteristic of erosion and corrosion damage. The lower inset image focuses on the blade tip 204, also displaying signs of wear. As the sensor network 102 detects wear of the structure 200, the sensor network 102 works in conjunction with the data acquisition and logging system 120 and the data analytics systems 122 to provide a more comprehensive monitoring of the wind turbine blade's condition. Indeed, the design of the sensor assemblies 104 allow for continuous monitoring of the blade's condition without the need for constant power, addressing the challenges of monitoring large, remote structures like wind turbine blades. This sensor system 100 enables early detection of erosion and corrosion, facilitating predictive maintenance and potentially extending the operational life of the wind turbine blades. In the depicted example, the sensor assemblies 104 include certain geometries useful in improving sensing techniques, as further described below.

[0070] FIG. 3A and 3B are perspective views illustrating example sensor assemblies 300, 302, that include certain geometries useful in sensing modalities that include non-planar surfaces, in accordance with some examples. The sensor assemblies 300, 302 are equivalent to the sensor assemblies 104 shown in FIG. 1 and FIG. 2. In FIG. 3A, the sensor assembly 300 includes two sensor devices 106 disposed on a substrate 304. In certain examples, the substrate 304 is a flexible substrate 304 that forms the base of the assemblies 300, 302. In some examples, the substrate 304 is non-flexible but has been formed to include a curved section 306. The substrate 304 can include a printed circuit board material, such as Polyester (PET) or Poly ami de(PI). The substrate 304 can then be fastened onto a desired surface, for example, using adhesives or by other techniques.

[0071] The substrate 304 is engineered to conform to curved surfaces, as shown by the curved section 306. While a single curve section is shown, it is to be understood that multiple curved surfaces can be provided, including sections that result in 360° coverage of a structure such as a column. The curved design allows the sensor assemblies 300, 302 to closely follow thecontours of structures like wind turbine blades, columns, and the like, ensuring more optimal contact and measurement accuracy.

[0072] In the depicted example, multiple sensor devices 106 are integrated into the substrate 304 and strategically placed along the curved section 306 to monitor different areas of a sensed structure, such as the wind turbine blade 202. The sensor devices 106 are designed to detect changes in electrical properties (such as resistance, inductance, and capacitance) as the material erodes or degrades, providing a measurable indication of the structure's condition.

[0073] A section 308 (e.g., t-section or tape or any other form) includes one or more electronics packages 310. The electronics packages 310 are communicatively coupled to the one or more sensor device 106 and can include the sensor processing and communications system 116 and / or the power supply 118. Also shown is a stagnation line 12. The stagnation line 312 typically defines a location where erosion loads are highest and then spread (e.g., chord-wise) from the stagnation line 312. As the structure 200, such as the wind turbine blade 202, experiences leading-edge erosive loading (due to rain, hail, snow etc.), the sensor device 106 sees the same loading conditions and changes the associated properties, such as resistance, capacitance, inductance, conductance, and so on.

[0074] In FIG. 3B, a third sensor device 106 is disposed in a section 314 (e.g., t-section or tape or any other form), in accordance with some examples. The use of the third sensor device 106 provides for sensing at a location, such as an upper and / or lower surface aft of the leading edge 206. FIG. 3B additionally illustrates an electronics package 316, that can include the sensor processing and communications system 116 and / or the power supply 118. The overall design of the sensor assemblies 300, 302 reflects their purpose of providing a more comprehensive, continuous monitoring of curved surfaces in harsh environments, such as those experienced by wind turbine blades. Indeed, the use of flexible structures and curved sections allows for easy installation and provides that the sensor device 106 maintain close contact with the monitored surface, enhancing the accuracy and reliability of the erosion and degradation measurements.

[0075] FIG. 4 is a top view of the sensor device 106 illustrating a placement of the stagnation line 312 about a middle of the sensor device 106, in accordance with some examples. As mentioned earlier, the stagnation line 312 marks a location where erosion loads are highest and then spread, for example, in directions 402, 404. Accordingly, sensor elements 110 further away from the stagnation line 312 should physically degrade at a lower rate when compared to sensor elements 110. The sacrificial layer 108 works in at least two forms. In a first form, there is change in resistance for each sensor element 110 in the sacrificial layer 108 that corresponds to the accumulated damage equivalent loading at the sensor device's location. Then, in a second form, sensor devices 106 along the sacrificial layer 108 (e.g., in directions 402 and / or 404) show the progress of the loading or severity of the loading away from the stagnation line 312. Accordingly, the sacrificial layer 108 could be sensing i) for mechanical loading (erosion, cracks, gaps / separation, fatigue) and / or ii) chemical loading (corrosion, gas leaks) exposure. The sacrificial layer 108 could be metallic or non-metallic, and can be manufactured by machining, molding or additive manufacturing. Data from the sensor elements 110 is transmitted to the electronics packages 310, 316 shown in FIGS. 3A, 3B via printed circuits, carbon nanotubes (CNF) elements, fullerene elements and the like. Based on the circuit arrangement, the sensor elements 110 act as resistors, capacitors, inductors, and so on.

[0076] FIG. 5A illustrates a perspective view of a wind turbine blade 502 equipped with an example of the sensor assembly 104, in accordance with some examples. The blade 502 is depicted as a three-dimensional structure with a curved, aerodynamic profile found in certain wind turbine blades. The sensor assembly 104 is shown mounted on the surface of the blade 502, positioned to monitor areas prone to erosion and degradation, including corrosion.

[0077] The sensor assembly 104 is shown as a thin, elongated structure conforming to the curvature of the blade's surface. This design allows the sensor to maintain close contact with the monitored area, providing for more accurate measurements of surface conditions. The perspective view of FIG. 5 A shows how, in some example, the sensor assembly 104 integrates withthe overall structure of the wind turbine blade 502, demonstrating its low- profile design that minimizes interference with the blade's aerodynamic properties. This placement enables continuous monitoring of the blade's condition with minimal impact on aerodynamic performance.

[0078] FIG. 5B illustrates a top view of a wind turbine blade 502 equipped with an example of the sensor assembly 104, in accordance with some examples. A leading edge protection 514, such as a film, a substrate, a paint, and so on is depicted as an outermost layer of the blade’s leading edge. This protective layer is designed to help shield the blade from environmental factors that can cause erosion and degradation, such as rain, hail, and airborne particles. Beneath the leading edge protection 514, the blade's composite structure 516 is shown.

