Preface As we all know, scale generates efficiency. In the past two or three decades, the increasing size of engineering projects has become an inevitable trend in modern industrial development. Pipelines with diameters greater than 500 mm are now very common in engineering projects, and their flow detection (especially for gases) has become increasingly urgent and needs to be addressed. There are many instruments that can measure gas flow. From a principle and manufacturing perspective, increasing the size should not be a problem. However, the size and weight of the instrument will increase exponentially with the diameter, and it will also bring other problems. For example, orifice plates, a well-known throttling device, are not only bulky when the diameter is large, but also have significant pressure loss and excessively high operating costs. Furthermore, the new ISO 5167 standard requires a straight pipe section of 30-40D upstream, which is difficult to meet on-site, making the use of such instruments unacceptable. From quantitative change to qualitative change, facing these difficulties, in the past two or three decades, the commonly used method has been insertion flowmeters, which use sampling principles and insertion installation methods to measure the flow velocity at only one or more points in the pipeline to estimate the flow rate. These instruments share the following characteristics: simple structure, convenient installation and maintenance, low price, good repeatability, but generally low accuracy. Because their principles are all sampling-based, it is essential to first understand the velocity distribution within the pipeline in order to correctly select the location and number of detection points. Industrial Pipe Flow The Ever-Changing Velocity Distribution in Pipes Engineering projects in various industries, based on their own process requirements, require the installation of various pipe fittings (such as valves, elbows, manifolds, reducers, filters, etc.) in the pipelines. Due to the numerous forms and combinations of these fittings, the velocity distribution within the pipes is also ever-changing and difficult to estimate. R. W. Miller (author of the American Handbook of Flow Measurement Engineering) believes that "velocity distribution is the main factor affecting the accuracy of flow measurement, and the various types of fittings in industrial settings, with their highly complex flow conditions, are not only difficult to describe but also difficult to simulate in the laboratory." Since most flow meters are related to velocity distribution, the flow field used for calibration should be consistent with the flow field under practical conditions for the calibration coefficient to be meaningful. This flow field is generally considered to be fully developed turbulence, which can be obtained as long as the pipe has a relatively long straight section (see Figure 1). Fully Developed Turbulence ■ Formation—Because all actual fluids have viscosity, they will drive and constrain the fluids in adjacent layers during the flow process. After a straight pipe section of about 30D (D is the pipe inner diameter), the velocity distribution will no longer change. If the Reynolds number Re < 2000, it is laminar flow; if Re > 4000, it is turbulent flow. Turbulence is more common in industry, i.e., fully developed turbulence. ■ Description—In the past century, many scientists have conducted a lot of tests and descriptions of fully developed turbulence. Among them, Nikuradse's formula for fully developed turbulence in a smooth pipe is the simplest. It is approximately expressed as: where Vi is the velocity at any point; Vm is the maximum velocity at the center; y is the distance from the velocity point to the pipe wall; R is the pipe radius; and n is an exponent related to Re. ■ Average Velocity Point—The average velocity point for fully developed turbulent flow in a smooth pipe can be derived from Equation 1. Equation 2 shows that the average velocity point in a circular pipe depends on three factors: 1) the length of the straight pipe section; 2) the Reynolds number Re; and 3) the roughness e. Therefore, its location is not fixed, unlike what some manufacturers claim, which states that the average velocity point can be obtained by measuring only one point in the pipe. The flow accuracy is ±0.5% to 1%, and according to ISO 7145, its accuracy can only reach ±3%; if the straight pipe section is short, the flow accuracy may even be less than ±5% to 10%. To accurately measure flow rate, flow conditioners require a relatively long straight pipe section, which is often unavailable in actual field conditions. Therefore, the International Organization for Standardization (ISO) has repeatedly recommended using more than ten types of flow conditioners. However, the author believes this is not the best approach because: 1) it increases costs, as the price