A multimodal piezoelectric matrix framework for high-density kinetic energy harvesting in industrial weaving looms

Abstract

The pressing demand to operationalize electrokinetic energy (EKE) applications in the weaving department of the textile industry is the motivation behind this large-scale study. In order to solve these problems, this paper introduces three new manners of energy conversion for the proposed electrokinetic phenomena. A novel approach to energy harvest optimization is represented by the first approach, i.e., the modified series–parallel piezo matrix. By appropriately arranging a number of piezoelectric devices in serial and parallel forms, this approach achieves unprecedented efficiency gains. Bidirectional linear generation is a second approach, and it reconsiders how to store energy in one way. This deep learning method achieves remarkable enhancement in energy harvesting and optimization with the aid of mechanical force from both forward and backward motions of weaving procedures. The third approach is a field tuning method referred to as unidirectional nonlinear power extraction force tuning. In order to maximize energy recovery, it focuses on specific instances in the weaving cycle where there is maximum kinetic energy. In conclusion, this project is a pioneering step to optimize utilization of EKE in the weaving section of the textile industry. Data-center efficiency, cost reduction, and sustainability are the primary focus areas. This research's value proposition in a nutshell is the conversion of wasted kinetic energy into useful forms with deep-reaching implications for the power landscape of textile industry and beyond.

Keywords

Textile unit electrokinetic energy DC–DC power conversion deep learning bidirectional linear generation

Optimization of electric power is essential in the dynamic environment of industrial operations to improve efficiency, cost, and responsibility for resource management.¹ The intelligent use of electric power is growing in significance as companies with greater purchasing concern themselves in maintaining both environmentally friendly and efficient operation. Figure 1 depicts an optimized industrial energy flow, where influencing factors are explored and improvement approaches presented. It reflects the complicated relationship between influencing factors of energy optimization in the industry.² As a comprehensive strategy, industrial electrical energy optimization is designed to schedule power consuming modes, exploit equipment performance, and minimize energy loss.^3,4 State-of-the-art technology such as smart meters, energy-efficient equipment, and data integration was covered. In the industrial sector, the objective of electrical energy optimization is to decrease power waste while enhancing the efficiency of equipment and controlling the pattern of power consumption. The use of big data, LEDs, and other energy-efficient technologies alongside smart meters for easy consumption can empower businesses to have a more comprehensive understanding of their trends in consumption. On the other hand, accentuating areas could be enhanced.^5-9

Figure 1.

Industrial energy flow optimization: the causality analysis and optimization strategies

The economic advantages of industrial electrical energy efficiency (EE) are beyond doubt. The cost reduction of using less energy also contributes to enhancing the financial status of an organization.^10-13 At the same time, transitioning from fossil fuels to green energy lowers the need for dirty fossil fuels, which is obviously a good thing for both the environment and also the future health of this particular industry. Energy saving in industry is fully synchronized with ecological objectives in the green world of today. Even when they generate their own electricity through renewable sources, such as wind or solar power, industries could do something positive for the environment and reduce their effect on the climate. A high-level diagram of the architecture is depicted in Figure 1. For the area of industrial energy optimization, Table 1 presents an international overview of renewable energy focus and industrial energy optimization. These data can help us understand the factors that influence which source is preferred in different countries. These nations are joining efforts to enhance industrial energy management by concentrating on renewables, hydropower, nuclear, and biofuels.

Table 1.

International survey of alternative energy sources and industry efficiency²⁰

Nations	Emphasis on renewable energy	Solar power capability	Wind resource	Energy combined with renewable energy (%)
India	Photovoltaics (PV) and wind	Outstanding	High	Targeting for 45% by 2030
United kingdom	Nuclear and wind	Adequate	High	Roughly 46%
United states	PV and wind	High	High	Around 25%
China	PV and wind	High	High	Goal: 30% by 2030
Japan	PV and nuclear	High	Adequate	Nearby 20%
Australia	Solar and wind	Outstanding	High	Over 25%
Canada	PV and hydro-power	High	Adequate	Around 18%
Russia	Nuclear and PV	High	Not Applicable	Nearby 6%
Germany	PV and wind	High	High	Over 50%
Brazil	Biofuel and hydro-power	High	Not Applicable	Major percentage

Deep learning algorithms can enhance electrokinetic energy (EKE) harvesting in the textile industry by predicting vibration patterns and optimizing energy extraction efficiency from weaving machines. Advanced models such as long short-term memory (LSTM) and neural networks analyze loom operating conditions, mechanical vibrations, and electrical outputs in real time. These algorithms help improve power generation stability, adaptive load management, and predictive maintenance of harvesting systems. Machine-learning-assisted monitoring also enables efficient tuning of piezoelectric and nonlinear energy-harvesting mechanisms under varying textile production conditions. Thus, intelligent data-driven optimization significantly improves the reliability and sustainability of self-powered textile manufacturing systems.

The power loom energy achievement is facilitating the goals of emission reductions and sustainability more broadly.^14-19 These energy alternatives are key factors in the construction of a more resilient industrial landscape and a healthier, ecologically balanced future. The diversity of approaches, which is affected by elements such as geography and access to resources, reflects a commitment to balancing industrial growth with the principles of protecting the environment, while also serving to accelerate a global movement toward more sustainable industries. In other words, the power usage of textiles is not confined to a single process; instead, it can be distributed across different cycles, with each cycle contributing its own cost to the overall gas consumption. Its huge electric power requirements are defined by the interaction between machinery and heating and by operational complexities.

Research contributions from this paper are as follows.

➢ Track down the kinetic power: Examine the complex and amazing kinetic energy generated by weaving-machine power looms. To measure exactly the remaining work in the loom.

➢ Innovate new approaches and emerging technologies of kinetic energy harvesting: Brainstorm new techniques to scavenge kinetic energy and convert it into electric power. Among these advanced approaches, the integration of piezoelectric materials has been studied.