[0079] The composite structure 516 forms the core of the blade, providing the strength and flexibility used during wind energy capture operations. The sensor assembly 104 is integrated into this layered structure, positioned to monitor the condition of both the leading edge protection 514 and the underlying composite structure 516. The placement of the sensor assembly 104 allows for continuous monitoring of the blade's condition, particularly in areas most susceptible to erosion and other forms of degradation. The top view of FIG. 5B illustrates how the sensor assembly 104 is incorporated into the blade's sections enabling the detection of changes in the blade’s condition, including both the blade's leading edge (e.g., leading edge 206) as well as the blade's main body, in accordance with some examples.

[0080] FIG. 5C illustrates a side view of a wind turbine blade 502 with an integrated sensor assembly 104, in accordance with some examples. More specifically, the figure provides a side view of the blade's leading edge (e.g., leading edge 206), showing an example placement and integration of the sensor assembly 104. In the depicted example, the wind turbine blade 502 is depicted in a simplified cross-sectional form, showing its aerodynamic profile. The leading edge of the blade, which is the part that first encounters the airflow, is shown as having a curved section 306. The sensor assembly 104 is shown mounted on the leading edge of the blade 502, with a curved section of the sensor assembly 104 having the same geometry as the curvedsection 306. Accordingly, the sensor assembly 104 is disposed conformal with the wind turbine blade 502, providing for full contact with the wind turbine blade 502 and thus acquiring more accurate measurements.

[0081] Turning now to FIG. 6A, the figure is a perspective top view of a sensor assembly 602 having a pre-engineered crack 604, in accordance with some examples. FIG. 6A illustrates a top view of a sensor assembly 104 featuring a pre-engineered crack 604. The sensor assembly 602 is equivalent to the sensor assembly 104 of previous figures. In the depicted example, the sensor assembly 104 is designed to mimic the failure behavior of the structure under observation, providing for a way to monitor and predict potential structural issues.

[0082] The pre-engineered crack 604 is intentionally manufactured to simulate the development and progression of cracks in the monitored structure. This crack 604 is engineered to respond to mechanical stresses and environmental factors in a manner similar to how actual cracks would form and propagate in the structure being monitored. In some examples, the pre-engineered crack 604 is located in a section 606. The section 606 can include one or more sensor devices 106. In other examples, the section 606 is a sensor device 106 and thus includes one or more sensor elements 110. As the crack 604 expands, the sensor assembly 104 will then continually monitor the crack expansion, and certain analyses, such as DEL analysis, multi-fidelity modeling and / or physics-informed modeling. By incorporating the pre-engineered crack 604, the sensor assembly 104 can provide early detection of potential structural issues, enabling proactive maintenance and enhancing the longevity of the monitored infrastructure.

[0083] FIG. 6B illustrates a bottom perspective view of the sensor assembly 602 of FIG. 6A, illustrating the pre-engineered crack 604, in accordance with some examples. As mentioned earlier, the crack 604 is designed to simulate the development and progression of cracks in the monitored structure. As the crack 604 develops and grows under applied stresses, it causes changes in the electrical properties of the conductive network within the sensor material, such as sensor elements 110 disposed insection 604, which can be measured and analyzed to provide insights into the structural health of the monitored component.

[0084] Two protrusions 608 are shown. These protrusions 608 are structural elements of the sensor assembly 104 that extend from the main body. In some examples, the protrusions 608 help to create specific stress concentrations and to facilitate the attachment of the sensor assembly 104 to the monitored structure. A valley 610 is formed between the two protrusions 608. In some examples, the valley 610 is designed to enhance the sensor's sensitivity to certain types of stress or strain, and it can serve as a channel for environmental factors that contribute to the degradation process being monitored. Accordingly, the example sensor assembly 104 of FIGS. 6A, 6B provides for continuous, power-efficient monitoring of structural conditions in various applications, such as wind turbine blades, offshore structures, and / or other infrastructure components.

[0085] FIG. 7A illustrates a side view of an example monitoring of a structure (e.g., pipe) 704 utilizing two sensor assemblies 602 attached to the structure 704 via an adhesive layer 702, in accordance with some examples. In the depicted example, the sensor assemblies 602 are positioned on opposite sides of the structure 704, allowing for more comprehensive monitoring of the pipe's structural integrity.

[0086] The adhesive layers 702 serve to affix the sensor assemblies 602 to the structure's surface, ensuring a more consistent contact and improved measurements. Arrows labeled 706 indicate the directions of tension (and / or compression) forces acting on the structure 704. The sensor assemblies 602 detect and measure both tensile and compressive stresses experienced by the structure 704. Accordingly, and as shown in FIG. 7B, the sensor assemblies 602 include sensor elements 708 suitable for use as strain gauges, such as uni-axial strain gauges.

[0087] In some examples as shown in FIG. 7B, the sensor elements 708 are based on the sensor elements 110. That is, the sensor elements 708 are manufactured using sputter silver materials, which are engineered to detect variations in electrical parameters such as resistance, inductance, and capacitance as the material erodes. In some examples, the material used forthe sensor elements 708 includes conductive nanoscale particles, such as carbon nanotubes (CNTs), graphene, or “Mxenes”, embedded in the sensor material enhances its sensitivity and functionality. As the sensor elements 708 are pulled or compressed, certain electrical properties, such as resistance, capacitance (c.g., between two sensor elements 708), inductance, and so on.

[0088] This configuration enables the capture of data on the structure's performance under various load conditions, potentially detecting early signs of fatigue, stress concentrations, or other structural issues. The dual-sensor setup also allows for comparative analysis between the two sides of the structure, which can be valuable in identifying asymmetric loading or localized damage.

[0089] FIG. 8A is a sectional side view illustrating a sensor assembly 602 disposed onto a structure 802, in accordance with some examples. Tn the depicted embodiment, the sensor assembly 602 is embedded into the structure 802 and the valley between the protrusions 608 is filled in, for example, using a filler 804 such as a low stiffness epoxy. Also shown is a detachable protective cover 806, which can be a steel cover. The cover 806 is shown as protecting a connector 808. The connector 808 is used to communicatively couple to the sensor assembly 602, for example, to capture sensor readings.

[0090] FIG. 8B is a sectional side view illustrating the example sensor assembly 602 of FIG. 8A but with the cover 806 removed. Also shown is a communications device 810 communicatively coupled to the connector 808. The communications device 810 is used to retrieve certain properties from the sensor assembly 602, such as resistance, capacitance, inductance, and so on, and transmit (e.g., via wired and / or wireless techniques) these properties to an external device 812, such as a tablet, a laptop, a cell phone, and so on or directly to a Cloud hosted API where the information will be stored in a database.

