Statistical processing of multiple measurements results

. The product quality is determined by improvements of equipment, technologies, and production arrangement, which directly depends on the accuracy of measurement information. To achieve high measurement accuracy, it is reasonable to automate measurement processes. In case of automation, some functions are performed by equipment computer programs. The processing of direct comprehensive measurements is a complicated process including multiple algorithms of computation and various hypothesis tests. Taking into account the complexity and duration of statistical processing of results of multiple measurements, this paper is intended to develop a software measurement suite to process direct multiple measurements. The software measurement suite is a number of tools and software programs operating together to accomplish the tasks related with acquisition of required parameters and measurement results. To achieve this goal, it was required to address the following issues: implementation of the advanced measurement technologies into the developed software measurement suite; using wireless transmission of observation findings; ensuring compatibility of the developed software with the proprietary software of the used measurement instruments; improving the quality of measurements; improving the quality of processing measurement results by minimising the human factor effects on the processing quality; minimising the time spend for processing of the obtained measurement results.


Introduction
Measurements are one of the most important factors in human life. Even in ancient times, people needed to know time, distance or mass. Measurements were required in all spheres of life including construction, trade, agriculture or science. In the modern society, their significance is even higher. Today, the scope of measured values is constantly growing, together with the complexity of measurements and processing of results. The measurement process has ceased to be a single-step action and now includes several stages, the primary of them include: measurement experiment, processing and interpretation of acquired information. In this connection, standards stipulating the procedures and methods of measuring were developed.
The product quality is determined by improvements of equipment, technologies, production organisation, which directly depends on the accuracy of measurement information where even a small error leads to defects, incorrect results and decisions in manufacturing. To achieve high accuracy, it is reasonable to automate measurement processes. In case of automation, some functions are performed by equipment computer programs. Automation and metrology address a wide range of issues that the metrology engineer faces every day: -improves manufacturing productivity and quality of manufactured products; -reduces the effects of the human factor when processing the results; -reduces the level of errors in measurements; -allows for accounting and systematisation of measurement tools. Growing requirements to the accuracy of processing the measurement results and automation level of processes dictated by the need to improve the quality of products and reducing time for testing have boosted the development of new equipment with integrated automation measurement and testing tools.