➢ System integration: Develop a strategy for integrating kinetic harvesting devices into existing power distribution and grid systems. Address the problems associated with synchronization and variability of energy generation.

➢ Performance assessment: Rigorous performance evaluation of the reliability and efficiency of the kinetic energy conversion technologies developed for the weaving power loom. We learn how they can reduce power bills and make the system more reliable.

➢ Economic and environmental effects: Compare the economics and environmental benefits of kinetic energy solutions in a weaving mill.

➢ Thes organization and description of the work are summarized in the following section, which discusses background data analysis and observational data analysis. We then deal with the execution methodology of the proposed research. Next, we develop a comprehensive analysis of experimental source details. Then we present an analysis of comparative results along with inferences and, finally, concluding remarks and future scope are presented.

Preliminary work

An exhaustive review of the field of industrial energy optimization and the challenges that are currently present is presented in this section. It narrows in upon the textiles industry and analyzes exactly how many challenges are furnished in the weave power loom sector to optimize energy utilization. Furthermore, this section discusses the need for power electronics interfaces designed for energy optimization in such power looms for weaving. We also point out the importance of deep learning algorithms to enhance the performance levels of renewable energy sources. The primary data observed at the neighboring power looms in respect of three types of firms, i.e., small, medium, and large, are presented in this section to facilitate the empirical implications. Future directions and remaining research gaps are discussed in the closing part of this section. In order to make the document readable and predictable, each of these parts has a dedicated section.

Investigating and optimizing energy use in the textile industry, data-driven transformations for sustainability

Tahir et al.¹⁷ investigate EE and the environmental complexity of the textile industry in Bangladesh and identify numerous factors that bear on promoting or throttling EE. They demonstrate the purposes behind EE programs and key barriers to implementation in textile companies.

Manikyala Rao and Singh¹⁹ studied the energy performance of the Indian textile industry from raw fabric to finished fabric. The industry is considering energy consumption at every stage of production due to its economic size, employment level running into millions, and as a major textile exporter. The work concentrates on weaving, yarn production, spinning, and drying in order to become more efficient through advanced technology and energy. The report also shows the effect on energy intensity validation of perform, achieve, and trade (PAT) schemes and further encourages adoption of the most sustainable energy utilization practices across various operations in industry. The exploration of weaving sustainability has taught us the following lessons.

Improvement in energy performance due to cogeneration in the textile sector.

There are fluctuations in Indian renewable energy data collection.

Evaluation of greenhouse gases (GHGs) in solar operated weaving industries.

Physics of weaving and the significance of kinetic energy

Saravanan and Albert²⁰ described the textile energy footprints of textile products regarding energy consumption in the food and textile sectors; the authors focused on textile energy impacts. Their research contributes to the understanding of energy extraction and environmental consequences of textile production. The paper sheds light on sustainability problems and opportunities in the textile industry through energy footprints.

Beeby et al.²² studied the effect of composite preform weaving parameters on weaving capability based on vibration characteristics. The authors used vibration characteristics to examine the effect of composite preform weaving parameters on weaving capabilities. Their research explores the complex link between weaving parameters and vibration behavior to better understand how weaving circumstances affect composite preform performance during weaving. The study optimizes composite material weaving procedures. The weaving process investigated the following factors:

power loom weaving electricity expenditure;

the importance of EKE in weaving;

efficiency issues and kinetic energy wasted.

Effective kinetic energy

Amarnath et al.²³ created a novel EKE harvester using a vibration rectification mechanism for self-powered applications in railways. This was ground-breaking work. The researchers presented an innovative vibration rectification-based EKE harvester for autonomous railway systems. Self-powered applications and new EKE utilization proved the study's relevance and support sustainable energy alternatives.

Additive manufacturing of laboratory waste into an energy harvester device for self-powered applications described by Dolez²⁴ epitomizes the frontier research in nano-energy. They used 3D printing to create self-powered energy-harvesting devices out of lab waste. This work aims for sustainability by converting waste into useful resources, which is the demand of this time for self-sustainable development and reflects a major leap toward eco-friendly technology.

Different techniques are utilized to obtain EKE from wind, water, and human motions. Wind turbines, hydropower generators, and piezoelectric devices convert kinetic energy into electricity.

EKE extraction efficiency depends on many factors. These include the moving object's velocity, surface area exposed to kinetic force, extraction system design, climate, and conversion technology.

Increasing EKE efficiency with power electronics interface

Saravanan and Albert²¹ used zero-voltage transition strategies to reduce switching losses in a Luo DC–DC converter. Innovative switching mechanisms improve energy conversion efficiency and power losses in DC–DC converters in their research.^24,25 They enhance power electronics technology by minimizing switching losses and improving DC–DC converter energy conversion efficiency.

Simple nonlinear current-mode control of a DC–DC Cuk-converter for low-cost industrial applications was proposed by Albert et al.²⁶ The research enhances the performance/cost ratio of the converter control method for use in industrial applications where cost is a major concern. Several researchers^27-30 have proposed optimizations of the DC–DC converter control methods in low-cost industrial applications.

Perspectives on EKE harvesting with power electronics interface

In this paper, a detailed comparative review of EKE harvester is presented in different environments with co-relation to the power electronics interface. The paper investigates the feasibility of storing extracted EKE. The following specific aims are proposed based on preliminary work.

Design a method to amalgamate EKE harvesting systems, i.e., modified piezo matrix system, linear generation system, and nonlinear generation approach for optimal collection and harnessing of wasted energy in weaving power looms.

Develop a novel DC–DC chopper configuration suitable for second-level power conversion in reducing conduction losses and integration into the power system.^31-34

Analyze and predict the stability of EKE production compared with traditional renewables with regression machine method.

Integrated EKE harvesting system for power-loom waste are a framework that fuses modified piezo matrix, linear generation, and nonlinear generation technologies to facilitate the efficient gathering and utilization of weaving power-loom waste EKE.