[0091] In some examples, the communications device 810 can also log data from the sensor assembly 602 for later analysis, analyze data on site, and provide other services, such as upgrading any firmware found in the sensorassembly 602, recharging any power supply of the power assembly 602, and so on. In the depicted example, the external device 812 is additionally communicatively coupled to a cloud 814. The cloud 814 then provides services including storage of data from the sensor assembly 602, analysis of data, sharing of data, and so on, as described further below. It is also to be noted that similar (or the same) services as those provided via the cloud 814 can be provided via the external device 812.

[0092] FIG. 9A is a block diagram of a sensor assembly 902 having a preengineered crack illustrating a 4-point bend setup, according to some examples. In the depicted embodiment, the sensor assembly 902 is shown disposed on a host structure 904, such as a column, a pipe, a support structure, and so on. The 4-point bend setup is a test setup useful in determining a modulus of elasticity in bending, a flexural stress, a flexural strain, fatigue, flexural stiffness, and the like, which can be used to calibrate the sensor assembly 902. As with other sensor assemblies discussed herein, the sensor assembly 902 is equivalent to the sensor assembly 104 and has the same of similar components. By using the 4-point bend setup as a calibration technique, the sensor assembly 902 can be calibrated to improve its sensing accuracy.

[0093] FIG. 9B is a perspective view of the now calibrated sensor assembly 902, according to some examples. In the depicted example, multiple (e.g., four) sensor assemblies 902 are disposed about a monitored structure 906. As the monitored structure 906, flexes, the sensor assemblies 902 will experience stresses that can then be monitored through time. As mentioned above, the monitoring can include analyzes performed by the data analytics system 122, including DEL, physics-based simulations, finite element analysis of impacts, fatigue damage, wear, and so on. By providing for calibrated sensor assemblies 902, the techniques described herein result in more accurate and reliable analytics of the monitored structure 906.

[0094] FIG. 10 illustrates a sensor erosion detection system 1002 having a leading edge protection (LEP) layer 1004 that is applied onto two example sensor assemblies 1006 and 1008, in accordance with some examples. Also shown are two sensor assemblies 1010, 1012 of the sensor erosion detectionsystem 1002 that are not covered by the LEP layer 1004. The sensor erosion detection system 1002 described herein enables for a monitoring of the health and adherence of the LEP layer 1004 via the four sensor assemblies 1006, 1008, 1010, and 1012. For example, the sensor erosion detection system 1002 provides for a sensor system that not only measures the erosion of the LEP layer 1004 but also tests the adhesion quality of the LEP layer 1004 over time.

[0095] The humidity sensor 1006 located under the LEP layer 1004, monitors moisture levels beneath the LEP layer 1004, which can indicate if the adhesive bond is compromised by the sensing of an ingress of moisture. Also positioned under the LEP layer 1004, is a sensor assembly 1008 that measures the pressure exerted on the structure’s surface through the LEP layer 1004. This measurement aids in tracking changes in the LEP layer thickness due to, for example, erosion.

[0096] The unprotected rain sensor 1010 detects the presence of rain and serves as a trigger for the sensor erosion detection system's active monitoring mode. That is, a power supply system 1014 operates on ultralow power usage and, in some examples, the rain sensor 1010 is used to trigger the power supply system 1014 to increase power delivery and engage a data analytics system 1016. In some examples, the power supply system 1014 includes a non-rechargeable battery that must last the life of the device or requires manual replacement. In other examples a rechargeable battery is used in conjunction with energy harvesting (e.g. solar cell), charging via a magnetic inductance system, and so on. In both examples, power must be conserved while recording data at critical times and intervals. The data analytics system is configured to operate the sensor erosion detection system 1002, including adjusting the power supply 1014 to operated in the desired mode.

[0097] When analyzing sensor 1006, 1008, 1010, and 1012 data, the surface pressure sensor 1008 acts as a reference point, measuring the direct impact of rain on the structure's surface. It allows for comparative analysis with the internal surface pressure sensor readings.

[0098] Under normal conditions, the sensor erosion detection system 1002 operates in an ultra-low power mode, with only the rain sensor active. When rain is detected, the sensor erosion detection system 1002 briefly activates to collect data from all sensors. The external surface pressure sensor 1012 provides a baseline for rain intensity, allowing for accurate comparison with internal sensor readings. By correlating these measurements with initial data from when the LEP layer 1004 was intact, the system can estimate LEP layer thickness reduction due to erosion. For example, as the LEP layer 1004 erodes, there is a higher surface pressure reading via the sensor 1008. The sensor 1012 includes an erosion protection, and thus doesn’t erode (or erodes much less) when compared to the sensor 1008. A comparison between the sensed pressures of sensors 1008 and 1012 is then used to determine how much erosion has occurred. The comparison can include the use of Al, such as by training a neural network with data at various levels or degrees of erosion of the layer 1004. Likewise, physical modeling, such as removing material at various depths from the layer 1004 and then measuring pressure differences between sensors 1008 and 1012. The humidity sensor 1006 is then used in detecting adhesion failures by identifying moisture ingress under the LEP layer 1004. After taking measurements via the sensor 1006, 1008, 1010, and 1012, the power supply system 1014 goes back to ultra-low power operations to wait until a next rain event. The sensor erosion detection system 1002 allows for continuous, power-efficient monitoring of both erosion progression and adhesion quality of the LEP layer 1004, enabling proactive maintenance and extending the operational life of the observed structure.

[0099] FIG. 11 illustrates a flowchart depicting a process 1100 for applying the techniques described herein, according to some examples. Although the example process 1100 depicts a particular sequence of operations, the sequence may be altered without departing from the scope of the present disclosure. For example, some of the operations depicted may be performed in parallel or in a different sequence that does not materially affect the function of the process 1100.

[0100] At block 1102, the process begins with selecting and positioning sensors. This block involves selecting and placing sensor assemblies 104 ona structure, such as one or more wind turbine blades, to monitor specific areas or components of the structure under observation. The sensor assemblies selected can include sensor assemblies having certain geometries useful in conformably fitting on the structure under observation.

