With the rapid promotion and application of robots in manufacturing, service and other industries, the reliability level of robot products has gradually attracted the attention of robot developers and users. How to evaluate the operational reliability level of robots is generally characterized by MTBF in the industry, such as MTBF 50000hrs in ABB Robot’s early (2014) public materials, and the MTBF of KUKA’s new krquantec robot’s 400000 operating hours. There are many evaluation methods for MTBF value of various types of robots, and there is no unified standard in the industry, which will be revealed one by one in the following.
1. What is MTBF
MTBF (mean time between failures) is the ratio of the total number of life units to the total number of failures under specified conditions and within a specified period. It is generally used to characterize Repairable Products. Calculation formula:
Where, MTBF — mean time between failures of this type / batch of products, h;
T_ I — cumulative working time of the ith product, h;
R_ N — the total number of failures in the working time of N sets of products of this type / batch participated in the statistics.
2. MTBF engineering boundary
For the concept of MTBF, the reliability field has been used to predict the reliability level of related military equipment in the 1950s. Until now, it has been nearly 70 years. Of course, there are some “displeasure” about the concept of product application in various industries, such as many people confuse MTBF with the general meaning life; how to evaluate the chip or electronic equipment with MTBF reaching nearly million hours. To solve these engineering puzzles, we must introduce another concept of “bathtubcurve” in the field of reliability. The definition of “bath curve” first appeared in the 1960s, which divides the failure rate curve into three stages in the product life cycle: early failure period (burn in period), normaloperatingperiod and wear out period.
In engineering, failure rate is usually defined as: the probability of failure in unit time after a certain moment of work for a product that has not yet failed. It is generally calculated by the following formula:
Where, △ R (T) is the number of products with failure within △ t time after time t; △ t is the time interval taken; n_ S (T) is the number of products without failure at time t.
According to the above definition, when the proportion of products with failure in △ t time is small, the treatment will be simplified in engineering. MTBF = 1 ／ λ, in fact, MTBF is slightly less than 1 ／λ. Regression to the bathtub curve above shows that MTBF represents the failure rate related to the ordinate rather than the life expectation. At the same time, in view of the high consistency of failure mechanism and the stability of failure rate in accidental failure period, MTBF is usually used to characterize the product reliability level at this stage in engineering. In fact, the engineering advantage of MTBF index lies in its engineering value to the user of the product, which is much higher than that of the product developer, because it is more convenient for the management of equipment operation time, maintenance and spare parts cost control.
3. Evaluation of robot MTBF
In fact, there are two kinds of work for MTBF quantitative calculation in reliability engineering
One is the prediction of reliability (MTBF) in the early stage of product design, the other is the evaluation of reliability (MTBF) in the later sample stage.
(1) MTBF forecast
It is predicted that the technology of robot MTBF has been relatively mature in China in the 1990s and will be applied in the industry. At present, there are general authoritative standards in reliability field, such as mil-hdbk-217, GJB / z299b, Bellcore, etc. We know that MTBF prediction is more used for comparison of architecture schemes and prediction of product foundation reliability due to its dependence on historical data and the deviation between calculated influencing factors and conditions. Therefore, the predicted value is difficult to be used as the reliability level evaluation value of actual robots.
(2) Evaluation of MTBF
The evaluation of robot MTBF depends on the actual operation data or test results. MTBF evaluation based on actual operation data has many problems in B2B business activities at present, which is often due to the high technical mutual trust between the two sides, and the evaluation is mostly large-scale and complex equipment, so it is difficult to form a standard. More MTBF evaluation in reliability engineering is based on reliability test.
We know that for MTBF evaluation based on reliability test, one of the keys is to make clear the test profile and conditions, in short, the quantitative relationship between the test profile conditions and the actual operating conditions. For the reliability evaluation of military equipment, every big country in the world has independent test standard methods, such as GJB899 of China and mil-std-781 of the United States. Due to the particularity of military equipment, these reliability evaluation standards have also stipulated the relationship between product test profile and MTBF quantitative evaluation. But for most civil products such as robots, it is difficult to form a unified user scenario, use conditions, maintenance conditions, etc. At present, some device level or module level products in China have formulated MTBF evaluation standards, including test profiles, such as GB / t29482.1-2013 and GB / t32245-2015.
The composition of robot products is more complex, including not only the traditional metal structure based mechanical actuator, but also the electronic and electrical drive control system. We know that reliability engineering for the above-mentioned systems, including the reliability assessment requirements mainly based on accidental failure, there are also life assessment requirements mainly based on wear-out failure. At present, robot industry is in the breeding and growth period, the economy and time cost are very valuable. For the robot products with tens of thousands of hours MTBF level in the industry, how to accurately and quickly evaluate the MTBF value is an urgent technical dilemma for the whole industry.
Editor in charge: CC