Novel DC–DC chopper circuits tuning the second-stage power conversion³⁵ have been developed and demonstrated in order to maximize the efficiency of energy extraction for a perfect match with the power system.

EKE production stability is compared with conventional renewable energy sources using regression machine analysis.

Performance benefits from integrating EKE harvesting and electro-kinetic generating systems are evaluated, including energy yield, power conversion efficiency, and energy waste reduction.

Optimization algorithms are developed to optimise kinetic energy harvesting and EKE production for optimal efficiency, including adaptive control mechanisms to adjust to weaving operations and energy consumption.

Deep learning algorithm for EKE harvesting in the textile industry

Recent studies on EKE harvesting in the textile industry have focused on utilizing piezoelectric materials and nonlinear vibration mechanisms to recover wasted mechanical energy from weaving machines. Researchers have reported that piezoelectric energy harvesters integrated with loom structures can effectively convert shuttle and reed vibrations into usable electrical energy for low-power industrial applications. Advanced machine learning and deep learning algorithms, particularly LSTM and neural network models, have recently been applied to predict energy generation under varying loom operating conditions. Several works also explored nonlinear power extraction circuits and hybrid converter topologies to improve voltage stability and harvesting efficiency in dynamic textile environments. Furthermore, intelligent monitoring frameworks using artificial intelligence (AI)-based predictive models have shown promising performance in optimizing energy management and fault detection in weaving systems. These studies indicate that combining high-density piezo-matrices, bidirectional vibration harvesting, and deep-learning-assisted prediction can significantly enhance sustainable energy recovery in modern textile industries.

Materials and methods

Optimizing power loom energy capture and efficiency with kinetic energy harvesting moving elements in weaving power looms may generate kinetic energy. Key components and locations for kinetic energy production are as follows.

The shuttle, which transmits the weft thread over the warp threads, moves rapidly, as illustrated in Figure 2. Piezoelectric or electromagnetic generators may harvest shuttle kinetic energy.

Mechanical drive systems that move harnesses and reeds create kinetic energy. Flywheels and regenerative brakes may collect this loom drive system energy.

As they move up and down, head frames may generate kinetic energy because they govern warp thread movement. Energy may be gathered via piezoelectric or linear generation.

The rollers that wound the woven fabric onto the cloth beam provide kinetic energy. Rotational kinetic energy harvesters, or mechanical devices, may collect this energy.

The shedding mechanism creates the shuttle's shed using moving elements. An appropriate system can gather kinetic energy from these pieces. Table 2 presents the bottom and top closed-loop power loom kinetic-energy utilities.

Both warp and weft thread tensioning systems use moving parts that generate kinetic energy. Energy may be harnessed to generate electricity.

Figure 2.

Front view of shuttle moment

Table 2.

Combined kinetic energy production from power looms

Types of power loom	Valid subclassifications	Source of kinetic energy	Energy extraction model
Power loom (top closed)	Regular power loom, drop and under-pick box	Portable open end read (X–Y)	Electric–piezo plates integration
		Two-way yarn carrier pusher	Linear power generation
		Open-wheel pipeline	Nonlinear and unidirectional energy extraction
Power loom (bottom closed)	Rapier loom, auto loom	Open-ended frame Open-wheel pipeline	Piezo plate multiarray integration Unidirectional nonlinear energy extraction

The weaving machine frame vibrates and moves during operation. Piezoelectric materials or vibration-based harvesters may create kinetic energy from these motions. Setting the beat-up or changing patterns in the loom may provide kinetic energy for power production in secondary motions. The behavior of piezoelectric material is governed by coupled electromechanical constitutive equations

S = S^{E} T + dE

(1)

D = dT + ε^{T} E

(2)

where S is strain, T is mechanical stress, $S^{E}$ is elastic compliance at constant electric field, d is the piezoelectric coefficient, E is electric filed, D is electric displacement, and $ε^{T}$ is the permittivity constant stress. Kinetic energy comes from the weaving shuttle, reed motion, and beam vibration. The equation is

E_{k} = \frac{1}{2} m v^{2}

(3)

where v is velocity of mechanical motion and m is vibrating mass. The mechanical vibration equation is

\overset{. .}{mx} + \dot{cx} + kx = F (t)

(4)

where m is mass, c is the damping coefficient, k is stiffness, x is the displacement factor, and F(t) is external force from the weaving operation. Voltage is generated via a piezoelectric strain mechanism, and the equation for piezoelectric voltage generation is

V = \frac{d_{33} Tt}{ε}

(5)

V_{total} = N_{s} V_{p} and I_{total} = N_{p} I_{p}

(6)

where ε is dielectric permeability, T is applied stress, t is piezoelectric thickness, and v is voltage generated from piezoelectric generation. Here N_s is the number of series elements in the modified series–parallel piezo-matrix output, N_p is the number of parallel elements in the modified series–parallel piezo-matrix output, and V_p and I_p are voltage per piezo and current per piezo. Based on power loom motion is forward and backward. The harvest electric power output is

P = \frac{V^{2}}{R}

(7)

E_{Total} = E_{forward} + E_{backword}

E_{T o t a l} = \int_{0}^{T} p (t) d t

(8)

The nonlinear power extraction model equation is

\overset{. .}{mx} + \dot{cx} + kx + α x^{3} = F (t)

(9)

where $α x^{3}$ is the nonlinear stiffness used for tuned harvesting. The harvested power is defined as

P_{harvest} = η F (t) \overset{\cdot}{x}

(10)

where η is electromechanical conversion efficiency. The energy conservation efficiency is

η = \frac{P_{electrical}}{P_{mechanical}}

(11)

The total energy from the piezo-matrix is

E_{matrix} = N_{s} N_{p} E_{p}

(12)

where E_p is energy per piezo-element. To extract kinetic energy from these sections, the dynamics of the weaving power loom must be carefully assessed and suitable kinetic energy-harvesting methods chosen for each application. Depending on the kinetic energy source and conversion mechanism, piezoelectric materials, electromagnetic generators, linear generators, and mechanical systems such as flywheels or springs may be used.