[0101] At block 1104, the process 1100 continuously monitors the sensed structure via the one or more sensor assembly 104. In some examples, the sensor assemblies 106 mirror degradation that is being experienced by the monitored structure without using power. As mentioned earlier, the sensor assemblies 104 are manufactured via materials and will change or degrade based on environmental conditions, such as material erosion, particulate buildup, material fatigue, cracks, and / or corrosion. That is, at the core of each sensor assembly 104 is a sensor device 106 consisting of multiple sensor elements 110. These strips may be arranged in various configurations to optimize detection capabilities.

[0102] At block 1104, the process 1100 collects sensor data. In certain examples, data collection occurs at a given schedule, such as daily, weekly, monthly, yearly, and so on. The data collection is performed by testing the sensor element 110 for certain electrical properties, such as resistance, capacitance, inductance, and so on. This involves the data acquisition and logging system 120 receiving and storing the transmitted data from all sensor assemblies 104 in the sensor network 102 disposed about the monitored structure.

[0103] At block 1106, the system analyzes the sensor data. The data analytics system 122 retrieves the stored data and performs advanced analysis, which may involve machine learning algorithms, statistical modeling, or other sophisticated techniques to identify trends, anomalies, or potential issues in the monitored structure. In some examples, the data analysis incorporates the concept of damage equivalent load (DEL).Through laboratory tests or field observations, a relationship is established between the environmental (e.g, erosive) damage equivalent load measured by the sensor and the damage equivalent load experienced by the surface or coating being monitored. This relationship allows for the conversion of sensor measurements into an assessment of the current damage equivalenterosion load on the material surface of interest. The relationship between sensor DEL and surface DEL can be established through empirical testing and curve fitting using mathematical relationships and / or machine learning algorithms. Physics-based simulations and finite element analysis of impacts, fatigue damage, and wear may also be employed to determine this relationship. Additional data such as weather conditions (e.g., rainfall rate, wind, temperature) and impact speed of droplets can be included in the analysis to account for non-linearity that may exist between different material types.

[0104] At block 1108, the system provides reports and alerts based on the analysis. If critical issues are detected, the system may generate alerts or notifications for relevant personnel. Additionally, the system may produce regular reports summarizing the structure's condition and any identified trends or concerns. This process enables continuous, energy-efficient monitoring of certain infrastructure, such as wind turbine blades, offshore structures, or industrial facilities. The use of multiple sensor assemblies and advanced data analytics allows for comprehensive and proactive maintenance strategies, potentially reducing downtime and extending the operational life of monitored structures.

[0105] At block 1110, the system uses the analyzed data to provide detailed reports and send alerts. If issues like erosion, corrosion, or structural stress are detected, notifications are sent to the maintenance team for timely action. The system also generates regular reports that summarize the condition of wind turbine blades or other monitored structures, tracking changes and trends over time. This continuous, energy-efficient monitoring helps prevent unexpected failures, supports proactive maintenance, reduces downtime, and extends the life of valuable infrastructure.

[0106] FIG. 12 is a diagrammatic representation of a machine 1200 within which instructions 1202 (e.g., software, a program, an application, an applet, an app, or other executable code) for causing the machine 1200 to perform any one or more of the methodologies discussed herein may be executed. For example, the instructions 1202 may cause the machine 1200 to execute any one or more of the processes or methods describedherein. The instructions 1202 transform the general, non-programmed machine 1200 into a particular machine 1200. The machine 1200 may operate as a standalone device or may be coupled (e.g., networked) to other machines. In a networked deployment, the machine 1200 may operate in the capacity of a server machine or a client machine in a server-client network environment, or as a peer machine in a peer-to-peer (or distributed) network environment. The machine 1200 may comprise, but not be limited to, a server computer, a client computer, a personal computer (PC), a tablet computer, a laptop computer, a netbook, a set-top box (STB), a personal digital assistant (PDA), an entertainment media system, a cellular telephone, a smartphone, a mobile device, a wearable device (e.g., a smartwatch), a smart home device (e.g., a smart appliance), other smart devices, a web appliance, a network router, a network switch, a network bridge, or any machine capable of executing the instructions 1202, sequentially or otherwise, that specify actions to be taken by the machine 1200. Further, while a single machine 1200 is illustrated, the term “machine" shall also be taken to include a collection of machines that individually or jointly execute the instructions 1202 to perform any one or more of the methodologies discussed herein. In some examples, the machine 1200 may also comprise both client and server systems, with certain operations of a particular method or algorithm being performed on the server-side and with certain operations of the particular method or algorithm being performed on the client-side.

[0107] The machine 1200 may include processors 1204, memory 1206, and input / output I / O components 1208, which may be configured to communicate with each other via a bus 1210. In an example, the processors 1204 (e.g., a Central Processing Unit (CPU), a Reduced Instruction Set Computing (RISC) Processor, a Complex Instruction Set Computing (CISC) Processor, a Graphics Processing Unit (GPU), a Digital Signal Processor (DSP), an Application-Specific Integrated Circuit (ASIC), a Radio- Frequency Integrated Circuit (RFIC), another processor, or any suitable combination thereof) may include, for example, a processor 1212 and a processor 1214 that execute the instructions 1202. The term "processor" is intended to include multi-core processors that may comprise two or more independent processors (sometimes referred to as “cores”) that may executeinstructions contemporaneously. Although FIG. 12 shows multiple processors 1204, the machine 1200 may include a single processor with a single-core, a single processor with multiple cores (e.g., a multi-core processor), multiple processors with a single core, multiple processors with multiples cores, or any combination thereof.

[0108] The memory 1206 includes a main memory 1216, a static memory 1218, and a storage unit 1220, both accessible to the processors 1204 via the bus 1210. The main memory 1216, the static memory 1218, and storage unit 1220 store the instructions 1202 embodying any one or more of the methodologies or functions described herein. The instructions 1202 may also reside, completely or partially, within the main memory 1216, within the static memory 1218, within machine-readable medium 1222 within the storage unit 1220, within at least one of the processors 1204 (e.g., within the processor’s cache memory), or any suitable combination thereof, during execution thereof by the machine 1200.