Analysis of current state of programming languages used for automation of measurements
Competitive products must be high-quality, which directly depends on many conditions, including the unity of measurements and metrological support of the entire production cycle. Today, it is the control automation at all production stages that ensures the required level of quality. Subjective evaluations and human factor are replaced by objective indicators of automatic instruments. PC-integrated automatically controlled instruments are especially important, which made it possible to quickly process a huge amount of measurement findings.
Automated measurement and testing stands allow quickly and identically test technical and economic characteristics of products.
The issue of measurement automation has been relevant for many years. The active development of automation started as early as in 1970s when microelectronic processors were introduced, digital equipment was implemented, software was developed in industry, and digital instruments and micro-computers emerged.
Development of modern manufacturing, technologies, researches, expanded use of mathematical methods in processing of results required specialized automated measurement tools. In this connection, the issue of measurement automation is still relevant and urgent today. Its development will make it possible to significantly improve the efficiency of metrological support at enterprises by relieving engineers from hand labor, enhance the accuracy of verification units and references, the reliability of certification and verification results.
Main goals of automation for entities are as follows: • scientific: increased accuracy of measurements and their processing results thanks to computers; • technical: increased reliability or safety of products thanks to more accurate indicators in production; • economic: reduced expenses for labour resources by replacing human labour with machine labour; • social: development of intellectual potential of personnel by assigning all routine operations to machines.
There are objective pre-requisites to implement verification automation of measurement instruments (MI): • Modern MI manufacturers offer instruments that can solve measurement tasks and have automatic control. It can be noted that most documents for various MIs describe possible programming and automated control options.
• Modern reference base also provides automatic control by computers.
• Sophistication of manufactured instruments leads to more complex verification procedures and, therefore, longer time for verification.
• Metrological laboratories need to store large amounts of information in hard copy (reports, certificates). Even in case of the systematised storage, searching for necessary documents in the archive is complicated and not always fast.
• Verifiers quickly get tired, which negatively affects the quality of verification. Automation brought significant changes to the life of metrological departments. Software suites [1][2][3][4][5] developed based on modern requirements of metrological units account for availability of instruments and their condition, can plan any type of metrological control and repair, control metrological servicing and maintenance, process statistical metrological data, generate requests for the condition and specifications of instruments, have an electronic log of data sheets, generate verification and calibration schedules depending on the type of measurements, can account consumables in them, and write off time of each metrology engineer for performing a certain type of activities, and can export all data into a Microsoft Excel form. Capabilities of these programs also include quick information search that reduces the time necessary to issue documentation for equipment.
Such programs make it possible to automate the metrology engineer workplace at a high level, so it is preferable to use them at workplaces related with accounting of measuring instruments in all industry branches [6][7][8]. They ensure acquisition of reliable information on the condition of measuring instruments and metrology equipment.
Automation makes the verification of measuring instruments four times faster. Storage of verification results in a database facilitates and accelerates the search of necessary information and simplifies generation of records and certificates. Automation of error calculation for measuring instruments makes it possible to prevent mistakes in verification and improve the reliability of verification results. This results in enhanced performance of a metrological unit and improved quality of MI verification.
Software suits are intensively developing. Implementing modern software suits to metrological units deals with optimisation of work processes, promotes improved manufacturing productivity, prevents personnel mistakes, makes the unit activities more transparent for management, and more accessible in acquiring information necessary to take operative and strategic management decisions, which makes automation of measurements a relevant task today [9][10][11].
Industrial controllers (PLC) are maintained by process personnel, so common programming languages of microprocessors and personal computers are not suitable for programming industrial controllers since they require special skills and knowledge from personnel. Involving third-party engineering companies for programming frequently attaches the final user of an industrial PLC controller to a certain contractor. Therefore, PLC programming requires more straightforward, simple and clear programing languages open for common use. In 1979, the International Electrotechnical Commission created a special panel of technical experts in PLC, including hardware, installation testing, documentation, and communication. The panel issued the IEC 1131 standard in 1982.
Section 3 of IEC 61131-3 provides for the use of five standard programming languages of PLC (industrial controllers); 1) functional block language (FBD); 2) relay logics language (LD); 3) sequential function charts language (SFC); 4) instructions language (IL); 5) structured text language (ST). These languages are selected because they can be simply and illustratively represented using control algorithms. IEC 61131 provided a foundation to create a unified school for training experts in PLC programming. This standard also made it possible to create hardware-independent libraries.
The above programming languages that can be used to automate the processing measurement results are complex software products requiring special knowledge and skills [12][13][14]. To maintain normal operation of such software, enterprises require at least one staffing position, an expert who would be engaged in the matters related with the operation of this software and elimination of its faults. User interaction with the software through its administrator complicates the process and slows it down.
The simplest programming language suitable for automation of measurement processing is Exсel, a part of the Microsoft Office suite, which is currently installed almost on any computer. This is flexible and easy to use software that can be applied by anyone who can operate other Microsoft products such as Word. In this case, it is not required to have an inhouse programmer or network administrator or to keep in touch with the software developer. Moreover, measuring instruments within the developed software and measurement suite have proprietary software that can transmit to the computer measurement data recorded into Exсel files.
The MS Excel tabular processor can: 3. Perform statistical analysis, forecasting (support of decision-making) and optimisation.
5 . Input passwords and set protection for some table cells, hide table fragments or an  entire table. 6. Visually represent data as diagrams, histograms, and charts. 7. Input and edit texts as one does with a textual processor, create pictures using Microsoft Office.
8. Carry out import/export, data exchange with other programs (including OLAP databases) and further processing. Support of XML format.
9. Create macro-commands, economic algorithms, own functions. 10. Create applications using the VBA programming language. A valuable opportunity of Excel is code-writing based on Visual Basic for applications (VBA). This code is written using an editor separately from tables. The electronic table is controlled by object-oriented code model and data. Using this code, the data of input tables will be processed instantaneously and displayed in tables and diagrams (charts). The VBA allows creating full-scale applied packages the functions of which greatly excel those of electronic table processing. The table becomes a code interface ensuring easy code modification and calculation control. In combination with the same programming language, the MS Excel tabular processor acquires a universal nature and can solve any task irrespective of its nature.
To ensure compatibility of the developed software with the proprietary software, it is reasonable to use MS Office Exсel to automate the processing of measurement results.