Efficient use of kinetic energy in weaving power looms

Kinetic energy might improve weaving power loom efficiency. We explore EKE extraction, power loom models, and selection criteria in this thorough entire research. In addition, we examine piezo plate layout to optimize energy extraction from certain components and discover the full potential of kinetic energy in weaving power looms.

EKE model selection criteria

EKE producers are selected based on their size and mass. Higher-weight components may need various energy extraction techniques for efficiency. Force of the moving object comes from moving objects such as wooden pillars or open-end readers which affect the EKE extraction model's compatibility. Different models are used depending on force magnitude.

Open-end head frame and read piezo plate configuration

In order to collect EKE effectively, an open-end head frame needs around nine piezo plates. The piezo plate structure is as follows: eight 4 × 2 piezo disks enhance energy collection in each piezo plate. For the 4 × 2 piezo plate design shown in Figures 3 and 4, each piezo cell produces 3 V AC, resulting in a total voltage of 24 V. Different setups and EKE input may cause voltage output variability.

Figure 3.

Piezo array integration with open head frame

Figure 4.

Piezo array integration with open end read (shuttle end)

Unidirectional nonlinear open-wheel power extraction

The majority of power looms use electric motors to weave. These may be three-phase squirrel cage induction or single-phase capacitance motors. This system relies on the electric drive shaft and large open-wheel pipeline as shown in Figure 5. Weaving requires the open-wheel pipeline. It contains weaving components and the production chain process unit. The spinning steel disk at the end of the open-wheel pipeline serves a unique role. It fixes design errors in the power loom's design printing unit but does not weave. Weavers manually spin this disk to identify and rectify design issues.

In order to troubleshoot steel disks, precise electrical input to drives is necessary. An electrical brake switch helps weavers control and analyze their work.

The open-wheel unidirectional nonlinear power extraction paradigm is similar to flywheel energy storage. Its one-way mechanical energy transmission system distinguishes it.

Built for unidirectional mechanical energy transmission and electrical energy extraction, the long open-wheel pipeline is laborious.

An AC or DC generator is strategically placed at the end of the conduit. Keeping the generator smaller than the power loom's electric drives reduces input voltage consumption.

The theoretical output voltage of a generator is typically 12 V DC or AC. Electric drives and Table 3 illustrate how power loom parameters affect the output.

Variations in small-scale power loom enterprises: one electric drive powers one loom, but medium and large firms usually link one electric motor to multiple power loom shafts.

The open-wheel unidirectional nonlinear power extraction concept is adaptable to both bottom and top closed-loop power looms. Bidirectional linear power generation is shown in Figure 6.

Figure 5.

Power loom weaving for open-wheel unidirectional nonlinear energy extraction

Table 3.

Models for EKE: electrical output parameters calculation

Industry cycle in textiles	Approximately how much power is used (%)
4 × 2 piezo array (single)	V = 24 V AC, I = 0.10 A AC, P = 2.4 W, E = 2.4 × 3600, E = 8640 Watt-hours
Unidirectional nonlinear power extraction model for single open wheel	V = 16 V DC, I = 1.8 A DC, P = 28.8 W, E = 28.8 × 3600, E = 103,680 Watt-hours
One linear power extraction model	V = 4 V AC, I = 0.013 A AC, P = 14 W, E = 0.044 × 3600, E = 158.4 Watt-hours

Figure 6.

Two-way linear power generation with yarn-filling carrier pusher

Maximizing efficiency in top closed-loop power looms

In contrast to the open-wheel unidirectional nonlinear power extraction method, the linear power generation model employs linear motion. The current-carrying coil remains fixed whereas Figures 7 and 8 demonstrate a linear field motion. The linear power generation model is tailored to the unique operating characteristics of top closed-loop power looms. Table 4 presents top closed-loop power loom energy estimate. The linear power-generating model typically produces 3 V AC theoretical output. The model's integration with power loom components produces this voltage output. The model works perfectly with the power loom's bidirectional yarn file carrier pusher, open-end X–Y read, and open-end headframe. By adding the linear power generation model to the top closed-loop power loom, it can collect more electrical energy from linear motion components. This improvement drives weaving energy economy and performance. Table 5 estimates bottom closed-loop power loom energy.

Figure 7.

Open-end head frame linear power generation

Figure 8.

Linear power generation with open end X–Y read

Table 4.

Top closed-loop power load energy estimation

Suitable EKE models	EKE components present	Energy estimate
X–Y read open end (piezo array combination)	Two (4 × 2) piezo array open-end read (single), total 2 reads. The total quantity is 4 (EKE)	E = 4 × 8640 E = 34,560 Watt-hours
Unidirectional nonlinear power extraction open-end model	One pipe line, the total quantity is one (EKE)	E = 1 × 103,680 E = 10,368 Watt-hours
Linear power extraction integration: bidirectional yarn filing process	Carrier pusher is 2, the total quantity is two (EKE)	E = 3 × 158.4 E = 475.2 Watt-hours
Linear power extraction model used in a piezo array for open-end (X–Y) readings.	Six end read single incorporate with 2 read, the total read quantity is 12 (EKE)	E = 12 × 44,200 E = 1580.5 Watt-hours

Table 5.