[0109] The I / O components 1208 may include a wide variety of components to receive input, provide output, produce output, transmit information, exchange information, capture measurements, and so on. The specific I / O components 1208 that are included in a particular machine will depend on the type of machine. For example, portable machines such as mobile phones may include a touch input device or other such input mechanisms, while a headless server machine will likely not include such a touch input device. It will be appreciated that the I / O components 1208 may include many other components that are not shown in FIG. 12. In various examples, the I / O components 1208 may include user output components 1224 and user input components 1226. The user output components 1224 may include visual components (e.g., a display such as a plasma display panel (PDP), a light-emitting diode (LED) display, a liquid crystal display (LCD), a projector, or a cathode ray tube (CRT)), acoustic components (e.g., speakers), haptic components (e.g., a vibratory motor, resistance mechanisms), other signal generators, and so forth. The user input components 1226 may include alphanumeric input components (e.g., a keyboard, a touch screen configured to receive alphanumeric input, a photo- optical keyboard, or other alphanumeric input components), point-basedinput components (e.g., a mouse, a touchpad, a trackball, a joystick, a motion sensor, or another pointing instrument), tactile input components (e.g., a physical button, a touch screen that provides location and force of touches or touch gestures, or other tactile input components), audio input components (e.g., a microphone), and the like.

[0110] In further examples, the I / O components 1208 may include biometric components 1228, motion components 1230, environmental components 1232, or position components 1234, among a wide array of other components. For example, the biometric components 1228 include components to detect expressions (e.g., hand expressions, facial expressions, vocal expressions, body gestures, or eye-tracking), measure biosignals (e.g., blood pressure, heart rate, body temperature, perspiration, or brain waves), identify a person (e.g., voice identification, retinal identification, facial identification, fingerprint identification, or electroencephalogram-based identification), and the like. The motion components 1230 include acceleration sensor components (e.g., accelerometer), gravitation sensor components, rotation sensor components (e.g., gyroscope).

[0111] The environmental components 1232 include, for example, one or cameras (with still image / photograph and video capabilities), illumination sensor components (e.g., photometer), temperature sensor components (e.g., one or more thermometers that detect ambient temperature), humidity sensor components, pressure sensor components (e.g., barometer), acoustic sensor components (e.g., one or more microphones that detect background noise), proximity sensor components (e.g., infrared sensors that detect nearby objects), gas sensors (e.g., gas detection sensors to detection concentrations of hazardous gases for safety or to measure pollutants in the atmosphere), or other components that may provide indications, measurements, or signals corresponding to a surrounding physical environment. The position components 1234 include location sensor components (e.g., a global positioning system (GPS) receiver component), altitude sensor components (e.g., altimeters or barometers that detect air pressure from which altitude may be derived), orientation sensor components (e.g., magnetometers), and the like.

[0112] Communication may be implemented using a wide variety of technologies. The I / O components 1208 further include communication components 1236 operable to couple the machine 1200 to a network 1238 or devices 1240 via respective coupling or connections. For example, the communication components 1236 may include a network interface component or another suitable device to interface with the network 1238. In further examples, the communication components 1236 may include wired communication components, wireless communication components, cellular communication components, Near Field Communication (NFC) components, Bluetooth® components (e.g., Bluetooth® Low Energy), Wi-Fi® components, and other communication components to provide communication via other modalities. The devices 1240 may be another machine or any of a wide variety of peripheral devices (e.g., a peripheral device coupled via a universal serial bus (USB) port), internet-of-things (loT) devices, and the like.

[0113] Moreover, the communication components 1236 may detect identifiers or include components operable to detect identifiers. For example, the communication components 1236 may include Radio Frequency Identification (RFID) tag reader components, NFC smart tag detection components, optical reader components (e.g., an optical sensor to detect one-dimensional bar codes such as Universal Product Code (UPC) bar code, multi-dimensional bar codes such as Quick Response (QR) code, Aztec code, Data Matrix, Dataglyph, MaxiCode, PDF417, Ultra Code, UCC RSS-2D bar code, and other optical codes), or acoustic detection components (e.g., microphones to identify tagged audio signals). In addition, a variety of information may be derived via the communication components 1236, such as location via Internet Protocol (IP) geolocation, location via Wi-Fi® signal triangulation, location via detecting an NFC beacon signal that may indicate a particular location, and so forth.

[0114] The various memories (e.g., main memory 1216, static memory 1218, and memory of the processors 1204) and storage unit 1220 may store one or more sets of instructions and data structures (e.g., software) embodying or used by any one or more of the methodologies or functions described herein. These instructions (e.g., the instructions 1202), whenexecuted by processors 1204, cause various operations to implement the disclosed examples.

[0115] The instructions 1202 may be transmitted or received over the network 1238, using a transmission medium, via a network interface device (e.g., a network interface component included in the communication components 1236) and using any one of several well-known transfer protocols (e.g., hypertext transfer protocol (HTTP)). Similarly, the instructions 1202 may be transmitted or received using a transmission medium via a coupling (e.g., a peer-to-peer coupling) to the devices 1240. In some examples, the machine 1200 and / or the one or more of the devices 1240 at virtual reality devices, augmented reality devices, and / or mixed reality devices suitable for presenting the visualizations described herein.

[0116] The techniques described herein provide for a hazardous environment robot system that provides for heat resistance in hazardous environments, such as a residential or commercial fire. A multi-layered approach to thermal protection and specialized component design is presented, resulting a in temperature-hardened hazardous environment robot system. The robot system incorporates a High Reflectivity Surface Coating (HRSC), a Passive Thermal Protection System (PTPS), and an Electronics Thermal Protection System (ETPS) to manage both external and internal heat loads more effectively.

[0117] Some example embodiments are listed below:

[0118] A sensor system wherein the sacrificial layer experiences physical and chemical degradation that matches the monitored structure, ensuring more accurate environmental and mechanical impact measurements.

[0119] A sensor network incorporating energy-harvesting power supplies configured for intermittent power to sensor assemblies, enabling autonomous, long-term operation and reducing operational costs.

[0120] Sensor assemblies incorporating nano-engineered materials embedded with conductive nanoscale particles that create pathways, altering electrical properties to detect structural changes such as erosion, fatigue, and cracks.

[0121] A data analysis platform utilizing multi-fidelity modeling, machine learning, and predictive algorithms to enhance predictions of structural health, optimize maintenance schedules, and facilitate proactive operational decision-making.

[0122] A wireless-enabled sensor network designed for remote monitoring, enabling seamless data transmission and integration with cloud-based platforms for real-time processing and comprehensive data storage.

[0123] Sensor assemblies with reference layers protected from environmental exposure to provide baseline data for comparative analysis, ensuring accurate assessment of structural health changes over time.

[0124] Flexible sacrificial layers within sensor assemblies that conform to non-planar surfaces such as wind turbine blades or offshore platforms, allowing precise monitoring on curved or multi-faceted structures.