of a flow conditioner is comparable to that of a flow meter; 2) it requires frequent cleaning, increasing maintenance; 3) even high-performance flow conditioners have significant permanent pressure losses, increasing operating costs; and 4) they are prone to clogging, and even partial clogging alters the flow velocity distribution, failing to improve accuracy. Since these factors are not ideal, why go through the trouble of using them repeatedly? Gas Flow Measurement Instruments for Large Pipelines Under the premise of advocating the construction of a resource-saving economy in China, this article introduces gas flow measurement instruments for large pipelines, excluding throttling devices with high pressure loss and excessive operating costs; it also does not recommend excessively expensive ultrasonic gas flow meters. It focuses only on insertion flow meters based on the sampling principle, which offer relatively high cost-effectiveness in industrial control systems. These can be divided into three main categories according to their sampling method: measuring point velocity. Any instrument capable of measuring flow velocity can be inserted into the pipeline and used as a flow meter. The following are some commonly used types: ■ Double Venturi Tube—This product was introduced by Taylar in the United States more than 40 years ago. It was once imitated domestically and used in thermal power plants, known as the "small trumpet tube." Domestic products have entered the market in the last ten years or so. It utilizes the acceleration at the throat of the outer Venturi tube to generate low pressure, which promotes further acceleration of the inner Venturi tube to obtain even lower pressure. Therefore, at the same flow velocity, a larger output differential pressure can be obtained, making it more suitable for measuring the gas flow rate at low velocities in large pipelines. ■ Thermal Type—Utilizing the principle of heat transfer, using a resistance temperature detector (RTD) as the sensing element, higher flow velocities carry away more heat, lowering the RTD temperature and changing its resistance value. The change in resistance value reveals the flow velocity and flow rate. Its greatest advantage is its ability to measure velocities below 5 m/s. Heat transfer is related to fluid mass, therefore the measured flow rate is mass flow. Its disadvantages include a gas temperature generally below 200°C and a response time exceeding 1 second. ■ Other—In theory, Pitot tubes, insertion vortex shears (Figure 2), and turbines can all be used to measure flow. Pitot tubes are used for industrial field calibration but are rarely used as industrial instruments. Insertion vortex shears are unreliable at low speeds and in vibrating pipelines; insertion turbines require significant maintenance due to rotating parts. The market share of these instruments has been declining significantly in recent years. Manufacturers of these instruments often claim their instruments are calibrated in wind tunnels, but this only calibrates the velocity, not the flow rate; the flow accuracy cannot reach the advertised ±1%. Linear Velocity Measurement Flow rate is estimated by measuring the velocity at multiple points along a line in a pipeline. This method is more accurate than measuring a single point, and offers stable and reliable installation. It is often the preferred instrument for detecting gas flow in large pipelines in industrial control systems. A typical example is the averaging pitot tube flow meter: differential pressure averaging pitot tube flow meter . Based on the pitot tube velocity measurement principle, it has undergone continuous improvement over the past thirty years. Currently, the following types are available in the domestic and international markets: ■ Rhombus-II—The earliest detection rod had a circular cross-section, but it was replaced by the Rhombus-I type due to the "resistance crisis." The Rhombus-I type was then replaced by the Rhombus-II type because the back pressure hole was prone to clogging. There are two main types: one is a detection rod composed of three profiles, introduced by Emerson more than ten years ago. Due to the large tolerance of the profiles, it is prone to leakage or excessive initial stress weakening its strength when the temperature changes, and is now rarely used. The other is an integrated structure, introduced by two or three German companies. It is reliable and can withstand higher temperatures, but it is more expensive. It can now be produced and used in the field in China. ■ Magazine Type – This type held a significant market share in China over the past decade. Manufacturers claimed that the rough surface of its head could control the boundary layer, thus improving accuracy. However, experts have demonstrated that the boundary layer's impact on accuracy is negligible compared to other factors. Its disadvantages include