Power loom bottom closed-loop energy estimation

Suitable EKE models	EKE components present	Energy estimate
Open head (piezo array combination)	Three wooden pillars with aluminum frame, 3 frames total. EKE model quantity is nine	E = 9 × 8640 E = 77,760 Watt-hours
Unidirectional nonlinear power extraction (open end model)	Single pipe line, the total quantity of EKE is one	E = 1 × 103,680 E = 103,680 Watt-hours
As a replacement for piezo array head frame and linear power extraction (open-end model)	Single aluminum frame contains 3 wooden pillars, 3 frames total. EKE model quantity is nine	E = 9 × 158.4 E = 1425.6 Watt-hours

Energy estimation in power loom applications: detailed analysis

This section analyses energy estimation in power loom applications, noting that these estimates depend on several parameters, including the loom's operation time and the EKE extraction models used. The deciding factor is the duration in power loom applications: loom operating hours affect energy estimation. Small-, medium-, and large-scale weaving industries are distinguished as follows. small-scale industries operate for 12 hours a day, limiting energy use. Medium and large-scale weaving industries run 24/7 shifts. This 24/7 operation requires a stronger and more reliable energy source. Figure 9 shows power electronics interface integration with a 4 × 2 piezo array and hardwood pillar. Figure 10 exhibits power electronics interface integration with a 2 × (4 × 2 piezo array) configuration and X-open-end read.

Figure 9.

Integrated power electronics interface with 4 × 2 piezo array and wooden pillar

Figure 10.

X-open-end read and power electronics interface integration involves a 2 × (4 × 2-piezo array) configuration

The top closed power loom energy estimate uses two EKE extraction models.

Matrix piezo array model. This model integrates 2 × (4 × 2) piezo arrays with open-end read. The key to weaving is that the open-end read travels constantly at 0.3 N, generating kinetic energy. The end of the open-wheel pipe is significantly modified in this variant. A PMDC generator replaces the rotating steel disk. This adaptation changes the energy extraction method, allowing mechanical energy to be directly converted into electrical energy.

Different energy extraction models. Bottom closed power looms energy estimate paradigm examines many EKE extraction models, each with its own characteristics. The open-end head frame is deliberately combined with the piezo array concept to gather EKE in a unique way.

Deep learning algorithm for EKE harvesting in textile industry

Deep learning algorithms play a significant role in improving EKE harvesting systems in the textile industry. These algorithms analyze vibration, motion, and mechanical stress data generated from textile machinery and wearable fabrics. Convolutional neural network (CNN) and LSTM models are widely used for predicting energy generation efficiency and optimizing harvesting performance. Deep learning techniques also help in fault detection, adaptive power management, and real-time monitoring of electrokinetic devices. Hybrid optimization models integrated with deep learning enhance energy conversion efficiency and reduce power losses in smart textile applications. Thus, deep-learning-based EKE harvesting systems contribute to sustainable energy generation and intelligent textile manufacturing. Table 6 presents the proposed deep learning algorithm parameters.

Table 6.

Deep learning performance parameters

Parameters	Description	Value/specification
Deep learning model	Algorithm used for prediction	CNN–LSTM hybrid
Input data type	Vibration and kinetic signals	Time-series data
Dataset size	Total collected samples	10,000 Samples
Training data ratio	Percentage used for training	80%
Testing data ratio	Percentage used for testing	20%
Batch size	Number of samples per iteration	32
Epochs	Total training iterations	100
Learning rate	Model learning speed	0.001
Optimizer	Optimization method	Adam
Loss function	Error minimization technique	Mean squared error (MSE)

Results and discussion

In this crucial section, the proposed work takes a step from the theoretical to the experimental: we describe the experimental facility used to measure and study the transformation of mechanical energy into electrical energy in a power weaving loom. This experimental work is based on cutting-edge techniques and instrumentation, using sophisticated machines for data collection. In addition, we stress the fact that the nature of deep-learning-based methods has transformed the accuracy and consistency of our experimental observations completely. This section presents an overview of the practical context of our research, bridging theory and practice. It gives insight into our creative and scrupulous method of EKE extraction through weaving.

Experimental modeling and performance evaluation of power loom for optimal extraction of kinetic energy

The three submodels within the test rig each have a unique configuration and interface to power electronics, which enables the submodel to extract EKE in an optimal manner. The matrix piezo array is an adaptable energy extraction device that incorporates these submodels. It is most famous for being able to capture electricity from mechanical vibrations. As a fully adjustable system, the arrangement also includes popular components such as the Open Winding unidirectional nonlinear power generator and Permanent Magnet linear generation.

Two different setups were also used in the experimental stage for matrix piezo array models: open-end head frame and open-end X–Y read. The open-end head frame was carefully fabricated with a mix of aluminum supports and wooden pillars to form a strong structure to provide for energy tapping. One special feature of this configuration is that the wooden pillars, where the interlacing thread process takes place, can move up and down in a continuous manner, therefore producing cyclic Magnetoelectric stress on the matrix piezo array shown in Figure 11. Every open-ended head frame has a T-model matrix piezo array (4 × 2), and nine wooden pillars are placed to enhance the energy removal.

Figure 11.

Piezo array model for open-ended head frame using a T-model

A matrix piezo array experiment for kinetic energy harvesting in power looms

During the experimental intervention, we used two different set-up combinations for matrix piezo array models: open-end head frame and open-end X–Y read. A solid piece of work for an energy extraction system, the open-end head-frame was meticulously built with an aluminum brace and hard-loom manning. What sets this particular formation apart is the fact that during weaving, vertical wooden shafts are also being continually moved along with the weft threads. This external configuration results in the matrix piezo array shown in Figure 12 being subjected to a periodic mechanical strain. Each open-end head frame is embedded with one set of 4 × 2 T-model matrix piezoelectric arrays, and nine wooden pillars are arranged in a particular pattern to obtain optimal energy extraction.

Figure 12.

T-model matrix piezo array open-end head frame testing notation

Subjecting the open-end X–Y read-based EKE extraction to the force of two layers of wood pillars allows for its assessment. Figure 13 displays the carefully measured output voltage, which provides useful information about the effectiveness of energy extraction. With the use of these data, the architecture of the open-end X–Y read may be adjusted to achieve maximum energy production. Figures 14 and 15 depict the testing frame for EKE extraction with second-level DC–DC conversion and for open-end X–Y read, respectively. The energy extraction method is built on the push–pull action of the wooden pillars on the matrix piezo plates. The excess energy is discharged from the capacitor and stored in the battery shown in Figure 16 once it reaches its maximum charge capacity. This experimental setup guarantees a steady flow of energy for many uses. As illustrated in Figure 17, this action typically produces an average power output ranging from 0.23 to 0.29 W. The collected energy is then channeled into an energy storing capacitor to be stored.