[0125] Pre-engineered cracks within sensor assemblies that enable measurement of the mechanical fatigue damage equivalent load by monitoring the sensor response to the crack propagation over time. An integrated system featuring rain detection sensors and external surface pressure sensors that activate monitoring during rain events to assess the health of leading-edge protection (LEP) layers and optimize energy consumption.

[0126] A comprehensive data analysis system that incorporates environmental conditions such as weather and impact forces for detailed assessments of structural degradation and predictive analysis of structural health.

[0127] Sensor assemblies with embedded communication and processing units that facilitate data transmission and on-site pre-analysis, reducing the amount of raw data transmitted to central systems.

[0128] protective covers on sensor assemblies designed to shield connectors and electronics from environmental damage, ensuring reliable long-term operation.

[0129] A damage equivalent load (DEL) analysis method calibrated through empirical testing to enhance the reliability of structural health assessments.

[0130] The damage equivalent load analysis further augmented by finite element modeling.

[0131] Nano-engineered sensor elements designed to change resistance, capacitance, and inductance properties as they erode, allowing for passive, continuous monitoring without an external power source.

[0132] Sensor assemblies equipped with modular designs that facilitate easy installation, maintenance, and replacement, improving operational efficiency.

[0133] A versatile sensor network adaptable for various environments and customizable to detect specific types of degradation, such as chemical corrosion, mechanical fatigue, and particulate buildup.

[0134] Configurable sensor networks capable of interfacing with multiple communication protocols, supporting seamless integration with existing infrastructure monitoring systems.

[0135] Adaptive sensor devices that respond to mechanical and chemical stresses, enabling versatile monitoring in diverse environmental conditions.

[0136] Protective coatings on sensor assemblies that provide enhanced durability against chemical exposure and physical wear, ensuring long-term reliability and stable baseline measurements.

[0137] Integration with cloud-based platforms for real-time processing, comprehensive monitoring, and historical data analysis, facilitating continuous infrastructure condition tracking.

[0138] Sensor systems capable of simultaneous multi-modal data collection to measure various structural impacts, such as stress, strain, and environmental exposure.

[0139] Energy-efficient operations achieved through the use of passive sensors that only activate under certain conditions, optimizing power consumption and extending sensor lifespan.

[0140] Advanced algorithms that integrate real-time sensor feedback, weather data, and historical trends for predictive analysis and proactive maintenance strategies.

[0141] Sensor systems designed with built-in redundancy features to ensure continuous data collection and reliable monitoring even in the case of sensor damage or failure.

[0142] Multi-layered sensor devices capable of distinguishing between surface erosion and deeper structural issues, providing comprehensive diagnostics of monitored assets.

[0143] Flexible monitoring solutions that include both contact and noncontact sensor types for diverse application needs, ensuring adaptability for various structural monitoring scenarios.

[0144] FIG. 13 illustrates machine learning engine for wind turbine blade leading edge material erosion rate prediction and cumulative damage index, in accordance with some embodiments. The machine learning engine may be deployed to execute at an edge computing device (NanoX sensor node (e.g., sensor 104) on the blade leading edge, mobile device), centralized hub (facility gateway or computer), or Cloud (server, data center). A system may calculate one or more weightings for criteria based upon one or more machine learning algorithms. FIG. 1 shows an example machine learning engine 1300 according to some examples of the present disclosure.

[0145] Machine learning engine 1300 uses a training engine 1302 and a prediction engine 1304. Training engine 1302 uses input data 1306, for example after undergoing preprocessing component 1308, to determine one or more features 1310. The one or more features 1310 may be used to generate an initial machine learning model 1312, which may be updated iteratively or with future labeled or unlabeled data (e.g., during reinforcement learning).

[0146] The input data 1306 may include:

[0147] - Simulation data: rain droplet size, rain impact velocity, rain intensity, blade speed, rate of erosion through depth and across length for the wind turbine blade leading edge material erosion scenarios.

[0148] - Laboratory test data: controlled water jet parameters (pressure, volume, distance from sample), test rig parameters (rain droplet size and locations, stationary or rotating arm speed), test sample material loss ratemeasured under controlled conditions (volume or mass loss, dimension measurement).

[0149] - Field data: weather conditions, erosion rates derived from the NEZP erosion sensor's sacrificial material loss measurement during the turbine operations, inspection and turbine blade controller data (blade type, material, speed, pitch angle).

[0150] In the prediction engine 1304, current data 1314 (e.g., real-time NEZP erosion sensor measurement, weather condition, blade speed, blade type) may be input to preprocessing component 1316. In some examples, preprocessing component 1316 and preprocessing component 1308 are the same. The prediction engine 1304 produces feature vector 1318 from the preprocessed current data, which is input into the model 1320 to generate one or more criteria weightings 1322. The criteria weightings 1322 may be used to output a prediction, as discussed further below.

[0151] The training engine 1302 may operate in an offline manner to train the model 1320 (e.g., on a server). The prediction engine 1304 may be designed to operate in an online manner (e.g., in real-time, at a mobile device, on a sensor edge computing device, etc.). In some examples, the model 1320 may be periodically updated via additional training (e.g., via updated input data 1306 or based on labeled or unlabeled data output in the weightings 1322) or based on identified future data, such as by using reinforcement learning to personalize a general model (e.g., the initial model 1312) to a particular user. Labels for the input data 1306 may include “mildly corroded”, “corroded”, “heavily corroded”, and so on.

[0152] The initial model 1312 may be updated using further input data 1306 until a satisfactory model 1320 is generated. The model 1320 generation may be stopped according to a specified criteria (e.g., after sufficient input data is used, such as 1,000, 10,000, 100,000 data points, etc.) or when data converges (e.g., similar inputs produce similar outputs).

[0153] The specific machine learning algorithm used for the training engine 102 may be selected from among many different potential supervised or unsupervised machine learning algorithms. Examples of supervised learning algorithms include artificial neural networks, Physics informed neuralnetworks, Bayesian networks, instance-based learning, support vector machines, decision trees (e.g., Iterative Dichotomiser 3, C9.5, Classification and Regression Tree (CART), Chi-squared Automatic Interaction Detector (CHAID), and the like), random forests, linear classifiers, quadratic classifiers, k-ncarcst neighbor, linear regression, logistic regression, and hidden Markov models. Examples of unsupervised learning algorithms include expectation-maximization algorithms, vector quantization, and information bottleneck method. Unsupervised models may not have a training engine 1302. In an example embodiment, a regression model is used and the model 1320 is a vector of coefficients corresponding to a learned importance for each of the features in the vector of features 1310, 1318. A reinforcement learning model may use Q-Learning, a deep Q network, a Monte Carlo technique including policy evaluation and policy improvement, a Statc-Action-Rcward-Statc-Action (SARSA), a Deep Deterministic Policy Gradient (DDPG), or the like.