low output differential pressure, small pressure measuring orifices, and a tendency to clog when the fluid contains dust, especially condensates, oils, or algae. ■ T-Type – This type features two rows of densely packed total pressure orifices less than 2 mm in diameter facing the flow direction. The low-pressure orifice is located on the T-shaped back, with a small diameter. Manufacturers claim that the numerous pressure measuring orifices allow for better "collection" of flow velocity distribution, achieving an unbelievable accuracy of ±0.7%. However, even if the pressure measuring orifices are densely packed into a single slit, it can only reflect the flow velocity along a straight line in the cross-section. How can accuracy be guaranteed when the straight pipe section is insufficient? Furthermore, due to the small size of the pressure measuring orifices, it is prone to clogging, just like the magazine type. Thermal Averaging Pit Tube Flowmeter The principle is the same as the single-point thermal averaging pitot tube flowmeter in the previous section, except that it is multi-point in structure, reflecting the velocity distribution at multiple points in the pipe, and thus calculating the flow rate. Comparing the two types of averaging pitot tube flowmeters, the thermal type has the advantage of high sensitivity, can measure the flow rate of low-velocity, low-temperature fluids, and directly reflects the flow velocity; while the differential pressure type, after averaging the total pressure within the sensing rod, may not reflect the average velocity due to the complex flow, so it must be corrected by calibrating the flow coefficient. It is expected that the thermal averaging pitot tube flowmeter will have significant development potential if its accuracy can be improved. Multi-point Flow Velocity Measurement at Cross-Section ■ Airfoil Flowmeter—This is an improved version of the classic Venturi tube, shortened in length, but still relatively bulky. ■ Total Flow Measurement Device—Multiple sensing tubes are inserted into the pipe cross-section. Multiple total pressure holes are drilled in the sensing tubes facing the flow direction, and multiple static pressure holes are drilled on the sides, providing more measuring points to reflect the velocity distribution at the cross-section. Although lighter than an airfoil, it is not accurate enough. ■ Thermal Averaging Ptole Flow Meter—Multiple thermal averaging pitot tube flow meters are inserted into the pipe to more comprehensively reflect the flow velocity distribution within the pipe. However, the flow velocity characteristics reflected by each thermal resistor may not be the same, and calibration and correction still need improvement. ■ Averaging Ptole Loop Flow Meter (Figure 3)—A patented product (patent number ZL200420061027.3) developed to address the shortcomings of averaging pitot tube flow meters, such as low output differential pressure, low accuracy, and neglect of the influence of pipe inner diameter on accuracy, which have existed for over thirty years. It uses dual Venturi tubes to measure low pressure, improving the output differential pressure, and employs multiple averaging pitot tubes to fully reflect the flow velocity distribution within the pipe, among other measures, improving the technical characteristics of averaging pitot tubes. This is attracting attention from domestic and international manufacturers and users. Calibration Flow rate is a derived quantity, influenced by many factors. It is necessary to correct for errors using coefficients obtained through calibration to obtain the correct flow rate value. Since most flow meters (except for volumetric and Coriolis flow meters, but these are generally used in pipes with diameters less than 0.2 meters) are closely related to the velocity distribution in the pipe, a relatively long upstream straight pipe section is required for their correct use. This means they should be installed in a fully developed turbulent flow environment. Therefore, the calibration device must also provide a fully developed turbulent flow environment; it serves as the flow field platform for calibration and application. Only in this way can the calibration coefficients be used for field instruments; otherwise, calibration is meaningless. Can a wind tunnel calibrate flow meters? Because it is difficult to measure the forces acting on an aircraft in motion, a relative method is used to keep the aircraft (or a scaled-down model) stationary. A wind tunnel generates an oncoming airflow to simulate actual flight conditions. This requires careful design and various measures to ensure that the velocity across the entire cross-section is equal (i.e., a uniform direct current field, see Figure 4). This is completely different from a fully developed turbulent flow environment and can only calibrate velocity meters, not flow meters. Some people believe that