Results
The development of the software measurement suite used the most advanced measurement technologies such as wireless transfer of measurement results and their automatic recording as a file into a computer.
Such measurements require: Instrumental component; Computer (laptop); MarComProfessional 5.1; Receiver (e-stick) connected to a computer USB port; Program for processing measurement results in automatic mode. Measurements include wireless data transfer from measuring instruments via a radio channel up to 6 m when using a receiver (e-stick) and up to 100 m when using FM2. Software can configure the interface by displaying the measured product and measuring instruments used, setting the representation form of the measurement results, e. g., in tabular and graphical form.
All measurements are taken with instruments having an Integrated Wireless interface that can output data to external devices using an Integrated Wireless system.
For correct software operation, the computer must meet the minimal system requirements given in Table 1, which depend on the PC model. It should be noted that all modern computers meet these requirements. DVD/CD drive Data acquired from measurements are transmitted in real time from the measuring instrument to the computer directly to MS Excel (Fig. 1). The transmitted data format is a textual file or code. All instrument data can be edited. In Excel, one can copy or transfer measurement results to various columns, lines, and sheets. It is also possible to carry out several measurement cycles.
Mahrcom Professional software provides the following capabilities: • Measured values can be directly transmitted to an MS Excel (Fig. 1)  Textual files can be edited or translated to another language. The software is developed by the manufacturer and delivered together with the tool and can be also downloaded from official sources of the manufacturer on a free-of-charge basis. It is prohibited to make any changes to the software functions in terms of software development.
The sequence of processing the results of direct multiple measurements consists of several stages. Stage 1. Giving the definite point estimates of the measurement results distribution law This stage calculates the mean arithmetic value of the measured quantity and standard deviation SD.
In accordance with the criteria, gross errors are excluded followed by a repeated calculation of estimates of mean arithmetic values and SD.
Stage 2. Finding out the measurement results distribution law or random errors Measurement results and calculations are used to build a histogram or a polygon of observation results distribution. The measurement results distribution law is evaluated by the form of dependencies built.
Stage 3. Evaluation of the distribution law upon statistical criteria For n>50 measurements, the Pearson criterion is used to identify the distribution law. For 50>n>15, the composite test is used to check the distribution law normality. For n<15, the normality of the experimental distribution is not checked.
Stage 4. Calculation of standard uncertainty by type A Stage 5. Calculation of standard uncertainty by type B Stage 6. Calculation of expanded uncertainty of measurement results Stage 7. Recording of measurement results It should be noted that the developed software measurement suite is intended to process the measurement results for 10 ≤ n ≤ 100.
According to the processing algorithm of multiple measurement results given above, the processing of results is automatic and includes all stages of the algorithm based on mathematical functions, Excel statistical formulas, and conditional formatting.
After the automatic processing of multiple measurement results, the measurement results are displayed in an Excel sheet (Fig. 2).

Conclusion
The software measurement suite was developed using the most advanced measurement technologies and wireless transmission of measurement results. It ensured compatibility of the developed software with the proprietary software of the measuring instruments used. The quality of measurement and process of measurement results was improved by reducing the human factor effects on the processing quality. Moreover, the time spent for the processing of measurement results was minimised.