[0154] Once trained, the machine learning model 1320 may output predicted erosion rate for current conditions, cumulative damage index (CDI) or damage equivalent loading (DEL) for leading-edge material, blade damage classification).

[0155] FIG. 14 illustrates a multi-agent agent system (MAS) 1400 for infrastructure monitoring in accordance with some embodiments. The MAS may be deployed across edge computing, hub and cloud layers.Accordingly, the MAS 1400 can calculate structural condition assessments and / or predictive maintenance recommendations based upon distributed processing via a multitude of agents as further described below.

[0156] In the depicted example, the MAS 1400 includes an edge computing agent layer 1402, a hub agent layer 1404, and a cloud computing layer 1406. While the sensor network is also shown, in some examples, the sensor network participates in the edge computing agent layer as sensor nodes. Indeed, the sensor node data may include sensor readings from NEZP sensors and / or general purpose sensors measuring erosion, corrosion, cracks, fatigue, elongation, breaks, gas presence, gaps, icing, temperature, humidity, leaks, debonding, strain and / or pressure. The edge computing agent layer1402 uses sensor data such as, erosion rates derived from sacrificial material loss, vibration amplitudes, change in conductivity due to leak or humidity or debonding or icing or gaps, and frequency shifts for internal damage detection from the NEZP sensors or GP sensors, for example after undergoing edge preprocessing (such as noise reduction, thresholding, and feature extraction (e.g., Fast Fourier Transform for vibration analysis, classifier)), to derive a preliminary data analysis (such as edge-level anomaly detection for identifying unusual erosion rates or vibration trends indicative of structural issues) via edge computing processing.

[0157] Edge agents employ edge processing, which focuses on low-latency processing, real-time decision-making, and distributed computation. Some examples of edge processing include:

[0158] Program Threads: Lightweight threads or processes running on loT devices (such as microcontroller, FPGA, neural decision processor e.g., Raspberry Pi, Arduino), and the like.

[0159] Local Al Models: TinyML or TensorFlow Lite models running directly on edge computing devices for tasks like detecting internal damage through vibration pattern analysis or classifying erosion severity based on sacrificial material loss.

[0160] Custom Firmware: Firmware updates sent to edge computing devices for specialized processing, such as analyzing erosion rates or blade imbalance based on real-time vibration data.

[0161] Resource-Constrained Processing: Performing quick filtering (e.g., noise reduction) on raw data streams before sending information to the hub, and / or flagging critical anomalies for immediate action.

[0162] The edge computing agent layer 1402 operates independently (e.g., at physical locations such as NanoX sensor nodes (e.g., sensor 104) installed on blade leading edges across the blade or inside blade or turbine components) of the hub agent layer and the cloud computing layer. Edge agents can be programmed to manage each sensor such as NEZP erosion sensor, NEZP crack sensor, general-purpose vibration sensor or can be programmed for specific role such as erosion detection agent, vibration analysis agent, turbine energy production agent.

[0163] In some examples, the hub agent layer 1404 operates at hub locations such as NanoX hubs located on turbine towers, which aggregate and preprocess data from all blade nodes. The hub processing included in the hub agent layer derives more advanced analytics such as by correlating erosion and vibration patterns across blades to detect system-wide anomalies or integration of vibration sensors along length to assess blade shape for deflection as well as data aggregation for on-demand retrieval of complete data set for a specified period for post assessment.

[0164] Hub agents function as intermediaries, aggregating, refining, and coordinating data between edge devices and cloud systems. They balance processing loads and provide fault tolerance.

[0165] Example Hub processing includes:

[0166] Clustered Threads: Multi -threaded processing in a local server or gateway to:

[0167] Aggregate data from hundreds of edge sensors.

[0168] Perform data deduplication or light Al inference (e.g., pattern detection across multiple blade sensor feeds video feeds).

[0169] Data Processing Nodes: Middleware software running on hubs for:

[0170] Protocol translation (e.g., MQTT to HTTPS).

[0171] Filtering, batching, and prioritizing data for cloud upload.

[0172] Edge Clusters: Small-scale Kubernetes clusters deployed on local servers to scale processing dynamically as more edge devices connect.

[0173] Message Brokers: MQTT brokers or RabbitMQ managing communication between edge devices and the cloud.

[0174] Distributed Databases: Regional databases for caching edge data (e.g., Redis or MongoDB) to reduce cloud dependency and enable local query execution.

[0175] The advanced analytics include predicting remaining useful life (RUL) for blades or identifying fleet-wide trends like storm-induced erosion and data aggregation combining data across turbines to improve fleet-wide condition insights and maintenance planning are then used by the cloud layer, via cloud processing. Cloud processing handles large-scale datastorage, advanced analytics, and hyperscaling. The clouds provides for more advanced (e.g., 100,000 nodes or more) AI / ML models and provide for a centralized data and analytics repository.

[0176] Examples cloud processing includes:

[0177] Hyperscaling Processors: Using GPUs, TPUs, or FPGAs for:

[0178] Deep learning model training on aggregated edge data and / or hub data.

[0179] Earge-scale simulations or optimization problems.

[0180] Serverless Functions: Event-driven functions (e.g., AWS Lambda, Azure Functions) triggered by incoming hub data to process or transform the data.

[0181] Global Load Balancers: Systems like AWS Elastic Load Balancer managing traffic from multiple hubs for high availability.

[0182] Central Al Models: High-complexity Al models (e.g., hyperscale neural networks and the like, transformer models, large recommender models) that require significant computational power.

[0183] Data Lakes and Warehouses:

[0184] Storing aggregated data for historical analysis (e.g., Amazon S3, Google BigQuery).

[0185] Running complex SQL queries or machine learning pipelines.

[0186] Orchestration Tools: Kubernetes or AWS ECS for deploying and scaling cloud applications across regions.