Figure 13.

A single wooden pillar’s force on the output voltage

Figure 14.

Free-flowing X–Y experimental setup for the reading-based EKE extraction

Figure 15.

EKE extraction using an open-ended head frame and a second-level DC–DC power converter

Figure 16.

Super-lift Luo converter power switch: (a) battery across voltage; (b) Pulse Width Modulation signal

Figure 17.

Output (a) current and (b) voltage super-lift Luo converter

Linear and nonlinear kinetic power loom kinetic energy harvesting experimental analysis

Figure 18 displays the uni/bidirectional linear power extraction. It shows the experimental setup for power production using wire loops with permanent magnets. This strategy is for straight-moving objects. The bottom closed-loop power loom has an open-end head frame for wooden pillar zig–zag motions, whereas the top closed-loop power loom has an open-end X–Y read mechanism for forward and backward movements and yarn filling. An X–Y carrier pusher generates bidirectional linear motions for power production. Instead, nonlinear power production works for top and bottom closed-loop power looms.

Figure 18.

Linear power extraction and bi/unidirectional converter using deep learning algorithms

Figure 19(a) shows an open-end wheel for unidirectional nonlinear power extraction in both kinds of power looms. These looms generate electricity as shown in Figures 19(a) and (b). The ultra lift Luo (ULL) converter performs second-level power conversion for the linear and nonlinear EKE extraction models. Figure 20(a) shows the output voltage of the ULL converter, and Figure 20(b) shows its output current. The ULL converter controller optimizes output voltage using feedforward artificial neural network methods. Finally, ULL converter PWM pulse power switch shows the power MOSFET PWM signal. This optimized energy transfer procedure maximizes energy conversion efficiency and minimizes power conversion losses.

Figure 19.

(a) Unidirectional and (b) bidirectional linear power extraction model output voltage

Figure 20.

(a) Output voltage and (b) current of the ULL converter

Examination of EKE extraction experiment design

Using open-ended head frames and X–Y read apparatus, the experimental design for EKE extraction details both linear and nonlinear methods, as presented in Table 7. There is a wealth of information on the setup's materials, geometric configurations, and operating settings in the table. In Table 8 we provide details about the next step of power conversion, including the components and metrics for conversion efficiency. Figure 21 provides a visual representation of the physical arrangement of the experimental setup, which helps in understanding the studies of EKE extraction methodology by showing the interconnected components, energy flow, and safety measures. It complements the tables provided. The stability study of the open-end X–Y read-based EKE and the open-end head frame is shown in Figures 22 and 23, respectively. The EKE stability study based on linear and nonlinear generation is presented in Figures 24 and 25, respectively. In Figure 26, we can see the results of the evaluation of the capacity of EKE -based power sources to offset the energy demand of the power loom.

Table 7.

Experimental EKE extraction source design

Equipment	Head frame in open model	End X–Y read in open model	Bidirectional and unidirectional power generation (linear)	Unidirectional nonlinear power extraction with open end wheel
A tool for extracting energy	4 × (4 × 2) piezo matrix	2 × (10 × 2) piezo matrix	Winding loop PM	Generation of micro PMDC
Model	Single-layer T-wooden pillar	Double-layered rectangular	Wooden pillar with push–pull actions	Rotary actions
Use in both directions	Not Suitable	Not Suitable	Suitable	Suitable
Raw kinetic energy level	Not constant	Not constant	Not constant	Constant
Standardized voltage output	22 V	30 V	2.9 V	12 V
Average current outcome	0.0201 A	0.012 A	0.03 A	1 A
Power of a single model on average	0.241 V	0.406 V	0.0425 V	12 V
Enhanced power conversion: second level	Super-lift converter	Luo converter	ULL converter	ULL converter

Table 8.

Specifics of the second level of power conversion

Particulars	Super-Lift Converter		Ultra-Lift Luo converter
Incorporated models	Open end X–Y read in open head frame model		Bidirectional and unidirectional linear power generation and open-end wheel nonlinear power extraction
	Theoretic	Real-time	Theoretic	Real-time
V_in (system input)	14 V	16 V	14 V	16 V
V ₀ (system output)	120 V	120 V	–180 V	–180 V
Capacitor	C ₁ = 500 µF	C ₁ = 2500 µF	C ₁ = 0.0875 µF, C₂ = 7.75 µF	C ₁ = C₂ = 44 µF
Resistive (R) load	150 Ω	200 Ω	100 Ω	100 Ω
Inductor	L ₁ = 7.5 mH	L ₁ = 16.8 mH	L ₁ = 2 H, L₂ = 150 mH	L ₁ = 100 mH, L₂ = 100 mH
Duty cycle (D)	0.85	0.70	0.95	0.55
Frequency switching	25,000 Hz	25,000 Hz	25,000 Hz	25,000 Hz
Control unit	Micro controller (PIC)		Micro controller (PIC)

Figure 21.

Experimental configuration for the extraction of EKE

Figure 22.

EKE stability investigation using open-end head frames

Figure 23.

Open-end X–Y read EKE stability analysis

Figure 24.

EKE stability investigation using linear generation

Figure 25.

Nonlinear unidirectional generation-based EKE stability analysis

Figure 26.

Power loom energy demand offset capacity

Modeling, prediction, and energy estimation for EKE extraction infrastructure

This research aims to evaluate the energy output of open-end head frame, X–Y read, uni/bidirectional linear generation, and open-end wheel unidirectional nonlinear power extraction EKE extraction models. The EKE estimation findings are noted in Table 9.

Table 9.

