Statistical Approach To Energy Efficiency Issue In Industrial Pneumatic Systems

Abstract

In this paper, some statistical methods to the energy efficiency issue in industrial pneumatic systems are presented. Since all compressed air (CA) systems are different, in industrial applications it may be difficult to analyse energy efficiency of a single CA system from a theoretical model-based viewpoint. Instead of modeling the industrial pneumatic system from first principles the energy efficiency of the system is analyzed based on time series measurements. The dependency between energy consumption, potential variable pressure and flow variable mass flow makes possible to draw conclusions of the energy efficiency based on the measurements of pressure and power/energy. In woodworking industry, e.g. sawmills, the functioning of the assembly line is relatively monotonic from day to day. Typically, CA is almost entirely used for linear motions. A double acting cylinder is the most common actuator in these applications. The length of the impact, the radius of the piston and radius of the cylinder are the essential key figures of this cylinder type. From these figures and basic dynamics of the cylinder, the air volume per impact can be obtained. When the average values and standard deviation of the air flow in each consumption point (cylinder) is known, statistical methods can be implemented for the total consumption of CA. According to the Central Limit Theorem, the total air consumption, i.e. the sum of consumptions in different consumption points approaches a normal distribution. Since the consumed air flow is proportional to the compressor’s energy consumption, the trend of energy can also be approximated as a sum of regression curve and normally distributed deviation component. By this result, a confidence interval for energy consumption can be determined. The result is also ready to be used in energy saving calculations in the future. Also, the adjustable pressure-difference in a two-point controlled compressor has an essential effect on the overall energy efficiency of the system. When this pressure difference is large, the system is not optimal from the viewpoint of energy efficiency. Typically the time series measurements of pressure are somewhat noisy, but some kind of conclusions of stationary behavior can be done. After possible pre-filtering process of the data, it becomes evident that the measurement data signal is typically a periodic shape with strict minimum and maximum values. The consumption of CA is typically dependent on pressure i.e. higher pressure increases consumption making the shape of pressure curve exponential. A stationary time series described by its mean, variance and autocorrelation function makes possible to identify the essential features of the dynamics of the CA system and the type of consumption. With the appropriate statistical methods, the results of short-time measurements can be applied for predicting the energy consumption trend in a long run. The combination of first principle modeling with measurements opens new methods in energy efficiency techniques especially when designing new pneumatic systems or the improving of existing designs.