[0187] The MAS 1400 thus provides for adaptive processing where analysis tasks can be shifted between edge, hub, and cloud layers to optimize resource usage and system performance. The MAS 1400 additionally provides for distributed computation across multiple layers to improve power usage, optimize bandwidth, and provide just-in-time (JIT) computation, which is particularly beneficial in remote environments where power and / or connectivity can be more limited.

Claims

1. CLAIMSWhat is claimed is:

1. A method, comprising: selecting and positioning one or more sensor assemblies on a structure, wherein the structure comprises a wind turbine blade, and wherein the one or more sensor assemblies comprise at least one sensor assembly that includes an internal power supply, at least one sensor assembly that is not powered, or at least one sensor assembly that is passively powered; continuously monitoring the wind turbine blade via the one or more sensor assemblies; collecting a sensor data representative of the continuously monitoring of the wind turbine blade from the one or more sensor assemblies; analyzing the sensor data to identify a trend, an anomaly, a potential issue, or a combination thereof, of the monitored wind turbine blade; wherein the analyzing the sensor data comprises using a machine learning algorithm, an artificial intelligence algorithm, _a statistical modeling, physics-based modeling, or a combination thereof; and automatically transmitting a notification if analyzing the sensor data identifies the anomaly or the potential issue.

2. The method of claim 1, wherein the at least one sensor assembly comprises a sacrificial layer and a reference layer.

3. The method of claim 2, wherein the sacrificial layer comprises a first plurality of sensor elements, each of the first plurality of sensor elements comprising a first strip of material, wherein the reference layer comprises a second plurality of sensor elements encased in a protective coating, each of the second plurality of sensor elements comprising a second strip of material.

4. The method of claim 1, wherein analyzing the sensor data comprises deriving a damage equivalent load (DEL) relationship between a first DEL of the at least one sensor assembly and a second DEL of the wind turbine blade.

5. The method of claim 4, wherein the DEL relationship comprises an environmental DEL relationship that measures a DEL erosive damage experienced by the wind turbine blade.

6. The method of claim 4, wherein analyzing the sensor data comprises adding additional data to the sensor data for analysis, the additional data comprising weather conditions and impact speed of droplets.

7. The method of claim 1, comprising: operating a rain sensor, a first pressure sensor, and a second pressure sensor in an ultra-low power mode wherein only the rain sensor is active and the first pressure sensor and the second pressure sensor are not active; receiving, from the rain sensor, an indication of rain; upon receipt of the indication of rain, operating on a standard power mode by: activating the first pressure sensor to receive a first pressure measure, wherein the first pressure measure is representative of rain impacting the first pressure sensor; activating the second pressure sensor to receive a second pressure measure, wherein the second pressure sensor is disposed under a protective cover; comparing the first pressure measure to the second pressure measure to derive a first erosion measure of the protective cover; and deriving a second erosion measure of the structure based on the first pressure measure and the and the first erosion measure of the protective cover.

8. The method of claim 7, further comprising operating a humidity sensor, wherein the humidity sensor is disposed under the protective cover, and wherein the ultra-low power mode further comprises operating only the rain sensor and not activating the first pressure sensor, the second pressure sensor, and the humidity sensor, wherein upon receipt of the indication of rain to operate on the standard power mode further comprises: activating the humidity sensor to receive a humidity measure; andderiving a measure of adhesion to the structure of the humidity sensor, the second pressure sensor, or a combination thereof, based on the measure of adhesion.

9. A near zero power sensor system, comprising: a sensor network comprising one or more sensor assemblies positioned on a structure to be monitored, wherein each sensor assembly comprises: a sacrificial sensor layer comprising a plurality of sensor elements configured to experience environmental degradation equivalent to the structure, wherein each sensor element comprises a conductive material that permanently changes electrical properties in response to physical degradation without requiring continuous electrical power; and a reference sensor layer comprising a plurality of reference sensor elements protected from environmental exposure.

10. The near zero power sensor system of claim 9, wherein each sensor assembly further comprises a power source configured to operate intermittently, powered only when measurements are taken for the sensor assembly.

11. The near zero power sensor system of claim 9, wherein each sensor assembly further comprises a sensor processing and communications system configured to be powered externally from the sensor assembly.

12. The near zero power sensor system of claim 11, wherein the sensor processing and communications system is configured to be powered via magnetic inductance, via radio frequency, via wired power, via solar power, via thermal power, or a combination thereof.

13. The near zero power sensor system of claim 9, further comprising a data acquisition and logging system configured to collect and record sensor data from the one or more sensor assemblies.

14. The near zero power sensor system of claim 9, further comprising a data analytics system configured to create a digital twin of the structure to bemonitored, the digital twin comprising a virtual representation that mimics the degradation of the structure to be monitored.

15. The near zero power sensor system of claim 14, wherein the data analytics system is further configured to: provide a predictive modeling of the structure to be monitored; provide an active learning for the digital twin; and provide a data analysis of sensor data to identify a trend, an anomaly, a potential issue, or a combination thereof; or provide a combination thereof.

16. A sensor system, comprising: a humidity sensor positioned beneath a leading edge protection layer and configured to detect a moisture ingress into the leading edge protection layer; a rain sensor configured to detect precipitation conditions; a first surface pressure sensor positioned without erosion protection and configured to measure direct environmental pressure impact; and a second surface pressure sensor positioned with erosion protection and configured to provide reference pressure measurements, wherein in operations, the sensor system is configured to: receive, from the rain sensor, and indication of rain; upon receipt of the indication of rain: activate the first pressure sensor to receive a first pressure measure, wherein the first pressure measure is representative of rain impacting the first pressure sensor; activate the second pressure sensor to receive a second pressure measure of structure: and compare the first pressure measure to the second pressure measure to derive an erosion measure of a structure.

17. The sensor system of claim 16, wherein the sensor system is configured to operate in an ultra-low power mode wherein only the rain sensor is active and the first pressure sensor and the second pressure sensor are not active.

18. The sensor system of claim 17, comprising a power supply system, wherein the power supply system is configured to enter a standard operating mode by activating the first pressure sensor and the second pressure sensor to derive the erosion measure, and to return to the ultra-low power mode after deriving the erosion measure.

19. The sensor system of claim 18, wherein the power supply comprises a solar cell, a magnetic inductance system, a wired power system, a solar power system, a thermal power system, or a combination thereof, and wherein the structure comprises a wind turbine blade.

20. The sensor system of claim 16, comprising a data analytics system configured to derive the erosion measure of the structure, wherein the structure comprises a wind turbine blade.

Images