EKE extraction estimate model

Energy estimate model	Category for deep learning regression	Discipline of deep learning methods	RMSE	MAE	Findings from the test
Optimal energy model for open head frames	Generalized linear modeling	Mateen 6/2	0.0038	0.0032	0.23
A model for energy prediction based on open-ended X–Y reads	Generalized linear modeling	Deterministic quadratic	00.0034	0.0024	0.24
A model for energy prediction based on unidirectional or bidirectional linear generation	Adapting a linear model	Linear, interfaces linear, stepwise linear	0.0036	0.0030	0.0320
Energy extraction based on open-end wheels that are unidirectional and nonlinear forecasting algorithm	Support vector machine (SVM)	SVM in cubic formation	0.040	0.036	12.5

Comparison of existing power-source-based and EKE-based power loom sectors

The outline of suggested methods in theoretical, experimental, and deep learning studies is described here. These visual representations also assess the practicality of an EKE weaving machine without external power sources. Average power loom estimation is as follows: unidirectional, 249 MW; bidirectional, 499 MW; UNPFT nonlinear, 54 MW. Then, the post-converter power is measured in Figure 27: super-lift Luo bidirectional base, 399 MW; and ULL converter nonlinear base, 408 MW. Figure 28 shows a comparison of proposed and existing research.

Figure 27.

Output: (a) raw piezo-voltage; (b) harvesting approaches; (c) power output

Figure 28.

Output voltage versus power loom speed for different harvesting

The current power-source-based power loom sector uses fossil fuels or grid electricity, resulting in high prices and carbon emissions. EKE-based power loom sector reduces energy use and environmental impact by using ambient energy.

Existing power source-based power loom sector uses expensive fuel and energy. EKE-based power loom sector reduces operational costs using free, renewable energy.

Existing power loom sector has a high carbon footprint and pollution from energy source combustion. EKE-based power loom sector generates clean energy.

Power source-based power loom sector has stable energy supply but is fuel price sensitive. Reliability uses ambient energy but provides stability. Centralized energy grids and infrastructure limit scalability in the existing power-source-based power loom sector due to their inability to adapt to local energy sources. In areas with abundant ambient energy, scalability minimizes grid expansion.

Energy-saving and automation benefits in the current power-source-based power loom sector. EKE-based power looms are driven by energy collection, storage, and control innovations.

EKE-based power looms represent a sustainable and cost-effective replacement for existing power sources. Location, energy, and technology determine its viability. This comparative research emphasizes cost savings, environmental impact reduction, and scalability of switching to EKE-based power looms. Feasibility depends on location and needs continual technical and infrastructural improvement. Proposed work performance matrices are presented in Table 10.

Table 10.

Performance matrices comparison

Parameters	CNN	LSTM	Hybrid CNN–LSTM
Prediction accuracy (%)	93.5	95.2	98.1
Training time (s)	120	150	180
Energy prediction error (%)	5.8	4.2	2.1
Convergence speed	Medium	High	Very high
Computational complexity	Medium	High	High
Power optimization efficiency (%)	89	92	96

Conclusion

There is hope for a more sustainable and energy-efficient textile industry thanks to studies that have focused on weaving power looms that use EKE. This research has helped bring the sector in line with global sustainability goals by demonstrating that EKE technology may significantly reduce energy usage and carbon emissions. Because of their scalability and low cost, EKE technologies are a good fit for weaving operations of all sizes. Future developments in electrokinetic generation are anticipated to increase the advantages even more, guaranteeing a revolutionary effect on the weaving power loom business and a greener textile sector in general. EKE-based weaving power looms have the ability to revolutionize the textile industry with their limitless potential. Some important areas of future focus are as follows: (i) lower power intake using renewable sources; (ii) intelligent weaving unit machines with advances in materials; (iii) energy storage solutions. In addition, strengthened to emphasize future research directions involving the large-scale industrial validation, multisource hybrid harvesting integration, adaptive nonlinear tuning mechanisms, smart energy management systems, and real-time predictive monitoring for textile machinery environments.

Footnotes

ORCID iDs

Johny Renoald Albert

Vanchinathan Kumarasamy

Ethics approval statement

Ethical approval does not apply to this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The datasets in this study are available from the corresponding author of the original dataset upon reasonable request.

References

Abualigah

Zitar

Almotairi

, et al. Wind, solar, and photovoltaic renewable energy systems with and without energy storage optimization: A survey of advanced machine learning and deep learning techniques. Energies 2022; 15(2): 578.

Krishnamoorthy

Albert

. Electricity theft detection in IoT-based smart grids using a parameter-tuned bidirectional LSTM with pre-trained feature learning mechanism. Electr Eng 2024; 106: 5987-6001.

Bin Abu Sofian

ADA

Lim

Siti Halimatul Munawaroh

Chew

Show

PL.

Machine learning and the renewable energy revolution: Exploring solar and wind energy solutions for a sustainable future including innovations in energy storage. Sustain Develop 2024; 32(4): 3953-3978.

Sathishkumar

Karthikeyan

Adaptive power management strategy-based optimization and estimation of a renewable energy storage system in stand-alone microgrid with machine learning and data monitoring. Int J Wavelets Multires Inf Process 2020; 18(1): 1941023.

Perera

ATD

Wickramasinghe

Nik

Scartezzini

. Introducing reinforcement learning to the energy system design process. Appl Energy 2020; 262: 114580.

Forootan

Larki

Zahedi

Ahmadi

Machine learning and deep learning in energy systems: A review. Sustainability 2022; 14(8): 4832.

Kiasari

Ghaffari

Aly

HH.

A comprehensive review of the current status of smart grid technologies for renewable energies integration and future trends: The role of machine learning and energy storage systems. Energies 2024; 17(16): 4128.

Punyavathi

Pandian

Singh

Bajaj

Tuka

Blazek

Sustainable power management in light electric vehicles with hybrid energy storage and machine learning control. Sci Rep 2024; 14(1): 5661.

Zitnick

Chanussot

Das

, et al. An introduction to electrocatalyst design using machine learning for renewable energy storage. arXiv preprint arXiv:2010.09435, 2020.