calibrating an insertion flow meter probe using a wind tunnel is sufficient, intentionally or unintentionally avoiding the influence of the pipe. As is well known, flow rate Q = pipe cross-sectional area × flow velocity V. Therefore, it can be said that an insertion flow meter, without being inserted into a pipe, is merely a velocity meter and cannot be considered a flow meter. This illustrates the crucial role of the pipe, which manifests in two aspects: 1) A relatively long straight pipe section is required to ensure the accuracy of the flow velocity; 2) An accurate pipe cross-sectional area is necessary. A wind tunnel can be used to calibrate the flow velocity, but not the flow rate. Gas Flow Calibration Device Basic requirements for the device: ■ Pipe diameter and shape are basically the same as the meter being calibrated; ■ Flow rate can be adjusted within a relatively large range (10:1); ■ A high-accuracy flow reference is required; ■ A relatively long straight pipe section is required to ensure sufficient development of turbulence (if space is limited, a flow regulator is recommended in the laboratory); ■ Other special requirements: such as operating conditions, actual flow, two-phase flow, pulsation, etc. How to handle field conditions: Large-diameter gas flow meters used in the field often have rectangular cross-sections, with varying sizes and aspect ratios, making it difficult for laboratories to meet all requirements. Secondly, there is often a shortage of straight pipe sections, meaning the meter is not installed in a fully developed turbulent flow environment. Even if calibrated in a device with these conditions, the flow coefficient may differ due to variations in the flow field, making direct application unreliable. In such cases, the velocity-area method must be used for field calibration. The velocity -area method is a classic method for measuring flow in pipes. However, due to its complexity, it is only suitable for field calibration and not for process industry testing. Because the velocity distribution within the pipe is uneven, the pipe can be divided into many unit areas Ai, and the velocity Vi on each unit area is assumed to be approximately equal. It is easy to understand that the more area divisions, the more accurate the flow, but this also increases the complexity and impracticality. ■ Accuracy—The accuracy of the velocity-area method depends on the following four factors: In the above formula: Due to the complex and asymmetrical velocity distribution in the field, and the possible presence of secondary flows and vortices, it is necessary to increase the number of measuring points to achieve higher accuracy. According to ISO 7194, for circular pipes using the velocity-area method, 48 measuring points result in Q < ±5%; 36 measuring points result in Q < ±7%; and 20 measuring points result in Q reaching ±14%. ■ Standards—Due to space limitations, the operation of the velocity-area method is omitted here. Readers can refer to the following standards: －ISO 3966-1977 Flow measurement in closed pipes—Velocity-area method using Pitot tubes. －ISO 7145-1982 Flow measurement in closed pipes—Method for measuring the velocity at a single point in a cross-section. －ISO 7194-1983 Flow measurement in closed pipes—Velocity-area method using Pitot tubes in circular pipes under vortex and asymmetrical flow conditions. - Chinese National Standard JJG835-1993 Verification Procedure for Flow Measurement Devices Using the Velocity-Area Method. Summary Straight Pipe Length is Crucial The length of the straight pipe section has a significant impact on the accuracy of flow meters, especially for insertion flow meters. However, this is difficult to guarantee on-site, and manufacturers often avoid this issue, unilaterally emphasizing the flow velocity accuracy of the instrument itself. Users should make rational choices. Measurement and Testing Have Different Focuses Flow meters differ in their intended use and required characteristics. For example, in trade measurement for logistics accounting, accuracy should be the top priority, and insertion flow meters generally cannot meet the requirements; however, their repeatability is better, and they can be used in the testing phase of industrial control systems. Consider All Factors and Select According to Needs This article introduces many insertion flow meters based on the characteristics of gas flow detection in large pipelines. Those measuring point velocity are simple and inexpensive, but inaccurate; those measuring multiple cross-sections have improved accuracy, but are bulky and inconvenient to install and maintain. In my opinion, the two types of averaging tubes for measuring lines (differential pressure and thermal) still have relatively simple structures and improved performance, making them high-performance products with a high cost-performance ratio in industrial control systems.