10.

Paşaoğlu

Habibnezhad

Machine learning enhanced hybrid energy storage management system for renewable integration and grid stability optimization in smart microgrids. J Therm Eng 2025; 11(4): 1267-1301.

11.

Abdullah

Park

Seong

Lee

Hybrid renewable energy system design: A machine learning approach for optimal sizing with net-metering costs. Sustainability 2023; 15(11): 8538.

12.

Hafiz

Awal

de Queiroz

Husain

Real-time stochastic optimization of energy storage management using deep learning-based forecasts for residential PV applications. IEEE Trans Ind Applicat 2020; 56(3): 2216-2226.

13.

Izadi

Shahafve

Ahmadi

Neural network genetic algorithm optimization of a transient hybrid renewable energy system with solar/wind and hydrogen storage system for zero energy buildings at various climate conditions. Energy Convers Managem 2022; 260: 115593.

14.

Areola

Adebiyi

Moloi

Artificial intelligence for optimizing solar power systems with integrated storage: A critical review of techniques, challenges, and emerging trends. Electricity 2025; 6(4): 60.

15.

Vanchinathan

Selvaganesan

Adaptive fractional order PID controller tuning for brushless DC motor using artificial bee colony algorithm. Results Control Optimiz 2021; 4: 100032.

16.

Bahlawan

Losi

Manservigi

, et al. Optimization of a renewable energy plant with seasonal energy storage for the transition towards 100% renewable energy supply. Renew Energy 2022; 198: 1296-1306.

17.

Tahir

Haoyong

Khan

Javed

Laraik

Mehmood

Optimizing size of variable renewable energy sources by incorporating energy storage and demand response. IEEE Access 2019; 7: 103115-103126.

18.

Kim

Park

Kim

Lee

Song

SC.

Optimization models for the cost-effective design and operation of renewable-integrated energy systems. Renew Sustain Energy Rev 2023; 183: 113429.

19.

Manikyala Rao

Singh

AK.

Optimization and control of renewable energy integrated cogeneration plant operation by design of suitable energy storage system. Energies 2022; 15(13): 4590.

20.

Saravanan

Albert

AS.

Enhancing energy efficiency in power looms: Utilizing regression machine learning for electro-kinetic energy assessment. J Text Inst 2025; 116(6): 1036-1061.

21.

Saravanan

Albert

JR.

Optimizing energy utilization in the weaving industry: Advanced electrokinetic solutions with modified piezo matrix and super lift Luo converter. Electric Power Components Syst 2023; 52(12): 2261-2287.

22.

Beeby

Cao

Almussallam

Kinetic, thermoelectric and solar energy harvesting technologies for smart textiles. In: Multidisciplinary Know-How for Smart-Textiles Developers. Woodhead Publishing, 2013; pp. 306-328.

23.

Amarnath

Mohite

Palaskar

Energy harvesting/storage textiles: Integrating sustainability into wearable technology. BTRA Scan 2025; 54(1): 16-27.

24.

Dolez

PI.

Energy harvesting materials and structures for smart textile applications: Recent progress and path forward. Sensors 2021; 21(18): 6297.

25.

Ramasubramanian

Sundarrajan

Rao

Reddy

Chellappan

Ramakrishna

Novel low-carbon energy solutions for powering emerging wearables, smart textiles, and medical devices. Energy Environ Sci 2022; 15(12): 4928-4981.

26.

Albert

Stonier

Vanchinathan

Testing and performance evaluation of water pump irrigation system using voltage-lift multilevel inverter. Int J Ambient Energy 2022; 43(1): 8162–8175

27.

Dong

Peng

Cheng

Wang

ZL.

Smart textile triboelectric nano-generators: prospective strategies for improving electricity output performance. Nanoenergy Adv 2022; 2(1): 133-164.

28.

Tian

Dong

Yang

, et al. Textile-based triboelectric nanogenerators for smart wearable systems: Comfort, integration, and application. Adv Mater Technol 2023; 8(8): 2201294.

29.

Peng

Wang

Shao

, et al. A review of recent development of wearable triboelectric nanogenerators aiming at human clothing for energy conversion. Polymers 2023; 15(3): 508.

30.

Bhuvanesh

Kumaresan

Subramaniyan

Sasikumar

Mukeshu

. Harvesting of electricity from handloom weaving machine coupled with dynamo. In: International Conference on Advances in Materials Research. Singapore: Springer Nature Singapore, 2019; pp. 459-466.

31.

Vanchinathan

Valluvan

Gnanavel

Gokul

Renold

. An improved incipient whale optimization algorithm based robust fault detection and diagnosis for sensorless brushless DC motor drive under external disturbances. Int Trans Electr Energ Syst 2021; 31(12): e13251.

32.

Babypriya

Johny

Renoald A

Shyamalagowri

Kannan

. An experimental simulation testing of single-diode PV integrated MPPT grid-tied optimized control using Grey Wolf Algorithm. J Intell Fuzzy Syst 2022; 43(5): 5877-5896.

33.

Albert

Ramasamy

Jerard

VJM

, et al. A symmetric solar photovoltaic inverter to improve power quality using digital pulsewidth modulation approach. Wireless Pers Commun 2023; 130: 2059–2097.

34.

Albert

Sinu

Karthikeyan

Gnanamurugan

PSO tuned reactive power compensation in unbalanced grid network using double boost-modified dual controller. In: Subirats

Gurung

Ningombam

Banerji

(eds) Advanced Computational and Communication Paradigms (ICACCP 2025) (Lecture Notes in Networks and Systems, Vol. 1761). Cham: Springer, 2026. https://doi.org/10.1007/978-3-032-13555-1_1.

35.

Sinu

Renoald

Albert J

Kumarasamy

. Multi-domain control framework for predictive solar irradiation system and reactive compensation in unbalanced distribution networks using AGC–PSO. AIP Adv 2026; 16(1): 015313.